EPA-650/2-75-064
July 1975
Environmental Protection Technology Series
STUDY OF CONCEPTS
FOR MINIMIZING EMISSIONS
fftO COKE-OVEN DOOR SEALS
-------
EPA-650/2-75-064
STUDY OF CONCEPTS
FOR MINIMIZING EMISSIONS
FROM COKE-OVEN DOOR SEALS
by
H. W. Lownie, Jr. and A. O. Hoffman
Battelle Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
Contract No. 68-02-1439
ROAP No. 21AQR-012
Program Element No. 1AB015
EPA Project Officer: R. D. Rovang
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
WASHINGTON, D.C. 20460
July 1975
-------
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental Research
Center - Research Triangle Park , OCiict; of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOM1C ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to
develop and demonstrate instrumentation , equipment and methodology
to repair or prevent environmental degradation from point and non-
point sources of pollution. This work provides the new or improved
technology required for the control and treatment of pollution sources
to meet environmental quality standards.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
Publication No. EPA-650/2-75-064
11
-------
ABSTRACT
The report gives results of a study aimed at minimizing emis-
sions from coke-oven door seals. It identifies problems associated
with the sealing of slot-type coke-oven end closures, and quantifies
them to a limited degree by test results presented in the report. It
analyzes coke-oven door-sealing systems - those which have been
developed in the past, as well as those currently in use - with respect
to individual strengths and weaknesses. It develops and critically
analyzes concepts to improve the seal design, and recommends
further development of the two most favorable concepts.
-------
EXECUTIVE SUMMARY
ORIGIN
IN JUNE OF 1974, BATTELLE'S COLUMBUS LABORATORIES
WAS AWARDED A RESEARCH CONTRACT TO "STUDY CONCEPTS
FOR MINIMIZING EMISSIONS FROM COKE-OVEN DOOR SEALS".
THIS CONTRACT WAS FUNDED JOINTLY BY THE ENVIRONMENTAL
PROTECTION AGENCY (EPA) AND THE AMERICAN IRON AND STEEL
INSTITUTE (AISI), AND WAS MONITORED BY EPA. NINE MONTHS
OF TECHNICAL PERFORMANCE WAS FOLLOWED BY REPORT
PREPARATION, REVIEW AND CRITIQUE, AND ISSUING OF THIS
FINAL REPORT.
SCOPE
THE MAJOR OBJECTIVE WAS TO DERIVE, ORIGINATE, AND
EVALUATE PRACTIAL CONCEPTS FOR SIGNIFICANTLY IMPROVED
SYSTEMS FOR RETROFITABLE SEALING OF COKE-OVEN DOORS.
PRIOR TO THE AWARD OF THIS CONTRACT, REPRESENTA-
TIVES OF EPA AND AISI DECIDED UPON THE CONTENT OF THE
TASKS TO BE COMPLETED AND THE FUNCTIONAL REQUIREMENTS
THAT NEW SYSTEMS MUST BE CAPABLE OF MEETING. IN
GENERAL TERMS THE RESEARCH TASKS INCLUDED:
A REVIEW OF EXISTING END-SEAL TECHNOLOGY
FIELD VISITS AND TESTS TO ASSIST IN DEFINING
THE CAUSES OF THE EMISSION PROBLEM
CONSIDERATION OF POSSIBLE AVENUES OF
TECHNOLOGY TRANSFER
DEVELOPMENT OF CONCEPTS FOR NEW SEALING
SYSTEMS
-------
DEVELOPMENT OF AN EVALUATION METHOD
THAT WOULD TAKE INTO CONSIDERATION THE
OPINIONS AND JUDGMENTS OF KNOWLEDGEABLE
PERSONNEL IN THE COKE-PRODUCING INDUSTRY
AND THE EPA
DEVELOPMENT OF A METHOD TO MEASURE
DOOR-SEAL PARTICULATE EMISSIONS IN ANY
FOLLOW-ON DEMONSTRATION PROGRAM.
RESULTS AND CONCLUSIONS
(1) COKE-OVEN DOORS RELEASE PARTICULATE EMISSIONS ONLY
BECAUSE THERE ARE GAPS OR UNSEALED SPACES BETWEEN
THE DOOR-MOUNTED METAL SEALS AND THE OVEN-
MOUNTED JAMBS (DOOR FRAMES). SOME (NOT ALL) OF
THESE GAPS "SELF-SEAL" OR ARE DAMMED SHUT BY CON-
DENSING OF SEMISOLID TARS SOME TIME DURING THE COK-
ING CYCLE. HOWEVER, THIS "SELF-SEALING" PHENOMENON
IS INEFFECTIVE IN PREVENTING EMISSIONS, AND COMES AT
A TIME IN THE COKING CYCLE AFTER THE MAJOR RELEASE
OF EMISSIONS HAS OCCURRED. IT IS NECESSARY TO HAVE
TIGHT DOORS PRIOR TO CHARGING OVENS WITH COAL.
DOOR AND JAMBS THAT HAVE NO GAPS AT THE SEALS ARE
FREE OF VISIBLE EMISSIONS OVER THE ENTIRE COKING
CYCLE.
(2) THE DOOR-MOUNTED SEALS OFTEN RECEIVE THE BLAME
FOR COKE-DOOR EMISSIONS, BUT THE FUNDAMENTAL
CAUSE OF THE EMISSION-RELEASING GAPS IS THE PRO-
NOUNCED DEGREE OF WARPAGE THAT HAS OCCURRED IN
VI
-------
DIFFERING DEGREES ON MOST (IF NOT ALL) OF THE 25, 000
OR MORE CAST IRON JAMBS IN OPERATION.
(3) EXISTING METAL SEALS ON COKE-OVEN DOORS WERE NOT
DESIGNED TO BE FLEXIBLE ENOUGH TO CONFORM TO (AND
THEREFORE SEAL) MORE THAN A MINOR AMOUNT OF JAMB
BOWING. INWARD OR CONCAVE BOWING OF JAMBS IS
PARTICULARLY TROUBLESOME. HOWEVER, IT IS A CON-
CLUSION OF THIS STUDY THAT IT IS FEASIBLE TO DESIGN
RETROFITABLE, SPRING-TYPE, METAL-DOOR-SEAL SYS-
TEMS THAT:
ADJUST TO A DEGREE OF JAMB BOWING FOUR TIMES
GREATER THAN EXISTING SPRING SEALS.
ARE SIGNIFICANTLY MORE RESISTANT TO HEAT DISTOR-
TION THAN EXISTING SEALS.
CAN ADJUST AUTOMATICALLY TO BOTH JAMB DISTORTION
AND DEFLECTION (I.E., ELIMINATE THE NEED FOR
MANUAL ADJUSTMENT).
A FOURFOLD INCREASE IN AUTOMATIC INWARD AND OUT-
WARD SEAL FLEXIBILITY IS REQUIRED TO EFFECTIVELY
SEAL MOST OF THE JAMBS IN OPERATION. INCREASED
HEAT RESISTANCE OF THE SEALING SYSTEM IS REQUIRED
BECAUSE GAP-FORMING THERMAL DISTORTION OF NEW
METAL SEALS HAS BEEN OBSERVED IN LESS THAN 3
MONTHS OF OPERATION.
(4) FORTY-FIVE CONCEPTS FOR NEW DOOR-SEALING SYSTEMS
WERE DEVELOPED BY METHODS AND APPROACHES DE-
SCRIBED IN THIS REPORT. THESE CONCEPTS WERE EVALU-
ATED IN STAGES. DURING BATTELLE'S FINAL EVALUATION,
CONSIDERATION WAS GIVEN TO (A) THE JUDGMENTS AND
VII
-------
COMMENTS MADE BY THE AISI AND EPA REPRESENTATIVES
DURING A PRELIMINARY EVALUATION, (B) THE TECHNICAL
SPECIFICATIONS THAT BATTELLE RESEARCHERS JUDGED
THAT NEW SEAL ELEMENTS MUST MEET, AND (C) THE
FUNCTIONAL REQUIREMENTS SPECIFIED BY THE SPONSORS.
(5) IN A SYSTEMATIC CONSIDERATION OF THE TECHNICAL
SPECIFICATIONS AND FUNCTIONAL CRITERIA, BATTELLE
RESEARCHERS GAVE A "PROBABLE" RATING ONLY TO CON-
TACT SEALS, I.E., TO SIGNIFICANTLY UPGRADED METAL
SEALS. THIS "PROBABLE" RATING PREDICTS A 90 TO 100
PERCENT PROBABILITY OF SUCCESSFUL DEVELOPMENT
AND PERFORMANCE IN MEETING ALL OF THE CRITERIA
DEVELOPED AND SPECIFIED.
(6) A "POSSIBLE" RATING (40 TO 90 PERCENT PROBABILITY OF
SUCCESS) WAS GIVEN TO VARIATIONS OF LUTED SEALS
DESCRIBED IN THIS REPORT AS THE APPLICATION OF
FOAMED-IN-PLACE SEALANTS EITHER BETWEEN THE
MATING SURFACES OR INJECTED INTO THE GAS PASSAGE
AFTER A DOOR HAS BEEN MOUNTED ON AN OVEN. THIS
RATING INDICATES THAT RESEARCHERS HAVE RESERVA-
TIONS ABOUT THE ABILITY OF LUTED-SEAL CONCEPTS TO
MEET SEVERAL OF THE TECHNICAL SPECIFICATIONS. IN
CONSIDERING THE DEVELOPMENT OF THIS APPROACH,
THERE ARE UNKNOWNS THAT CAN BE EVALUATED ONLY IN
AN EXPERIMENTAL PROGRAM. ALL OTHER CONCEPTS AND
CONCEPT FAMILIES WERE GIVEN LOWER RATINGS.
(7) THE DEVELOPMENT OF A COKE-DOOR EMISSION-TEST
METHOD TOOK THE PATH OF ENCLOSING A COMPLETE COKE-
OVEN DOOR WITHIN A SEALED HOOD AND PASSING THE
viii
-------
EXHAUST THROUGH A FILTRATION UNIT TO COLLECT PAR-
TICULATES. THE METHOD WAS TESTED SUCCESSFULLY
ON AN OPERATING COKE-OVEN DOOR.
(8) OVERALL, IT WAS FOUND THAT THERE HAS BEEN VERY
LITTLE TECHNICAL EFFORT WITHIN THE COKE-PRODUCING
INDUSTRY TO COLLECT AND ANALYZE BASIC DATA RELAT-
ING TO THE CONDITIONS, PERFORMANCE, AND PROBLEMS
OF COKE-OVEN END CLOSURES OF WHICH DOORS AND
JAMBS ARE A PART. THIS IS THE REASON FOR THE WIDE
DIFFERENCES OF OPINION THAT EXIST ON THIS SUBJECT.
THE EFFORT MADE IN THIS RESEARCH PROGRAM TO DEFINE
THE PROBLEM REPRESENTS ONLY A START OF WHAT
BATTELLE BELIEVES IS REQUIRED IN THE WAY OF TECHNI-
CAL ANALYSES. FURTHER STUDY AND ANALYSES ARE
NECESSARY TO AID IN RATIONAL DEVELOPMENT OF EFFEC-
TIVE NEW SEALING SYSTEMS.
RECOMMENDATIONS
(!) MORE-FLEXIBLE AND MORE HEAT-RESISTANT METAL SEALS
SHOULD BE DEVELOPED FURTHER IN A FOLLOW-ON PRO-
GRAM. DESIGN WORK SHOULD BE PRECEDED BY EXPERI-
MENTAL EFFORT TO ANALYZE THE TEMPERATURE DISTRI-
BUTION AND THERMAL-STRESS PATTERNS IN EXISTING
SYSTEMS AND DESIGNS. THESE ANALYSES WOULD SERVE AS
VALUABLE INPUTS TO THE DESIGN AND MATERIAL-SELEC-
TION PROCESS AND WOULD ALSO SERVE AS A BENCHMARK
FOR EVALUATING THE PROJECTED PERFORMANCE LIFE OF
DESIGNS FOR NEW SYSTEMS.
-------
(2) BECAUSE THE BASIC EMISSION-CAUSING PROBLEM IS THE
DISTORTION THAT HAS OCCURRED (AND IS CONTINUING) AT
OPERATING JAMBS, THE FACTORS CAUSING THIS PROBLEM
SHOULD BE ANALYZED QUANTITATIVELY. A TECHNICAL
ANALYSIS SHOULD INDICATE WHAT STEPS CAN BE TAKEN IN
DESIGN AND MATERIALS TO DEVELOP A MORE DIMEN-
S1ONALLY STABLE JAMB FOR BOTH NEW COKE-OVEN BAT-
TERIES AND REPLACEMENT OF SOME JAMBS AT EXISTING
BATTERIES.
(3) BECAUSE THE CONCEPT OF USING SEALANTS OF SOME
VARIETY BETWEEN THE DOOR AND THE JAMB SURFACE HAS
THE POTENTIAL FOR TOTALLY ELIMINATING DOOR EMIS-
SIONS, THE APPLICATION AND PERFORMANCE OF VARIOUS
SEALANTS SHOULD BE TESTED IN A LABORATORY
ARRANGEMENT.
-------
TABLE OF CONTENTS
Page
CHAPTER I. INTRODUCTION I-1
Background and Antecedents 1-1
Scope and Purpose . 1-2
Research Staff 1-5
Special Acknowledgments 1-5
Qualifying Statements 1-6
CHAPTER II. SUMMARY AND RECOMMENDATIONS . . II-1
Results of Battelle's Final Evaluation II-2
Cause of the Coke-Door-Emissions Problem ... II-5
Development of Emission-Test Methods II-5
Recommendations for a Follow-On Approach
and Program II-6
CHAPTER III. REVIEW OF THE EVOLUTION OF
PERTINENT COKE-MAKING TECHNOLOGY AND
COKE-OVEN SEALS Ill-1
Pertinent Coke-Making Technology III-l
Historical Review of Coke-Oven Door Seals .... III-6
,CHAPTER IV. INVESTIGATION AND DISCUSSION OF
THE CAUSES OF THE EMISSIONS PROBLEM IV-1
Introduction and General Statement of the Problem. . IV-1
Summary and Conclusions IV-2
Fundamentals of Demountable Industrial Seals . . . IV-7
Laboratory-Test Results IV-10
Field-Test Results IV-18
Discussion of Heat Warpage of Metallic Components . IV-43
Flexibility and Conformability Limitations of
Existing Sealing-Edge Designs IV-48
Technical Specifications Tha.t New Sealing
Systems Should Meet IV-54
Judgments and Opinions on Existing End-Closure
Sealing Systems IV-57
XI
-------
TABLE OF CONTENTS
(Continued)
CHAPTER V. POTENTIAL TECHNOLOGY TRANSFER . . V-l
Technical Objective V-l
Expanded Objective V-l
Literature Search . V-2
Survey of Resilient Sealing Materials and Resins . . V-3
Small-Group Technology Transfer and Innovation . . V-4
CHAPTER VI. CONCEPTUALIZATION OF SEALING
METHODS VI-1
«t
Approach to Concept Generation . ....... VI-1
Preliminary Evaluation of Sealing Methods . '. . . VI-1
Classification of Sealing Methods ..'..... VI-2
Sealing-Concept Families VI-4
Seal Concept Family 1 - Hot Zone Seals VI-5
Seal Concept Family 2 - Luting Seals ...:.. VI-9
Seal Concept Family 3 - Resilient Seals VI-17
Seal Concept Family 4 - Inflatable Seals VI-29
Seal Concept Family 5 - Contact Seals . . . . . VI-39
Seal Concept Family 6 - Noncontact Seal VI-54
CHAPTER VII. EVALUATION OF FAMILIES OF
SEALING CONCEPTS VII-1
Initial Judgmental Evaluation of Concept Families
by AISI/EPA/BCL VII-1
Battelle's Evaluation of Concept Families Via Com-
parison With Seal-System Specifications and
Functional Requirements ......... VII-13
CHAPTER VIII. DEVELOPMENT OF AN EMISSION
EVALUATION METHOD VIII-1
Comments on the Measuring Method VIII-1
Discussion of Equipment and Procedure ..... VIII-Z
Laboratory Test Runs VIII-12
Coke-Plant Test Run Vin-12
XII
-------
LIST OF FIGURES
Page
Figure III-l. Early Method of Charging Coal to
Coke Ovens IU-2
Figure III-2. Location of End Flues in an Early
Coke Oven III-Z
Figure 111-3. Location of "Free Space" in a By-Product
Coke Oven III-3
Figure 1II-4. Gas Flow During By-Product Coking
For Various Net Coking Times .... III-4
Figure III-5. Reported Variation in Gas-Channel Pressure
Over the First 60 Minutes in a 6-Meter
Oven III-5
Figure III-6. Luted Coke-Oven Door of Early Design . III-8
Figure III-7. Luted Coke-Oven Doors Prior to 1919 . . Ill-8
Figure III-8. Koppers Luted "Stopper" Door .... III-9
Figure IH-9. Luted Door Design in Use About 1929'. . IH-9
Figure 111-10. Luted Door Design in Use About 1946 . . 111-10
Figure III-11. "Otto" Self-Sealing Door Design, Prior
to 1915 III-ll
Figure III-12. Self-Sealing Door Placed in Operation
in 1946 m-12
Figure III-13. Self-Sealing Door for Coke Oven Placed
in Operation About 1951 111-13
Figure III-14. "Turtleback" Door for a Coke Oven
Placed in Operation About 1953 .... HI-14
Figure III- 15. Coke-Oven Door Design for 6-Meter
(20-Foot) Coke Oven Placed in Operation
About 1968 Ill-15
Figure 111-16. Coke-Oven Door Design for a 1968
4.5-Meter (15-Foot) Battery 111-15
Figure III-17. Knife-Edge Seal Designs for Coke-Oven
Doors Ill-16
Xlll
-------
Figure IV-9.
Figure IV-10.
e
LIST OF FIGURES
(Continued)
Figure III-18. The Goldschmidt Vented Coke-Oven
Door Ill-18
Figure III-19. Recent Coke-Oven-Door Design Used on
a Battery Under Construction During
1974 111-20
Figure IV-1. Photograph of Leakage of Emissions
From a Single Door IV-3
Figure IV-2. Photograph of Heavy Emission From
Doors in a Coke-Oven Battery .... IV-3
Figure IV-3. Microstructure of a Metal Sample Taken
From Near the Top of the Side Member
of a Discarded Jamb, Close to the Seal
Contact Area IV-13
Figure IV-4. Microstructure of a Metal Sample Taken
From the Center of the Top Member of
the Discarded Door Jamb IV-13
Figure IV-5. Basic Components Used in Laboratory
Metal-to-Metal Sealing Tests . .... IV-16
Figure IV-6. The Metal-to-Metal Testing Equipment
in Operation IV-16
Figure IV-7. "Streaming" of Emissions From a Test
Door IV-ZO
Figure IV-8. Sketch of the Side and Front View of a
Seal-Edge Buckle at the Upper Latch
Area of a Pusher-Side Door IV-23
A Sketch Showing the Permanent Outward
(Away From the Oven) Distortion of the
Sealing Strip Above the Upper Latch . . IV-24
Sketch Showing the Location of an Experi-
mental Water-Cooling Passage in the
Cross Section of a Jamb IV-27
xiv
-------
LIST OF FIGURES
(Continued)
Page
Figure IV-11. Sketch Showing the Location of Thermo-
couples Attached to the Seal Edges of a
Door at the Chicago Test Site IV-27
Figure IV-12. A Photograph of the General Thermo-
couple Installation on a Test Door . . . IV-30
Figure IV- 13. A Photograph Showing the Installation of
a Spring-Loaded and a Fixed Thermo-
couple Mounted Outboard of a S-Shape
Seal Design IV-30
Figure IV-14. Sketch Showing the Location of Thermo-
couples Located Inside and Outside of the
Seal Edge and the Location of Spring-
Loaded Contact Thermocouples .... IV-31
Figure IV-15. Photograph Showing the Installation of
Sheathed Thermocouples Attached to the
Outside of a Sealing Strip IV-31
Figure IV-16. Plot of Temperatures of Selected
Thermocouples During the Second Cycle
Recorded IV-33
Figure IV-17. Drawing of a. Collapsible Straight-Edge
Used in Obtaining the Profile of Jambs
and Sealing Edges IV-35
Figure IV-18. Using a Straight-Edge to Measure the
Profile of a Sealing Edge IV-35
Figure IV-19. Vertical Profile of a Jamb and Door-
Edge Seal at the Chicago Test Site . . . IV-37
Figure IV-20. Vertical Profile of a Test Door on a
10-Year-Old Jamb, and the Profile of a
Leaking Door on its Mating Jamb . . . IV-38
Figure IV-21. Plot of the Horizontal Displacement of
Jambs on Four Coke Ovens IV-40
Figure IV-22. Photograph of a Thermally Warped
Leveler Bar IV-44
xv
-------
LIST OF FIGURES
(Continued)
Figure IV-23. Full-Scale Shape and Dimensions of the
Seal-Edge Components of Coke-Oven
Doors IV-49
Figure IV-24. Seal-Deflection Pattern on Point Loading
a New S-Type Seal Arrangement. . . . IV-5Z
Figure V-l. "S"-Type Metallic Seal V-3
Figure VI-1. An Illustration of the Logical Develop-
ment of the Concept Classification
System VI-3
Figure VII-1. Evaluation Worksheet VII-6
Figure VII-2. Full-Scale Cross Section of a Gas
Passage Area Showing the Installation of
an Example Curved-End Cantilever
Spring Seal VII-20
Figure VIII-1. Coke-Oven Hood Held by Researcher
(D. Hupp) in Folded Position Prior to
Mounting on Buckstays or Simulated
Coke Oven. Magnetic Strips Show on
Back of Hood Section VIE-4
Figure VIII-2. Hinged Hood Being Pulled into Place
on Buckstays VIII-4
Figure VIII-3. Hood Pulled up to Cover Over Half of
Cavity Between Buckstays VIII-5
Figure VIII-4. Hood Clamp Attached to Buckstay Flange
and Fixed in Open Position VIII-5
Figure VIII-5. Hood Clamp Attached to Buckstay Flange
and Fixed in Closed Position - VIII-6
Figure VIII-6. Overall View of Hood Clamp in Closed
Position VIII-6
Figure VIII-7. View of Hood-Top Cover Plate Showing
Pressure Tap, Gas-Exit Port, Length-
Adjustment Feature, and Plate-Support
Hooks VIII-7
xvi
-------
LIST OF FIGURES
(Continued)
Page
Figure VIII-8. Hood-Top Cover as Viewed Looking
Upward From Between Buckstays . . . VIII-8
Figure VIII-9. Hood Air Supply Before Installation
Shown Attached to Compressed-Air
Supply Line Containing Pressure Gage,
Pressure Regulator, and Filter .... VIII-8
Figure VIII-10. Overall View of Sampling Equipment
Used in Taking Samples and Making
Measurements VIII-10
Figure VIII-11. Closeup of Hi-Vol Sampler and its Flow
Recorder, Adsorber Column With Flow
Meter and Pump, and the Manometer
Board VIII-10
Figure VIII-12. Hi-Vol Sampler With Coarse (Upper)
and Regular (Lower) Filters Being
Installed VIII-11
Figure VIII-13. Koppers High-Bench Coke-Oven Battery
From Coke Side of Battery VIII-14
Figure VIII-14. Oven Doors and Buckstays on Coke Side
of Battery VIII-14
xvn
-------
LIST OF TABLES
Page
Table IV- 1. Summary of Metal-to-Metal Seal Tests With
and Without Fillers IV-17
Table V-l. Digest and Summary of Ideas Generated at
the Battelle-Columbus Brainstorming
Session V-7
Table V-2. Digest and Summary of Ideas Generated at
the Battelle-Northwest Brainstorming
Session V-9
Table VII-1. Total Base Ratings VII-10
Table VII-2. Rating of Concepts VII-11
Table VII-3. Comparison of Ranking of Concepts for
Selected Criteria VII-12
Table VII-4. Identification Key for Concept-Family
Evaluation VII-14
Table VII-5. A Summary of Battelle's Evaluation of the
Sealing-Concept Families Based on Selected
Specifications and Functional Require-
ments VII-15
Table VII-6. Range of Jamb Distortion Measured on
Eight Jambs at a Battery Having Serious
Door-Emissions Problems VII-17
Table VII-7. Summary of Example Spring Calculations . . VII-24
xvi 11
-------
CHAPTER I
INTRODUCTION
This is the Final Report prepared by Battelle's Columbus
Laboratories concluding a 9-month study and research program
entitled "A Study of Concepts for Minimizing Emissions from Coke-
Oven Door Seals". The work reported herein was sponsored by the
Control Systems Laboratory of the Environmental Protection Agency
(EPA) and by the American Iron and Steel Institute (AISI). The
opinions, evaluations, judgments, and recommendations expressed
are strictly those of the participating Battelle-Columbus staff.
Background and Antecedents
In January of 1970, Battelle-Columbus issued a formal report
on coke-plant emissions control to the National Air Pollution Control
Administration.'*' This report was the first step in defining and
reviewing the vast problem of curtailing airborne emissions from
coke ovens. Battelle researchers recommended that the solution to
air-emission problems, which all coke-oven operators have in
common, can best be achieved by group action and joint contributions.
At that time, joint action was in its early stages, and later was ex-
panded to include both technical and funding contributions by both
EPA and AISI. The present project is another example of combined
effort toward a common goal.
(1)"Final Report on Evaluation of Process Alternatives to Improve Control of Air Pollution from
Production of Coke", by T. M. Barnes, A. O. Hoffman, and H. W. Lownie, Jr. j prepared for
the Division of Process Control Engineering, National Air Pollution Control Administration,
United States Department of Health, Education, and Welfare under Contract PH 22-68-65.
Available from the National Technical Information Service as Document PB 189266.
1-1
-------
In January of 1974, the Batte Lie-Columbus Laboratories
responded to an EPA Request For Proposal (EPA RFP No. DU-
74-A039) dealing with a research program to study, innovate, and
evaluate concepts for minimizing emissions from coke-oven door
seals. It was understood that the research was to be sponsored and
financed jointly by EPA and AISI and that the research was to be
monitored by EPA.
The contract for this research was awarded to Battelle-
Columbus in June, 1974. The stated period of performance included
9 months of technical effort to be followed within 30 days by the
submission of a Proposed Final Report for review by the Sponsors.
This Final Report is the result of the technical effort and the review
of the Proposed Final Report which was dated March 26, 1975.
Scope and Purpose
The scope of work as stated in the joint contract with EPA and
AISI was as follows:
"The Contractor shall undertake a scientific and engi-
neering investigation to define technology to eliminate
emissions due to leakage from slot type coke oven and
closures. The product of the investigation will be tech-
nical descriptions of one or more clearly defined tech-
niques to meet the above objective with background infor-
mation sufficient to support future full scale demonstra-
tion work."
"The technique or techniques recommended must be
capable of meeting, to the greatest extent possible, the
following functional description:
1. Capable of being retrofitted to all current and con-
templated slot-type coke ovens, encompassing all
oven heights and construction types.
2. Compatible with existing door handling and oven end
working machinery.
3. Operability and reliability commensurate with pres-
ent coke oven practice.
1-2
-------
4. Dependability and repeatability of operation.
5. No creation of additional or different environmental
problems.
6. No adverse effects on product quality.
In order to meet the project objectives, it is expected
that the investigation consider the basic problems asso-
ciated with coke oven end emissions and examine past
and present attempts to control this source or similar
sources. Using this background information, conceptual
schemes for meeting the objectives can be brought to a
point where actual technological development can be
initiated. It is not anticipated that construction or test-
ing of individual conceptual schemes will be required in
the present investigation. However, if fabrication of a
seal section or component for shop testing is necessary
to establish the practicality of a particular concept,
such work shall be performed. "
The major purpose of the research program was to derive and
originate practical concepts on which further research/engineering
and testing of improved coke-oven seals can be soundly based. As
part of the project input, Battelle-Columbus was expected to review
existing end-seal technology, to consider possible avenues of
technology transfer, and to develop a critique method that would
take into consideration the opinions and evaluation (of concepts) of
experienced coke-plant superintendents and EPA personnel.
In addition, the scope included development of a method for
measuring door-seal emissions. The purpose of this task was to
provide a basis for future evaluations of the performance of different
end-seal configurations in any follow-on demonstration project.
In accordance with the EPA Request For Proposal, the study
was divided into six tasks. The timing of these tasks overlapped and
most were conducted, at least partially, simultaneously. This
report follows, in general, the sequence of tasks as defined in the
Contract. The identifications of the Tasks were as follows:
" 1. Problem Definition - Conduct an analysis to determine the
conditions which effect emissions from oven end seals. Vari-
ous battery sizes and configurations both in current use and
contemplated shall be considered. Techniques such as
1-3
-------
interviews with plant operators, literature surveys, and
in-plant measurements may be utilized to define the
problem.
2. Review of Existing End Seal Technology - Catalogue and
analyze various sealing schemes that have been put into prac-
tice and/or proposed. The schemes shall be defined by means
of literature reviews, industry interviews, in-house informa-
tion, or other means at the disposal of the contractor.
3. Review Potential Technology Transfer - Define other technical
areas which have similar sealing requirements.
4. Develop Conceptual Methods - Based on the background infor-
mation developed above, develop concepts for sealing techniques
to meet the functional goals of the system. For the more
promising systems, prepare conceptual engineering drawings
and working descriptions adequate for further analysis and
development.
5. Develop Test Methods - In order to secure useful information
from the follow-on demonstration effort, it is necessary to
develop methods for measuring door seal emissions or (as a
minimum) methodology for comparing the performance of
different door seal configurations.
6. Critique of Conceptual Ideas - Prepare a critique of each
system developed in 4 above. The critique should address
itself to the functional goals of the system and should define
the specific problems to be overcome for the scheme to oper-
ate successfully."
Early in this program, the Chairman of the AISI Task Force
on Coke-Oven Door Seals polled the AISI coke-producing plants to
determine (a) whether individual plants had any information, data,
or contributions to make to this program and (b) whether each plant
was willing to have a visitation by the Battelle-Columbus researchers.
It is understood that all plants were willing to have research visitors
and several plants requested a visitation. Contact was established
with most coke-producing companies, and visits and return visits
were made to those plants that were in a position to contribute
important information dealing with the objectives of this program.
1-4
-------
Some individual plants are experimenting with improved end-
seal closures, mainly in terms of adding elastomeric or refractory-
type guard seals either inside or outside the primary, standard
metal-to-metal seal. The results of these programs were examined
and evaluated.
Research Staff
The Battelle-Columbus professional staff most directly con-
cerned with major inputs to this study included the following:
H. W. Lownie, Jr., Program Manager
A. O. Hoffman, Project Leader
J. M. Allen
J. J. Grimm
R. E. Maringer
R. L. Paul
J. B. Purdy
E. E. Reiber
J. Varga, Jr.
The Sponsorship personnel most directly concerned with this study
were:
Mr. John G. Munson, Jr., Chairman, AISI Task Force on
Coke-Oven Door Seals
Mr. Calvin Cooley, AISI Headquarters, Washington, D.C.
Mr. Richard D. Rovang, Project Officer for EPA
Mr. George G. Bennett, Contract Administrator for EPA
Mr. M. P. Huneycutt, Contracting Officer for EPA.
Special Acknowledgments
Early in this program it became apparent that many coke-plant
superintendents were much interested in contributing their experi-
ence, thoughts, and time to this program. This helpful attitude
and the resulting inputs were of great assistance in reaching the
objectives of this program. These individuals chose not to be men-
tioned by name or affiliation, but Battelle-Columbus researchers
appreciate their extra effort on behalf of this program.
1-5
-------
The number of "other contributors" is large and is, therefore,
not included. However, this problem-solving program attracted
the interest and the contributions of other Battelle staff members
experienced in many fields and based in two of Battelle's research
laboratories. To these "other" contributors, we also express
appreciation.
Qualifying Statements
Some of the comments and judgments in this report could be
construed as being critical of the door-sealing designs and
approaches of the builders of coke batteries. In this regard, it
should be appreciated that (a) Battelle has been working toward a
new set of performance standards, i.e., what was good then is not
acceptable now; (b) Battelle has been working in what is a hindsight
position relative to the designers and builders of existing seals; and
(c) builders may be actively developing their own designs for future
ovens. If they are doing so, these designs were not revealed to
Battelle researchers.
In hindsight, everyone can recognize mistakes or compromises
that were not effective. In this regard, it should be recognized that
the concepts presented in this report are aimed at retrofitability,
which may or may not be an objective of coke-oven builders. Battelle
researchers hope that coke-oven builders will review this report and
will make known to our Sponsorship (and Battelle) any errors or
omissions that they deem important. Battelle takes the position that
all viewpoints should be considered in solving this air-pollution
problem.
A group of steel-industry coke-producing experts known as the
Task Force on Coke-Oven Door Seals of the AISI Technical Commit-
tee on Coke-Oven Practice, along with cognizant staff members of
EPA, assisted Battelle researchers in the review and evaluation of
concepts and in a review of this report in its proposed final form.
The efforts of all reviewers are acknowledged with thanks. However,
their participation and assistance should not be construed as approval
of or concurrence with findings and recommendations expressed in
this report. The judgments and conclusions are to be regarded as
Battelle's contribution and responsibility.
1-6
-------
CHAPTER II
SUMMARY AND RECOMMENDATIONS
Based on data and conceptual inputs derived and collected from
sources within Battelle and within the coke-producing industry,
Battelle researchers selected the 45 concepts described and pre-
sented in this report. These were then classified into six concept
families (concepts with a common base) and each family was eval-
uated. As a first step, evaluators from EPA, AISI, and Battelle
met at Battelle-Columbus and by joint action developed an ordered
and ranked list of 15 criteria for use in the evaluation process.
Evaluation of the concept families was then completed privately by
each individual evaluator using a weight of 1 through 5 for each
criteria for each concept family. The total score for any one con-
cept family was the sum of the multiplications of ranking and the
weighting.
Overall, the results of the combined evaluation/critique did
not indicate any outstanding preference or consensus. However,
three concept families were rated low by all groups of evaluators.
This judgmental evaluation was followed by continued research
effort to gain further insights into the practical aspects of the vari-
ous concept families. In this effort, the comments of the AISI and
EPA representatives gave the researchers useful inputs. Battelle's
final evaluation was based on seven technical specifications derived
by Battelle plus the six functional requirements specified by the
Sponsorship. The seven technical specifications are summarized
as follows:
Must withstand the 200 to 315 C (400 to 600 F) tempera-
ture pattern in the seal location for prolonged periods
of time.
II-1
-------
Must withstand occasional short periods of up to 430 C
(800 F) without being destroyed.
Must have automatic gap-closure capability; i.e., no
need for manual adjustment of any kind.
Should have increased gap-closure capability (up to
four times that of existing seals).
Must be resistant to corrosion and chemical attack.
Must be total-failure proof; i.e., must not be sus-
ceptible to the possibility of complete and/or sudden
failure during operation.
Should not introduce new cleaning problems that are
not clearly solvable.
Results of Battelle's Final Evaluation
In the final evaluation of the six concept families described
in this report, Battelle researchers gave only one family a "probable"
rating (90 to 100 percent probability of overall successful develop-
ment and performance) in each of the technical specifications and
functional requirements. This preferred-concept family is that of
metal contact seals. These would be upgraded versions of existing
metal contact seals. Stated another way, Battelle researchers
believe that (a) existing sealing systems (although often called in-
adequate) are really marginally successful and can be greatly im-
proved, and (b) there is no other concept or concept family that
approaches this sealing approach in terms of probability of success-
ful performance in this difficult and hostile environment.
A stylized sketch of one of the contact-seal concepts, taken
from the family of contact seals shown in Chapter VI, is as follows:
II-2
-------
Adaptor frame
Contact Seal
Movement of seal when door is
closing scrapes jamb surface
clean
Jamb
Door-
clJ
\
A technical evaluation of the family of metal contact seals in
terms of increased sealing flexibility, adaptability, and improved
heat resistance established to Battelle's satisfaction that this ap-
proach is feasible (Chapter VII). Overall, it is judged that this
concept family has the greatest potential for successful develop-
ment of an effective and broadly acceptable oven-door-sealing
system.
Only one sealing concept was rated as "possible" (40 to 90 per-
cent probability of overall successful development and performance).
This "possible" rating was given to "luted seals". In this report
the luted-seal family includes the applications of foamed or unfoamed
sealants placed on the jamb or door before a door is mounted on an
oven, and also includes foaming materials that are injected into the
gas passage after the door is mounted on the oven. A stylized
sketch of one of the luted seals is as follows:
II-3
-------
Foam
(closed pores)
Existing brickwork
Existing jomb
Existing door
Gap of varying thickness
resulting from jomb
worpage fills with foam
Sealing frame
(added)
The luted-seal concept had appeal to evaluators representing
the AISI and EPA. Battelle researchers are also aware that sealants
have the potential for completely eliminating emissions from coke-
oven doors. In addition, there is a possibility that a simple pro-
cedure for foam sealing of doors could be developed. However,
this remains only a possibility at this time and was rated as such.
Battelle's reservations were on the technical-specification points
of increased gap-closure capability and avoidance of new cleaning
problems. Also, there were reservations in the functional require-
ment relating to dependability and repeatability. The only way
possible to evaluate and/or bypass the uncertainties of this luting
approach would be through a laboratory testing program or through
a combined laboratory and coke-plant testing program.
It is a well known and demonstrable fact that the contact seals
as represented by existing doors today do not truly seal emissions
through intimate metal-to-metal contact. When doors are emission
free, sealing has been accomplished through the contact with deposits
condensed at the sealing surfaces. In a broad sense, the fluid and/or
semisolid deposits can be thought of as luting compounds. Using this
idea, today's contact seals are "self-luting seals". Ultimate sealing
capacity for contact seals (developed according to the recommended
contact-sealing approach) might require a luting compound. There-
fore, if development of sealing devices according to the recommended
concept should prove eventually to be less sucessful in sealing emis-
sions than required, development of luted seals would be a logical
and rational continuation of the development effort.
II-4
-------
Cause of the Coke-Door-Emissions Problem
Investigation of the causes of visible emissions from coke-
oven seals shows that the door-mounted seals receive the blame for
emissions, but that the fundamental cause of the problem is the
continuing thermal warpage and distortion of the cast iron, oven-
mounted jambs. While there are also design and warpage problems
in existing door-mounted seals, the degree of warpage on many
jambs is in excess of the limits that some existing seals have been
designed to handle. Emissions are the result of gaps between sealing
edges and the warped jambs; i.e., gaps that either are not filled
with previously deposited tar sealants or are only slowly filled with
tar sealants during the coking cycle. The fact that gaps are only
slowly dammed or filled is a result of the varying composition of
the gases and vapors reaching the seal location and the varying con-
densation rate of the tars on the jamb surfaces. During the period
of high internal pressure at the beginning of every coking cycle, the
major portion of the gases reaching the jambs is steam. This steam
content does not permit rapid self-sealing of gaps by the tar compon-
ent of these gases. Later in the coking cycle, higher-boiling-temper-
ature tars reach the jambs and seal areas and the steam content is
lower. Self-sealing of gaps is rapid near the bottom of ovens where
the jamb is cooler and is slower at the top of the oven where the
jamb is hotter.
It is judged, however, that replacement of the unknown number
of jambs that are warped excessively with new jambs of conventional
design is not a solution to this problem. It is not a solution because
new jambs would promptly begin to warp. This judgment was taken
into consideration by Battelle researchers in presenting their
recommendations.
Development of Emission-Test Methods
The objective of this portion of the research program was to
develop the equipment and a method for measuring the amount of
particulate emissions from coke-oven seals. This equipment and
procedure forms a basis for techniques to be used in the evaluation
of new seals developed in any follow-on program.
II-5
-------
The principle selected for collecting emissions involves a
sealed hood placed over a complete coke-oven door with an exhaust
leading to a filtration unit to collect particulates. Although this
basic principle may have some inherent disadvantages for pro-
longed collection of emissions, it does lend itself to a quantitative
measurement (by weight) of coke-oven-door emissions during the
early portion of a coking cycle. The EPA has an interest in further
development of this equipment to collect all of the emissions and
gases over an entire coking cycle. However, there is concern over
the increase in temperature of the door resulting from the full-time
hooding of the door. Further development work on this aspect is
being separately funded by EPA.
For the purposes of the present program, the selected equip-
ment and method of measuring emissions were demonstrated at a
commercial coke plant. It was concluded that the equipment and
method can be used to measure quantitatively the particulate emis-
sions from experimental coke-oven-door seals in any follow-on
project.
Recommendations for a Follow-On Approach and Program
This section deals with Battelle's recommendations and
suggestions for a follow-on approach and a follow-on program.
Introductory Comments
A point of view, or relative position, is inherent in all things
involving judgment or a difference of interest. Battelle's point of
view places emphasis on developing a rather detailed technical
definition of any problem so that subsequent work can be completed
rationally. From this viewpoint, it is Battelle's suggestion that
additional research/investigation/analysis should be completed
prior to beginning the design of contact seals. The technical causes
of coke-door emissions are more complex than is readily apparent.
Other points of view are based on differences in interest,
approach, and background. Battelle does not discount other points
of view and hopes that every possible approach is taken to the solu-
tion of the problem. For example, another point of view might be:
II-6
-------
"Call it empirical if you want, but let's attack the problem by build-
ing test doors - see what happens then go from there". This
approach could give fast results and Battelle hopes that companies
will assign additional technical personnel to such projects to attempt
to get fast results. However, empirical approaches do not lend
themselves to extrapolation nor do they contribute to developing
answers to such questions as "If it works at Plant A, why doesn't
it work at Plant B"? Regardless of the various approaches that will
be taken, Battelle suggests that continued technical analysis would
contribute significantly to the solution of the problem. Battelle's
recommendations are presented from this point of view.
Recommendations
Battelle's major recommendations concerning a follow-on
approach/program are as follows:
(1) Because the door-emissions problem is both urgent
and serious, it is recommended that research/
development of new, retrofitable metal contact seals
should be emphasized. This approach is expected
to seal most, if not all, of the existing warped
jambs in the coke-producing industry. This program
should be taken to the point of demonstration, and
should be followed by the preparation of retrofit
specifications and operational manuals. The be-
ginning of the design phase should be preceded by
a technical program designed to analyze the stresses,
temperature fluctuations, and dimensional shifts of
existing end-closure components. It is suggested
that a comparison of the operating stress level in
new seals with existing baseline conditions will indi-
cate whether the new seals will have a superior
longevity. This approach should eliminate the need
for prolonged testing to establish seal life.
(2) Because the real cause of the coke-door emissions
problem is the warping of the door jambs, it is
recommended that research be directed to quanti-
fying and analyzing the factors causing this warpage.
This investigation should produce recommendations
of materials/designs/procedures that will result in
11-7
-------
dimensionally stable jambs for (a) new ovens,
and (b) replacements, and probably for upgrading
existing jambs.
(3) Because luted seals (of one kind or another) have
the possibility of completely eliminating door
emissions, it is recommended that a laboratory/
plant exploratory experimental program be
initiated to define further the problems and possi-
ble solutions. It is suggested that if there is
demonstrable progress, the entire approach (luting)
should then be evaluated in a feasibility/cost/bene-
fits analysis relative to the progress being made
in development of metal contact seals.
II-8
-------
CHAPTER III
REVIEW OF THE EVOLUTION OF PERTINENT COKE-
MAKING TECHNOLOGY AND COKE-OVEN SEALS
Various factors in the production of coke have a direct influence
on the problem of leaking coke-oven doors. This chapter provides
some information concerning these factors, as well as historical back-
ground on the development of coke-oven doors and door seals.
Pertinent Coke-Making Technology
The features of coke manufacture in coke ovens that can influence
the leakage of coke-oven doors can be divided into two groups: (1) those
that relate to operating practice, and (2) those concerned with gas
evolution during coking.
Coke-Making Operating Practice
The first by-product coke ovens were placed in operation in the
United States in 1892 at Syracuse, New York. "''r The early coke ovens
were not charged to their full volume as is the normal operating prac-
tice today. Leveling mechanisms had not as yet become part of pushing
machines. Coal charged into an oven found its natural angle of repose
as shown in Figure III-l. W The coal was charged in this manner to
minimize the formation of "black ends" (or what is known today as
"green coke") because of the location of the end flues as shown in
Figure III-2. I ' Subsequent evolution of design of coke-oven doors is
discussed later in this chapter.
* References are at the end of the chapter.
HI-1
-------
FIGURE III-l.
EARLY METHOD OF CHARGING COAL
TO COKE OVENS
Vertical Heating Flue
Buckstay
>oor
FIGURE III-2. LOCATION OF END FLUES IN AN
EARLY COKE OVEN
Horizontal cross section.
Maintaining the "free space" above the charged coal is an im-
portant factor in minimizing gas pressure within a coke oven. The
"free space" is illustrated in Figure III-3, which is approximately to
scale for a 4-meter (13-foot) oven. The "free space" is about
0. 3 meter (1 foot) high and is that space available for gas flow be-
tween the leveled coal and the roof of the oven. This space must
be kept open to allow passage of the gases from the coking coal
to the ascension pipe and on into the collecting main. If for any
reason an oven is charged with coal in excess of the optimum
volume, the excess coal will result in a decrease in the volume of
III-2
-------
free space with an accompanying restriction to the passage of gases
and a resulting increase in gas pressure. This situation may be over-
come if the steam-aspiration system has the capability of increasing
suction in the ascension pipes. Overcharging of an oven may also
result in forcing coal into the bottom of the gas offtake to the ascension
pipe. Because of the high temperatures at this location in the oven,
small amounts of coal may fuse to the hot refractory lining the gas
offtake, resulting in restriction of the gas passage and causing in-
creased back pressure in the oven.
Gas Main
Refractory
Door Plug'
Free-Space
Top of Leveled Coal
Coke Oven Floor
FIGURE III-3.
LOCATION OF "FREE SPACE" IN A
BY-PRODUCT COKE OVEN
During the coking cycle, carbon can build up on the roof of the
coke oven and cause a decrease in available free space and, again, an
increase in oven gas pressure. This is a particular problem at some
plants. The carbon forms in fine filaments hanging down from the roof
(similar to stalactites formed on the roof of a cave) and can severely
lower the volume of free space. Removal of the roof carbon is usually
done by allowing the oven to remain empty with the doors on the oven.
One or more charging-port lids are usually opened a small amount to
permit air to enter the oven and to burn the roof carbon. During these
periods the coke-oven doors are exposed to high temperatures in the
III-3
-------
range of about 925 C to 1150 C {1700 to 2100 F). It is believed that if
this decarbonization period is prolonged, it can result in warpage of
the jambs and door seals on an oven.
Evolution of Gases
Factors that contribute to the problem of keeping coke-oven doors
Leak-tight are the larger-than-average generation of gas at the start of
the coking cycle, coupled with the restrictions in the upward flow of gas
through the coal. This combination of high gas volume and restriction
to flow results in the development of relatively high gas pressure at the
door seals.
The amounts of gas generated throughout the coking cycle for
different coking times are illustrated in Figure III-4. (*) in referring
to Figure III-4, it should be noted that the gas-flow rates for coking
times of 15 to 17 hours (the usual range for blast furnace cokes) vary
between 40 and 60 cubic meters per minute (25-35 cubic feet per
0 I 2 5456 T « 9 10 II 12 13 14 19 16 IT IS 19 20 21 22 2324 2426 !
Coking rune, hours
FIGURE UI-4. GAS FLOW DURING BY-PRODUCT COKING
FOR VARIOUS NET COKING TIMES
III-4
-------
second) during most of the coking cycle. The period of highest pres-
sure is at the start of the coking cycle, during which time gas pres-
sures in the coke-oven-door gas channels have been reported to peak
as high as 710 mm of water (about 28 inches of water or about 1 psi).
The time-pressure relationship for this unusually high-pressure peak
is shown in Figure III-5. (4) The high-pressure peak may correlate
with a brief but unusually high volume release of gas just after charg-
ing. However, it is suspected that the pressure in the gas passage is
elevated at the beginning of a coking cycle because of the inability of
the gas to flow upward through the finely pulverized coal. With a
restricted upward flow of gas through the coal, gas can flow laterally
in high volume into the gas channels. This lateral flow continues until
the coal against the heated wall cokes and develops fissures. Once
fissures have developed, numerous paths are open for the gases to
travel up the coke/coke-oven wall interface,1 resulting in a drop in
pressure in the gas channel.
Average Gas Channel Pressure
2 4 6 8 10
15
20 30 40
Minutes After Start of Charge
50
60
FIGURE III-5. REPORTED VARIATION IN GAS-CHANNEL
PRESSURE OVER THE FIRST 60 MINUTES
IN A 6-METER OVEN<4>
III-5
-------
Historical Review of Coke-Oven Door Seals
The operators of early by-product coke ovens (prior to 1920)
were not concerned with the same types of operating problems as the
coke -plant operators of today. From an early publication (1932), the
following statement describes the requirements of a good coke-oven
door of that
"A good coke oven door must close tightly so that no air
will enter the oven chambers and damage the quality of
coke and gas. It must be a good heat insulator, so as to
assure economical operation. The door must also be
constructed of such material that its frame will not be
attacked by the heat. "
The requirements for a good door at that time were similar to those for
today's coke plant, except that the early coke -plant operator was con-
cerned with the leakage of air into the ovens, while today's operator
although still concerned with leakage into an oven is also concerned with
the leakage of steam, gases, and vapors out of the ovens and with the
air -pollution problems such leakage creates.
Any discussion pertaining to coke-oven door seals cannot be
limited to the seals alone. The coke-oven door, door jamb, seal, and
refractory brickwork immediately adjacent to the door jamb all play an
important part in maintaining tight coke-oven doors. A coke-oven
battery is not a stable structure. The high temperatures occurring in
the flues and oven chambers, combined with the periodic exposure of
the ovens to atmospheric temperatures during the pushing operations,
cause movement throughout the battery. This movement is sufficient
to initiate opening of joints in the refractory brick and opening of seals
between the refractory brick and door frames. Once the joints have
opened and gases can leak from the oven to the door-jamb area, the
coke -plant operator's problems increase. The escaping gases can
ignite and result in severe damage to doors and door jambs, to the
extent that they must be replaced.
Luted Doors
Doors on the first by-product coke ovens were luted to effect the
seal between the door and door frame. The earliest of such designs is
III-6
-------
shown in Figure III-2. The door was a very simple shape with a thin
lining of refractory brick. There was considerable distance between
the last flue of the ovens and the door, with the result that a consid-
erable amount of uncoked coal (green coke) was produced.
An attempt was made to resolve the green-coke problem by
bringing the coke-oven door into the oven so that its inner surface was
in line with the end flue as illustrated in Figure III-6.' ' This design
may well have been established at the turn of the century. The metal
door frame fitted against the refractories, and the final seal was
accomplished with luting clay applied between the edge of the door and
the adjacent refractory. This type of door did not perform as expected
because the short distance between the door and the end flues caused
the door to overheat.
Further design changes were made in an attempt to resolve the
problem of overheating of the doors. Two of these designs are shown
in Figure III-7. *2' In each of these designs, resolution of the overheat-
ing problem was attempted by increasing the thickness of the refractory
lining in the door. However, in each of these designs the distance be-
tween the coke-oven door and end flues was unchanged and there was no
improvement in door operation or service life.
Another change in the design of luted coke-oven doors is illus-
trated in Figure III-8. *2' The metal part of the door frame was moved
toward the outside of the oven and the refractory liner took on the shape
of the "refractory plug" used on today's coke-oven doors.
Later luted-door designs are illustrated in Figures III-9 and
III-10. (5) The door design in Figure III-9 makes use of a roll-formed
steel shape to contain the refractories on the ends of the ovens. Struc-
tural steel channels are used as buckstays, which is the design retained
from the earlier luted doors. Luting clay placed between the metal door
frame and the roll-formed steel shape accomplished the seal for con-
trolling leakage. This is one of the earliest designs containing a channel
which can be referred to as a "gas passage". The gas passage is a
means of exhausting gases from the bottom of the coal charge at the
doors to the free space above the coal. Whether this was the intent of
designers in that day is not known. In some subsequent illustrations
the gas passage is present, while in other examples the gas passage
is absent.
The door design illustrated in Figure III-10 shows the same com-
ponents as the design in Figure III-9, except that the roll-formed steel
HI-7
-------
Luting Clay
FIGURE III-6. LUTED COKE-OVEN DOOR OF EARLY DESIGN
Luting Location
FIGURE III-7. LUTED COKE-OVEN DOORS PRIOR TO 1919
III-8
-------
FIGURE IU-8. KOPPERS LUTED "STOPPER" DOOR
Refractory Plug Retainer
Luting Locatio
Roll-Formed Steel Shape
Buckstays
FIGURE III-9. LUTED DOOR DESIGN IN USE ABOUT 1929
III-9
-------
section at the end of the refractories has been replaced by a heavy cast
structure which also incorporates the door jamb as a distinct component.
This is also the first luted door with wide-flange beams as buckstays.
FIGURE HI-10. LUTED DOOR DESIGN IN USE ABOUT 1946
Self-Sealing Doors
The first mention of "self-sealing" as applied to coke-oven doors
appears to have been in a patent survey published in 1915. W The self-
sealing coke-oven door was an "Otto" design that is shown in Fig-
ure III-11. ' ' One of the main features of this design was an integral-
frame jamb casting that enclosed the entire end of an oven between the
buckstays. This Otto door employed many of the design features that
are in use today. The rough adjustment of the door was made by the
bolts designated as "5", which are a part of present-day mechanisms.
"Fine adjustment" within the various door sections was accomplished
by the bolts designated as "11". Quoting from the literature(2):
III-10
-------
r
_
'/"
r
-
s
1
3
L|J
J
1
1
J
>
0
^
)
»
1
1
>
0
01!
O
£
T(O
_S=-i_
Ro)5
3
B^ !
=(o)=
2
- - -
^p)=
^"^
yO)
Vcr
ryi
fO
o
d
(
J
a
r
r
j
j
n
Ll
/
f3
r
r-
B
ff
«C3
FIGURE III-11. "OTTO" SELF-SEALING DOOR DESIGN,
PRIOR TO 1915
III-II
-------
"The door has three links which enable it to bend in an
inward as well as in an outward direction. Along the
edge of plate "6" there is a flexible rim "7" bearing a
sealing frame which consists of a Tee-steel-section. "
The use of adjusting bolts, the flexible links, and the sealing frame are
design concepts still in use today. The designers of this early self-
sealing door had a great deal of confidence in their design as shown
from the following statement^':
"It is easy to change the sealing frame though this will
not be necessary under no circumstances. " (This is a
correct quote. )
Figures III-12 through III-16 illustrate the various designs of
self-sealing doors that have been placed in operation from 1946
through 1970. Figure III-1.2 is an illustration of a self-sealing door
FIGURE III-12.
SELF-SEALING DOOR PLACED IN OPERATION
IN 1946
111-12
-------
for a 4-meter (13-foot) coke oven. ^ The door jamb is removable and
is bolted to the buckstay backing plate. The seal component has the
cross section of a junior structural channel; the channel web providing
flexibility in the seal which is pressed against the jamb by a spring -
loaded assembly.
A coke-oven door design for a 4-meter (13-foot) coke oven placed
in operation about 1951 is illustrated in Figure III-13. The door has a
removable jamb, and a special spring-loaded S-shaped seal. The buck-
stays for this coke oven are much heavier than those shown in Fig-
ure III-12.
FIGURE III-13.
SELF-SEALING DOOR FOR COKE OVEN
PLACED IN OPERATION ABOUT 1951
III-13
-------
A "turtleback" coke-oven door for a 4-meter (13-foot) battery
placed in operation about 1953 is illustrated in Figure III-14. (5' The
name "turtleback" comes from the design of the door casting which
has a heavy part projecting away from the center of the door. Con-
trast this design with the doors shown in the preceding figures. A
flat sealing ring is used with the adjusting load applied by a bolt pass-
ing through the door casting.
FIGURE III-14. "TURTLEBACK" DOOR FOR A COKE OVEN
PLACED IN OPERATION ABOUT 1953
Two coke-oven door designs for ovens placed in operation about
1968 are illustrated in Figures III-15 and III-16. The design shown in
Figure III-15 has rolled sections as the main door structure with in-
serted knife edges in heavier supporting sections for the seal. The
knife edge is attached to a flexible diaphragm which also acts as the
retainer for the refractory door plug. The knife edge is loaded by
bolts passing through lugs welded to the rolled-steel door structure.
The door jamb is heavier than in other designs and is held in place by
UI-14
-------
FIGURE III-15. COKE-OVEN DOOR DESIGN FOR 6-METER
(20-FOOT) COKE OVEN PLACED IN
OPERATION ABOUT 1968
FIGURE 111-16. COKE-OVEN DOOR DESIGN FOR A 1968
4.5-METER (15-FOOT) BATTERY
III-15
-------
clamps bolted to the buckstays. Figure III-16 illustrates another door
from a 4. 5-meter (15-foot) battery placed in operation during 1968.
The coke-oven door jamb is bolted to the buckstay backing plate per-
mitting removal and replacement. Spring-loaded pins are used to load
the S-shaped seal.
Knife-Edge Designs
Some knife-edge designs are evident in the figures in the preced-
ing section. Several other designs are illustrated in Figure III-17. ' '
Fixed -
Edge. Seal
FIGURE III-17. KNIFE-EDGE SEAL DESIGNS FOR
COKE-OVEN DOORS
III-16
-------
The two bottom designs shown are predominant in the vast majority of
coke ovens in the United States. These are often described in terms of
the name of the builder and also in descriptive terms. For the pur-
poses of this report, Battelle terms the bottom-left design as the
"fixed-edge design" and the bottom-right design as the "S-shaped seal
design".
Although the exact reasoning behind the designs has not been
stated, the thought is generally expressed that a narrow edge on the
sealing surface would cut into a tar deposit on the door jamb and pro-
vide improved sealing. Another line of reasoning suggests that the
knife edge offers a small area (total area of the thin knife edge) to
make leak-tight and, therefore, lowers the force needed to make a good
seal.
Coke-Oven Door Design Developments
In the past 2 years considerable work has been done on two devel-
opments. One is the use of vented refractory plugs in the coke-oven
doors and the other is the use of water-cooled door jambs on the coke
ovens. However, the basic ideas of these developments were patented
before 1932.<2)
Vented Refractory Door Plugs
Patents were issued before 1932 for doors with flues. German
Patent Number 86, 145 was issued to Dr. Otto and Company, German
Patent Number 97,480 to J. W. Neinhans, and U. S. Patent Number
100, 774 to F. Wolff for coke-oven doors with flues. The patents de-
scribed the flues as a means of introducing heat to the door plug for
the purpose of lowering heat losses from the oven chamber through
the door. The vents were not identified as useful to carry gases
from the bottom of the coke oven to the free space above the coal.
In 1953, the idea was reported of venting coke-oven door plugs
as a means for conveying the gases and vapors formed at the bottom of
coke ovens to the free space of the oven. The concept was patented by
Friederick Goldschmidt and became known as the "Goldschmidt Door",
which is illustrated in Figure III-18. (') A year later, 12 coke plants
were reported to be operating in the United Kingdom with the vented
III-17
-------
10
*+vl
SECTION A B.
FIGURE 111-18. THE GOLDSCHMIDT VENTED COKE-OVEN
DOOR(?)
III-18
-------
door plugs. ^ The following statements were reported with reference
to the operating situation:
"The heavy gassing and leakage, so often noticeable when
an oven is first charged are absent. So also is the egress
of heavy tarry matter and pitch, generally noticeable
around the leveler door and the bottom of the door seal. "
"The work of cleaning and scraping the self-sealing doors
was reduced to a minimum and in some cases did not
appear to be necessary."
"There was complete absence of leakage from the Goldschmidt
door both when the ovens were first charged and throughout
the coking time. "(8)
However, the preceding quotes must be qualified by the fact that no pub-
lished reports concerning the Goldschmidt door appeared in subsequent
years to verify the claims.
The vents in the refractory plug of the coke-oven door provide an
easy passage for the crude gas from the bottom of the ovens at the doors
to the free space above the coal, thereby lowering the pressure at the
coke-oven door seal. The vented plug has received renewed attention
by coke-plant operators in the United States as evidenced by two papers
delivered at the April, 1974, AIME Ironmaking Conference. ^, J0)
Work is continuing at several coke-oven plants throughout the country
to develop the design concepts for efficient use of vented coke-oven
door plugs.
Water-Cooled Door Jambs
U.S. Patent Number 725,471 was issued to E. A. Moore before
1932. No further information was found pertaining to the possible appli-
cation of this concept. During the course of field visits and field trials
for this study, the application of water-cooled jambs was observed at a
coke-oven battery. Details of these studies are discussed elsewhere in
this report (Chapter IV).
Ill-19
-------
Most Recent Coke-Oven-Door
Design Concept
The most recent design concept up to the time of the writing of
this report is being used on a 5-meter (16.4-foot) coke-oven battery
under construction during 1974 in Canada. The design is illustrated
in Figure 111-19. f11) The following elements are used: (1) heavy
buckstays, (2) heavy buckstay backing plates, (3) heavy door jambs
bolted to the backing plate for ease of door-jamb removal, (4) spe-
cially-designed refractory shapes to minimize paths for gas leakage
to the doors and jambs, (5) blanket woven refractories placed in
critical refractory-shape joints to permit movement of refractories
and still restrict the passage of gases, (6) flexible door diaphragm,
(7) knife-edge seals, and (8) a vented refractory door plug.
\
FIGURE III-19.
RECENT COKE-OVEN-DOOR DESIGN USED ON A
BATTERY UNDER CONSTRUCTION DURING 1974
III-20
-------
References for Chapter III
(1) Sheridan, E. T., and DeCarlo, J. A., "Coal Carbonization in
the United States, 1900-62", U. S. Bureau of Mines Information
Circular 5251, U. S. Department of the Interior, p. 4 (1965).
(2) Gluud, W., and Jacobson, D. L., "International Handbook of the
By-Product Coke Industry", The Chemical Catalog Company,
Inc., New York, New York, pp. 346-360 (1932).
(3) Brown, W. T., "By-Product Oven Operation Under Present
Conditions, Part U", Blast Furnace and Steel Plant, 3_0 (2),
pp. 219-223 (February, 1942).
(4) Muller, J. M., "Gary Coke-Oven Door Development Program",
Paper presented at the Joint Meeting of the Eastern and Western
States Blast Furnace and Coke Oven Association, 43 pp.
(October 25, 1974).
(5) Voelker, F. C., Koppers Company, Inc., communication to
John Varga, Jr., Battelie's Columbus Laboratories (September 9,
1974).
(6) Barnes, T. M., Hoffman, A. O., and Lownie, Jr., H. W.,
"Evaluation of Process Alternatives to Improve Control of Air
Pollution from Production of Coke", PB No. 189266, National
Air Pollution Control Administration, Department of Health,
Education, and Welfare, 166 pp. (January 31, 1970).
(7) Cellan-Jones, G., "A New Self-Sealing Coke-Oven Door", Coke
and Gas, _!£, pp. 221-222 (June, 1953).
(8) Cellan-Jones, G., "The Go Id Schmidt Door", Coke and Gas, Jjb,
pp. 307-310 (August, 1954).
(9) Komac, T. , "Vent Door Linings - Do They Work?", AIME Iron-
making Proceedings, 3_3, pp. 391-396 (1974).
(10) Cameron, A. M. , "Factors to be Considered When Designing a
Coke-Oven Door", AIME Proceedings, 3_3, pp. 397-399 (1974).
(11) Cameron, A. J., Communication to R. Paul, Batte lie's
Columbus Laboratories (November 1, 1974).
m-21
-------
CHAPTER IV
INVESTIGATION AND DISCUSSION OF THE CAUSES
OF THE EMISSIONS PROBLEM
This chapter deals with the tests and judgments made by Battelle
researchers in their efforts to "define the problem". Also included
are sections "Technical Specifications That New Sealing Systems
Should Meet" and "Judgments and Opinions On Existing End-Closure
Sealing Systems".
Introduction and General Statement of the Problem
Overall, "the" problem to be solved is that almost all coke-
oven doors release hydrocarbon emissions (smoke and volatiles) into
the atmosphere during the early part of the coking cycle and even
during the entire coking cycle. Equipment which may seal reasonably
well when new degrades because of the rough and severe service con-
ditions so that leakage becomes worse as the equipment and machin-
ery become older. The personnel in the coke-producing industry
(management and labor), EPA, and other organizations concerned
with worker health want this leakage problem eliminated in existing
and future coke batteries. In the present research, "emissions" are
judged primarily as the presence of visible emissions.
The complexity of the problem is a function of many factors that
are dealt with in this chapter. However, an overall conclusion in this
research program is that door leakage cannot be eliminated consis-
tently only by overhauling the present end-closure components or by
increasing the overall maintenance effort on present designs.
IV-1
-------
Technical personnel from coke-producing companies, who have
assignments to eliminate the emission problem, have often made
statements such as "We have just completed an overhaul of all of the
sealing and latching mechanisms, and emissions were lowered. How-
ever, as you can see the emission rate is already climbing".
This inability to hold an effective seal for a reasonable length of
time was confirmed at a plant where Battelle researchers were moni-
toring a test door. A rebuilt door was placed on a coke oven and the
new s-eal eliminated visible emissions for the period of the test
(1 week). However, upon, revisiting the test door 11 weeks later, it
was found that this door was leaking significantly without any evidence
of damage to the seal. There was, however, evidence of thermal
warpage of the seal in the area of leakage.
It was not a part of Battelle's scope of work on this program to
develop a quantitative method for evaluating the range of door-leakage
emissions in the coke-producing industry. However, one assigned
task was to develop a method for evaluating the leakage from future
test doors or from future doors incorporating new sealing concepts.
This work is reported later in Chapter VIII.
The overall degree of door leakage varies considerably from
one coke battery to another. There is a variability depending on bat-
tery age, type of seal design, and the past and present level of re-
pair and replacement. There is little agreement between observers
on quantity designations for door emissions. However, for purposes
of illustration of the problem of visible emissions, Figures IV-1 and
IV-2 are examples of visible leakage from a single door and heavy
visible emissions from doors in an entire battery.
Summary and Conclusions
The research, experiments, tests, and observations on which
this section is based are described later in the chapter. However,
to orient the reader early to the conclusions resulting from the data
that were collected, the results are summarized here before the
details and data are presented.
(1) By its very nature, a coke-oven battery is not a dimensionally
stable structure. It is reasoned, but not proven, that there
IV-2
-------
FIGURE IV-1. PHOTOGRAPH OF LEAKAGE OF EMISSIONS
FROM A SINGLE DOOR
(Photographed from bottom of doors doors are vertical. )
FIGURE IV-2. PHOTOGRAPH OF HEAVY EMISSION FROM
DOORS IN A COKE-OVEN BATTERY
IV-3
-------
is some thermal deflection of battery trim and end-closure com-
ponents (jamb/seal/door) on every coking cycle.
(2) There is evidence of thermal warpage of most metal components
at the ends of coke ovens. This warpage is caused by thermal
cycling, under constrained conditions, coupled with warpage due
to occasional temperature excursions*. Battelle has no direct
evidence - such as thermocouple records - illustrating temper-
ature excursions, but has collected evidence of internal oxida-
tion of cast iron jambs at the top of the ovens. This oxidation,
which leads to growth in the volume of cast iron, indicates that
the gray iron castinga in these locations have been repetitively
heated above 425 C (about 800 F). "Normal" jamb temperature
readings in this location are about 260 C (about 500 F).
(3) In measuring jambs, it is evident that there is severe thermal
warping, and there also is some evidence of thermal buckling of
the door-mounted seal edges of the S-shaped seal design. Pre-
sumably, such seals had been subjected to a temperature
excursion. *
The major problems, however, are the bowing of the jambs and
the inability of the various designs of seal-edge arrangements to ac-
commodate themselves to jamb distortion, particularly inward jamb
distortion. The bowing of jambs is the inward and outward displace-
ment of the jamb's seal-mating surface relative to a reference line
drawn from the top to the bottom of a jamb. The data included in this
chapter indicate a possible maximum, inward displacement of 17. 5 mm
(0. 7 inch) over a span of 2. 4 m (8 feet) and a maximum outward dis-
placement of 8 mm (0. 28 inch) over a span of 2. 4 m (8 feet). There
are no data available indicating the range of jamb bowing throughout
the coke-producing industry. It is believed, however, that almost all
jambs have been distorted to some degree.
(4) Because the vast majority of coke batteries in the United States
are of the Koppers and Wilputte design, the retrofitability con-
sideration thai is basic to this program is mainly concerned with
coke batteries (specifically the end-closure components) of these
two designs. Because these end-closures differ in design, each
is treated and discussed separately. The major concern in this
comparison is the ability of the designs to accommodate to
warped jambs. In this regard, it should be appreciated that
"Temperature excursions" are periods during which the temperature of the part exceeds the normal
and usual operating temperature by a large amount. A temperature excursion is a temporary condition
which may be of short or long duration.
IV-4
-------
neither design was built to accommodate the degree of warpage
that now exists on most operating coke ovens. Each design has
flaws and strong points, some of which are discussed later.
Relative to the fixed-edge seal (see page III-16), the S-shaped
seal is a more flexible design inasmuch as it can automatically accom-
modate itself to outward bowing of the jamb, and to continued outward
bowing of the jamb with time. Required inward deflection of the seal
(to enter an inward bow on the jamb), on the other hand, is supposed to
be accomplished by the combination action of (a) the backward deflec-
tion of the entire seal when mounting the door, and (b) the counteract-
ing inward force of the spring-loaded plungers pressing against the
back of the sealing edge. The required inward deflecting forces now
provided against the seal are insufficient to accommodate a deep valley
or a narrow-pitch, wide-amplitude inward displacement of a jamb.
A complicating factor in this judgment is that S-shaped seals in
operation are under high compressive strain because the thermal ex-
pansion of the hot seal is restrained by the numerous mounting bolts.
Whereas the S-seal in an ambient-temperature test stand will deflect
outward easily, it has been noted that many S-shaped seals in opera-
tion splay sideways between the two latching mechanisms. That is,
they elect to splay to the side plus deflection backwards. Specific
answers in this regard can be obtained only after completing a stress
analysis (strain-gauge) program.
The fixed-edge seal design consists of a sealing-edge mounted
inboard on a relatively inflexible beam or seal-edge holder. This
assembly is set into accommodation to the jamb shape by manual
tightening or loosening of bolts pressing against the back of the seal
assembly. The stiffness of the seal-edge holder makes it improbable
that this type of seal can accommodate extreme instances of inward
and outward bowing on jambs. In theory, the deflection bolts should
require only occasional adjustment. However, bolt adjustment is a
difficult and artful operation. In addition, the seal-deflection bolts
become loose with time.
(5) Visible emissions from coke-oven doors are almost always the
result of a gap between the seal edge and the deposits on the
jamb (or the jamb surface itself). Minor physical damage to
the seals (bends and dents) were observed; but, for the purposes
of this summary, visible emissions are caused by open spaces
IV-5
-------
between the seal edge and the jamb. Gaps are the result of many
conditions, some of which are as follows:
Excessive inward jamb displacement or bowing which the
seal cannot accommodate.
Removal of jamb deposits which formerly sealed a gap
(filler material removed).
Uneven or low sealing force on the door during the mounting
procedure. (This problem has been described in a recent
paper*.)
Warpage of the seal edge because of thermal buckling. The
formation of this gap is usually accompanied by a backward
deflection of the entire sealing edge above the upper latch.
Hour glassing** of the jamb to the point where part of the seal
edge does not strike the jamb surface.
Placement of a fixed-edge seal off the spot; i. e., either left
or right of the original adjustment spot, meaning that the seal
can be against a different configuration of the jamb.
Accumulation of hardened deposits in the door gas passage to
the point where the seal is physically backed off of the jamb.
Accumulation of the corners of the jamb of hardened deposits
that prevent the stop bolts from touching the jamb itself.
(6) Any gap between the seal and the jamb (or jamb plus deposits) will,
early in the cycle, release visible emissions of combined low-
boiling-point hydrocarbons and uncondensed steam. This mixture
will not "self-seal" gaps.
Later in the cycle when the steam volume has decreased and the
hydrocarbons and tars start to condense on the jamb, gaps will close
in stages, starting generally from the bottom of the door. The tem-
perature of jambs increases from the bottom to the top of the jamb
and, therefore, "sealing" (condensing) tars arrive at the bottom gaps
before they plug off the top gaps. In many instances observed, the top
gaps never seal.
(7) Cleaning of the jambs is important only if the deposits become
hard. It is believed that jamb deposits become hard only slowly
at "normal" jamb temperatures. It is known, however, that
Muller, J. M., "Gary Coke-Oven Door Development Program", Paper Presented at the Joint Meeting
of the Eastern and Western States Blast Furnace and Coke Oven Association, October 25, 1974. This
paper can be obtained by writing to the United States Steel Corporation, 600 Grant Street, Pittsburgh,
Pa. 15230.
"Hourglasslng is the distortion of a jamb such that the distance between the two sides is less in the
center of the jamb than at the ends of the jamb.
IV-6
-------
heating of coal-tar pitch to 340 C (650 F) will result in rapid exo-
thermic (heat releasing) reactions that emit gaseous hydrocarbons
from the tar and leave behind an adhering form of hard carbon. It
is indicated, but not proven, that thermal excursions can be the
cause of hard deposits on jambs.
(8) It was judged that new sealing systems should meet the following
technical specifications:
Must withstand 200 to 315 C (400 to 600 F) for prolonged
periods without deterioration.
Must withstand occasional short temperature excursions to
430 C (800 F), without being destroyed.
Should automatically seal gaps, i. e., no manual adjustment
required when placing any door on any oven on one side of a
battery.
Must have increased gap-closure capability up to four to
five times that of the flexible S-shape seal now in operation.
Must withstand corrosion and chemical attack of hot gases and
liquids.
Must be total-failure proof, i. e., not susceptible to the possi-
bility of complete failure during any cycle.
Must not present new and insolvable cleaning problems.
(9) Very little information is available on various basic measure-
ments and parameters within the coke-producing industry. Be-
cause of lack of data (and subsequent analysis), there is a wide
range of opinions within the industry. Within the funding of this
particular task, Battelle researchers could not examine any
facet of the basic problem in depth. However, the effort ex-
pended in this task approached the problem at its core.
Fundamentals of Demountable Industrial Seals
In many batch-type industrial applications, there is a need for
demountable seals on doors and covers that are closed and opened as
required. Coke-oven doors are one example of a large, demountable-
sealing application that has particular problems.
Sealing in any application is accomplished by achieving and
maintaining a continuous barrier (fluid or solid) against the transfer
of materials between the seal element and the mating coupling
IV-7
-------
surfaces.* With existing coke-door sealing designs, the problems are
both in achieving and in the maintaining of a continuous barrier.
Some coke-plant operators are of the opinion that surface tension
plays a role in the sealing of existing coke-oven doors. An example of
this type of seal is shown in the following sketch.
High-pressure
side
Solid ^Liquid
In this type of seal the two solids are wet by a liquid and the liquid will
remain in place (seals) if the differential pressure is low and the gap
is small, and if the surface tension of the liquid is high. The equation
for calculating the workable gap distance is as follows**:
D = 26/Pj, where D is the gap in centimeters, 6 is the
surface tension of the liquid sealing material in dyne/cm,
and PI is the differential pressure in dyne/cm^.
(1 atmosphere = 1 x 10& dyne/cm^.)
If the organic liquids that condense at the metal seals of standard
coke-oven doors have a surface tension of about 30 dyne/cm, then the
calculated gaps that these liquids would seal (by surface-tension ef-
fects) are as follows:
Differential Pressure Calculated Gap Sealed
Across the Gap by Organic Liquids
100 mm H2O (4 inches) 0. 006 cm (0. 002 inch)
700 mm H2Oi(l psi) 0.0009 cm (0.0003 inch)
Although Battelle researchers have been told by coke-plant op-
erators that gaps of 0.006 cm (0. 002 inch) should not leak, the re-
searchers doubt that surface-tension effects do much sealing at coke-
oven doors, at least during the initial period of high pressure at the
door seals. As described later in the test-results portion of this
chapter, the initial 100 mm I^O pressure developed in some 3.65-
meter (13-foot) coke ovens is by volume more than 50 percent
Daniels, C. M., "Aerospace Cryogenic Static Seals", Journal of the American Society for Lubrication
Engineers, April, 1973, p. 157.
**Roth, A., Vacuum Sealing Techniques. Pergamon Press, 1966.
IV-8
-------
uncondensed steam. The calculations for 700 mm t^O internal pres-
sure were included in the foregoing listing because it is being reported
that 6-meter ovens encounter this high pressure at the start of the
coking cycle.
Most coke-door seals depend on pressing a door-mounted edge
strip against the mating face on the oven-mounted jamb (door frame).
These are described as being "self-sealing" seals. This designation
is used because most areas of leakage on door seals eventually stop
leaking after some variable period of time. This may be due to a
surface-tension effect when high-boiling-temperature liquids condense
at the seals, but Battelle's field tests indicate that visible emissions
from doors stop after (a) the internal pressure at the door has dropped
and (b) condensed and partially solidified tars build up on the jamb in
front of the leak and block off the emission flow. *
The light pressure of the door-mounted edge strip against the
mating face on the oven-mounted jamb does not present a solid barrier
to the leakage of vapors and gases, Without metal deformation,
m^tal-to-metal contact represents only a contacting of the high points
on the two surfaces as illustrated below in the schematic cross section
through a metal-to-metal joint without a gasket.
To a gas or a vapor, the valleys.between the contacting high
points are an escape route. The rate of leakage between the two mat-
ing surfaces is partially a function of the surface finish of these sur-
faces. However, testing at Battelle, simulating coke-oven seal condi-
tions of temperature and mating pressures has shown that there is gas
leakage between clean, undeformed metal contact surfaces at all
levels of practical industrial finishes.
To develop a true, demountable, static seal it is necessary to
fill the roughness of the mating surfaces. This filling operation can
be accomplished with (a) compressible (one-use) gaskets; (b) resil-
ient, compressible materials (rubber and elastomer materials); and
(c) compression of relatively plastic coatings on either one or both of
*See the Field-Test Results later in this chapter.
IV-9
-------
the metal mating surfaces. A sketch showing the "filler" action of
gaskets and compressible materials is shown below in the schematic
cross section through a metal-to-metal joint with filler material.
Gasket or Filler Material
Battelle researchers have seen new metallic sealing strips con-
tacting old jamb surfaces without the release of any visible emissions
on any portion of the coking cycle. In this instance, the deformable or
filler material between the mating metallic surfaces was the coating
of hot, soft tars remaining on the jamb surface from previous coking
cycles. This filler effect is in operation even when the jamb has been
carefully scraped to what would appear to be a clean metal surface.
The filler in this instance is a thin residual film of tar on the jamb.
Instances have also been observed where a thick layer of material on
the jamb was able to seal a relatively large gap between the sealing
edge and the jamb when the seal touched and/or compressed the jamb
deposit.
In summary, -effective demountable seals need a barrier be-
tween the two mating surfaces to prevent leakage. In coke-oven doors
that are leak-free from the start of the cycle, the barrier is residual
tars on the jamb. If, however, there is between the sealing edge and
the jamb a gap that is not filled with tar, the gap will leak heavily
until later in the cycle when heavy tarry materials condense on the
jamb and plug or dam the leak.
It is apparent that with improvements in the metal-to-metal
concepts for sealing coke-oven doors, there will still be a need for a
compressible or pliable material on the door jamb. If steps are
taken to eliminate the deposition of tars on the jambs, then steps will
have to be taken to develop a suitable substitute material.
Laboratory-Test Results
Examination of a Discarded Jamb
As indicated by measurements taken in the field-test program,
(page IV-34), cast jambs made of gray iron bow, hourglass, and
IV-10
-------
probably twist. Bowing is the inward and outward distortion of the
jamb from a reference line between the top and bottom of the jamb.
Hourglassing is the tendency for the jamb opening to become smaller
at the middle of the jamb as referenced to the ends.
During the course of this program, a coke-producing company
shipped a discarded coke-side jamb to Battelle for examination.
The first saw cut into the discarded jamb was at the coal line on the
side piece of the jamb sprang inward a distance of about 38 mm
(about 1-1/2 inches). This indicated that the as-received jamb
was under considerable residual stress (in the direction of hour-
glassing). If it is assumed that this particular jamb was stress
relieved following manufacture, or was stress relieved in service,
then the residual stresses noted were the result of nonuniform
plastic deformation during use.
Complete cross sections of the jamb were removed for metal-
log raphic examination. These sections were from the center of the
top member (top of jamb), and from the top, middle, and bottom of
one side member.
Brinell hardness numbers (BHN) were determined for various
positions on the sections. The hardness values were in the range of
155 to 170 BHN for all areas in the side sections except in the seal-
mating surface location on the top section. This area had apparently
been overheated in service with a resultant lowering of the hardness
from 160 to 120 BHN. The entire section from the top member was
softened down to a level of 110 to 134 BHN. These hardness readings
lead to the conclusion that the softer areas were annealed while in
service. At temperatures of 600 C (1100 F) or lower, such softening
as occurred requires a total of weeks of exposure to such temperatures.
Samples for metallographic specimens taken from the four sec-
tions were polished and etched for examination under a microscope at
magnifications ranging from 100 to 500 diameters. From each sec-
tion, samples were taken from the inside edge of the jamb (the portion
directly facing the door plug and parallel with the oven bricks and
called the "hot face") and from the outboard end of the jamb wing. All
of the specimens taken from the sections at the bottom and the middle
of the jamb were typical gray iron (flake graphite in a matrix of pearl-
ite). This structure was probably quite similar to the microstructure
of the entire jamb before it was placed into service.
IV-11
-------
The microstructure of the hot-face location from the top of the
side member, as shown in Figure IV-3, showed that the metal in that
area had been subjected to cyclic heating to temperatures that could
cause "growth". Structural evidence of this condition is the network of
oxide particles around the graphite flakes, partial replacement of the
graphite with iron oxide, and the decarburization of the matrix as indi-
cated by the lower amount of pearlite.
The microstructure of the hot face of the top cross piece shows
that the metal in that area had experienced severe growth of a degree
that can occur in unalloyed gray iron only at high temperatures over an
extended period of time. Figure IV-4 shows that the graphite had been
almost entirely replaced with iron oxide and that considerable internal
oxidation had occurred throughout the matrix. Also, there was vir-
tually no carbide phase (pearlite) remaining in the matrix. This struc-
tural condition extended through most of the section.
This metallographic study established that the upper part of the
door jamb, and particularly the portion of the jamb near and where
the door seals, was severely affected by the mechanism that is com-
monly referred to as "growth".
Growth in cast iron is a nonreversible increase in volume that
results from microstructural changes such as occur from the cyclic
heating of the cast iron to elevated temperatures. There is a signifi-
cant increase in growth rate if the temperature extends above the
range in which the transformation reactions occur. This range is
700 C and higher (about 1300 F and higher). The start of internal oxi-
dation leading to growth is at a temperature of about 425 C (about
800 F). Growth is further accelerated by strong oxidizing conditions
or by the presence of sulfur-bearing gas in the environment. When
this nonreversible increase in volume occurs, it is inevitable that the
parts will develop internal stresses, warp, and ultimately become
unfit for service.
Three conclusions were drawn from the work on this discarded
jamb:
(1) The top portion of this jamb (and possibly many more
jambs in service) was heated numerous times above
425 C (800 F) and may have been cycled to higher tem-
peratures for appreciable periods of time.
IV-12
-------
. ».
JU^*^ *5'"aP?r
» ----
500X Picral Etch
FIGURE IV-3. MICROSTRU.CTURE OF A METAL SAMPLE TAKEN
FROM NEAR THE TOP OF THE SIDE MEMBER OF
A DISCARDED JAMB, CLOSE TO THE SEAL
CONTACT AREA
500X
Picral Etch
FIGURE IV-4. MICROSTRUCTURE OF A METAL SAMPLE TAKEN
FROM THE CENTER OF THE TOP MEMBER OF
THE DISCARDED DOOR JAMB
IV-13
-------
(2) A metal other than unalloyed gray iron should be
considered the coke-oven jambs.
(3) Temperature excursions can and do occur at least to
some coke-oven jambs and door seals.
Temperature data summarized in a following portion of this chap-
ter show that the highest jamb temperatures measured by Battelle re-
searchers were about 340 C (650 F). Although these temperatures
were taken on the end of the seal edges, subsequent sampling with a
contact thermocouple did not show any top-jamb temperatures over
260 C (500 F).
It was concluded that all of the temperatures taken by Battelle
were outside of the periods when high-temperature excursions occur on
coke-oven jambs and door seals. It is not known for sure when these
temperature excursions occur, but they probably occur during periods
when doors are kept on empty ovens for prolonged periods of time.
These periods can occur when the charging machine malfunctions and
the ovens are pushed "ahead"; i. e., ovens are pushed empty ahead of
the charging machine in a plan that the charging machine will (and
does) eventually catch up, i. e., no production tonnage is lost. Another
possible factor is prolonged decarbonization of ovens to remove roof
carbon, either with or without coke in the oven.
It is believed that temperature excursions play a major role in
warping coke-oven jambs and also, in some instances, coke-oven
seal edges. With a door on an empty oven, the entire length of the gas-
passage area* and the general seal-edge area and jamb are subjected
to direct and reflected radiant heat over the entire height of the door.
The level of thermal stresses developed during these temperature ex-
cursions can be high and damaging. A disturbing consideration is that
jambs and edge seals may be overstressed during the period a coke-
oven battery is first put on line. That is, the initial heating of new
equipment (with no coal in the ovens) may cause the initial warpage.
Once distortion has started, it becomes easier to continue the
distortion.
The gas-passage area is a channel on the door side of a seal strip.
IV-14
-------
Laboratory Metal-to-Metal Sealing Results
At the start of this program there was some question from the
supervisors of host coke plants and from the Battelle researchers as to
"What is the mechanism of sealing at coke-oven doors"? The decision
was made to attempt to answer this question in both field trips and in
laboratory experiments.
For the laboratory tests, a miniature coke-oven seal {with a
full-size seal edge) was made using a circular seal pressed against a
steel block as shown in Figures IV-5 and IV-6.
It was arranged that this test rig could be heated to the reported
coke-oven jamb temperature of 260 C (500 F) and could be given the
same calculated average mating pressure as coke-oven seals (11 to 18
kg/cm or 65 to 100 pounds per linear inch). All tests were completed
with an internal gas pressure of about 175 mm of H^O (about 7 inches
of water).
The first experiments were to determine if leaks could be mini-
mized or eliminated by improving the surface texture of both the seal-
ing edge and the mating block. As expected, over a range of 6450
nanometers down to 1600 nanometers (250 to 62 microinches) the leak
rate decreased; but even with the relatively smooth finish, there still
was considerable gas leakage. This confirmed the sealing theory out-
lined previously.
The next series of experiments was to test various "fillers" for
possible use as sealing materials. The results of these tests are
shown in Table IV-1.
Fluid coke-plant tar and many other materials can be used as
sealing materials in the laboratory test rig.
Comparing these laboratory results with those developed in
field tests, it was concluded that:
(1) If the seal edge of existing coke-plant seals compresses
or even "touches" tar or tarry materials on the jamb,
a complete seal is made because the pressure differen-
tial across the seal edge is not high enough to force out
the sealing material.
IV-15
-------
Part simulating the coke-door
sealing edge.
(Edge dimensions are the same
as on existing doors.)
Part simulating the jamb. Opening
is for introducing gas pressure.
FIGURE IV-5.
BASIC COMPONENTS USED IN LABORATORY
METAL-TO-METAL SEALING TESTS
(This equipment was also used to test various potential
filler materials. )
FIGURE IV-6. THE METAL-TO-
METAL TESTING
EQUIPMENT IN
OPERATION
(Test rig is under simulated coke-
oven-seal sealing pressure and
was heated to coke-oven jamb
temperatures by an electric heat-
ing furnace.)
IV-16
-------
TABLE IV-1. SUMMARY OF METAL-TO-METAL SEAL TESTS
WITH AND WITHOUT FILLERS
Equipment temperature 260 C (500 F), Mating Pressure 18 kg/cm,
Internal Pressure 175 mm of H_0.
Material
Comments
Fine Finish
Rough Finish
Steel to Steel
Coal-Tar Filler, Commercial
Coke-Plant Tar
Roofing Pitch
Corn Syrup (thick)
Molasses
Malasses and Lime
Paper Mache (wet)
Silver Seal Compound
(a)
Asbestos Sheet, 1/8-Inch
(b)
Leaks slowly
No leaks
No leaks
Hardens, then leaks
Boils, no leaks
Boils, no leaks
Hardens, no leaks
Leaks
No leaks
Leaks slowly
leaks at high rate
No leaks
No leaks
Bakes hard, leaks
slowly
Boils, no leaks
Boils, no leaks
No leaks
Leaks
No leaks
Leaks
(a) Silver Seal Compound is commercial high-temperature sealing
material. Laboratory tests at 260 C (500 F), in air, indicate
that this compound hardens in about 4 hours.
(b) Increased mating pressure eliminates leaks with this material.
IV-17
-------
(2) If there is any gap (no filler material) on any portion
of the seal edge/jamb interface, there is leakage,
with the rate depending on the size of the gap and the
internal pressure.
(3) Gaps permitting emissions are either plugged or
dammed off only after condensation of adhering viscous
liquids or semisolids begins at the seal/jamb interface.
(4) This test rig, or a modified test rig, could be used to
evaluate seal materials for new concepts, existing
self-sealing doors, or the older luted-door designs.
(5) The use of filler materials should be considered as a
concept for totally sealing doors, i. e., for complete
elimination of emissions. The range of possibilities
includes sealants usable for only one cycle, sealants
with a life of many cycles (with and without cooling),
and others.
Field-Test Results
Early in this project, it was decided to complete week-long visits
at selected plants to obtain project data. At a meeting with represen-
tatives of the sponsorship, it was decided that Battelle should concen-
trate its measuring and evaluation efforts on the size and type of the
vast majority of coke plants in the country; i. e., the standard 3. 4-
meter to 4.0-meter oven (11 feet, 2 inches, to 13 feet, 2 inches) of
the Koppers and Wilputte design. Testing was completed at a battery
in Chicago and a battery in Youngstown, Ohio. The results of these
testing programs are grouped and reported in terms of specific areas
of interest.
Observations and Comments on
Visible Emissions From Doors
Observations Made During the Test Program in Chicago. In
Chicago, the investigators were working with a newly rebuilt coke-
side door equipped with a fixed-edge seal design.
IV-18
-------
Individual fixed-edge door seals are "matched" to a. particular
jamb configuration by adjusting the diaphragm-deflection bolts the
first time the door is latched to an oven. In theory, after this initial
adjustment the seal edge continues to assume the shape of the jamb
when pressed against the same jamb and continues to seal cycle after
cycle; i.e., the subsequent adjustments are supposed to be minor.
When the test door was first latched to an oven, the workers
spent 37 minutes tightening and adjusting the diaphragm-deflection
bolts. During this period, the empty oven was under coke-oven-
gas back pressure, and the tightening program was aimed at stop-
ping burning leakage.
On the first cycle of the test door (i.e. , after the coal was in
the oven), a small amount of visible emissions became apparent. An
additional 10 minutes spent in tightening selected diaphragm-deflection
bolts resulted in lowering the visible emissions to a trace level. In
about 30 minutes after the last adjustment, all visible emissions had
stopped.
On the second cycle, the initial level of visible emissions was
about two or three times the amount of the initial level on the first
cycle. The major emission site was in the vicinity of the top latch on
the right side. Subsequent jamb-distortion measurements indicated
that this was the location where the jamb had a depression (inward
bow).
On the third cycle, the initial level of visible emissions was
minor, i.e., about the same as on the first cycle after the second
adjustment of the door. Those visible emissions that were released
were again at the right side at the upper latch.
As the pressure built up in the test oven on the fourth cycle, a
rather heavy leak developed on the left side. This is the opposite side
from the leakage in the first three cycles. The leakage on the left
side was fairly heavy down by the bottom latch with a "horizontal
streamer" at the top of the door on the left side as shown in Fig-
ure IV-7. The term "heavy" is used with reference to the first cycles
on this test door.
IV-19
-------
FIGURE IV-7. "STREAMING" OF EMISSIONS
FROM A TEST DOOR
Photographed from bottom of
doors.
Measurements of the position of the door on the oven indicated
that in this fourth cycle the door had been positioned to the extreme
left limit of the door guide. With reference to the previous cycles,
the door was found to be positioned further left on the jamb by about
3 mm (1/8 inch). It was reasoned that because of the door-position
shift, the seal edge was now against a slightly different jamb "shape"
than it had been adjusted for. In 3 hours, the visible emission from
the door decreased to a wisp. The internal pressure in the oven at the
bottom of the door after 3 hours was only about 5 mm of HoO.
It was noted that each cycle of the oven resulted in a slight gen-
eral increase in the thickness of the semisolid deposits on the jamb.
However, in an emission location, the deposits on the jamb formed in
a single cycle were visibly thicker. The thickness of deposits on the
door and seal area remained almost constant or increased in thick-
ness only slightly. It was reasoned that visible emissions from loca-
tions where there is a gap between the seal and the jamb are sealed
by the more rapid buildup of deposits on the jamb in front of the gap.
The first three cycles on this test door showed visible emissions from
the right side by the upper latch, the location of an obvious gap between
the seal and the jamb. The deposits on the jamb in this location were
relatively thick by the start of the fourth cycle and, with the shift in
the door position to the left, the seal edge pressed into the thick
IV-20
-------
deposit on the jamb. This entry of the seal edge into the thick deposit
effectively stopped the emissions from this location on the jamb.
Examination of the jamb after the fourth cycle showed the imprint of
the seal edge in jamb deposits at this location.
A few general observations made at this plant were found to be
true also for other coke plants. It was noted that the rate of buildup
of deposits in the gas passage on the door was low. In plants where
the deposits in this location are both thick and in some instances hard,
it seems reasonable to conclude that no cleaning of the passage had
been done for months. There is, however, one reservation to this
conclusion. It was noted that in almost all cycles some coal falls
between the door liner (door plug) and the oven wall and enters the
gas passage on the door. In such cases, sticky clumps of coal/tar
mixtures can accumulate rapidly. In this regard, it should be noted
that no coke plants operating with 3.4-meter to 4. 0-meter ovens are
equipped with mechanical cleaning equipment; i.e. , all cleaning is
done manually. Manual cleaning is both awkward and difficult and is,
therefore, usually not done thoroughly, if at all.
Another observation of interest is that on different locations on a
coke door, the "self-sealing" of gaps (by deposition of tars on the
jamb) proceeds at different rates. Visible emissions from the bottom
half of any door "self-seal" rather rapidly (up to about 60 minutes)
even though this is in the higher pressure sector of the door. On the
other hand, visible emissions from the top portion of the doors often
take hours to "self-seal" and can continue visible emissions for the
entire cycle. This is true although the pressure inside the oven at
these delayed-sealing locations is controllable at a positive pressure
level as low as 6 mm of t^O (about 1/4 inch). Temperature measure-
ments indicate that there is an increase in jamb/seal edge temperature
from the bottom to the top of the door. It was concluded that "self-
sealing" is a function of deposition rate of tars and that where the
equipment is hotter at and near the top of doors, the deposition rate
and the related "self-sealing" rate are retarded.
Observations Made During the Test Program in Youngstown, Ohio.
In Youngstown, the investigators were working with a rebuilt pusher-
side door equipped with an S-shape seal design.
Door seals of the S-shape type (see bottom-right sketch, Fig-
ure III-17) are intended to adjust automatically to any oven on a battery,
providing the warpage of the jamb is not too pronounced. This adjust-
ment or accommodation is obtained by spring-loaded plungers pressing
IV-21
-------
against the back of the S-bend. The theoretical adaptability of being
able to place a door with a flexible seal on any oven on one side of a
battery is a desirable feature and would be a desirable feature of any
door developed from a new concept. This approach permits routine
repairs and maintenance of the door at the end of the battery without
adding to the emissions problem. It is suggested here that a leaking
door be replaced on the next cycle with a replacement door.
The research team observed the first four cycles on this rebuilt
door and noted that, in all instances, the entire seal edge around the
door was completely free of visible emissions from the moment the
door was placed on the oven. Throughout the cycles there were some
slight visible emissions from the chuck door.
The buildup of tars in the gas passage over four cycles was
small. In fact, the sheen of the stainless steel surface of the gas
passage was still visible. The buildup of deposits on the jamb was
only a trace amount. Again, however, it was noted that some coal
entered the gas passage and formed sticky clumps that partially
blocked the gas passage. At this plant, the doors are manually cleaned
almost every cycle.
When this particular test door was at Battelle for the mounting
of thermocouples, an 0. 8-mm (1/32-inch) hole was drilled through
the upstanding seal edge, near the bottom of the door, to serve as a
"known-size leak". On the fourth cycle of this test door, this hole
was opened prior to the charging of coal into the oven. The 0. 8-mm
hole streamed visible emissions for 35 minutes, with the volume of
emissions being an approximate function of the internal pressure in
the gas passage. Final "sealing" occurred rapidly, just as if the hole
had been suddenly plugged with high-viscosity tar. Unfortunately, with
the door on the oven, it was not possible to reach the hole and to reopen
it to determine the rate of sealing later in the cycle. This test lends
credibility to the hypothesis that the first volatiles to reach any sealing
strip are not capable of sealing small gaps. Stated another way, the
high-pressure period at the start of a coking cycle is also the period
when the volatiles released in the oven do not seal leaks.
At this battery, it was noted that on every cycle a nearby door
had very heavy visible emissions coming from the upper-right latch
area and all across the top of the door. Because the location of these
emissions was at the hottest part of the door, there was only a minor
decrease in emission over an observed 3-hour period. It is indicated,
as previously noted, that "self-sealing" does not occur very rapidly at
the top of doors.
IV-22
-------
After a period of 11 weeks, the research team returned to
Youngstown to examine the emission-control performance of the test
door. The door now had a heavy visible emission from a localized
area at the left upper latch. Examination of this location at the end
of the cycle indicated that the seal edge had heat buckled and opened
a gap. Sketches of the side and front view of the buckled seal strip
are shown in Figure IV-8. From a front view, the buckled strip had
a slight "S" shape. From the side view, there was a depression (gap)
of about 2.4 mm (3/32 inch) over a span of 150 mm (6 inches).
15 cm (6 )
Seal
strip
y-2.4 mm gap
/ (3/32")
Side View
Front View
FIGURE IV-8.
SKETCH OF THE SIDE AND FRONT VIEW OF A
SEAL-EDGE BUCKLE AT THE UPPER LATCH
AREA OF A PUSHER-SIDE DOOR
(The wave in the buckled section is similar to the
wave in a leveler bar discussed in a later section
of this chapter. )
An examination was then made of about 12 doors at other locations
on this battery. For about half the doors examined, the heat war page
effect shown in Figure IV-8 was present on both sides of the door at
about the upper latches. Further examination indicated that in all
instances where the seals were buckled at the sides, the S-shaped
IV-2 3
-------
sealing strip above the upper latches was permanently deformed out-
ward from the jamb. This was true regardless of the condition of the
springs pressing the plungers against the back of the S-shaped sealing
arrangement. This was also true of the test door which had only been
in operation for about 3 months. A side view of the deflection of the
S-shaped sealing strip on the test door is shown in Figure IV-9.
This measurement was taken with the door off the oven. The points
of buckling may also be locations of mechanical-stress concentrations
superimposed on existing thermal stresses.
J~
Seal
strip
1.6 mm deflection
(1/16")
/i
I Approximate latch
location
FIGURE IV-9.
A SKETCH SHOWING THE PERMANENT OUTWARD
(AWAY FROM THE OVEN) DISTORTION OF THE
SEALING STRIP ABOVE THE UPPER LATCH
The pattern of outward deflection of the upper seal coupled with
the buckling of the seal strip near the upper latches is not believed to
be the result of any mechanical damage to the door. Instead, it is
believed to be another example of the heat warpage of door /seal/ jamb
components under thermal cycling coupled with the restraint of move-
ment of these parts. Heat warpage is discussed in a later section of
this chapter.
IV-2 4
-------
In the return visit to this coke plant, it was observed that this
company was now using 3-mm-thick (1/8 inch) asbestos sheet as a
gasket for heat-warped chuck doors. In many instances, this use of
a complete square of asbestos paper between the chuck door and the
casting on the coke-oven door made these chuck doors free of visible
emissions and free of deposits on the chuck door and on the mating
face of the casting mounted in the coke-oven door. When the gasket
did not result in a tight seal, the sealing edge on the chuck-door seal
edge was so distorted that the resulting gap was wider than the thick-
ness of the asbestos paper. This was confirmed by placing the gasket-
ing material on the casting and closing and then opening the chuck
door. When the sealing edge did not mark the paper in all portions
of the gasket, the chuck doors leaked.
This chuck-door gasketing procedure was rated as an approach
that, under further development, could become a standard procedure.
Of concern, however, is the fact that the procedure increases the
temperature of the chuck-door-casting housing. It is probable that
the cost of the gasketing material can be appreciably lowered below
the cost of the asbestos paper used in these experiments.
Temperature Measurements of
Jambs and Sealing Edges
Early in this project it was decided that continuous measurements
of temperatures of jambs, doors, and seals could contribute to better
understanding the warpage problems of coke-oven end closures. To
be able to install thermocouples on sealing strips and in the door
liners, it was decided to work with "new" doors; i. e. , doors just
recently overhauled with the installation of new seal equipment and a
new door liner (door plug).
Battelle researchers expressed a desire to do temperature re-
cording at an experimental water-cooled, coke-side jamb being tested
by a steel company in the Chicago area. Arrangements were made for
tests at this location. For additional testing, several selections of
battery locations were offered, and it was decided to complete these
tests at a battery located in Youngstown, Ohio.
It should be pointed out in this introduction to the temperature -
measuring tests that all coke-plant supervisors contacted had strong
feelings against giving researchers permission to drill holes into
IV-2 5
-------
operating jambs for the permanent installation of thermocouples.
This was an unfortunate constraint because a major interest of this
program was to determine the temperatures of the interior surfaces
of jambs; i. e. , the surfaces exposed to radiant heat. In this regard,
Battelle researchers have suggested to various coke-plant super-
intendents that the installation of thermocouples inside of new jambs
(prior to installation) would give them a record of the temperature
variations that are caused by variations in operating practices.
With the limitation that thermocouples could only be installed
on doors, jamb temperatures could only be approximated by (a) in-
stalling spring-loaded thermocouples that press against jambs, and
(b) installing thermocouples fastened on the seal-edge sides close to
the contact point between the seal and the jamb. This limited the
temperature measurement to the sealing-strip location on the jamb
which is, in most instances, about 4 cm (1.6 inches) away from the
location which is expected to be the hottest portion of the jamb. This
limitation should be kept in mind when examining the following data.
Summary of the Temperature Data Obtained at a Chicago Battery.
Prior to the arrival of the Battelle research team, this coke plant had
3 months previously installed an experimental coke-side jamb having a
single internal water passage going up the right side of the jamb, across
the top, down the left side, and across and out the bottom of the jamb.
The location of the water passage in the jamb is shown in cross section
in Figure IV-10.
This jamb, but without water cooling, was in position on an oven
for about 2 months, after which it was noted by the coke-plant operators
that the jamb was already hourglassing (warping by necking inward).
Water was then connected to the cooling passage for about 1 month
prior to putting on the test door. Unfortunately, on this first trial of
a water-cooled jamb, no records were kept on the dimensional changes
of the jamb both without and with water cooling.
Seven sheathed thermocouples were fastened to the sealing edge
in the locations shown in Figure IV-11. Also on this figure are the
temperature data obtained from these seal-edge locations both with
and without water cooling.
IV-2 6
-------
Seal mating surface
Water passage
FIGURE IV-10. SKETCH SHOWING THE LOCATION OF AN
EXPERIMENTAL WATER-COOLING PAS-
SAGE IN THE CROSS SECTION OF A JAMB
1
1
p<
61 cm(24>
'<
91 cm (36")'
:<
61 cm (24">
J) <£
2) (S
^
._
@
5)
43 cm(l7")
48 cm (19")
91 cm (36")
|
V
o
Thermocouple
Position
1
6
2
7
3
9
4
5
Average Temperature, C(F)
Water On
196(390)
188(370)
149(300)
138(280)
149(300)
101(215)
149(300)
149(300)
No Water
260(500)
263(505)
260(500)
238(460)
232(450)
227(440)
243(470)
204(400)
FIGURE IV-11. SKETCH SHOWING THE LOCATION OF THERMO-
COUPLES ATTACHED TO THE SEAL EDGES OF A
DOOR AT THE CHICAGO TEST SITE
Also shown are the average temperatures by loca-
tion depending on whether jamb water cooling was
or was not being used.
IV-2 7
-------
At the different points of measurement, the use of cooling water
lowered the temperature between 55 and 125 C (100 to 225 F) depending
on the temperature of the water; i. e., distance from the water inlet.
The steady-state thermal loading of the cooling water was about 230 kcal
per minute (900 Btu/min).
As was expected, the highest temperatures on the seal edges were
above the coal line on the test door.
An interesting feature of the water-cooled jamb is that close to
the water passage (at the inner corner of the jamb - Point A in
Figure IV-10), the jamb-surface temperature was low enough to
permit hand touching. This observation gave credibility to the potential
feasibility of sealing concepts using water-cooling to prolong the life of
resilient or inflatable seals. A drawback that must be considered, how-
ever, is that all of the cooling water would have to be treated, recycled
water; i. e., water that would not deposit an insulating film inside the
cooling passages. Also, even a brief interruption in the supply of cooling
water could ruin expensive special seals.
A thermocouple positioned in the door liner 25 mm (1 inch) from
the hot face and 350 mm (14 inches) above the bottom sealing strip re-
sponded as expected. With the entry of coal into the oven there was a
drop in the hot-face temperature to about 300 C (600 F) in about 5 hours.
Following this decline, there was a steady increase in temperature to a
peak of about 480 C (900 F) at the end of the cycle.
An observation of interest was the recording of a thermocouple
positioned in the gas-pas sage space between the retaining steel on the
door liner and the oven walls; i. e., about 100 mm (4 inches) inside the
door diaphragm and 0. 9 m (3 feet) up from the bottom of the oven. This
thermocouple peaked at about 480 C (900 F) about 2 hours after coal was
dropped into the oven. Following this peaking, there was a steady de-
cline in temperature even though the thermocouple was "seeing" an
increasing temperature of coal and coke. It was concluded that the
thermocouple was, at least partially, measuring the first temperature
of the hot volatiles and steam entering the gas passage. One hypothesis
has it that the steam and volatiles are finding a horizontal passageway
along the bottom corners of the oven.
Summary of Data Obtained at a Youngstown Battery. In this in
instance, it was possible to send a rebuilt pusher-side door (new Mas-
rock liner and new S-shaped seal with spring loading) to BCL for the
installation of thermocouples and pressure taps.
IV-28
-------
The goals of this temperature-measuring program were to:
Determine the inside and outside seal-edge temperatures
with particular emphasis on the chuck-door area
Obtain jamb-surface temperature close to the sealing-
edge thermocouples
Determine the range in seal-edge temperatures over the
entire door for several cycles
Obtain data on any thermal excursions that occur during
the decarbonization procedures
Use sheathed thermocouples to probe the temperature in
the gas passage.
A total of 18 thermocouples was installed inside and outside the
sealing edge along both sides of the door. Two were spring-loaded
contact thermocouples positioned to press against the jamb. Figure
Figure IV-12 is a photograph of the thermocouple and connecting-
board installation. An example of the contact thermocouples and the
outside seal-edge thermocouples is shown in Figure IV-13. The
locations of the seal-edge thermocouples in terms of distances from
the top and bottom sealing edge are sketched in Figure IV-14.
Figure IV-15 is another view of the installation of the fixed strip-
edge thermocouples.
Additional thermocouples were installed at the hot face of the
Masrock door plug. These were located 86, 254, and 376 cm (34,
100, and 148 inches) down from the top of the plug.
Pipe nipples were installed leading into the gas passage to serve
as internal-pressure taps and also as passageways for temperature
probing of the gas passage, including the space between the liner and
the oven walls. These taps were installed at the top, middle, and
bottom of the door, on both sides.
Summary of Temperature Data. The highest temperatures
recorded were always at the top and top sides of the door at the end of
the coking cycle. The final peak temperature on the test-door seal
IV-2 9
-------
FIGURE IV-12.
GENERAL THERMOCOUPLE INSTALLATION
ON A TEST DOOR
(During testing all thermocouple lead wires were disconnected from
the door during the period when a door was off the oven. )
FIGURE IV-13.
INSTALLATION OF A SPRING-LOADED AND A FIXED
THERMOCOUPLE MOUNTED OUTBOARD OF A
S-SHAPE SEAL DESIGN
IV-30
-------
Thermocouples on inside of seal edge
Contact thermocouple
-Contact thermocouple
*- Thermocouples on outside of seal edge
FIGURE IV-14. LOCATION OF THERMOCOUPLES INSIDE AND
OUTSIDE THE SEAL EDGE AND THE LOCATION
OF SPRING-LOADED CONTACT THERMOCOUPLES
FIGURE IV-15. INSTALLATION OF SHEATHED THERMOCOUPLES
ATTACHED TO THE OUTSIDE OF A SEALING STRIP
IV-31
-------
edges varied from cycle to cycle depending, it is believed, on the
length of the decarbonization period. At this plant, the decarboniza-
tion of the oven roofs is accomplished by dampering off the valves
leading to the collecting mains and opening the two standpipe covers
and one or more of the charging ports. This is done prior to the
pushing of the oven.
The final temperatures for four cycles of two seal-edge
thermocouples located on the inside and outside of the edge and 14 cm
(5-1/2 inches) down from the top of the sealing-edge cross piece were
as follows:
Temperature at End of Cycle
Cycle Inside Thermocouple Outside Thermocouple
Number C F C F
1 274 525 263 505
2 346 655 335 635
3 235 455 229 445
4 282 540 271 520
These data show variations between supposedly identical cycles.
The second cycle recorded had a particularly high final temperature.
Various temperatures on the door over the length of this cycle are
graphed in Figure IV-16.
It is not known whether or not the general upsweep in tempera-
ture at the end of this particular cycle was the result of a prolonged
decarbonizing period. However, other cycles recorded on this same
oven showed lower finishing temperatures with the next cycle remain-
ing steady at 235 C (455 F) for the top thermocouples for the final
6 hours of the cycle.
The contact thermocouples consistently gave lower readings than
the nearby seal-edge thermocouples and were judged to be inaccurate,
as contact thermocouples often are.
It is believed that the highest temperature that an oven door/
door seal/jamb can reach occurs during production-upset periods when
a door is latched on an empty oven for a long time. During this period
when a door is on an empty oven, portions of the sealing equipment
and jamb are exposed to direct, high-temperature heat from the oven.
From a research viewpoint, it would have been very desirable to have
IV-32
-------
15-
12
n
a)
a
w a
I "8
E n
10-
9-
6
7J
300-
-Hot-face temperature of the door
plug 0.86 m (34 in.) from the top
0
400-
300-
200
Inside-seal thermocouple
14 cm (5-1/2 in.) from the
top. This is the reading
shown in the foregoing listing.
-Outside-seal thermocouple
1.6 m (63 in.) from top.
This is not shown in the
foregoing listing.
0 4 8 12 16 20
Elapsed Time From Cool Drop, hours
FIGURE IV-16. PLOT OF TEMPERATURES OF SELECTED
THERMOCOUPLES DURING THE SECOND
CYCLE RECORDED
*A11 thermocouples were disconnected about
15 minutes prior to removing the door from
the oven.
obtained such temperature data, but the dictates of production and
absence of a breakdown during the test period made this impossible.
Sheathed thermocouples were inserted through the middle pressure
taps into the gas passage of the door and also further into the gap be-
tween the door liner and the oven walls. The temperature data were
as follows:
IV-3 3
-------
Thermocouple Position
Time From Opposite Distance Past the Interior Jamb Face
Start of Interior 7 cm 13 cm 18 cm
Cycle Jamb Face (3 in.) (5 in.) (7 in.)
15 minutes -- 515C (960F) 549C (1020F) 565C (1050F)
25 minutes 565C (1050F)
40 minutes 354C (670F) 426C (800F)
These measurements confirmed other data that the temperature in the
gas passage is highest at the beginning of the cycle and falls with time.
It is possible that this high heat input into the gas passage early in the
cycle can account for the relatively high temperatures of the seal
strips at the beginning of each cycle. (See Figure IV-16).
Unfortunately, it was not possible to install thermocouples in any
of the jamb surfaces themselves to determine (a) peak temperatures
and (b) any excursions in temperature. If this is not done by any one
of the coke-producing companies, it is recommended that it should be
done on new jambs in any follow-on project.
Dimensional (Warpage) Measurements
At both test locations, measurements were made of the vertical
profiles of the jambs (i. e. , the inward and outward bowing or distor-
tion of the jamb) and the profile of the seal edge on the door.
The measuring equipment was a Battelle-designed, 3. 8-meter
(12 foot, 6 inch) collapsible straight-edge. Figure IV-17 is a sketch
of this portable straight-edge and Figure IV-18 is a photograph show-
ing the use of this equipment on the seal edge of a coke-oven door.
Measurements Completed at the Chicago Site. Starting with a
rebuilt door and a coke-side jamb which had been in operation for
about 3 months, measurements were taken of the vertical seal-edge
profile and the jamb profile. The presentation of the collected data
in Figure IV-19 has foreshortened the vertical dimension, but has
retained the actual horizontal displacement dimensions of the end-
closure components. All designations as to sides (right or left) refer
to the sides when the door is on the oven and observed from the bench.
IV-34
-------
FIGURE IV-17. DRAWING OF A COLLAPSIBLE STRAIGHT-EDGE
USED IN OBTAINING THE PROFILE OF JAMBS
AND SEALING EDGES
(All distances are referenced to a taut wire between
the standoff posts. )
FIGURE IV-18. USING A STRAIGHT-EDGE TO MEASURE
THE PROFILE OF A SEALING EDGE
IV-35
-------
These data indicate that prior to adjusting the diaphragm-
deflection bolts on the door there was a calculated 4. 8-mm (3/16-inch)
gap in the vicinity of the right side of the upper latch. In the section
of this report describing the emissions from this test door, it is noted
that the emissions from this door were mainly from this gap.
Because of the lack of clearance between the door and the outward
extension of the jamb, it was not possible to measure actual gaps be-
tween the seal edge and the jamb. All of the gap dimensions indicated
above are, therefore, calculated dimensions assuming that there is no
thermal deflection of the door and seal edge upon becoming heated. In
this regard, it is to be expected that when a door at ambient tempera-
ture is first placed on an operating oven, the door frame and seal edge
will bow outward; i. e., the portions of the door and seal edge above
and below the latches will tend to deflect outward from the jamb, and
the center portion between the latches will deflect toward the jamb.
The foregoing data then are only an indication of the problems in mat-
ing a seal edge and a warped jamb. Various coke-producing plants
have attempted with only limited success to measure the deflection of
doors when they are first heated. The bowing characteristic resulting
from the heating of the door has, however, been confirmed.
Measurements Completed at the Youngstown Site. The same
types of profile measurements were completed at this site with a re-
built door on an old pusher-side jamb. In addition, a nearby door
which was releasing heavy visible emissions was also measured. At
this site, the seal profile was measured with the seal and door at op-
erating temperatures. The data collected, presented in the same
form as in Figure IV-19, are shown in Figure IV-ZO.
Over an observation period of about 5 days, the test-door jamb
seal was completely free of visible emissions on all cycles, and the
nearby leaking door was releasing emissions heavily at the upper
latch on the right side as well as across the top of the door.
In the door-latching procedure used at this plant, the S-shaped
seal is deflected outward until the stop bolts on the door corners seat
on the jamb. This deflection is normally about 3 mm (about 1/8 inch).
On first touching the jamb (prior to the deflection of the seal), the hot
test door had a calculated maximum gap (between the seal and the
jamb) on the left side of about 5 mm (about 3/16 inch). Upon deflecting
the S-shaped seal the 3 mm (about 1/8 inch) to seat the corner stop
IV-36
-------
LEFT SIDE
RIGHT SIDE
Seol Edge Jamb Seal Edge Jo
3 08mm
/ U/32")
08 mm 1 + 32mm
d/32") I 11/8")
\t 1.6mm
Ij (I/ 16")
24mm
(3/32") *
/
2.4mm /
(3/32") 1
1 2 mm I
13/64') \
\
mb
J24mm
(3/32")
7.0 mm
(9/32")
5.6 mm
(7/32")
Results of Straight-Edge Measurements
4 Omm
(5/32")
4 4 mm
111/64")
4 8 mm
(6/32")
'92mm
(23/64")
Touch
I 2 mm
(3/64")
1.6mm
(1/16")
3.2 mm
(I/a")
Touch
4.4mm
(11/64')
Calculated Gaps Remaining on First Touching of Jamb and Seal Edge
0 8 mm
(1/32")
14 8mm
(3/16")
Calculated Gaps Remaining With All Four Corners Touching
FIGURE IV-19. VERTICAL PROFILE OF A JAMB AND DOOR-EDGE
SEAL AT THE CHICAGO TEST SITE
(All vertical distances have been foreshortened, but
the horizontal displacements are actual dimensions.
The center and bottom presentations are calculated
values rather than measured values. )
IV-37
-------
Top
Teil Door
Leaking Door
L.Seol L.Jomb R Seal R Jamb L Seal L Jomb R Seal R Jamb
am
32mm
d/B")
1 6mm
d/16")
1 6mm
d/16") |
1
i
1 6 mm
11/16")
1 6mm
11/16")
16mm -i
J
3 2 mm
11/8")
3 2mm
d/6")
1 6mm
(1/16")
1 6 mm
d/16")
1 6 fTIITI
d/16")
1 6mm
d/16"]
32mm
11/8")
32mm
ll/8'l
6mm
1/16")
\
1 6mm \ 6 4 mm
d/16") 1(1/4")
32mm -1 J4 8 mm
d/8") / 13/16 )
6mm + j 3 2 mm
1/16") / 11/8")
f
Bottom
Vertical Profile of the Door Jamb and Seal Edge on a Test Door and on a Leaking Door
I 6mm
d/16")
f 4 3mm
I 13/lfa")
Touch
[32 mm
(1/8*)
16mm
11/16")
Touch
t 6mm
d/16")
1.6mm
11/16")
Touch
Toucn
T32 mm
d/8")
4 8mm
(3/16")
48mm
(3/16')
Touch
Touch
I'4 8mm
(3/16)
16mm
d/16'}
f 4 8mm
/ (3/16')
Touch
(1/8 )
Calculated Gap Dimensions Remaining on First Touching of Jamb and Seal Edga
loucn
3 2 mm
(l/8">
Touch
1 6 mn>
11/16")
Touch
loucn
1 6mm
tl/16")
16mm
d/16")
Touch
Touch
foucn -i
1.6 mm 4
0/16")
16 mm
d/16")
Touch I
Touch '
Touch
\ 4 8mm
(3/16")
1 6mm
11/16")
4 8mm
13/16")
Touch
Calculated Gaps Remaining With All Four Corners Touching
FIGURE IV-20. VERTICAL PROFILE OF A TEST KOPPERS DOOR
ON A 10-YEAR-OLD JAMB, AND THE PROFILE
OF A LEAKING DOOR ON ITS MATING JAMB
(Only the horizontal dimensions are actual dimensions.
The center and bottom presentations are calculated
values.)
IV-38
-------
bolts, the calculated maximum gap (prior to the springback of the seal
into the gap) was about 3 mm (about 1/8 inch). Because the seal
showed no evidence of leakage at any time, it is apparent that the
springback of the loaded S-shaped seal successfully closed and sealed
the 3-mm (about 1/8-inch) gap.
The leaking door had gaps (on seating the corner stops) calcu-
lated to be about 5 mm (about 3/16 inch), and these were not sealed by
springback deflection of the sealing edge.
Random checks of jajnbs on the coke side of this battery indicated
a wide variation in the degree of jamb warpage. One heavily warped
jamb was bowed inward about 11 mm (7/16 inch) at the middle of the
jamb.
Jamb Distortion Data From Steel Companies. It became appar-
ent during this study that most companies have not measured jamb
distortion, or they regard their data as being unreliable and are there-
fore reluctant to release these data. However, in one instance,
rather complete data were released on measurements on four ovens
on both the pusher and the coke sides. These data were obtained by
establishing a vertical reference line and then measuring the horizontal
distance to the bottom and top of the jamb plus the distances to the
midpoint and quarter points.
On both sides of the ovens, there was an outward displacement
of the top of all jambs relative to the bottom of the jamb. This is
common to most coke batteries (if not all) because of the outward dis-
placement of the top of the battery itself. On the coke side, the
average outward displacement was 6 cm (about 2.4 inches) with a
range of 4. 4 to 7. 3 cm (1.7 to 2.9 inches) over the four ovens. On
the pusher side, the average outward displacement was 5. 8 cm (about
2. 3 inches) with a wider range of 4. 2 to 8. 3 cm (about 1.6 to 3.3
inches).
Given the horizontal distance from a vertical reference plane or
line to various points on a jamb, it is possible to map the depressions
and outward bowing in the jamb surfaces. These data are plotted in
Figure IV-21. The vertical dimensions in this figure are not drawn
to scale, however, the indicated horizontal displacements from the
dashed reference lines are actual dimensions reported in 32nd1 s of
an inch.
IV-39
-------
1
0
1 1
-1
0
2
0
0
-1
0
0
3
0
-22
-8
-8
0
4
0
-12
-8
-2
0
(In 32 nd's of on
inch)
Left-Side Jamb Measurements for
four Ovens on the Pusher Side
1
_0°
0
2
0
+1
+0
+ 7
0
3
0
-8
-II
-1
0
4
0
+ 6
0
0
0
Left-Side Jamb Measurements for
Four Ovens on the Coke Side
1
0
+ 4
+8
0
2
0
-1
*3
*5
0
3
0
-15
-12
-9
0
4
0
-10
+4
0
Right-Side Jamb Measurements for
Four Ovens on the Pusher Side
1
0
0
-8
-9
0
2
0
+2
0
+6
0
3
0
-4
-15
0
4
0
+ 9
+ 3
+ 3
0
Right-Side Jamb Measurements for
Four Ovens on the Coke Side
FIGURE IV-21.
PLOT OF THE HORIZONTAL DISPLACEMENT
OF JAMBS ON FOUR COKE OVENS
[Vertical distance of jambs is not to scale. The
measured horizontal displacements from a
reference line drawn from the top to the bottom
of each jamb are actual dimensions in 32nd1 s
of an inch. Negative numbers (left lines) indicate
inward jamb bowing (depressions).]
IV-40
-------
These data indicate that there had been considerable jamb dis-
tortion at the coke battery supplying these data. Because this partic-
ular battery was experiencing relatively heavy door emissions at the
time of Battelle's visit, it is believed that the indicated jamb warpage
is about an upper limit relevant to the warpage to be expected through-
out the industry.
These data also indicate that there is much more pronounced in-
ward bowing than outward bowing. This is also indicated in the mea-
surements taken by Battelle researchers.
Another conclusion is that there is no particular pattern of dis-
tortion on any one jamb; i. e. , each jamb appears to develop an in-
dividual profile for each side of each jamb.
Internal-Pressure Measurements and Collection of Volatiles.
In the measuring programs in Chicago and Youngstown, measure-
ments were made of the gas pressure in the gas-passages of the test
doors. All of these measurements were made near the botton of the
door.
At the Chicago test site, the internal pressure peaked at 100 mm
H2O (about 4 inches H2O) about 2-1/2 minutes after the start of the
leveling operation. At this time, the pressure fluctuations were be-
tween 25 and 100 mm (1 and 4 inches). In 30 minutes, the pressure
was down to a steady 12 mm (about 1/2 inch). In 80 minutes, the
pressure was down to 6 mm (about 1/4 inch), which was the same as
the pressure at the top of the oven.
At the Youngstown site, the peak pressure was 125 mm of H2O
(about 5 inches H2O), but after 80 minutes the pressure was 20 mm
(about 0. 8 inch).
Throughout most of the coke-producing industry there is an in-
terest in installing a vertical vent or duct in the door liner to lower
the pressure at the seals during the initial high-pressure period in
the oven. This approach is described on pages 111-17 through III-19.
Overall this innovation has merit, but it is judged (by Battelle) that
(a) it cannot completely eliminate the pressure at the door seal, nor
(b) can it lower the emissions that are presently visible at the top or
low-pressure portion of some existing doors. The installation of
IV-41
-------
door vents will tend to decrease emissions from the bottom of leaking
doors, but much of the observed emissions are from the top third of
the door where the vented plug cannot significantly decrease the in-
ternal pressure. It is for this reason that the vented-plug concept was
not included in the collection of concepts in this report.
At the Chicago test site, some first attempts were made to con-
dense emissions deliberately released through a 6-mm (1/4-inch)
pipe installed in the bottom of a door gas passage. For the Youngstown
test, a water-cooled copper column was built for condensing these
emissions. Condensation .was started while the leveling operation was
being completed, and was continued with frequent collector-bottle
changes for the next 40 minutes. In all of the samples collected, there
was a dark, heavy liquid layer and a clear liquid layer. The clear
liquid layer was analyzed and found to be water. The dark liquid is
believed to be low-temperature coal tar dissolved in solvent-type vol-
atiles. The water was the major portion of the early samples and de-
creased in relative volume toward the end of the sampling period.
In general, this information lends support to the hypothesis that
the majority of the pressure rise developed at the oven doors at the
start of the cycle is contributed by steam. The amount of steam con-
densed decreased with time as the water was driven out of the coal.
Given the high percentage of steam and probable presence of only low-
boiling volatiles in the steam at the start of the cycle, it is not sur-
prising that gaps or damage points on the seal system do not "self-
seal" during the initial high-pressure period of the coking cycle.
"Self-sealing" occurs later in the cycle when condensation of higher-
boiling point components begins in the seal/jamb area. From this
viewpoint, the last leaks to seal would be those where the temperature
of the jamb and/or seal is the highest. This would appear to be cor-
rect because leaks near or at the tops of the doors "self-seal" very
slowly. It is judged that, to eliminate emissions completely from
coke-oven doors, it will be necessary to have the door "gas tight"
upon latching the door and before the coal is charged. It was judged
that water-cooling could improve the rate of "self-sealing", but that
it alone is not a direct solution to the emission problem.
IV-4Z
-------
Discussion of Heat Warpage of Metallic Components
At coke batteries, gross warpage and distortion of backstays,
door jambs, and frames for chuck doors frequently is evident. In ad-
dition, Battelle researchers have noted evidence of heat warpage on
the sealing strips on operating coke-oven doors. Many coke-plant
supervisors believe that door frames also may be warping over a
period of time.
A literature search did not reveal that any stress analysis of
buckstays, tie rods, buckstay springs, or door-seal components
(jamb/door/seal) has ever been reported. A study and analysis of the
stresses (thermal and mechanical) of coke-oven end closures was
beyond the scope of this present project, but it is an element to be
considered in evaluating the performance of test components in the
future. Instrumentation of test components (and also the present end-
closure designs) could reveal from the start whether or not new pro-
totype equipment is dimensionally more stable than present structures.
Some measure of stability is required to forecast relative performance.
Because of the absence of data on the mechanical forces acting
on the end-closure components, the mechanism of warpage can only
be hypothesized. However, it is judged that the warpages seen at
coke-oven end closures are examples of the adverse effects of pro-
longed thermal stresses and particularly of thermal cycling.
As an introduction to the subjects of thermal stress and the warp-
age effects of thermal cycling, a literature search was made in an
effort to find an applicable industrial example. Fortunately, a recent
report described and analyzed a heat-warpage problem that occurred
in a leveler bar at a coke plant in Pennsylvania. * In this case, a new
leveler bar of structural carbon steel became distorted so that "the
front end raised approximately 160 mm (6.3 inches) starting 6 meters
(about 20 feet) back from the nose. At a distance of about 4 meters
(about 13 feet) from the nose, the two upper sections of the bar
buckled". Figure IV-22 is a photograph of the buckled leveler bar.
This leveler bar had to be discarded because it would no longer line up
vertically with the chuck doors of the ovens.
Stoltz. J. H.. "Coke Charging Pollution Demonstration", EPA-650/2-74-022, Prepared for the AISI/
EPA, March, 1974.
IV-43
-------
It is worthy of note that the S-shape form taken at the point of
buckling of the leveler bar (Figure IV-22) is also the shape seen at the
point of buckling of sealing edges at some coke plants (Figure IV-8).
It is this shape that lends credibility to the conclusion that the buckling
is the result of heat warpage and not a result of mechanical damage.
FIGURE IV-22.
PHOTOGRAPH OF A THERMALLY
WARPED LEVELER BAR
Point of distortion is generally the
S-shaped portion.
Jones & Laughlin Steel Corporation research personnel investi-
gated the leveler-bar distortion problem and concluded that the failure
was caused by a differential thermal-expansion effect resulting from
nonuniform heating of the bar while the bar was restrained from bend-
ing downward. Because the bar was restrained from bending downward
(by the coal in the oven), the higher-temperature upper surface suf-
fered plastic (permanent) compression. When the bar was cooled, the
top of the bar was shorter than the bottom and the bar reacted to this
differential in length by deflecting upward. It was calculated that a
IV-44
-------
temperature difference of only 160 to 240 C (about 320 to 460 F) be-
tween the top and bottom of the bar could have accounted for the warp-
age. It was J & L's recommendation that the leveler bar be made of
structural steel having a high yield strength at elevated temperatures.
While it is well known that warpage can occur during thermal
cycling without the part or section being under restraint, the applica-
tion of restraint results in patterns of warpage different from what
one would "normally" expect. * The leveler bar, for example, upon
cooling bent in the direction of the original heat source because it had
been restrained from bending downward during the period when the
overheated top of the bar was expanding.
Every coke-oven battery is in cyclic operation and all of its
components are subjected to thermal gradients and thermal cycling.
In addition, most of the battery components are restrained in one way
or another. It is not surprising, therefore, that warpage and distor-
tions become evident after a period of time. The apparent differences
in warpage effects from plant to plant are probably a function of the
differences in the thermal gradients and the differences in the amount
and degree of thermal cycling.
An examination of existing metallic-seal mechanisms indicates
that not enough attention may have been given to thermal-expansion
effects in the development of the designs. A new sealing strip begins
to attempt to expand thermally, as soon as it is placed on the oven.
The fact that the seal strip and its mounting have a long dimension
(4 to 6 meters - 13 to 20 feet) exaggerates expansion effects in the
vertical direction. If, for example, the edge of an austenitic stainless
steel sealing strip reaches an average temperature of 200 C (about
400 F), then the edge, which is always hotter than the rest of the seal-
ing mechanism, attempts to expand a distance of about 12 mm (about
0. 5 inch) over a length of about 4.3 meters (14 feet). It is believed
that the temperatures go much higher than an average of 200 C during
the period when the doors are on empty (but hot) ovens during periods
of charging delays. The sealing strips, however, are either bolted or
welded to the support element with no provision to permit thermal
expansion. This may be the partial cause for the "wrinkling" noted on
door seals, and is the probable reason why there are reports that ex-
perimental thinner edges rapidly developed a scalloped shape. The
Bergman. D. J. "Some Cases of Stress Due to Temperature Gradient", Transactions of ASME. July,
1954. p. 1011.
IV-45
-------
concepts for improved sealing of coke-oven doors (Chapter VI) do not
detail designs aimed at resolving the vertical thermal-expansion prob-
lem. However, this must be taken into consideration in any develop-
ment program.
The heavy H-beam-section buckstays at each end of each oven are
designed to provide the required support (compression) for the refrac-
tory oven and regenerator walls. In coke-plant operations, the buck-
stay flange against the battery bricks is warmer than the exposed flange.
However, during periods when the emissions from the adjoining door
or doors are burning, the .buckstays can become very hot with the out-
side flange remaining at the lowest temperature. Battelle researchers
believe that the outward bowing of buckstays is largely the adverse re-
sult of door fires only. This judgment is reinforced by the fact that
many buckstays that have bowed outward from the battery have devel-
oped a clearance between the back flange and the bricks; i. e., there is
no continuing outward pressure from the oven bricks or deposited car-
bon. Also, in many instances, the jamb lugs were not actively press-
ing against the buckstay. It is judged that this bowing is an example of
thermal ratcheting, i. e., distortion by the mechanism of thermal cycling
alone. The occurrence and effects of thermal ratcheting are discussed
in several papers.* Thermal ratcheting is an incremental increase in
the total strain at the end of each thermal cycle, resulting in a contin-
uing increase in deformation. Based on this, solving of the emissions
problems at the seals (and from behind the jambs) could eliminate the
problem of bowing buckstays.
For the most part, the jambs of coke-oven end closures are com-
plex, rigid, one-piece shapes made of gray cast iron or ductile cast
iron. These jambs are subjected to unknown mechanical loads and to
variable thermal stresses and thermal cycling. Not much has been re-
ported about the physical loading (pushing of the oven bricks) against
jambs, although it is suspected that this is a variable quantity from
door to door and from plant to plant. Jambs are either partially locked
behind the buckstays or are fastened to the buckstays with lugs. One
school of thought in the coke-producing industry is that there is consid-
erable force acting outward against the jamb (and therefore against the
buckstay) resulting from the "growth" of carbon behind the jambs. To
Parkes, E. W., "Structural Effects of Repeated Thermal Loading". Thermal Stress. Pitman Publishing
Corporation, 1964.
Miller. D. R., "Thermal-Stress Ratchet Mechanism in Pressure Vessels", Transactions ASME, 81 D,
1959, p. 190.
IV-46
-------
a degree, the Battelle researchers discount the effect of carbon
growth behind the jambs. Instead, it is believed that thermal cycling
of the buckstay and jamb can open up gaps behind the jambs, and that
these gaps result in (a) emissions and (b) formation of carbon deposits
behind the jambs. These deposits are not believed to grow per se,
but, rather, the carbon deposit may increase in thickness as the re-
sult of further cycles of relaxation, emissions, and carbon deposition.
However, overall, the mechanical forces acting on the back side of
the jamb cannot be discounted as a contributor to jamb warpage by
creep. The reasoning here is that if an expansion cycle can form gaps
(and carbon buildup) behind the jamb, a contraction cycle can result in
considerable force on the jamb and buckstay.
Thermal Stresses and Geometry
It is well known that sectional discontinuities can contribute to
uneven distortion during cycles of heating. Abrupt changes in the
cross sections of the jambs are sectional discontinuities. Sectional
discontinuities occur on jambs at (1) corners, (2) latch-plate attach-
ment points, and (3) features provided for mounting the jambs on bat-
teries. The effects of sectional discontinuities on jamb distortion are
similar to those which could be produced by any form of inhomogeneity.
For example, assume that a round steel bar perfectly homogeneous in
every respect is heated uniformly from end to end and also from around
the periphery. If this bar were suspended vertically from one end, it
could be expected to expand by axial elongation and by uniform increase
in diameter. It would remain a straight round bar. But, if a sectional
discontinuity were built into the shape of the bar as illustrated below,
the bar would deflect by axial bending.
If the uniform round bar were heated uniformly while both ends
were constrained from moving linearly, the bar would bend in a
smooth, mathematically continuous curve. The bar with the enlarged
portion (sectional discontinuity), however, would probably bend in two
smooth curves from each end toward the center. The smooth curves
would be joined by a relatively sharp change in direction of the curve
at the enlarged portion of the bar.
IV-47
-------
In a similar manner for jambs, the sectional discontinuities oc-
curring at the corners, latch-plate attachment points, and features
for mounting the jamb on the oven can produce changes in direction in
the otherwise smooth curve of a distorted jamb. At the latch-plate
attachment points there are abrupt changes of size and shape of the
jamb as well as sharp corners and bolt holes. At the mounting fea-
tures there are abrupt changes in shape, also associated with holes.
Flexibility and Conform ability Limitations
of Existing Sealing-Edge Designs
Apparently there has always been a history of jamb warpage over
time. This is suggested by the fact that battery designers included a
semiflexible metal sealing strip and sealing arrangement and also in-
corporated high power in the door-handling equipment in an attempt to
force the sealing strip to conform to warped jambs. This section deals
with the limitations in the present sealing designs in terms of conform-
ability to the degree of jamb warpage that exists on operating ovens.
Again it should be noted that the original designers only considered
compensating for or adjusting to some limited amount of jamb warpage
before they would (and do) recommend installing new jambs. From one
viewpoint, this can be considered an oversimplified approach. At
least it is oversimplified until such time as a nonwarping jamb design/
material is developed.
Neglecting for the moment the complicating effect of the rein-
forced corners on existing coke-oven sealing arrangements, the seal-
ing action on the tall sides of the ovens is one of attempting to make a
steel beam conform to the hills and valleys of bowed (inward and out-
ward) jambs. The shape and dimensions of these beams and the in-
tended directions of flexibility are shown in Figure IV-23. As shown,
the height of the upstanding component in both seals is more than the
width. These components are, therefore, stiffest in the direction in
which flexibility is desired. The thick edge-holder design is signifi-
cantly stiffer than the thinner design.
IV-48
-------
Portion of S-Shape
Seal Design
Fixed-Edge Seal Design
4.8x19 mm
(3/l6"x3/4")
c
o
£ >.
3 mm.
11/8")
24 mm
(15/16")
i
u
0>
t.
15
o
in
Of
O
,
1 |
*> A
£ 36.5 mm T
~ d-7/16") |
^
28.6 mm
U-l/8")
1
/~\
». .j
^X
^.x"^*'
s^ S^
<-!9 mm>
(3/4")
i
25.4 mm
(1")
» I
/I
FIGURE IV-23. FULL-SCALE SHAPE AND DIMENSIONS OF
THE SEAL-EDGE COMPONENTS OF COKE-
OVEN DOORS
Material is carbon steel and/or stainless
steel.
It is probable that the existing dimensions of the sealing sections
were developed partially by way of operational requirements. One
probable influencing factor, for example, is that heat expansion would
buckle thinner sections more easily. It is not known if shorter (less
stiff) sections were ever tried.
The dimensions of the fixed-edge holder leads to the judgment
that this seal would require more deflection force to conform to warped
jambs. With this type of seal, it is worthwhile to consider what jamb
"shapes" and applied forces must do to the beam-sealing arrangement.
In Example 1, page IV-50, consider a straight-line seal brought into
contact with a jamb bent sharply inward at the top. As increased force
is applied to produce contact with the top and bottom, the center of the
seal tends to spring away from the jamb.
IV-49
-------
In Example 2, sufficient force has been exerted at top and bottom
to close the seal. However, the distorted center portion of the jamb
has not been contacted even though the seal has been bent toward the
jamb.
Example 3 illustrates an extension of Example 2 in which addi-
tional force has been exerted at the center of the seal to produce con-
tact. This could cause the seal to bend outward from the jamb near
top and bottom.
Force
Force-
Force
Jamb
Seal
center
Jamb
Bottom
Example I
ForceJ Bottom
Example 2
These examples illustrate several basic facts about the defor-
mation of a stiff sealing edge:
(1) If the distortion of the jamb occurs over a short dis-
tance, deformation of the seal to conform to the jamb
must also occur over a short distance. The distance
between peaks or valleys on a warped jamb will be
referred to as "pitch" in further discussions.
(2) If the distortion of the jamb is a great variation from
the nominal sealing plane, then the seal must also be
deflected far from its nominal (or relaxed) plane if it
is to conform to the jamb. Distance of the warped
surface of jamb from a nominal flat surface is called
"amplitude" in further discussion.
IV-50
-------
It is conceivable that by careful adjustment of bolts between the
door and the seal, the seal might be deformed to conform to a dis-
torted jamb. However, the following factors make this a difficult op-
eration, often impossible with present designs:
(1) The workman cannot see the effect of his adjustment
directly nor can he directly see the reason for the
leakage.
(2) The spacing or location of adjustment points may not
be favorable for proper adjustment of the seal.
(3) The force required to deform the seal to match the
jamb may be greater than the force provided by the
door-latch mechanism or the door-machine latch-
drive mechanism.
(4) The pitch of the distortion of the jamb may be so short
that it is not possible to force the seal because of its
stiffness to conform by bending within its elastic limit.
(5) The amplitude of the distortion of the jamb may be so
great that it is not possible to force the seal because
of its stiffness to conform by bending within its elastic
limit.
The difficulty of attempting to seal a leaking door by adjusting
the shape of the seal by pressure from bolts is familiar to every
bench workman on a battery. The leak seems to escape adjustment
and move around the door as screws are tightened. This phenomenon
is illustrated by the preceding three examples of door-seal curvature
and is also related to the spacing or location of the adjustment points.
Point (3) (above) that the force required to deform the seal may
be greater than the force provided by the door-latch mechanism is
possibly the most important. The following discussion explains this
situation.
The fixed-edge "diaphragm-pan" seal deforms as a flat sheet or
as a broad-leaf spring. Bolts are provided along the edges to permit
selective adjustment of the seal-edge contour. As a bolt is adjusted
inward (toward the hot side) the seal edge is deflected inward.
IV-51
-------
However, as one bolt is adjusted inward, the seal edge will deform
proportionally to the strength or stiffness of its design and the seal will
be raised away from the adjacent bolts. The deformed shape of the
seal edge will not, however, fit the shape of the warped jamb in the
general case. Bringing other neighboring bolts to bear on the seal edge
will not solve the problem because the seal will only move further in-
ward. At this point, the problem might seem to be simply that the
bolts can only push inward on the seal and cannot pull outward on it.
The fact is that portions of the seal edge adjacent to the point where the
adjustment was desired can be moved too far inward so that they hit the
jamb where they'should nqt. Therefore, when the door is closed, some
latch pressure is required to deform the seal edge locally adjacent to
the point where adjustment is required. This requires more force than
the force per linear inch of seal specified (average 11 to 18 kg/cm, 65
to 100 Ib per linear inch) for sealing, and therefore the sealing force
per inch of seal is diminished.
The S-type sealing edge is inherently more flexible than the fixed-
edge design, but it also has an inherent flexing problem. However,
with this design the problem is limited to inward flexing, i. e., the lim-
ited ability of an S-type sealing edge to enter a major depression on the
jamb mating surface.
The inherent "outward" flexibility of the S-type or "floating-edge"
seal design was established in tests on a rebuilt door and seal. The re-
sults of one of these tests are shown in Figure IV-24.
Load for 3-mm(l/8-in.) Deflection
268kg
630 pounds
1 \
, 1
' 1
I .
- 1 I
r~T~r- rn
* i
, '
\
1 i
0.000 0.004 0.013
0.067 0.125 0.055
Deflection in Inches
0.019 0.005 0.000
FIGURE IV-24. SEAL-DEFLECTION PATTERN ON POINT LOADING A NEW S-TYPE SEAL ARRANGEMENT
Door and seal were in a horizontal position and the downward deflection is equivalent
to an outward deflection on an operating door.
IV-52
-------
In this experiment, a force of only 270 kg (630 pounds) deflected
the seal edge downward (outward from the jamb) a maximum distance
of 3 mm (about 1/8 inch). The effect of this single-point application of
force caused decreasing deflection about 0.6 m (about 2 feet) on either
side of the point of application. It was concluded, disregarding ther-
mal effects, that the S-type sealing design can easily conform to out-
ward displacement of the jamb mating surface. In this regard, the
maximum outward displacement of existing jambs in the data available
to Battelle was 7 mm (0. 3 inch) over a pitch of about 2.4m (about
8 feet).
However, as previously noted, data on existing jambs indicate
that there are jamb depressions (inward bowing from a reference line)
of as much as 17.5 mm (0. 7 inch) of amplitude over a pitch length of
2.4 m (about 8 feet). With this seal design, it is worthwhile to ana-
lyze whether conformity with a deep depression on the jamb is possible.
Some designs of S-seal arrangements have 44 points of spring
loading around the periphery of the seal, acting against the back of the
upstanding seal strip (see Figure III-17). There is apparently a his-
tory of spring relaxation with time and temperature, but, with a new
or rebuilt door, the sealing edge is under some "back pressure" at
all times; i. e., either on or off the oven. In mounting an S-type seal
arrangement on an oven, it is standard practice to attempt to force
(and latch) the door on the oven with enough inward force to seat the
stop bolts on the door against the four corners of the jamb. This seat-
ing action normally results in a backward deflection of the entire seal-
ing edge (if the jamb is completely straight) of about 3 mm (about 1/8
inch). The effect of this deflection is to increase the spring pressure
against the back side of the seal edge. However, there is a limitation
to the amount of inward deflection this increased force against the back
side of the seal strip can cause.
The distance that the seal will deflect inward into a depression on
the jamb is a balance of the applicable (and decreasing) spring force
and the resistance of the seal material and design. However, the max-
imum inward deflection conceivable is 3 mm (about 1/8 inch); i. e., a
return of the edge to the plane that existed prior to putting the door on
the oven. It appears that in order for the S-shaped seal arrangement
to conform to deep depressions or valleys on warped jambs, it would be
necessary to increase the initial overall deflection via the force intro-
duced by the door-handling machine. For example, if the door stops
IV-53
-------
were set at about 6 mm (about 1/4 inch) and sufficient door-handling-
machine force were available to deflect the entire sealing edge this
distance, then the S-seal supported by new or unrelaxed springs would
have a chance of contacting the bottom of a 6 mm (1/4-inch) depression
in the jamb.
It was concluded that there are inherent mechanical limitations
in the conformability of both of the two major designs of door-sealing
mechanisms. The "fixed-edge81 type is particularly inflexible, but it
can be forced to deflect into jamb locations having pronounced inward
bowing. Once "set", however, this design cannot automatically adjust
to the degree of variations in the jamb shape that are believed to occur
from cycle to cycle. These variations can come about by lateral varia-
tions in the positioning of the doors on jambs and by continuing warpage
and flexure of jambs. The S-shaped seal has considerable flexibility
in conforming to outward bowing of jambs, but has limitations on con-
formability to inward bowing. However, a desirable feature of the
S-seal approach is the spring action which automatically adjusts to
minor variations if these variations are within the limits of the design.
It was judged that the designs and materials of fabrication of metal-
to-metal contact seals can be improved to minimize emissions.
This discussion of different existing designs of seals does not
take into consideration the further complications introduced by ther-
mal expansion and thermal deflection. A complete analysis will re-
quire data on heat-caused dimensional shifts and data on levels of
stress and strain in the components.
Technical Specifications That New
Sealing Systems Should Meet
Study of the causes of emissions from existing metal-to-metal
coke-oven sealing systems gave insights into the technical specifica-
tions that new sealing systems should meet to minimize emissions and
to be considered practical retrofitable solutions. To present these
technical specifications, it is necessary to outline the limits of the
system being discussed. For the purpose of evaluating the sealing
concepts that were developed, Battelle researchers are here defining
the sealing system as the existing jambs, doors, and any sealing com-
ponent, including sealants. Within these limits, the technical specifi-
cations for an effective and practical sealing system are as follows:
IV-54
-------
Specification Number and Name
(1) Temperature Tolerance
(2) Heat-Excursion Tolerance
(3) Automatic Gap-Closure
Capability
(4) High Gap-Closure
Capability
Specification (and Comments)
Must withstand the 200 to 300 C
(400 to 600 F) temperature pattern
in the seal location for prolonged
periods of time without deteriora-
tion or dimensional changes. (See
Figure IV-16 for source of temper-
ature data. ) As used here, the pro-
longed period of time can be years
with expensive metal seals or
shorter periods if the seals or seal-
ants can be easily replaced and are
less costly.
Must withstand occasional short
periods (perhaps as much as 4 hours)
at 430 C (800 F) without being
destroyed. Upsets in the operating
conditions at coke batteries (equip-
ment failure) can result in such
temperature excursions.
With an eye to the future, an ideal
sealing system should permit any
door to be placed on any jamb on
one side of a battery; i. e. , no need
for manual adjustment of any kind.
This specification facilitates rapid
replacement of leaking doors from
spare or rebuilt stock, with mini-
mum effect and minimum loss of
time.
The system should seal most of the
gaps caused by the heat war page of
the jambs. Because the warpage
data in Figure IV-21 were from a
battery having a particularly seri-
ous leakage problem, it is thought
that this is about the upper limit
on jamb warpage to be expected.
IV-55
-------
Specification Number and Name
Specification (and Comments)
(4) High Gap-Closure
Capability (Continued)
(5) Resistance to Corrosion
and Chemical Attack
(6) Total-Failure Proof
Based on these data, it is indicated
that the system should seal an in-
ward bow on the jamb face having a
maximum "depth" of about 13 mm
(about 0. 5 inch). The span of the
inward bowing to be sealed is a
minimum of 1. 8 m (about 6 feet).
The system must also accommodate
(and seal) a maximum outward
jamb bowing of about 6 mm (0. 25
inch) over a span of 1.2 meters
(about 4 feet) or more. There are
some minor combinations of both
inward and outward distortions on
individual jamb uprights. The
most flexible metal sealing systems
now in operation can accommodate
outward bowing, but have a 3 mm
(1/8 inch) limit in terms of enter-
ing an inward distortion.
The sealing system must withstand
attack and corrosion by heated gases
and liquids for prolonged periods of
time without failure. Jambs and
seals are exposed to steam, heated
organic solvents, coal-tar vapors
and liquids and, in some instances,
corrosive chlorides from the coal.
Some data indicate that the gases
can at times be at a higher temper-
ature than the present metal seals
(200 to 300 C, 400 to 600 F).
The sealing system should not be
susceptible to the possibility of
complete and/or sudden failure
during operation.
IV-56
-------
Specification Number and Name Specification (and Comments)
(7) Avoidance of New Cleaning Preferable would be a sealing sys-
Problems tern that includes the elimination
of a cleaning problem, or which
functions better with the present
cleaning methods. Acceptable is a
sealing system that introduces no
new cleaning problems that are not
clearly solvable.
In addition to these specifications, new sealing concepts should
satisfy the six functional requirements. These include retrofitability,
dependability, and the other criteria listed in the Introduction of this
report (Chapter I) and in Chapter VII where the concept families are
evaluated.
Judgments and Opinions on Existing
End-Closure Sealing Systems
In meetings with the Sponsors to review the draft of this report,
Battelle researchers pointed out that fewer quantitative data (temper-
atures, stress levels, metal flexure, etc.) had been collected and de-
veloped on operating conditions at end closures than had been expected
and desired. There is actually very little information within the coke-
producing industry on various basic measurements and parameters.
Battelle's efforts to define the problem, as presented in this chapter,
represent only a start of the research and development work required
in this field.
During the review of the draft of this report, the Sponsors asked
Battelle researchers to add the following to this final report:
Judgments and opinions on the existing seal systems
A listing of the unknowns that Battelle researchers en-
countered and which could not be answered within the
funding of this task
The investigator's opinions on the special problems of
the new sealing systems on the taller ovens (6 meters
and higher).
IV-57
-------
This section is responsive to the foregoing three points, and deals
mainly with judgments and opinions based on incomplete quantitative
data. Because of the incompleteness of the data, even among the
Battelle investigators (who have a wide diversity of backgrounds) there
is no unanimity of judgments and opinions. The situation within this
group is similar to the diversity of judgments and opinions held by op-
erators of different commercial coke plants. A diversity of opinions
will continue to be held by various "experts" until the base of quantita-
tive data is enlarged. The following judgments and opinions are offered
to spur continued discussions on the problem and further collection of
quantitative data in research programs.
The coke-oven sealing system includes all of the metal armor on
the ovens, the doors and their components, the door-handling and
latching equipment and procedures, and the cleaning methods used (or
not used). Comments on these elements are as follows.
Jambs
Some removable types of jambs are held against the oven brick-
work by lugs attached to the buckstays, and other types are almost
permanently positioned because they are locked partially behind the
buckstays. It is common to hear that there is considerable force
against the buckstays either directly from the jamb positioned behind
it or from the jambs through the fastening lugs to the buckstays. This
force, the variation and distribution of this force, and the possible
contribution of this force to jamb warpage are not known. These
forces, however, can be measured and should be measured in a
follow-on program.
It appears important that the jamb maintain some pressure on
the brickwork regardless of the dimensional movements in the jambs,
buckstays, and ovens. This does not appear to be possible with the
jamb-bolting procedures used on the majority of existing ovens. Con-
sideration should be given to fastening jambs to the buckstays with
spring loads pressing on the jamb from the outside, with perhaps re-
silient refractory insulation cushioning the movement on the oven side
of the jamb.
The fact that jambs are heat warped introduces consideration of
lowering the temperature of the jamb using insulation and also con-
sideration of analyzing the jamb warpage to develop dimensionally
IV-58
-------
stable jamb design/materials. Any change that would minimize or
eliminate the temperature fluctuation in this critical component of the
sealing system would be a contribution to stabilization. There is very
little information on the temperature fluctuations in existing jambs,
and Battelle researchers have no information on the temperature pat-
terns on the newer, taller ovens. Coke-plant superintendents will not
as yet permit researchers to "sink" permanent thermocouples into
existing jambs because of fear of cracking of the jambs. It is probable
that internal thermocouples can be installed in new test jambs of ex-
isting and upgraded designs. Battelle researchers believe that re-
corded thermocouple readings of the temperature at the top of jambs
on an entire battery would give the coke-plant superintendent insight
into what is going on at each oven for every hour of the day. This
procedure would help to identify the operating practices that are caus-
ing temperature excursions.
Discussions of the possibilities of insulating jambs with resil-
ient refractory materials (Carborundum's Fiberfrax, Johns-Manville's
fiber products, etc. ) behind jambs, brings forth expressions of con-
cern about the possible leakage from behind the jambs. Leakage
would depend upon the degree of compression of the insulating mate-
rial, the distance the vapors have to travel horizontally through the
insulation, and whether the movement of the jamb (during thermal
cycling) will be compensated for by the "bounce" in the resilient re-
fractory. If necessary, foamed (closed-pore) refractory composi-
tions could be developed for this application. Regardless, resilient
refractory of any kind should prove more effective than the simple
rope packing presently being used behind hot, flexing jambs. Labora-
tory tests would give insights into whether resilient refractories are
capable of preventing leakage.
Even with the development of more flexible sealing elements,
some badly warped jambs likely will have to be replaced. Replace-
ment of any jamb is expensive and it could be considered a waste of
time and money if the new jamb either starts warping or releases
emissions from behind the jamb. This introduces consideration of
either substituting upgraded jambs or testing overlay plates on warped
jambs (fastened over a layer of insulation) to present a new surface
to the sealing edge. This, however, is only a concept at this stage,
and further consideration of this approach would require a detailed
thermal analysis. This overlay approach has particular appeal for
those jambs that are fixed behind the buckstays.
IV-59
-------
Consideration of lowering the temperature of jambs has to take
into account how much heat is arriving at the jamb from the flues and
how much is radiant and flame-input heat arriving at the exposed in-
terior surfaces of jambs. It is, for example, important to know
whether widening of the door plug (to give a narrower gap between the
plug and the furnace walls) would lower the jamb temperature. A
thermal model (mathematical) of the operating end of a coke oven could
contribute significantly to any seal-development or heat-conservation
program.
Operating jambs are either approximately "L" shaped in cross
section or, in some of the newer ovens, some of the jambs are sturdy-
looking square posts. Cases of extreme warpage have been seen on
both types of jambs. The heat stress in these jambs is a function of
the thermal gradient (as influenced by thickness and thermal conduc-
tivity) and other factors such as the material's coefficient of thermal
expansion, the modulus of elasticity, and Poisson's ratio. Gray cast
iron has a relatively high thermal conductivity, but its high-temperature
tensile strength is low. It is noted with interest that some of the build-
ers of coke ovens are recommending ductile iron jambs which develop
a higher thermal stress (under identical shape, size, and operational
conditions), but presumably can tolerate the higher stress. It is one
of Battelle's recommendations that the design/materials relationships
for jambs be analyzed in a mathematical and physical modeling pro-
gram. A possibility for consideration as a new jamb design/material
is high-strength steel plates having improved resistance to creep and
relaxation at operating temperatures. Again, however, applied-solid-
mechanics specialists prefer to arrive at a solution by analysis rather
than indulging in conceptualizing and theorizing. This also holds true
for the analysis (and prevention) of the hourglassing problem observed
at many coke-oven batteries.
Earlier in this chapter, it was noted that the layer of semiliquid
tar deposited on the jamb is the filler material that actually forms the
true seal between the metal parts. Stated another way, if successful
steps were taken to prevent the tars from reaching the jambs, the
metal-to-metal existing seals would leak gas even where there is firm
metal-to-metal contact. This is, for example, one of the concerns
with the luted-seal concept family described in Chapter VI. With no
tars coming to the jambs, it would be necessary to develop either a
temperature-resistant, flexible coating for the jamb or to take steps
to lower the operating temperature of the jambs so that a switch could
IV-60
-------
be made to resilient seals. Unfortunately, Battelle's search for a coat-
ing material that will not harden upon being heated to existing jamb
temperatures in air has not been successful. On the other hand, if the
temperature of the jambs were lowered, the tars presently depositing
on existing jambs would not harden and cleaning would be simplified.
With upgraded metal seals that are more reliable in terms of contact-
ing the jambs, the ovens could be emission free and the amount of
tars deposited on the jambs, therefore, would decrease. The amount
of tar deposited could be further decreased by (a) minimizing the gas-
passage area, (b) using a wider door plug (narrower gap), and (c) using
a vent in the plug.
As it stands now, there is little doubt that mechanized jamb-
cleaning equipment will have to be installed on the batteries that do not
have it. Prior to their installation, however, it is suggested that a
test program be completed to evaluate the cleaning results after ar-
ranging to give the cleaning personnel improved equipment and im-
proved working conditions. Improvement can be made in protecting
the workers from radiant heat and in getting the workers closer to the
surfaces needing cleaning. With the present manual cleaning meth-
ods, the bench man has difficulty in cleaning the important top section
of the jamb and his cleaning time is limited by the amount of radiant
heat he will or can tolerate. Heat-shields and steps (or elevators)
can be installed on the door-handling equipment. It is suggested that
the door-machine operators (especially on the pusher side) are in the
best position to clean the top portion of jambs. This would be a simple
job if a heat shield could be swung into the open oven to allow men to
approach the jamb. This cleaning operation has been performed by
Battelle personnel (wearing protective equipment), and it was judged
to be a simple operation. Cleaning of the tops of the jambs is partic-
ularly important in terms of preventing carbon formation in the corners
where the stop bolts on some door designs seat. It is not uncommon
to see hard deposits in jamb corners. This raises the question of
whether the stop bolts are actually seated at some plants.
Doors/Seals
Discussions with coke-plant superintendents and coke-plant build-
ers indicate that some think that doors are flexed inward during the
mounting and latching procedure to match the contour of the jamb like
two bananas nested together in a bunch. It is agreed that when the back
IV-61
-------
side of a jamb and the oven side of the door casting are first heated,
they expand differentially and flex in the same direction, i. e., the
jamb can go concave (if it started as a straight casting) and to a degree
the door casting will approach this shape. Also, the data that Battelle
researchers collected on jamb warpage indicates overall that there is
more inward distortion (development of a concave surface) than outward
distortion. However, Battelle1 s calculations indicate that the 5400 kg
or more of latching force (12, 000 pounds and higher) applied to a door
casting on a 3. 9-meter oven (12. 8 feet) cannot deflect the door casting
more'than a minute amount. Because of this, after the doorstops
seat on the jamb (on one door design), additional application of force
may be wasted and is perhaps even damaging. However, various coke-
plant operators disagree with this statement, saying that extra force
can be helpful in lowering the emission rate from a door. The door-
flexure situation is believed to be different on the 6-meter and taller
ovens where it has been stated that the doors are very flexible, and
according to some opinions too flexible. Overall, a flexing door can
be damaging to the door brickwork or door refractory materials. The
flexure that is required should be in the sealing arrangement rather
than the door. Battelle has no information on door flexure due to heat
and/or applied forces other than to state that these movements can be
measured and analyzed.
Battelle1 s calculations (which are really estimates) indicate that
the maximum outward force that the peak gas pressure inside the oven
can exert on the door is only about 2 to 10 percent of the total latching
force. In addition, the coal inside the oven may exert some small
force against the inside of the door. The remaining latch force is in
theory transmitted to the seal edges, at least on the fixed-edge design
of seal. In the S-seal design, the inward force on the door is effective
in flexing the seal only until such time as the corner stops on the door
seat on the jamb. This point is discussed in other sections of the re-
port. In addition, with the S-seal design there are maintenance prob-
lems in keeping the thrust bearing on the door in working shape so that
they can transmit the applied screw-down force.* In general, Battelle
researchers are recommending that the sealing-edge arrangement be
made more flexible and that only sufficient force be put on the door/
seal to have the sealing edge enter any depression on the jamb surface.
Mullet. J. M., etal., "Gary Coke-Oven Door Development Program", Presented at the Joint Meeting
of the Eastern and Western States Blast Furnace and Coke Oven Association on October 25, 1974.
Copies of this paper can be obtained from the United States Steel Corporation, 600 Grant Street,
Pittsburgh, Pennsylvania, 15230.
IV-62
-------
Even low-pressure contact of the sealing edge and the jamb will prevent
emissions, but there must be contact. This is particularly true on a
freshly cleaned jamb.
However, improvement in the sealing-edge flexibility can be
retrofitted into both types of sealing-edge designs now in existence.
As stated in the earlier section on "Technical Specifications That New
Sealing Systems Should Meet", it is important that retrofitted designs
should have automatic gap-closure capability; i. e., require no manual
adjustment of any kind. This can be more easily accomplished with a
spring-seal design. It can, however, be accomplished with the fixed
seal (see Figure IV-23, page IV-49) if (a) the seal support is redesigned
to achieve flexibility and (b) if independent mechanical force can be ap-
plied to the back of the seal support after the door is mounted on the
oven. It is considered probable that this approach would involve expen-
sive retrofit, but equally distributed force against the back of the
flexible seal edge will drive the seal into contact with the inward bows
on the jamb.
In the judgment of some coke-plant operators, the most difficult
cleaning problem on coke-oven end closures is the manual cleaning of
the entire height of the gas passage. This may be correct at some
plants, but of those plants observed by Battelle researchers (in the
test programs), it was judged that gas-passage cleaning would be sim-
plified if (a) coal could be kept out of the gas passage, (b) the hot tar
collected at the bottom of the gas passage was removed on every cycle
immediately upon opening the oven, and (c) the amount of tar entering
the gas passage and collecting at the bottom of the gas passage could
be decreased. In fact, the hot tar could probably be drained off some
time in the cycle.
A severe cleaning problem, and one contributing to additional
problems, is the progressive buildup of carbon and/or carbon/pitch
on the sides of the door plug. This material is particularly difficult
to remove manually. If tight-fitting doors could be used on ovens, it
is judged that this would minimize (a) coal entry into the gas passage,
(b) the gas-pressure buildup at the seals, and (c) the amount of tars
and volatiles that enter the gas passage. Ways should be investigated
to operate with tighter door plugs. This would include development of
methods for occasional nonmanual cleaning of door plugs, improve-
ment of the patching methods on the ends of ovens, and upgrading the
door-handling machinery to control accurate placement of the doors
on ovens. In addition, there is a probability that this approach would
1V-63
-------
lower the jamb temperatures or minimize the temperature excursions
on door jambs.
Door-Handling Machinery
It is a functional requirement of this program that the concepts
developed be "compatible with existing door-handling and oven-end
working equipment". Although this is a requirement, it is suggested
that the capability of such existing equipment should be upgraded.
The swinging of the door and the inward/outward motion of the
door-handling equipment should be interlocked to permit only one mo-
tion at a time. Battelle observers have witnessed door plugs being
damaged (by striking the latch hooks) by operators attempting to swing
and place the door all in one motion.
At other locations, Battelle observers have seen operators at-
tempt to come to a stop on an electric-eye spotting point without
having any slow-travel capability built into the equipment. The "slop"
or loose tolerances in the power train resulted in the operator over-
shooting the target from one side to the other. It is suggested that
the existing power trains be used only for large movements along the
battery, i. e. , to bring the equipment "into the ball park". Auxiliary
equipment should be used to spot the equipment exactly. This could
be done in various ways including the installation and use of a hydraulic
piston that attaches itself to the rail to obtain a firm support. It is
appreciated that coke batteries shift dimensions with time. However,
accurate placement of doors on jambs can be developed so as to elim-
inate or minimize damage to the doors and seals.
It is normal to see door-machinery operators spend up to a min-
ute attempting to increase the latching force on screw-type latches by
repetitively jogging the tightening motor. This operation could be up-
graded with the installation of larger motors and the use of replace-
able, calibrated, slip-type clutches. Also, with the screw-type
latches there is concern with the problem of erratic door-latch pres-
sures resulting from damaged thrust bearings. Test equipment can
be developed to indicate the latching force without the use of special
latch hooks. Battelle personnel, for example, estimated latching
forces by measuring the depth of the impression made on strips of
lead taped to the latch bars.
IV-64
-------
Overall, it is judged that the door-handling machinery can be
and should be upgraded to (a) prevent damage to the doors and seals,
(b) spot the door exactly, and (c) install and latch the door as per
specification. Such upgrading will contribute to better sealing and
to containment of potential emissions from the ovens.
IV-65
-------
CHAPTER V
POTENTIAL TECHNOLOGY TRANSFER
Sealing problems are common in many applications over a wide
range of conditions. It was considered possible that advanced tech-
nology in other fields might be adapted to become effective sealing
methods for coke-oven doors.
Technical Objective
The technical objective of this task was to search other in-
dustries and other technologies for possible elements of technology
and design which might be transferred to conceptualization of new
coke-oven door-sealing systems.
Expanded Objective
In addition to making a search of possible transferable sealing
technology from other industries and other technologies, Battelle-
Columbus in its research proposal outlined the desirability of con-
ducting small-group conferences for idea generation/technology
transfer. The participants in these sessions were to be members of
Battelle's professional staff who are recognized for their combined
innovative ability and knowledgeability in various fields of technology.
Further, it was foreseen that (a) some sealing concepts would
center on the use of resilient materials to effect a seal and (b) conditions
V-l
-------
at coke-oven doors may be beyond the limits of various commercially
available materials. It was for these reasons that those researchers
assigned to this technology-transfer task were also asked to explore
the thermal capabilities of new elastomers, resins, and other non-
metallic materials.
Literature Search
A major portion of this search was done by machine using the
key-word approach. This search included the following sources:
Source
Coverage
Engineering Index 10 years
NASA A, B, N, and X Series 1962 to present
NTIS Report Literature 1964 to present
DoD Report Literature 10 years
AEC Related Literature 6 years
Foreign Literature Search 10 years
Air Pollution Abstracts 5 years
Applied Science and Technology 12 years
British Coal Utilization Research 10 years
Association
Coke Review 22 years
In general, the return on this investment of time was low. The
results of the searches suggest that certain conditions at coke-oven
seals are unusual if not unique. These conditions include the large
ratio of height to width of the doors, temperatures over 260 C (500 F),
and exposure to hot tars and solvents.
The most promising leads developed were inflatable elastomer
seals used on large doors in gas-tight rooms and the patented metallic
"Sn seals developed by Trouvay and Cauvin of France for high-pressure
and high-temperature piping connections where misalignment is a
potential problem.
V-2
-------
Inflatable rubber-type seals were discussed with the two manu-
facturers in the United States. The concepts that were developed are
shown in Concept Family No. 4 (Chapter VI). The "S"-type metallic
seals of Trouvay and Cauvin (Paris, France) are shown in Figure V-l,
Spring Seal
Seal Contact Points
FIGURE V-l. "S"-TYPE METALLIC SEAL
The Trouvay/Cauvin approach is of interest because this company
has stated that these seals (a) are demountable, (b) can function at
temperatures up to 600 C (1100 F), and (c) do not require close toler-
ances in the mating surfaces. Variations of this approach have been
considered to some degree in Seal Concept Family 5 (Chapter VI).
This French company has not responded to letters asking for informa-
tion about the royalty considerations involved in using either this seal
or a variation of this seal.
Survey of Resilient Sealing
Materials and Resins
A search for high-temperature elastomers and reinforced resins
for possible use as a flexible contact strip was conducted mainly by
contacting government agencies and their suppliers. The Air Force
Materials Laboratory and other agencies have been supporting re-
search to increase the temperature resistance of elastomers and
plastic composites.
It is not unusual to read announcements stating that "a new
plastic that can withstand temperatures to 430 C (800 F) and is stable
at 370 C (700 F) has been developed by. . . . ". Investigation in all
V-3
-------
instances revealed that stability of plastics is rated on a relative short-
term basis; sometimes as long as 2000 hours, but often much shorter.
The consensus that developed is that there is no plastic that will
not deteriorate at 315 C (600 F) or even at 260 C (500 F). This is in
agreement with the information developed during tests at coke ovens
where Teflon (rated stable at 260 C) deteriorated and flowed under
pressure in a few month's time. It was concluded that if plastics
had an application on coke-oven seals, it would be necessary to water -
cool these materials.
Small-Group Technology Transfer and Innovation
Technology transfer means literally the use of existing technology
in a new and novel application. Achievement of such a transfer requires
both a sufficient knowledge of the existing technology to appreciate its
ramifications and limitations, and a sufficient knowledge of the prob-
lem area to be able to recognize where the existing technology might
apply. Today's technology is so broad and complex, and is developing
so rapidly, that it should not be expected that a single mind could bring
to bear the vast scope desired. Further, different minds view a given
problem from different perspectives. Often, the solution to a problem
comes from a totally unexpected quarter. Alternatively, a problem
may change dramatically in character if one (in the words of medicine)
attacks the source rather than treats the symptoms.
For coke-oven doors, the problem is one of sealing the doors to
prevent the leakage of hydrocarbon emissions into the atmosphere
surrounding the ovens. A large part of the present program has been
aimed at the details of this leakage. What is it? How much is there?
Where does it come from? Why do leaks develop? What is the environ-
ment a successful seal must endure, etc? This information was re-
quired not only to enable us to pass judgment on the pros and cons of
proposed seals, but also to provide some guidance as to where to look
in the existing technology for possible solutions.
As information developed, it was fed to various members of the
research team so that a scan of the literature covering seals would be
more meaningful. At the same time, realizing the difficulties associ-
ated with recognizing potential solutions in the literature, and realizing
V-4
-------
the possibility that an ideal solution could be missed in such an ap-
proach, an additional method was adopted. This was a variation on
the "brainstorming" idea.
Basically, brainstorming amounts to assembling an appropriate
group of technical people, proposing the problem, and letting the dis-
cussion proceed with no negative comment allowed. This generally
produces a large number of ideas, most of which are eventually
judged worthless when viewed as solutions to the problem. However,
even the worst of the ideas sometimes stimulates a cross-link and
generates a novel useful idea. Negative comments during the session
tend to suppress novel thoughts. Therefore, negative comments are
outlawed. Ideas are evaluated later.
The first innovative session was held at Battelle's Columbus
Laboratories (BCL). It involved 18 people with broad and diverse
backgrounds. Participants were selected on the basis of the research
team's personal knowledge of their creative abilities. A brief sum-
mary of the problem was presented, along with a movie of coke ovens
with leaking doors. Discussion was then initiated on the problem,
with a moderator to keep order. In this session some dozens of ideas
developed. Subsequent evaluation by the moderator and by the
mechanical-design specialists on the project narrowed these concepts
and turned some of these into an engineer ing-concept presentation.
Interestingly enough, even the idea list itself was enough to spur con-
tinuing creativity during the subsequent evaluation sessions. Thus,
several of the proposed seals were conceived after the innovative
session, but could be traced in concept to the discussions. Of par-
ticular interest was the recognition during the innovative session that
a seal close to the tip of the plug might be ideal in that it could prevent
condensible gases from ever reaching the conventional seals. This,
of course, could eliminate the deposits of coal tar at the conventional
seals and would greatly relieve current cleaning and reseating
problems.
This concept of "hot-zone" sealing was discussed further with
other team members during evaluation sessions. It was at that time
realized that, if an appropriate foam could be found, this might pro-
vide precisely the sealant necessary. Thus, a search began for a
suitable foam, something not considered at all in earlier stages of the
investigation. We believe that we may have identified such a foam.
It is patented and preparations are currently being made by a manu-
facturer to bring it to the market place within the next year. Specific
V-5
-------
details on this foam have been requested from the manufacturer*,
but are not yet available.
A second innovative session was conducted at Battelle's Pacific
Northwest Laboratories (BNW) in Richland, Washington. There,
eight members of the senior staff, with widely varying backgrounds,
selected for their creativity and independence of thinking, were
assembled. The procedure followed was the same as at BCL, except that
that the same moderator introduced some of the ideas discussed at
BCL when appropriate to avoid redundancy or to expand the discussion.
While some redundancy was anticipated, it came as a surprise
that most of the ideas presented were clearly different from those
generated at BCL. At one point, the concept of a "hot-zone" seal
was introduced and there followed a series of ideas relative to the
use of coke breeze as a sealant, along with the idea of redesign of the
oven to permit coke breeze to be poured in next to the plug during the
loading of coal into the slot. This is a radically different approach
which no amount of literature searching would have revealed.
The procedure at BNW differed from that at BCL in that a second
meeting was held the following day. The purpose of the second meet-
ing was to become more specific relative to materials and concepts.
Negative criticism was permitted. Again, creativity continued and
new thoughts developed as the previous day's concepts were refined.
The results of these meetings were later discussed with the BCL
project team, and selected ideas were converted to design drawings.
A digest and summary of ideas generated at the two Battelle
laboratories are provided in Tables V-l and V-2. As the reader
examines Chapter VI of this report (Conceptualization of Sealing
Methods), it will be apparent that some of the ideas on these lists
were expanded into entire concept families and into several variations.
Van Leer (V.S). Associated with Royal Packaging Van Leer of Holland.
V-6
-------
TABLE V-l. DIGEST AND SUMMARY OF IDEAS GENERATED AT THE
BATTELLE-COLUMBUS BRAINSTORMING SESSION
A. Concepts Involving a "Cold Seal"; Fairly Adaptable to Existing Door
or Jambs with Minor Modifications
1. Lute internally, in a groove, combined with pressure sealing* '.
2. Gun groove shut with pitch-clay, or clay-molasses foam.
3. Mount seal retainer on jamb; push knife edge into
(a) lead or copper
(b) asbestos/Teflon
(c) spring-loaded fiberglass
(d) silicone rubber - water-cooled retainer.
4. Coat parts with Teflon for easier cleaning.
5. Spray jamb with mold-release compounds (silicone, graphite in
oil, flint powder in binder, edible release compound now sold
for household use, etc.).
B. Concepts Involving a "Cold Seal", But Requiring Moderate to Major
Redesign of Doors or Jambs
1. Change seal every time with quick-release mechanism.
2. Design labyrinth seal.
3. Use a double seal with the inside seal tight and the outside
seal more-or-less loose. Collect emissions at top of seal.
A. Use a second door (like a garage door) covering the main door.
Collect emissions.
5. Use a corona-discharge wire outside the sealing edge to collect
emissions.
6. Suck out gas from between two seals.
7. Electrically heated knife-edge to cut into some sealing
material such as thermoplastic tar.
8. Cool sealing edge to accelerate condensation of leaking
vapors.
9. Water-cool jamb to prevent warpage.
10. Use thin Teflon tubing, water cooled, as sealing edge.
Pressurize with water to cool and provide seal.
11. Provide Teflon tube with tail for easy replacement by
shaping the tail into a groove.
12. Use wedge-shaped door with chevron-shaped seals.
13. Provide auxiliary pressure for sealing by hydraulic or
pneumatic means.
V-7
-------
TABLE V-l. (Continued)
14. Provide more sping in knife edge so that it can follow
contour better. Use hardened steel to provide more rugged
sealing.
IS. Introduce "self-aligning" into door to minimize position-
shifting.
16. Bolt "add-on" water cooling to jambs.
17. Use twin seals with water flowing between seals.
C. Major Redesign and Operational Concepts
1. Provide seal in gas passage to prevent condensibles from
ever reaching existing seals.
*During the preparation of this Final Report it was learned that
on April 1, 1975, Mr. Albert Calderon was issued U. S. Patent
No. 3,875,018 dealing with a form of internal luting of coke-
oven doors. The Calderon method was not directly evaluated by
Battelle researchers, but a general evaluation of luting
approaches is included in Chapter VII.
V-8
-------
TABLE V-2. DIGEST AND SUMMARY OF IDEAS GENERATED AT THE
BATTELLE-NORTHWEST BRAINSTORMING SESSION
A. Concepts Involving a "Cold Seal"; Fairly Adaptable to Existing
Doors with Minor Modifications
1. Add foaming agent to tars as they condense. This would
fill space quickly to seal off leaks.
2. Use silicon foam (good to 600 F or better) as soft re-
usable seal.
B. Concepts Involving a "Cold Seal"; But Requiring Moderate to
Major Redesign of Doors
1. Self-cleaning or disposable condenser, possibly in gas
passage but probably at edge of door; maybe a water
labyrinth.
2. Reorient the seal 90 degrees to battery face. Achieve
seal by side pressure either from articulated door or
from pneumatic or hydraulic pressure from the sides.
3. Same as 2, but use Flexatallic seal.
4. Swing-wing door f^/ \+^] to apply pressure.
5. Put second door (flexible) over plug-holder door.
6. Seam-weld shut each cycle.
C. Concepts Involving a Gas-Passage Seal; Fairly Adaptable to Exist-
ing Doors with Minor Modifications
1. Disposable nose plug of coke breeze, maybe made into
blanket or partially bonded together by coal tars or
bentonite clay.
2. Foamed carbon in gas passage, possibly incorporate breeze.
3. Suitable preformed foam to abrade and jam into gas passage
as door closes.
D. Concept Involving a Gas-Passage Seal but Requiring Moderate to
Major Redesign of Oven
1. Pour coke breeze adjacent to doors along with coal. Coke
breeze will fill gas passage and act as trap for condensibles.
V-9
-------
TABLE V-2. (Continued)
E. Major Redesign and Operational Concepts
1. Have only one door on bottom of slot.
2. Bury jam in ceramic to minimize temperature gradient
and, therefore, minimize warping.
3. Gunite repair of plugs and ceramic surfaces.
F. Materials Possibilities
1. New 1-Mo ductile cast iron may be superior to existing materials.
2. Check boiler grates as possible materials with long-term
dimensional stability in thermal-cycling environment.
V-10
-------
CHAPTER VI
CONCEPTUALIZATION OF SEALING METHODS
Improved emission control at the end-closures of coke ovens
will require an improvement in many of the elements of end-closure
systems. The following discussion of concepts deals principally
with the concepts regarding seals and sealing techniques.
Approach to Concept Generation
Candidate concepts relating to oven-sealing systems were
generated by idea conferences and consideration of current literature
and developments in other industrial fields. To assure that ingenuity
would not be stifled, a "from-the- bricks -out" attitude was adopted
for these conferences.
Preliminary Evaluation of Sealing Methods
As seal concepts were accumulated from various researchers,
they were catalogued according to the extent they modified the exist-
ing oven designs. Thus, a concept which replaced door and jamb,
"from the bricks out", was called a "Category One" concept. A
concept which merely added a seal feature to the existing door (or
jamb) was called a "Category Five" concept. As more concepts were
collected, it became apparent that a somewhat different system of
classification of concepts was required to facilitate evaluation of
concepts.
VI-1
-------
A preliminary evaluation of seal concepts was performed by
Battelle researchers of various technical backgrounds, already
familiar with the research work. In this culling operation, individuals
were presented with word descriptions and illustration representing
49 seal concepts. Each evaluator was asked to place each concept
in one of three groups, namely:
Group 1: Believed to offer greatest benefits and greatest
promise for successful outcome of develop-
mental research
Group 2: No strong feelings for success or failure of
developmental research
Group 3: Believed to offer no unique advantage and to
hold little promise for successful outcome of
developmental research.
As the results of this evaluation exercise were studied, two conclu-
sions became quite apparent.
(1) There was consensus among the evaluators regarding
the concepts placed in Group 3.
(2) Groups 1 and 2 did not reveal clear lines of dis-
crimination.
Therefore, it was concluded that this preliminary evaluation could
effectively lower the number of candidate concepts by eliminating
Group 3. Additionally, as a result of this study of concepts, an
important insight into classification of concepts for further evaluation
was realized.
Classification of Sealing Methods
Working collectively with Groups 1 and 2 from the preliminary
evaluation, it became apparent that these concepts could be regrouped
into six families. Figure VI-1 illustrates the logical development of
this classification system. Each of the concepts generated could be
placed in one of the six classes of concept families.
This grouping of seal concepts into concept families now revealed
strengths and weaknesses in each family. Some seal concepts were
combined or modified to incorporate more desirable features. Some
new ideas were sought to extend the number of design possibilities
VI-2
-------
Oven
door
seals
Conventional
zone
seals
Hot zone
seals
Concept I
Contact
seals
Concept 4
Non-contact
Concept 6
CONCEPT TREE
FIGURE VI- 1.
AN ILLUSTRATION OF THE LOGICAL DEVELOP-
MENT OF THE CONCEPT CLASSIFICATION
SYSTEM
illustrating the possible latitude of each family. The purpose of this
system of concept classification was to present concept families,
each of which could be represented by numerous design possibilities,
for further evaluation. This was done because
(1) Detailed design study of specific experimental seal-
ing hardware was not within the scope of the present
program, and
(2) Success of an overall program of such scope and im-
portance should be based on selection of a concept
approach which allows many design alternatives for
experimental development.
VI-3
-------
Sealing-Concept Families
This section includes all of the seal concepts classified into
families as shown in Figure VI-1. Each concept family is described
by illustrations of several design possibilities with related functional
explanations. Additionally, each concept family has been synthesized
into a complete sealing system. Written descriptions of all major
portions of a complete sealing system are included with each concept
family. The exceptions to this approach are where there is no re-
quirement for a modification of existing equipment and/or procedures.
For example, only one concept has a requirement for a modification
to existing oven brickwork. In all instances, existing latch mecha-
nisms in good condition are compatible with all concepts. A measure
of good condition is that latching forces would be distributed to the
four latch points and the torque of the latch-drive mechanism must be
transmitted effectively to the latch plates. No concept requires
greater precision in door positioning than present practice. It is
suggested, however, that modifications should be made to improve
the precision of door positioning. This would include lowering of
door-position tolerance with the addition of replaceable guide pads
and upgrading the door-machine(s) power train to permit improved
positioning control. This could be accomplished with the addition of
an auxiliary power unit on the existing equipment.
In general, it is assumed that if a concept can be developed to
seal coke-oven doors, it can also be developed to seal the small
chuck door(s) on the pusher-side doors. In some instances, it would
be necessary to consider ceramic insulation or metallic heat shields
to lower component temperatures and attendant heat loss. The
adaptation of some concepts might be limited by the cooling which can
be provided, either by air or water. Overall, however, the emis-
sions from existing chuck-door seals can be minimized by maintain-
ing the original design objectives. Hinge pins should fit the hinge so
that doors are held in alignment with the jamb surface and with the
latch mechanism. The latch mechanisms should be free to move so
that latching forces are adequately transmitted to the seal. Latch-
driving mechanisms should be capable of producing designed forces
to close and latch the doors properly. Minor modifications should
be made to the chuck-door equipment to improve the serviceability
of these features on an as-needed basis.
VI-4
-------
This chapter describes concept families in the way that they
were first presented to the AISI/EPA/BCL judgmental evaluators.
This first evaluation was followed by a more technical evaluation by
Battelle researchers. It is in this final evaluation (Chapter VII) that
details and discussions of materials, problems, and possibilities are
included.
Seal Concept Family I - Hot Zone Seals
This concept includes seals located in the hot zone between the
door plug and the oven wall, outboard from the end flues, or between
the door plug and the coal charge. The intent of this concept is to
close off the door-plug clearance area so that (1) coal will not be
admitted to that area or to the so-called gas channel, (2) gas pressure
will be lowered in the gas channel, and (3) volatile vapors will be
significantly lowered in the gas channel, thus minimizing condensa-
tion products in the region of the conventional (present) seal. Present
materials developments admit the possibility of suitable materials.
Frangible foamed ceramics developed for other industrial applica-
tions possess some properties similar to those required in coke-oven
application of this concept. One example might be the frangible glass
foam produced by the Pittsburgh Corning Corp.
Several design possibilities illustrating this concept are shown
in Drawings 1-1, 1-2, and 1-3. The following discussion of the con-
cept refers to these illustrations and reveals relationships between
the concept and conventional designs of ovens and operating equipment.
Oven Brick Work. Modifications to oven brick work are not
required except by the design possibility illustrated by Drawing 1-3,
in which a hole through the oven roof is required behind each door
plug.
Door Jambs. Existing door jambs of various designs are com-
patible with this concept, except that extreme "hourglass" distortion
would restrict door placement and removal. Presently acceptable
degree of flatness of the jamb is not detrimental to application of the
concept.
VI-5
-------
A .haped plug leal of begged or eolded coke flnea could be placed on
the door plug autoeatlcally while the door la on the door machine.
Aa the door ! Installed on the oven the ahaped plug of fln«» could
b« fretted away by Che jeefc or oven well to Bike e tight fit.
Subsequent cleaning of the plug could be accoepllahed with >echanl>ed
ecrepere. The original door eeel featurea ere retained ee e eecoodary
eeel.
l-l
-------
0
^
5
E=
I k .. '1
ill! ,^
Frangible cerulc Mtcrl*! would be faaaad In plec* round th« door plug
by autoaatAd oparatlon while th« door U on th« door MchliM. A* th«
door U r«pl*c«d th« fo«« c«r««ic would fr«t «v«y on Cb* ov«n wall to
ka tight fit. When the doot Lm removed th« r«m«lnlng tamm would
b* «cr«p«d away If n«c«»»«ry and naw aatartal would b« foaaMd In place.
Th« original door a«al fsaturaa ar« retained a a«coo4*ry «!.
7
"ZJ
W
-------
HH
00
la thla concept coke flnea ere dropped through e (nev) hole In the oven
roof, directly behind the door plug, ee the coel le charged. The coke
flnea ere puahed, by the Increaalng height of the coel charge, tightly
egelnat the door plug. Thla produce! a firm, denae plug of flnea vhlch
do not contain volatile! to exclude oven gee end volatile! froa the
conventional aeel area. Flnaa nay fuae together during the cycle, to
be puahed out with the coke. Plug aide! oould be ecreped cleen (e
needed by mechanical icraperi.The original door (eal faaturee are
retained aa a aecondary aeal.
OOOfc.
1-3
-------
Seal Components. Existing door seals of various designs in
good condition are compatible with this concept and would be retained
as a secondary sealing mechanism. The concept would perform the
primary sealing function and the present door-seal would be exposed
to lower temperatures and to lower quantities of oven gases and
particulates.
Oven Doors. Existing doors of various designs are compatible
with this concept and would be retained without modifications, if in
good condition.
Cleaning. Sealing media which might remain in the oven would
be pushed out with the coke. Sealing media and other accumulated
deposits of the coking process which would cling to the door plug
would be cleaned periodically by conventional means, preferably
mechanized, including scraping or high-pressure water jets. Clean-
ing of the oven wall would not be required. However, the sill should
be shoveled clean each cycle. Jambs and conventional door seals
will require occasional cleaning by conventional means, preferably
mechanized. Accumulations in this area are minimized by the pri-
mary seal which excludes intrusion of volatile vapors and
particulates.
Seal Concept Family 2 - Luting Seals*
This concept involves seals located in the conventional zone of
existing metal-contact seals. The concept performs the sealing
function by the use of formed-in-place material which has the ability
to deform to fill and close potential leak paths. Somewhat different
than the luting done at older design coke-oven doors, this concept is
directed mainly to locating sealants between or in some instances
inside of the mating metal surfaces. The concept minimizes the need
for accuracy of fit between adjacent metal members of the seal.
' Luting has the present meaning in the coke-producing industry of sealing older-design doors by
plastering the outside of the mating surfaces with a clay mixture. In this study, the luting concepts
consist generally of placing (and compressing^ a sealant between the two mating surfaces. It was
decided not to attempt to coin or originate a new name for this sealing approach because it could
develop that the sealant will be applied or introduced either inside or outside of the mating surfaces
either before or after a door has been placed on an oven.
VI-9
-------
Present materials developments admit the possibility of suitable
sealants. Frangible foamed ceramics and elastomeric materials
developed for other industrial applications possess some properties
similar to those required in coke-oven applications of this concept.
Several design possibilities illustrating this concept are shown
in Drawings 2-1 through 2-6. The following discussion of the con-
cept refers to these drawings.
Door Jambs. Existing door jambs of various designs are com-
patible with this concept, except that extreme "hourglass" distortion
would restrict door placement and removal. The presently accept-
able degree of flatness of the jamb is not detrimental to application
of the concept.
Seal Components. Existing door seals are not retained in this
concept, but are replaced with more rugged components. Surface-
to-surface contact of metal components as required in existing
knife-edge seals is not required in this concept. However, in the
design possibilities shown in Drawings 2-5 and 2-6, surface-to-
surface contact could occur at several locations along the jamb.
Oven Doors. Existing doors of various designs are compatible
with this concept and, if in good condition, would require only simple
modification to accept the new seal component(s).
Cleaning. The cleaning required by this concept is minimized
by the ability of the sealing media to conform to surface irregulari-
ties including those formed by previous deposits of sealing media.
Sealing media and other accumulated deposits of the coking process
which would cling to the door, the jamb, and the sealing components
in excessive amounts would be cleaned periodically by conventional
mechanized scraping.
VI-10
-------
1. Foamed-ln-place material would be applied to the channel frame by
mechanical wans mounted on the door machines. Llkevlae,
mechanical means would be required to remove the material when It
Is no longer functional.
2. Preformed foam shaper would be formed at an off-battery site, brought
to the door machines, stored In suitable facilities, and applied to
the channel by mechanical means.
G>ROOTe.O TO tv\AKE B-AT
IF
2-1
-------
r
to tb«
l IMB« oonud on tb* door chlno
vdualul IHIU voald b« n^ulrcd to T«KT* tb* iHttrUl «fc4n it
! DO lonf«r foncttoiMl.
2. Pr«for«d fora ih>p«T vould b« foc«»d «t an off-b«tt«ry lit*, bToocht
t. th. do«-chin... .tor- 1. .^t*U f-ctllU... -d .Wll- ..
ch< cbum.1 by mcbmlc.l Mn..-
-------
SOMC. Og.V>OS\T5> Ov/ER
1. Poomed-ln~place material would be applied to the channel frame by
mechanical means noun ted on the door machlnea. tdkevlie,
mechanical aeant would be required to remove the material when It
la no Longer functional. ' <
2. Preformed foam shaper would be formed at an off-battery site, brought
to the door machine*, atored In suitable facilities, and kppllled to
the channel by mechanical neana.
'".] ' 'I :
L.J
W/LPUTTEl-
2-3
-------
AV Be. sv*m*
WILPUTTETj
'2-4
-------
1. This concept employs e foaoed-In-place material. Foam preforms
are not applicable,
2. Mechanized means vould be mounted on the door machines to a'pply
foaj] to the door while it is off the oven.
3. Cleaning uould be accomplished by mechanized means mounted on the
door machines.
in
c
2-5
-------
1-4
I
1. This concept employs a foamed-ln-place material. Foam preform*
are not applicable.
2. Mechanized means would be mounted on the door machines to apply
foam to the door while it Is off the oven,
. Cleaning would be accomplished by mechanized means mounted on th
door machines.
-------
Seal Concept Family 3 - Resilient Seals
This concept involves seals located in the conventional zone of
existing knife-edge seals. The intent is to use continuous shapes of
resilient materials mounted to the door or jamb. These shapes are
fabricated and molded of various materials in a variety of cross-
sectional designs to form a seal with mating surfaces which may be
somewhat irregular. Similar devices for many industrial applica-
tions are commercially available from many standard as well as
custom designs. Some available materials and some new materials
under development possess properties which could be suitable for this
application.
Several design possibilities illustrating this concept are shown
in Drawings 3-1 through 3-11. The following discussion of the con-
cept refers to these drawings and reveals relationships between the
concept and conventional designs of ovens and operating equipment.
Door Jambs. Existing door jambs of various designs are com-
patible with this concept, except as follows:
(1) Extreme "hourglass" distortion would restrict door
placement and removal.
(2) If a seal component is to be added to the jamb (see
Drawings 3-7 and 3-8), this component could be
shimmed and grouted to re-establish flatness of a
warped jamb. Presently acceptable degree of
flatness of the jamb is not detrimental if no seal
component is to be added to the jamb. The resil-
ient seal could conform to a warped jamb.
Seal Components. Existing door seals of various designs could
be compatible with this concept and could be retained as a secondary
sealing mechanism. However, existing seals would be replaced with
similar features which would provide secondary sealing as well as
protection for elastomeric seals. The resilient seal would require
protection from (1) radiant as well as conductive heat, (2) mechanical
damage, (3) oven-gas chemicals, and (4) contact with hot coke. Means
for accomplishing these requirements are specifically noted in the
sketches of design possibilities.
VI-17
-------
\ \ \ \\
IViC.TAV.LATV OVA TOOO
II
I
t'
00
BEFORE. APPLICATION)
In this design possibility, the sesllng
material es supplied In rolls would be applied
to the existing seals by automated means while
the door Is on the door machine. The same
operation could remove the worn or damaged
seal just ahead of the new material.
BEINJG APPLIED
KOPPE.RSSEAL
AFTER APPLICATION
M ETA1- FOI L
15ESILIEKJT
MATERIAL
COMPOSITE.
EESILIEWT SEAUKJG
MATERIAL
iUPPUIED IM ROLLS
INJSTAUl-ATIOKJ TOOUJS
COUl-D BE RO\_I_EES
RESIL.IEMT
SEAL.IU&
MATERIAL
BEFORE APPUCATION
BEIKJG APPLIED
VdlLPOTT^L
AFTER APPLICATION
AND
-------
r
METAL FOIL
-HELD IN PLACE
BY TAR LIKE MATERIALS
RESI LI ENJT
MATEKIAL
Ii
I
>
vO
COMPOSITE RESILIENT
-SEALING MATERIAL,
SUPPLIED IN ROLLS
In thll design poeilblllty, the lelllng aeterlel
» >up|ilted In roll! nould b. epplled to the
eil.tlng «e»l. by autoeuted neeni while the door
1<
-------
JAMB FRAME
(ADDED)
SH1M AND GEOOT
TO CORBECT
JAMB WARPAGE
COOl_lWGi WATER
COUUD BE ADDED
TO EEDUCE TEMPERATURE
FOR THE ELASTOMER
JAMB
.....L.Zo
r\7y^^JTS^-V^ v^N-
*~ s s i is s~ s s~
I-M
o
VARI ES ' nnr«
ACCORDS DOOK
TO JAMB
FU AT MESS
\
t
a
SEAL. FEAME
(ADDED)
VIEW SHO\W1NG DEFUECTIOM
DUE. TO M1SAUIOKJMEKIT
This design possibility employs an elaetomeric
strip which Is applied Co the door by
mechanized means while the door is on the door
machine. The strip, which could be supplied
in rollst could be formed into * continuous
loop by tapering the ends and joining with e
quick-setting cement. This operation would
also be mechanised.
-------
JAMB FRAME
(ADDCD)
71 C.AP VARIES
t\J ACCOKDKOCi
> . TO JAMB FI-ATNESS
SHIM AMD
TO CORRECT
JAMB WAHF-A£.e
COOLJNCi WATER COULD
BE ADDED TO REDUCE
TEMPERATURE FDR TUB. ELASTOMER
SEAL. FRAME
SCM-
This design possibility eaploys an elsstosieric
trip wbich Is applied to the door by
Mchanised aeons while .the door is on the door
achine. The strip( which could be supplied
in rolls, conld.be forwed into a continuous
loop by tapering the ends end joining with e
quick-setting eeneat. This, operation would
also be Mcbamlced.
SHOWING DEFL-ECTIOW
DUE TO MISAUOJMEMT
3-4
-------
DEFLECTOR, OPTION Al_
5eAl_
OPTIOMAL-
II
ALTERNATE. ARKAMCIEMEWT
£|F EXISTING SEAL. IS OMITTED)
TUBE SHIMMED TO
CORRECT FOR WARPED
JAMBS
- DIAPHRAGM (MEW)
WATER COOL-ED TUBE
WELDED IN PLACE.
This design posaibility employs a water-cooled cube as the sealing surface against
which the resilient seal closes. The cooling system requires a water supply,
water treatment, water-supply cooling, pumping equipment, and a distribution
system. Optionally, the existing knife-edge seals could be retained as a shield
for the resilient. Also, optionally, a deflector can be used to shield the
resilient seal. A back-pressurizing gas can be added to the system for added
protection for the resilient seal. A gas supply and distribution system would
be required.
3-
5
-------
BR1CKUORM
DE.FUECTOR,
OPTIONAL
is)
OJ
ALTERNATE
FORM WO. I
(IF EXISTING SEAL IS QMITTED)
TUBE SHIMMED TO
CORRECT FOR WARPED
JAMBS
WATER COOLED TUBE.
WELDED IN PLACE
RESILIEUT SEAL
.EXISTING SEAL,
OPTIOMAU
This, design possibility employs, a water-cooled tube as the sealing surface against
which the resilient seal closes. The cooltna system requires a water supply,
water treatment, water-supply cooling, pumping equipment, and a distribution
system. Optionally, the existing knife-edge seals could be retained as a shield
for the realllent. Also, optionally, a deflector can be used to shield the
resilient eesl. A back-pres»urlling gas can be added to the system for added
protection for the resilient seal. A gas supply and distribution system would
be required.
ALTERNATE
FORM kJO.Z '
(IF EXISTING, SEAL IS OMITTED)
L...
3-6
-------
.SHIM AND iROUT TO
COBBECT WARPED JAMBS
h-1
I
Im thle deeleji poaalblllty a realllaat e»t«rUL it efelied
t* « M^«r OB eh* door by aechaalBod ^MIU. 1C !
ntaiood in place by «a dhoaiv* vblch COMBS on cb*
>t«rl*l. Cl««nlng could b* accOKBliilMd by i^clunlud
tcrmpflas of the aet«l ml >ttrfac«. ««t«r coo Hoc c*o
b* added to th« acriwr aoimtcd on the Jo*, tbtcr cocUn(
require* -water ettpply* water treataant, water aupply
cooling, pua^lnc eqolpaHttt and a dlatrlbutlon ayatea.
RESILIEMT SEAU
ADHEsive ONE SIDE
ONUY, WATeeiAU MAYBE
COMMERCIALLY AVAIL.ABLE
STOCK SHAPE OK
FORMED INJ PL.ACE
KOPPERS
3-7
-------
i
ro
Ul
WIM AND CiROUT TO
CORRECT WARFED JAMBS
RESILIENT SEAL.
ADHESIVE ONE S»DE
ONLY, MATERIAL MAYBE
COMMERCIALLY AVAILABLE.
STOCK SHAPE OK
FORMED IK) PLACE
In this deilgn possibility a resilient material la applied
to member on the door by nechjnlced oaana. tt 1*
retained In place by an adhealve vhlch COMB* on the
aterlal. Cleaning could be accomplished by echanlaed
crapping of the metal aeal surfaces. Water cooling can
be added to the iseri>er aounted on the jaab. Water cooling .
requires a water supply, water treetwnt, water supply
cooling, punping equlpasnt and a distribution system.
WIUPUTTE
-------
SEAL WITH TOUG.H SKIM
FOE WE.AC AMD EESIUIEKJT
CORE FOE 5EAl_l MO,
COMPUIAMCE.
COLJUD BE. 5UPPL.1ELD INJ
ROL.L. OR AS FABRICATED LOOP.
CAM BE. WATER COOL.ED
SKJAP - I M
FEATURE
This design possibility uees a resilient component
j i .. .r»<.l. The outer material is
composed or two materials. i»"* «wt=»
tough and wear resistant while the Inner material la
more spongy. The crosa-sectlon form has bulbs at
the corners to as.tat In retaining the seal and also
In removing It when It needs to be replaced. An
.alternate possibility would mount the resilient seal
on the door rather than on the jamb. Hater cooling
could be provided.
KOPPC.ICS
-------
M
-v)
WTTH TOUCH .SKI M
AktOF .maiURN
UltUG
COUL-D *er SUPPLIED IW
*~ ' VtftBRICATeO UOQR
' cooi_eQ
SNAP-IN
FEATURE
Thl« dailgn potilblllcy u««. « r««tll«nt component
COBpOMd Of CVO KeXlall. Th. OUt.T MtUlll It
tout* and «« r««l»t«nt wUU th« lnn«T ««t«rl«l it
ma* tpongy. Ttaa cro««-»«ctloo font kM bulb* it
tlw comer, to mice in ntalnlni th« >ul ad «lw
In renovlng It when 1C M*d> Co b« rapUecd. An
Itfxnace po..lbllity xrald BDunt th. mlUrat «!
on th. door r.th.r than on th. J«*. tf«t«T eeelln*
could b« provided.
.1-10
-------
ts>
00
SEAL, WITH TOUGH SKIKJ
FDR WEAR AWD RESIUBUT
CORE FOR 3EAUW6
COMPLIANCE, COUL-D AC
.SUPPLIED IKJ ROl_L-tOR
AS FABRICATED UOOP-
Th«
«
-------
Oven Doors. Existing doors of various designs are compatible
with this concept and would be retained with slight modifications
such as:
(1) Elimination of existing seal components and plugging
of any resulting openings.
(2) Addition of new seal components by bolting or
welding.
(3) Addition of holes for injection of back-pressurizing
gas.
Cleaning. Cleaning required by this concept would be accom-
plished by scraper mechanisms mounted on the door machine if metal
seal areas are to be cleaned. However, if the resilient seal material
should require cleaning, automated, high-pressure water jets might
be utilized for both door and jamb areas, on both metal surfaces and
the resilient seal material. Cleaning would be minimized by inboard,
secondary-seal features which would exclude gross intrustion of
particulates.
Seal Concept Family 4 - Inflatable Seals
This concept includes inflatable seals located in the cold zone
or the conventional sealing area relative to the door and jamb. The
intent of this concept is to utilize continuous frame shapes of elasto-
meric tubes, mounted to the door or jamb, and which may be inflated
to form a seal. These frame shapes are fabricated and molded of
various materials and in a variety of cross-sectional designs so that,
when inflated, the cross section is extensible (as an accordion) to
form a seal with mating surfaces which may be somewhat irregular.
Similar devices for many industrial applications are commercially
available with standard as well as custom designs. Some available
materials may possess properties which could be suitable for this
application. Liquids or gases could be used for inflating the seal.
Several design possibilities illustrating this concept are shown
in Drawings 4-1 through 4-8. The following discussion of the concept
refers to these illustrations and reveals relationships between the
concept and conventional designs of ovens and operating equipment.
VI-2 9
-------
I
OJ
o
JAMB
Wll-PUTTE SEAL
AS SECOXJAEV
SEAU
DOOR
This design possibility requires a utter supply, water treatment facility
and provisions for cooling the vater, as veil aa circulating pimps. Alao
required are pumpa for the inflating ttdia. If back-presaurlelng ass la
used gas supply facility and pumping oeans are required, as well aa
distribution plugging for all required media. Seal could be replaced
PRIMARY SEAL manually vlth a shield plug'lnserted into the oven.
INFLATED WITH
WATEE OE AIR -/
RETAIN IMCi
COOL.ING
VWATER
JAM 6 FRAME .
ADDED AROOUD
DOOR
AMD
EAMP
FRAMED COMPl_ETEL.Y
AROUKJD THE JAMB
coui_D BE SHIMMED
AND &BOUTED TO MAKE
FLAT IF JAMB IS BADLY
HOUR GLA5SE.D
COOI_D INJECT BACK
PSESSUBIZ.IMC, CiAS
THCOUGH HERE
TO
W/IPtlTTg-
4-1
-------
n pli-bln* for all required M
-------
Thti duica poMlblllty nfolTM mtu nmlj, Mtar tnttantt Utility
nd prortiloM (or cooling tlu Mt«T, well clrtuUtlnf ptvp*. Alto
ntolnd an f««* for th« InfUtlac aidla. If tack-prutnxiiliis |u U
| imd gu >upply faclllc; nd pwqilng nu in nqulnd, u v«ll »
distribution pluM>ln« for ill nqulnd wdU. S*al could b* r«pl»c«J
MntKll; with ihUld pluj ln»«rt«d Into the ovn.
'SECONDARY SEAL. PROTECTS
INFLATABLE 5EAL. FROM RADIANT
HEAT, COKE CONTACT. AND
MCCHAMICAL. DAMAGE
<
M
I
INFUATED
COO1_IM<1
WATER
COCJUD INJECT BACK
PRESSUR.IZ.IUCj dAS
THRDO6H HERE TO
CHANNEL BETWCCKi
TWO SEALS
-RETAIN i M<*
GROOVE.
COULD BE SHIMMED
AND (SHOUTED TO
MAKE FLAT IF JAMS
IS BADLV HOUR GLASSED
PROTECTIVE AND ALIGNMENT
RAMP FRAMED COMPLETELY
AROUND THE JAMB
-------
<
t-l
1
PRIMARY SEAL-
IMFL-ATED W|TH WATER OR AIR
rSECONDARY SEAL.
\ COM Pl_ETE COKJTACT
U M W ECE.SSARY
X
This design possibility requires a source of Inflating media. If the
media is a gas, air might be used from a pressure bottle on the door
machine. The door seal could be inflated by using a hove like when
filling an auto tire. If water it used to cool the elastomer, a water
supply, voter treatment and cooling facility as well as circulating pump*
ad distribution plumbing are required. If back-pressurising gas is
used, gas supply facility and pumping means are required, as well as
distribution plumbing. A variation of this design might place the seal
components on the Jamb rather than on the door. The seal could be
replaced manually when door la on the door machine.
-IMJECT BACK
PRES5URIZIM
-------
BLOW INERT GAS
INTO SEAL. WHILE
SEALING TO KEEP
AREA GLEAM, BLOW
COMPRESSED AIR
THRU SAME HOLES
TO COOL. RUBBER
SEAL, WHILE NOT J
.SEALING '^
nrumKlclD| « ll
u»ed fr* (apply facility and pacing Mm in ra^tdnd. «!! u
0 C
dlitrlbutlon plmbtng foe ill raqnlnd dU. EMl eoild be nfUeed
o
ally vltb ' ibUld plug ln>«rt*d into th* o*n.
ELASTOMER SEAL.
(DEFLATED)
WATER COULD BE CIBOULATED
AT LOVU PRESSURE WHILE
DEFLATED FOR COOLI NK1
-ELASTOMER SEAL
INFLATED WITH AIR OR WATER
KOPPEgS
-------
<
1-4
I
01
Thl. deaign poMlbllltjr require. a eourca of Inflating Bed la, poialbly
«lr or wtar. If He li uaed prenure bottle might ba urrled on the
SHI M AMD GROUT door Mchina «nd the door « M *"'t'r 1§ ""* to co°l the !"«»".
Mter lupply. wear trMCnent «nd cooling facility *j veil > circulating
famft and dutrlbutlon pliablng tra required. Ihe lul could be replaced
only »lth the door horliontal In aalntenance itatloa.
INFLATED WITH AIR
OR WATER TO
DEFLECT METAL
SEAL,
SEAL.
ll
KOPPERS
4-6
-------
<
n
(JO
&EOUT TO LEVEL
UP WABPE.D >IAMB>
p
IK)P«_ATABLC SEAL.
TUB^ DEPORMS METAL
SEAL. AMD ACSO
MAKB.S AM ELASTOMeeiC
COMTACT SEAL.
WATER
\MFI_ATEO WPTH
AIE OR "WATEE
Thii design pOBjilblllEy reqolraj « aoorce of Inflaciqg «*££, ponlblf
it 01 vitcr. It «ir it oiad * pre«ur« battle nl»he b< urrt«d on tb*
oaoi Mchtu «nd th« door w«l could be tainted by luing > tiote Ukt
Am (tiling «s tuto tin. It v*t« ii u»d to cool th« el«»toner, *
MCer lupply, wter treitnent »d cooling Utility ji mil i< clreuUelng
pjap* nil dt«trtbucioa plunblng *.n required. The «»1 emit b* »pl«C«4
onlj «lth tlu door horiiooul IB > ailntwunc* >uttoa.
KOPPERS
4-7
-------
<
M
I
SHIM AND GROUT
TO' CORRECT JAMB WARPA6E
INFLATED WITH AlB
OC WATER TO DEFLECT
METAL SEAL
TWM
IhH de.lgn po.ill,lllty require! ,ourc« of Infl.tlng medl., po«lhlT
llr or water. If air 1« iu«d i pre.sur. bottle might ba carried on tha
door uchtu aw! tha door >
-------
Door Jambs. Existing door jambs of various designs are com-
patible with this concept, except as follows:
(1) Extreme "hourglass" distortion would restrict door
placement and removal.
(2) If a seal component is to be added to the jamb (see
Drawings 4-1 and 4-3) this component could be
shimmed and grouted to re-establish flatness of a
warped jamb. Presently acceptable degree of
flatness of the jamb is not detrimental if no seal
component is to be added to the jamb. The
inflatable tube could conform to a warped jamb.
Seal Components. Existing door seals of various designs could
be compatible with this concept and could be retained as a secondary
sealing mechanism. However, existing seals would be replaced with
similar features which would provide secondary sealing as well as
protection for the elastomeric tube. The elastomeric tube would
required protection from (1) radiant as well as conductive heat, (2)
mechanical damage, (3) oven-gas chemicals, and (4) contact with hot
coke. Means for accomplishing these requirements are specifically
noted, where provided, in the sketches of design possibilities.
Oven Doors. Existing doors of various designs are compatible
with this concept and would be retained with slight modifications such
as:
(1) Elimination of existing seal components and plugging
of any resulting openings.
(2) Addition of new seal components by bolting or welding.
(3) Addition of holes for injection of back-pressurizing
gas.
Cleaning. Cleaning required by this concept would be accom-
plished by scraper mechanisms mounted on the door machine if metal
seal areas are to be cleaned. However, if the elastomeric tube
should require cleaning, automated, high-pressure jets mounted in
the door machine could be used. High-pressure water jets might be
utilized for both door and jamb areas, on both metal surfaces and
elastomeric tubes. Cleaning of the elastomeric seal is minimized
VI-3 8
-------
by inboard, secondary-seal features which would exclude gross intru-
sion of particulates. Between the secondary seal and the elastomeric
tube, a back-pressurizing gas might be injected to more completely
exclude intrusion of oven vapors to the elastomer, thus preventing
coal-volatiles from condensing.
Seal Concept Family 5 - Contact Seals
This concept involves seals located in the conventional zone of
the existing knife-edge seals. This concept is similar to existing
knife-edge seals in that it employs metal-to-metal contact. One of
the contacting metal members must be sufficiently flexible to conform
to the irregularities of the other metal member, usually the jamb. The
same contacting metal member must also possess adequate spring
characteristics to transfer latching forces properly to its sealing
edge.
Experiments (Chapter IV) have demonstrated that bright,
thoroughly clean, metal sealing members with typical surface finishes
will not seal coke-oven gas pressures. A period of time is required
for volatiles to condense and to form a seal from deposits of tars.
Design possibilities illustrating this concept are illustrated
in Drawings 5-1 through 5-14. The following discussion of the con-
cept refers to these illustrations and reveals relationships between
the concept and conventional designs of ovens and operating equipment.
Door Jambs. Existing door jambs of various designs are com-
patible with this concept, except that extreme "hourglass" distortion
would restrict door placement and removal. Moderate degree of
flatness of the jamb is not detrimental to application of the concept.
A method of improving flatness of the sealing surface of existing
jambs is shown in some of the figures.
Seal Components. This concept requires the replacement of all
existing seal components with components of new design. A possible
exception is the sealing surface on the jamb, as discussed under door
jambs.
VI-3 9
-------
II
I
o
1. bpUcoot of thU Ml li nail operation raqulriiic powr
much. lapliCMHnt could be ecOBplUhed with the door oa the
door McbliM if hMt rtleldlog It provided for th. op»r«tor.
2, Door clot Ing aovownt It limited by exUtloi corner ttopt.
5-1
-------
F<
L..
1. fayllcMMnt of thir Ml U maatl opmtloo ra«alrlas
noch. (pUuant eauM k« cco^ll>h«d with ch* door on
door chloi If hMt (hUldlat U prort
-------
SEAl
oooa \«> O-OS.\VAC»
1. (novel of worn or deaeged rail «d replecement with e aw .Ml
Id be don* by aechenlcel muu vhlle the door 1« on the door
nchlne.
2. Door clo.lng oovenent ! Halted by exUtlng comer »ope.
5-3
-------
1. Kraoval of von or duigcd iul >nd nplworat «lth n» »ul
vould b« don* by uhanlcal p»«M «blU th« door H on the door
chloi.
2. Door closing aovUHnt ! ItAlced by Mlotlog coraar Itopa.
5-4
-------
I. ItpUoMst at 0,1, .»il 1. rani opmtlan
wraeh. lapUewit could b. iceo^ilUIud «lth th* door oa th*
door MChlne If hut ihleldlng 1> provide (or ch« opwMor.
5-5
-------
s
den whin u b«t AUUtic u rnrUfi fox th« otmtor
5-6
-------
<
ii
i
Is* <
I ' I
JAMB
cssvkg^
^
^_
DOOR
J
1. Replacement of thle teal li * gtknuil opcritton nqulrlng powr
vreach. Repliceaent could tw ecconplutud with the door on the
door chine If hut .hlsldlng 1> provided for cbe operator.
5-7
-------
II
I
-J
nt of thli >ul li i oinuel openclon requiring paver
Replacement could be ccoepllihed with Che door on the
nchlne If heat ihleldlng li provided for the operetor.
5-0
-------
3
00
r
s-a
-------
l-l
SOCH A'b CORUER'a-TOTflyV,
\fjiLPu~rrf-
5-10
-------
r"
ACTION
MUSTV eft ADOSO OR -^
-------
<
}<
1
Off. RS.MCVE.P TO AOQUST
COMTIMUOU'5 SE.AV- MEMBER- SEE
FOR.
VAAS MORE. CORKIER.
5-\o
-------
CO E N EES FORM ED
STAMP| NG,
<
II
Ol
STRAIGHT SECTIONS
ROl_l_ FOEMED Of?
MADE IN PRESS BRAKE.
-MOUMTIWCS HOL.E.S
(WHERE RE-Q'D. )
SEAL
WEI-D SEAI- ONL.Y
ISOMETRIC VIEW SEAL. ASSEMBLY
FOR DESI6N COMCEPTS 5-9,5-IO,5-11,5-IZ.
5-1.1
-------
I
Ul
U)
5-14
-------
Oven Doors. Existing doors of various designs are compatible
with this concept and would be retained with slight modifications such
as:
(1) Elimination of existing seal components and plugging
of any resulting openings.
(2) Addition of new seal components by bolting or
welding.
Cleaning. The purpose of cleaning required for this concept is
to achieve a smooth surface but not a bright surface. Cleaning would
be accomplished by scraper mechanisms or high-pressure water-jet
devices. Scraping might be used for cleaning the jamb and high-
pressure water jets for cleaning the seal member. The cleaning
mechanisms and devices would be mounted on the door machine.
Seal Concept Family 6 Noncontact Seal
This concept includes seals located in the cold zone or in the
conventional sealing area. The intent of this concept is to use pres-
surized gases in the seal area to retain oven vapors in the oven and
to exclude coke-process condensates and particulates from the sealing
surfaces. Labyrinth seals, in which intimate contact of mating mem-
bers is not essential, would retain oven gas pressures by means of
injected-gas pressure within the labyrinth. This injected gas must
create within the labyrinth pressure which exceeds the internal gas
pressure in the oven. Thus the sealing gas will leak into the oven,
as well as out to the atmosphere. Therefore, the sealing gas should
be atmosphere compatible and also should not affect adversely the
quality of coke-oven gas or the coking process. Nitrogen might satisfy
these requirements. To minimize the amount of gas consumed in this
concept, improvements to jamb flatness are desirable. Additionally,
elastomeric seal strips placed outboard from a metal labyrinth can
eliminate sealing-gas leakage to the atmosphere. In this arrange-
ment the sealing gas leaks only into the oven, excluding oven gas and
volatile condensates from the elastomeric seal and seal area. Several
design possibilities illustrating this concept are shown in Drawings 6-1,
6-2, and 6-3. The following discussion of the concept refers to these
drawings and reveals relationships between the concept and conven-
tional designs of ovens and operation equipment.
Vl-54
-------
01
01
^^\ L ,^_ f??~\
i.'y. *;-.: -.*J*f'S
SHIM AMD GiROUT TO \-EVEX.
DOOR SEAL HWVE. AOOCO
Steel seel trail etteched to both door and Job ere brought
together to fan « labyrinth. More chuull ««y be deilfned
thn ihown In the llluitratlon. Clou fit 1« not eiientlel
zctpt ci> reduce the mount of bick-preiiorUing gai required.
The g» will relic the preiiure In the libyrtnth 10 thet oven
gii will be excluded, thue the leel eree le kept cleen. A (*
eupply. compreiiori. end distribution plue<>lng ere required.
0-1
-------
A ateel aeal frame la attached to the door to retain on elestomerlc seal. This
seal baa a firm outside covering to resist mechanical damage and a softer
Interior to be compliant to warped, rough Jambs. A portion of the aeal frame
closes to a portion of the Jamb. Back pressurizing gas Is injected between
this (partial) metal seal and the elastomeric seal. This gaa includes the oven
gas and particulates from the area of the elastomeric seal, keeping it clean.
A Jamb frame might be shimmed and grouted onto the old Jamb to level out the
varpaget thus minimizing the amount of gas needed. A gas aupply, compressors,
and distribution plumbing are required.
i
i
I
On
l
COPPERS
6-2
-------
A ateel seal frame i* attached to the door to retain an lastooeric seal. Thl»
teal his a firm out tide covering to resist mechanical damage and a softer
Interior to be compliant to warped, rough Jamba. A portion of the aeal frame
clo
-------
Door Jambs. Existing door jambs of various designs are com-
patible with this concept, except as follows:
(1) Extreme "hourglass" distortion would restrict door
placement and removal.
(2) Presently acceptable degree of flatness of the jamb
is not detrimental if no seal component is to be
added to the jamb. The elastomeric seal strip could
conform to a warped jamb. However, if a seal
component is to be added to the jamb (see Drawing
6-1), this component could be shimmed and grouted
to re-establish flatness of a warped jamb.
Some jambs may become distorted to an extent that they are
unfit for use as a sealing component or are incapable of effective
sealing against the brick work. Replacement jambs should be im-
proved to yield a longer effective life.
Seal Components. Existing door-seal components of various
designs would be replaced with new seal components on the doors or
jambs. New components of labyrinth seals would incorporate fea-
tures for injection of sealing gas and retention and protection of
elastomeric seal components. The elastomeric strip would require
protection from (1) radiant as well as conductive heat, (2) mechanical
damage, (3) oven-gas chemicals, and (4) contact with hot coke.
Oven Doors. Existing doors of various designs are compatible
with this concept and would be retained with slight modifications such
as:
(1) Elimination of existing seal components and plugging
of any resulting openings.
(2) Addition of new seal components by bolting or welding.
(3) Addition of holes for injection of back-pressurizing
gas.
Cleaning. Cleaning required by this concept would be accom-
plished by mechanized scraper devices mounted on the door machine
if metal seal areas are to be cleaned. However, if the elastomeric
strip should require cleaning, automated, high-pressure water jets
VI-5 8
-------
mounted in the door machine would be used. High-pressure-water
jets might be utilized for both door and jamb areas, on both metal
surfaces and elastomeric tubes. Cleaning of the elastomeric seal is
minimized by inboard, secondary-seal features which would exclude
gross intrusion of particulates. Between the secondary seal and the
elastomeric strip, the back-pressurizing gas would exclude intrusion
of oven vapors to the elastomer, thus preventing coal volatiles from
condensing.
VI-59
-------
CHAPTER VII
EVALUATION OF FAMILIES OF SEALING CONCEPTS
The families of sealing concepts described in Chapter VI were
evaluated by two methods. The first was a judgmental procedure
(using a numerical scoring method) performed by knowledgeable
representatives of EPA and AISI and by Battelle personnel familiar
with the work. The second evaluation took into consideration the
various comments made by evaluators in the first evaluation, and
then each concept family was analyzed by Battelle researchers with
respect to the seven sealing-system specifications described in
Chapter IV. The output from this second evaluation determined
Battelle's recommendations.
Initial Judgmental Evaluation of Concept
Families by AISI/EPA/BCL
In this evaluation, a broad base of evaluations was sought to
apply the collective knowledge and experience of those people who are
intimately experienced with the operation and management of coke-
plant facilities and who are trained and experienced in conducting
research and development programs and in related fields of physical
science. The following paragraphs discuss the organizational steps
leading to the evaluation, the evaluation procedures, and the analysis
of the results.
Evaluation Criteria
Evaluation of concept families to select preferred concepts for
further development required establishment of logical and relevant
VII-1
-------
evaluation criteria. In this sense, "preferred concepts" are defined
as those which hold the greatest promise for
(1) Successful accomplishment of the requirements of
an oven-sealing system, and
(2) Ultimate successful development through experi-
mental design and research to a point of industry
usefulness.
Establishment of evaluation criteria to define the requirements
of an oven-sealing system involved the efforts of a committee of the
joint A1SI-EPA sponsorship and the Battelle research team. From
experience with previous similar research programs, BCL research-
ers could suggest typical criteria. However, concentrated effort
of the full committee was required to develop acceptable phraseology
so that the criteria related specifically to coke ovens. Following are
the criteria developed for the evaluation of sealing-concept families.
(1) Relative Effectiveness to Lower Emissions (Stop Smoke).
Emissions from coke-oven doors are the result of deterioration of
sealing conditions. With this in mind, does this concept have essen-
tial elements with potential for (a) stopping or drastically lowering
coke-oven emissions, and (b) improving one or more conditions which
contribute to leak-producing deterioration?
(2) Operating and Maintenance Costs. The cost of replacing
sealing components includes the costs of materials, fabricated com-
ponents, and labor for replacement and related efforts. It also in-
cludes the cost of operating the facilities and equipment required to
replace the components. Costs will also be affected by whether the
sealing components may be replaced while the door is on the door
machine at the oven or whether the door must be removed to a ser-
vice area. How does this concept compare with others?
(3) Capability of Retrofit. Can this concept be adapted to
existing battery designs of various types or variations? (a) Can it
be retrofitted to present types of ovens with minimum effort; i.e.,
by the use of clip-on devices or by the use of drilled or tapped holes
in existing components, or does it require more extensive modifica-
tion or redesign of components? (b) Is the concept adaptable to
VII-2
-------
different types of ovens without requiring significantly different varia-
tions of the concept? (c) Does the concept hold promise of working
on ovens of any practical size? (d) Will the concept function without
changing existing door machinery, will it require minor changes, or
will it require extensive changes of and/or additions to the door
machinery?
(4) Life. Will the sealing components of this concept, other
than expendable seala'nts if used, have the dependability and repeat-
ability to be effective for jb months? Six months is a period common
to some schedules for repair and replacement of existing seals
independent of major repair schedules involving plugs, etc. Six
months is chosen for this criterion so that all evaluators will have
identical, objective criteria.
(5) Environment and Safety. What is the effect of this concept
on operator safety? Can this concept perform without creating new
environmental problems when the door is on the oven or at other
times in the cycle?
(6) Effect on Cycle/Schedule. There are periods at any coke-
oven battery when the charging and pushing operations (including door
handling) are completed in the shortest possible time; i.e., there is
a departure from time-spaced operations. Does this concept increase
the "shortest possible time"? Will this concept permit the pushing/
charging function to be accomplished with no increase in the time
presently required?
(7) Sensitivity. What is the sensitivity of the concept to damage,
error, abuse, or nonstandard operating conditions? Will component
failure be the result of deterioration, or is there a possibility of com-
plete failure of a component during a coking cycle?
(8) Relative Cost Installed. What is the cost of acquisition and
initial installation of the concept using a 100-oven battery as a base
for comparing one concept with others? Costs may include but not
be limited to materials, fabrication of components, labor, downtime,
VII-3
-------
new or modified ancillary equipment, and provisions for utilities.
Engineering to retrofit to a specific battery will add to the cost.
(9) Effect on Coke Quality. Will this concept be free of adverse
effects on coke quality?
(10) Operation Complexity and Operator Options. Both man-
agement and workers'appreciate that best results are obtained when
a procedure is straightforward and uncomplicated with option-related
decisions. Are operator procedural options limited effectively by
the concept to avoid procedural errors?
(11) Availability of Expendable Materials or Components. Are
expendable materials, if used, readily available and are they likely
to remain so? Are expendable components readily and simply fabri-
cated, or do they require complex fabrication methods and facilities
which sometimes lead to delays?
(12) Maintainability. Can expendable items be replaced by
using familiar and durable tools? Can any of this work be accom-
plished while the door is on the door machine with proper heat shield-
ing for the maintenance crew? Can existing levels of skill of
maintenance personnel be used in the setup and maintenance of this
concept or will higher-level skills be required?
(13) Cleanability. Will the cleaning provision described in this
concept provide adequate cleaning without penalty of time or
reliability?
(14) Operator Skill Requirements. Can this concept be effec-
tively operational with the present levels of operator skill?
(15) Cost for Development. Development costs include engi-
neering, material investigations, pilot models, instrumentation,
field tests, field modifications, etc. These are the costs to develop
VII-4
-------
and prove the concept before it is engineered for retrofit to specific
batteries. How do these costs compare with those for other concepts?
Weighting of Evaluation Criteria
Weighting factors for the evaluation criteria were established
by mutual agreement between AISI, EPA, and Battelle personnel.
The weighting factor indicates the relative importance of the criteria.
The weighting factors are shown below (see also Figure VII-1).
Weighting Factor
(1) Relative Effectiveness to Lower 10
Emissions (Stop Smoke)
(2) Operating and Maintenance Costs 10
(3) Capability of Retrofit 9
(4) Life 8
(5) Environment and Safety 8
(6) Effect on Cycle/Schedule 7
(7) Sensitivity 6
(8) Relative Cost Installed 6
(9) Effect on Coke Quality 6
(10) Operation Complexity and Operator 4
Options
(11) Availability of Expendable Materials 4
or Components
(12) Maintainability 4
(13) Cleanability 2
(14) Operator Skill Requirements 2
(15) Cost for Development 1
Scoring Values
Scoring values used by evaluators were on a scale of 1 to 5 de-
fined below.
VII-5
-------
COMCEPTS
NO. I DESCRIPTION!
HOT ZOME SEALS __
COMMEMTS
LUTED SEAL.S
RESILIENT SEALS
INFLATABLE SEALS
COWTACT SEALS
WON-COM TACT SEALS _
FIGURE VII-1. EVALUATION WORKSHEET
-------
Score Definition
1 Poor: Most certainly believed to be unacceptable by
evidence, prior practice, or closely related
experience.
2 Fair: Believed to be unacceptable by somewhat related
evidence or experience, but not demonstrated by
prior practice.
3 Average: Not believed to be unacceptable by evidence or
experience but no strong feeling for accepta-
bility.
4 Good: Believed to be acceptable by somewhat related
evidence, experience, or prior practice.
5 Excellent: Most certainly believed to be acceptable by
evidence, prior practice, or closely related
experience.
For each concept grouping, each criterion was judged relative
to:
1. The above definitions of the scoring values, and
2. The other sealing concepts.
At a meeting with the evaluators, it was emphasized that concepts
were to be judged and not designs. Each evaluator was asked to
visualize the extent to which he believed it possible that skillful, ex-
perienced designers could create successful hardware systems repre-
senting the concept.
A Listing of the Evaluators
The establishment of evaluation criteria and the weighting
factors assigned to these criteria was completed by committee action.
However, the evaluation of concepts was done, with one exception, by
individual evaluators acting independently. The EPA elected to sub-
mit a consensus evaluation acting through the Project Officer. A
listing of the evaluators and their affiliations follows:
VII-7
-------
Evaluator's Name
A. M. Cameron
L. G. Gainer
T. R. Greer
D. S. Gregg
A. O. Hoffman
L. M. Hoopes
R. E. Maringer
J. C. McCord
J. G. Munson, Jr.
R. L. Paul
R. G. Phelps
E. E. Reiber
R. D. Rovang, et al
J. Varga, Jr.
Organization
Algoma Steel Corp., Ltd., SaultSte. Marie,
Ontario, Canada
U. S. Steel Corp., Clairton, Pennsylvania
Jones & Laughlin Steel Corp., Pittsburgh,
Pennsylvania
Empire-Detroit Steel Corp., Portsmouth,
Ohio
Battelle's Columbus Laboratories, Columbus,
Ohio
Republic Steel Corp., Youngstown, Ohio
Battelle's Columbus Laboratories, Columbus,
Ohio
Bethlehem Steel Corp., Lackawanna, New York
U. S. Steel Corp., Pittsburgh, Pennsylvania
Battelle's Columbus Laboratories, Columbus,
Ohio
Inland Steel Co., Indiana Harbor, Indiana
Battelle's Columbus Laboratories, Columbus,
Ohio
U. S. Environmental Protection Agency,
Research Triangle Park, North Carolina
Battelle's Columbus Laboratories, Columbus,
Ohio
Scoring Procedure
Each evaluator at EPA, AISI, and BCL considered each concept
relative to each criterion and marked a score in the appropriate
block (indicated by arrows) of the evaluation work sheet. Figure VII-1
is an example of this evaluation work sheet. Although estimates of
costs were not given, relative estimates were made by each evaluator
with the smallest cost requirements being most acceptable. Gener-
ally, the criteria relating to cost were evaluated by considering the
cost aspects of all concepts together at the same time. Evaluators
VII-8
-------
were asked for comments relating to concept alternatives. Upon
completion of each individual evaluation, the work sheets were
analyzed at Battelle.
Ranking of Concepts by
Each Evaluator
(1) Each individual score on each work sheet was multi-
plied by the appropriate weighting factor to produce
a base rating.
(2) Base ratings were totalled for each concept.
(3) Total base rating for each concept is reported for
each evaluator in Table VII-1.
(4) Ranking of concepts in order of preference was
established for each evaluator, as illustrated in
Table VII-2.
(5) Concepts are ranked for selected significant criteria
for comparison in Table VII-3 to aid interpretation
of other statistical results.
Interpretation of Statistical Results
(1) The averages in Table VII-1 do not indicate any strong
preference or even a consensus in favor of any single
family of concepts.
(2) All groups of evaluators had a low level of overall
confidence and feeling for the applicability of Concept
Families 1, 3, and 4 (Hot Seals, Resilient Seals, and
Inflatable Seals). Their rankings by the various
groups were almost completely in the 4th to 6th posi-
tions in a list of six.
If only the top three concept families are considered, their
rankings by the various groups of evaluators were as shown
on page VII-13.
VII-9
-------
TABLE VII-1. TOTAL BASE RATINGS
Group
and
Evaluator
AISI
A
B
C
D
E
F
G
H
AISI Ave.
AISI RANK
CONCEPTS
Hot Zone
1
207
234
161
292
300
253
289
318
257
4th
Luted
2
246
301
177
293
249
337
309
333
281
1st
Resilient
3
294
259
121
244
212
332
223
254
242
5th
Inflatable
4
243
286
121
218
210
322
178
223
225
6th
Contact
5
379
316
146
264
300
293
283
226
276
2nd
Noncontact
6
320
248
185
278
318
312
314
215
274
3rd
EPA
I
EPA RANK
BCL
J
K
L
M
N
BCL Ave.
BCL RANK
Overall Ave.
OVERALL RANK
218
5th
248
416
369
298
163
299
6th
269
4th
304
1st
253
389
366
321
247
315
4th
295
2nd
227
3rd
326
339
311
329
290
319
3rd
268
5th
219
4th
311
338
269
323
266
301
5th
252
6th
240
2nd
366
371
342
361
341
356
1st
302
1st
215
6th
303
355
286
252
339
327
2nd
289
3rd
VII-10
-------
TABLE VII-2. RATING OF CONCEPTS
Group
and Hot Zone
Evaluator 1
AISI
A 6
B 6
C 3
D 2
E 2
F 6
G 3
H 2
Frequency 2 3X
3 2X
3 3X
AISI Rank 4
EPA RANK 5
BCL
J 6
K 1
L 1
M 6
N 6
Frequency 1 2X
6 3X
BCL Rank 6
Ove ra 1 1
Frequency 1 2X
2 3X
3 2X
5 IX
6 6X
Overall
Rank 4
CONCEPTS
Luted
2
4
2
2
1
3
1
2
1
1 3X
2 3X
3 IX
4 IX
1
1
5
2
2
5
5
2 2X
5 3X
4
1 4X
2 SX
3 IX
4 IX
5 3X
2
Resilient
3
3
4
6
5
4
2
5
3
2 IX
3 2X
4 2X
5 2X
6 IX
5
3
2
5
4
3
3
2 IX
3 2X
4 IX
5 IX
3
2 2X
3 5X
4 3X
5 3X
6 IX
5
Inflatable
4
5
3
5
6
5
3
6
5
3 2X
5 4X
6 2X
6
4
3
6
6
4
4
3 IX
4 2X
6 2X
5
3 3X
4 3X
5 4X
6 4X
6
Contact
5
1
1
4
4
2
5
4
4
1 2X
2 IX
4 4X
5 IX
2
2
1
3
3
1
1
1 3X
3 2X
1
1 5X
2 2X
3 2X
4 4X
5 IX
1
None on tact
6
2
5
1
3
1
4
1
6
1 3X
2 IX
3 IX
4 IX
5 IX
6 IX
3
6
4
4
5
2
2
2 2X
4 2X
5 IX
2
1 3X
2 3X
3 IX
4 3X
5 2X
6 2X
3
VII- 1 1
-------
TABLE VII-3. COMPARISON OF RANKING OF CONCEPTS FOR SELECTED CRITERIA
IS)
CONCEPTS
Criteria
All Criteria
Effectiveness to
Reduce Emissions
Operating Cost
Life
Sensitivity
Availability of
Materials
Maintainability
Operator Skill
Weighting
Factor
-
10
10
8
6
4
4
2
Hot-Zone
1
4
4
5
3
2
3
3
2
Luted
2
2
2
3
1
1
6
1
1
Resilient
3
5
2
4
4
4
4
5
4
Inflatable
4
6
1
6
5
6
5
6
6
Contact
5
1
5
1
2
5
1
2
3
Noncontact
6
3
3
2
2
3
2
4
5
-------
Evaluation Ranking by Concept-Family Name and Average
Group 1st 2nd 3rd
AISI Luted (281) Contact (276) Noncontact (274)
EPA Luted (304) Contact (240) Resilient (227)
Battelle Contact (356) Noncontact (327) Resilient (319)
Again, there was no concensus and no firm conclusions could be drawn
from these independent evaluations. Subsequent "debriefing" of the
Battelle evaluators indicated that their higher rating for contact seals
was influenced by (a) the experience of seeing a totally tight, existing
metal-to-metal sealing system, and (b) their judgment that a metal-
to-metal contact seal could be developed to overcome the shortcom-
ings of the existing seals.
(3) Examination of the selected criteria given in
Table VII-3 indicates that, on the average and in
terms of effectiveness to reduce emissions, the
evaluators gave a higher ranking to luted seals
than to contact seals. It is indicated (in Table VII-3)
that concerns over "Operating Cost" and "Availability
of Materials" kept the luted-seal family of concepts
from getting a higher scoring value, especially from
the AISI evaluators.
Battelle's Evaluation of Concept Families Via Comparison
With Seal-System Specifications and Functional Requirements
The foregoing judgmental evaluation of concept families by the
AISI/EPA/BCL groups was completed at about the beginning of the
last quarter of this research program. This evaluation was followed
by research effort to gain further insights into the practical aspects
of the various concept families. In this effort, the comments of the
AISI and EPA representatives gave the researchers useful inputs.
For example it was learned that some of the AISI evaluators regard
the addition of water cooling to the end closures of coke-oven doors
as being somewhat impractical if additional problems are created
by any interruption in the flow of water.
VII-13
-------
Battelle's Evaluation and Comments
As a means for presenting the results of the final evaluation,
an identification key was developed to indicate the consensus on each
of the factors used to grade the various concept families. The key
is given in Table VII-4.
TABLE VII-4. IDENTIFICATION KEY FOR CONCEPT-FAMILY EVALUATION
Level of Confidence for
Successful Final Devel-
opment
Range
and Performance
of Probability
HI ^
^^^^^^B
d
Probable Possible Unlikely
100-90% 90-40% 40-10%
Ixxx 1 I ? I
Not Possible Unknown
<10% Unknown
The results of Battelle's final evaluation are shown in Table
VII-5. The evaluation was in terms of both the technical specifications
and the functional requirements. In this table, an asterisk indicates
that there is pertinent discussion in the following text.
The evaluation summary shown in Table VII-5 shows that the
Battelle evaluators have a high level of overall confidence in the
successful final development and performance of only one concept
family; i.e,, contact seals. Battelle's use of the term "contact
seals" can be considered to be the same as significantly upgraded
metal-to-metal seals. In this second and more critical evaluation
(see following comments), most of the concept families were down-
graded in rating relative to the first evaluations completed by the
AISI/EPA/BCL. Only contact seals received a 95 to 100 percent
probability rating for each of the specifications and requirement
criteria. Stated another way, Battelle researchers believe that
existing sealing systems, although often called inadequate, are really
marginally successful, and can be improved.
The luted-seal concept family appealed to the AISI and the EPA
evaluators. Battelle researchers are aware that foamed-in-place
sealants have the potential for completely eliminating emissions from
coke-oven doors. In addition, there is a possibility that a simple
procedure for foam-sealing of doors could be developed. However,
this remains only a possibility at this time. Therefore, the luted-
seal concept was downrated by the Battelle investigators. Battelle
researchers are, however, recommending further research and
testing of this concept family.
VII-14
-------
TABLE VII-5 A SUMMARY OF BATTELLE'S EVALUATION OF THE SEALING-CONCEFT FAMILIES
BASED 01 SELECTED SPECIFICATIONS AND FUNCTIONAL REQUIREMENTS
(Jl
Hoc-Zone Seals
(,) Temperature Tolerance9 (Ac
V o 200-300 C (400 co 600 F)
Operating Level9)
Concept Fully Number and Name
Luced Seala Reilllenc Seali
2 3
Inflatable Seal!
6
LZI
Contact Ssala
5
Noncontacc Seala
6
(2) Heat Excursion Tolerance9 __
Short Periods Without ^B Essl U5SLI |jJ iisssi ^B
Destruction9)
(3) Increased Gap-Cloaure Capa-
bility9 (Can Close a 12 7 nm-
(0 5- Inch) Inward Bow on a
Warped Jnob')
(4) Automatic Gap-Closure Capa-
bility9 (No Manual Ad-
luacnenti9)
(5) Resistance to Corrosion and
Chemical Attack9 (Withstand
Heated Solvents. Tars9}
(6) Total-Failure Proof9 (Suscepti-
ble to Sudden Failure9)
(7) Avoids New Cleaning Problems?
Technical Specif lea clona
Summary Racing
Functional Requirements In Question Form
(8) Retrofttable9
(9) Compatible with Existing Work-
Ing Machinery
^^<^ *
^
'
LZI
LZI
LZI
LZI
LZI
' LZJ
^ LZI*
^ ^
^ LZI
Fallible Unlikely
-!
^
n
|xxx|
LZI
[xxx|
Not Ponlble
si
^
*
FroDable
|xxx|
n
^
|xxx|
Not Fallible
S
^
auiate With Existing Seals9 ^^ »SSi 1 1 | 1 ^H £§£.1
(11) Dependable and Ropeaeable9
(12) Avoids Any Additional En-
vironmental Problems7
m
a
(13)
Avoids Adverse Effects on
Product Quality9
Evaluation Sumaary Baaed om
Functional Requirements
OVERALL EVALUATION
Level of Confidence In Successful
LZI
Unknown
en
UNKNOWN
^
Foiilbli
POSSIBLE
o
Unlikely
o
UNLIKELY
n
Unlikely
1 XXX 1
NOT POSSIBLE
a
Frobable
PROBABLE
*
^
Fallible
NOT POSSIBLE
Future Developaent and Perforaance
-------
The four other concept families were given essentially low
ratings. It is suggested that these concept families be re-evaluated
if the temperatures of the end closures are lowered (or if a company
elects to test water-cooling of the equipment) and if new heat-resistant
materials become available.
Contact Seals (Rating: High Probability
of Successful Development and
Performance
The contact-seal concepts shown in Chapter VI (pages VI-40 to
VI-53) all emphasize seal-element mechanical flexibility in one way
or another. The objective of this flexibility is to enter and seal sec-
tions of jambs having a degree of inward bowing that cannot be
sealed by existing seal elements. Statistical information on the
range and distribution of jamb-distortion measurements across the
coke-producing industry are not available. However, the measure-
ments given below in Table VII-6 (taken from Figure IV-Z1, page
IV-40) are believed to be an example of the upper limit or maximum
level of distortion to be expected. This belief is based on the high
level of door emissions being encountered by the company that sup-
plied these data.
It is possible to consider a flexible metal seal that would con-
form automatically to the indicated gross distortion range of from
-0.69 to +0.28 inch, or a total range of about 1 inch (about 25 mm).
This, however, is regarded as being impractical in terms of proba-
ble costs/benefits determinations. The specifications, therefore,
were lowered to a maximum of 16/32 inch (12.7 mm) for an inward
jamb bow and 8/32 inch (6.3 mm) for a maximum outward bow. Even
these specifications are believed to be on the high side of the majority
requirements.
All of the flexible seals shown in the contact-seal family of
concepts assume that (a) the seals will be "loaded" by deflecting the
seals against the jambs as the doors are mounted on ovens, and (b)
there will be no need for helper or auxiliary springs. The amount
of deflection of the corners of the spring seals will depend upon the
setting of the corner stop-bolts on the doors. Ideally, door-stop
settings should be uniform for every door on a battery. This, how-
ever, requires a rather large amount of allowable flexibility in the
VII-16
-------
TABLE VII-6. RANGE OF JAMB DISTORTION MEASURED ON EIGHT JAMBS AT A
BATTERY HAVING SERIOUS DOOR-EMISSIONS PROBLEMS
(Results are presented In 32nds of an inch measured inward (-) or outward (+)
from a straight line drawn between the upper and lower jamb corners.)
Maximum Directional Distortion
Per Jamb Side
Jamb
Number
1
2
3
4
5
6
7
8
Left
5
1
22
12
5
0
11
0
Side
0
0
0
0
0
7
0
6
Right
(-)
0
1
12
10
0
0
15
0
Side
(+)
8
5
0
4
6
8
0
9
Maximum Directional
Distortion Per Jamb
(-)
5
1
22
12
5
0
15
0
(+)
8
5
0
4
6
8
0
9
Maximum
Total Horizontal
Displacement
In 32. ids
13
6
22
16
11
8
15
9
In Inches In MM
0.41
0.19
0.69
0.50
0.34
0.25
0.47
0.28
10.4
4.8
17.5
12.7
8.6
6.35
11.9
7.1
All Jambs -22/32 +7/32
-15/32 +9/32 -22/32 +9/32
-0.69 inch + 0.28 inch
Average 0.39
9.9
spring-seal element. For example, if the specification distances for
both inward and outward bowing occur on the same side of a jamb,
then the maximum seal deflection required is 19 mm (0. 75 inch).
This 19-mm deflection of a seal could be set as a development
target, but there is no evidence that any extreme combination of
inward and outward bowing exists on any one jamb. If, on the other
hand, the seal-deflection limit were set at 13 mm (0.5 inch), then
with door-stop adjustments, 7 out of the 8 jambs previously described
could be sealed. As one example, the door-stop setting on Jamb 4
(in Table VII-6) could be set at 10 mm (3/8 inch) at all four corners.
Then a maximum seal deflection of 13 mm (1/2 inch) would accommo-
date both the inward and outward (3 mm) distortion on this jamb.
This is considered only a minor deviation from the specification that
there be no need for manual adjustment of upgraded seals.
Taking the view that the example jambs are upper-limit exam-
ples, it was decided to set a 13-mm (0.5-inch) seal deflection as a
practical requirement in the following spring calculations. This
VII-17
-------
represents a deflection distance four times that considered possible
in the existing spring-type seals, and would probably seal the vast
majority of existing coke-oven doors.
Some of the contact-seal concepts have the flexible seal angled
toward the door plug (pages VI-44 and -45), and some seals are angled
away from the door plug. Internal pressure in the oven will attempt
to push the latter type away from the jamb in a minor way. For a
peak pressure of 100 mm (4 inches) of t^O at the bottom of the seal
on an.oven, the internal pressure on the seal is about 0.96 KPa
(0. 14 psi). To compensate for this backpressure, it is estimated
that it will be necessary to have a minimum contact pressure of
0. 18 kg/cm (1.0 pound per linear inch) at the bottom of an inward
bow. This would require an additional corner deflection of only
0. 5 mm (0.02 inch) for most of the spring seals that can be considered
for this application. Because this is only a minor increase in re-
quired deflection in the jamb corners, the effect of internal pressure
was disregarded in the following calculations.
The amount of force required to mount future doors, to the point
where the door-stops seat on the jamb, will vary depending on the
spring-seal configuration, the spring thickness, the amount and direc-
tion of distortion of the jamb, and other considerations. With highly
flexible sealing elements, the mounting force is expected to be much
lower than present mounting force. It is probable that with more-
flexible sealing elements, doors will be latched with only enough
force to seat the corner stops'. Calculations of the expected effective
latching force are included in the following paragraphs.
Example of Spring-Seal Calculations. Because one of Battelle's
major recommendations centers on the development of more-flexible
and (more-heat-resistant) metal-seal elements, a simplified example
calculation was completed to judge the feasibility of this approach.
The purpose of this feasibility calculation is to estimate the
probability that sufficient required mechanical deflection (13 mm)
can be achieved (within reasonable bending-stress limits) with a
spring thickness that will be capable of resisting both abuse and
thermal stresses.
There is a wide choice of types of springs that can be considered
for this application. Some of these are:
VII-18
-------
Simple cantilever spring (parabolic shape).
Tapered-end cantilever springs (ends tapered to achieve
a circular arc).
Constant-strength cantilever springs (thickness tapered
toward the loaded end).
Leaf springs (shown on pages VI-48 to -52).
Curved-end cantilever springs.
The last four types of springs listed above are designed to
achieve an increased amount of deflection for a given limit of allowable
bending stress. Of these types, Battelle researchers elected to eval-
uate feasibility using a single, curved-end cantilever spring of con-
stant cross section. This type of spring is shown in Figure VII-2.
This figure also shows the full-scale dimensions of the gas passage in
one type of door/jamb relationship into which a spring seal is to be
fitted. For calculation purposes (and as shown in Figure VII-2), the
curved-end spring is taken to have a 3-inch straight section and a
1-inch radius on the curved section. The effective length of the
spring is 6. 14 inches (from the loaded end to the point where the
spring is fastened).
The basic equations for deflection and maximum bending stress
for curved-end cantilever springs are taken from Alexander Blake's
article in Machine Design, July 23, 1959, page 151. These equations
in U. S. units are as follows:
Spring-End Load on Spring End (Ib) x
Deflection = (Radius of Curved End-In.)3 x Constant*
(Inches) Modulus of Elasticity (psi) x Moment of Inertia (in.4)
Maximum 6 x Load on Spring End (Ib) x
Mechanical = Effective Spring Length (in.)
Stress (psi) Spring Width (in.) x (Spring Thickness, in. )2
For the purpose of this example calculation, the maximum allow-
able bending stress was taken to be half of the yield strength of any
candidate high-performance spring material at 320 C (600 F).
This constant is taken from a table in Blake's paper. The constant vanes with the magnitude of
the ratio of the straight portion of the spring length to the radius of the curved end.
VII-19
-------
Jamb
Deflection direction
Door
FIGURE VII-2.
FULL-SCALE CROSS SECTION OF A GAS-PASSAGE
AREA SHOWING THE INSTALLATION OF AN
EXAMPLE CURVED-END CANTILEVER
SPRING SEAL
VII-20
-------
Age-hardened Incoloy Alloy 903 has a yield strength of 150, 000 psi
at this temperature. Many alloys (including iron-based superalloys)
have equivalent or higher yield strengths at this temperature. Data
for Incoloy Alloy 903 were used in the following calculations pri-
marily because of this alloy's low coefficient of thermal expansion.
This characteristic has no effect on the deflection performance of a
spring, but it is very favorable in terms of lowering the thermal
stresses in springs operated in heated applications. The importance
of a low thermal expansion is discussed in later paragraphs.
Assignment of a maximum allowable mechanical stress of only
75, 000 psi is a conservative approach. However, many other factors
must be considered, such as spring relaxation, creep under load,
etc. Selection of a relatively low stress limit minimizes these other
problems.
Assumptions. In the spring-seal calculations, the assumptions
(along with some comments) are as follows:
(1) The curved-end spring data are:
Effective length of spring (contact point to
fastening point) is 6. 14 inches.
Length of the straight portion of the spring
is 3. 0 inches.
The radius of'the curved end is 1.0 inch.
The constant (in the deflection equation) is
50, as taken from Blake's paper.
The spring thickness is a constant over the
length of the spring.
(2) To simplify the calculations, the width of the spring
seal is taken to be 1.0 inch. In practice, the width
of the spring will be about equivalent to the height
of the jamb.
(3) Maximum allowable mechanical bending stress is
75,000 psi (fiber stress).
(4) It is assumed that elementary be am-deflection theory
for small deflections is also correct for "large"
deflections. The deflections being considered in the
VII-21
-------
subsequent calculations are large deflections. In
practice, springs with a large ratio of width to thick-
ness have an increased flexural rigidity over and
above those calculated. On the other hand, a spring
being subjected to a large deflection is more flexible
than calculated. It will require a detailed analysis
to relate these factors for curved-end springs, but
if simple cantilever springs of constant thickness were
used in this application, the large ratio of width to
thickness (jamb height/spring thickness) would in-
crease flexural rigidity about 10 percent*. On the
other hand, the large deflection (say 13 mm or
0. 5 inch) would decrease spring rigidity by 30 percent
or more. Overall, it is expected that the spring seals
will be more flexible, per unit of stress, than cal-
culated. Simple cycling-bending test equipment has
been developed for characterizing the mechanical
properties of spring materials, including high tempera-
ture performance. **
(5) The modulus of elasticity is taken at 21, 800, 000 psi
at 320 C (600 F). Many alloys have a modulus at this
temperature in the range of 24, 000, 000 to 26, 000, 000
psi. Changes in modulus, however, have only a
minor effect on spring thickness.
(6) Spring-seal deflection due to the internal gas pressure
against the back side of the seal is disregarded.
(7) Distortions (in or out) on the jamb are all gradual;
i.e., there are no abrupt changes in distortion along
the height of the jamb. The maximum change ex-
pected is about 4. 3 mm (0. 17 inch) inward deflection
per foot of jamb height.
(8) Sealing can be effected by metal contacting the
residual jamb deposits, even at low contact pressure.
(9) Seal deflection is directly proportional to the load
on the end of the spring.
Wahl. A. M., Mechanical Springs, McGraw-Hill Book Co., 1963.
Weissman. G. F., Wonsiewicz. B. C., "Characterization of the Mechanical Properties of Spring
Materials", Journal of Engineering for Industry, August, 1974.
VII-22
-------
Using the two foregoing equations, various calculations were
made to show the relationships and trends among four interrelated
elements. The results are presented in Table VII-7. Only those
results within the boundaries of interest are shown.
The example calculations indicate that:
(1) From a mechanical viewpoint, it is highly probable
that a spring sealing element can be developed to
deflect 13 mm (0.5 inch) and still have a thickness
that will resist being damaged. This deflection
can be accomplished without exceeding a mechani-
cal bending stress of 75, 000 psi.
(2) As would be expected, large deflections (0.75 inch)
with this simple type of spring (within the space
limitations described) require very thin seal ele-
ments. Also, thin springs cannot exert much
contact force on the jamb.
(3) If the maximum required deflection on the door
seals of all the ovens on a battery is less than
13 mm (0.5 inch), the deflection distance (door-
stop setting) can be lowered with a proportional
general decrease in the bending stress level in
the seal elements; i.e., the design safety factor
would be increased.
Calculations 11 through 15 in Table VII-7 were made to estimate
the total spring-loading force required for individual doors. Along
any warped jamb there will be a variation in the contact force at the
seal depending on the incremental deflection of the seal. On a per-
fectly straight jamb with assumed dimensions of 0. 6 x 3. 6 m (2 feet
by 12 feet), a 13-mm (0.5-inch) deflection of the seal at each corner
with a 2. 8-mm (0. 111-inch) thickness of curved-end spring seal
would require a minimum door-latching force of 3810 kg (8400 pounds).
On a jamb with considerable inward bowing, the required minimum
door-latching force would be decreased. For example, Jamb 7 in
Table VII-6 has a maximum inward bowing of 8.7 mm (11/32 inch) on
the left side of the jamb and a maximum 11.9 mm (15/32 inch) inward
bowing on the right side. The average amount of inward bowing is
about 3. 7 mm (1/8 inch, 0. 147 actual). Under these conditions the
required latching force is 2676 kg (5900 pounds) with a 13-mm (0. 5-
inch) seal-element deflection in the jamb corners. The range of
VII-2 3
-------
TABLE VII-7. SUMMARY OF EXAMPLE SPRING CALCULATIONS
Selected Spring
Calculation Deflection
Number MM
1 10.2
2 10.2
3 10.2
4 10.2
5 12.7
6 12.7
7 12.7
8 12.7
9 19.0
10 19.0
Inch
0.40
0.40
0.40
0.40
0.50
0.50
0.50
0.50
0.75
0.75
Selected Unit
Loading Calculated
Pounds/
Spring
KG/ CM Linear Inch MM
0,9
1.8
3.6
6.2
0.9
1.8
3.6
4.5
0.9
1.8
5
10
20
35
5
10
20
25
5
10
Calculated Unit
11 12.7
12 10.2
13 5.1
14 2.5
15 0.5
0.50
0.40
0.20
0.10
0.02
Loading
4.5
3.6
1.8
0.9
0.2
25
20
10
5
1
1.8
2.2
2.8
3.4
1.6
2.1
2.6
2.8
1.4
1.8
Thickness
Inch
0.070
0.088
0.111
0.135
0.065
0.082
0.103
0.111
0.057
0.0715
Selected
Spring
2.8
2.8
2.8
2.8
2.8
Thickness
0.111
0.111
0.111
0.111
0.111
Calculated
Bending Stress
kPa
259,000
328,000
413,000
487,000
300,000
378,000
479,000
517,000
391,000
496,000
Psi
37,600
47,600
60,000
70,700
43,600
54,800
69,450
75,000
56,700
72,000
Calculated
Bending
517,000
413,000
207,000
103,500
21,000
Stress
75,000
60,000
30,000
15,000
3,000
-------
contact pressure in this instance would be 4. 5 kg/cm (25 pounds per
inch) across the top and bottom of the jamb decreasing to 0.28 kg/cm
(1.6 pounds per inch) at the bottom of the jamb section bowed inward
11.9 mm (15/32 inch).
Thermal Stresses and Thermal Buckling. As noted in Chapter
IV, there is some evidence of thermal buckling on the upright sealing
edges of one type of existing seals. This buckling is believed to be
due to excessive stress resulting from the restraint of the thermal
expansion of these seals. These seals are all bolted into place on
the doors at ambient temperature and are then placed against heated
jambs. There is also a possibility of concentrated mechanical load-
ing at the point of buckling, which would further contribute to
buckling.
It is not possible to analyze the thermal-stress aspects of the
present and upgraded metal-to-metal seal designs without getting
detailed field data (temperatures and strain measurements) and
completing an analysis and modeling program. However, some
general comments can be made about thermal stresses and thermal
buckling in spring-seal elements. For example, for a given thermal
stress, the ability of restrained sheet material to resist buckling is
a function of the square of the thickness. Doubling the thickness will
increase the resistance to buckling by a factor of four. Therefore,
thin sealing strips are to be avoided.
In the comparison of different metals and alloys, a lower
thermal stress is favored by a low coefficient of thermal expansion,
a higher thermal conductivity (a lower temperature gradient for the
same geometry), a lower modulus of elasticity, and a larger number
for Poisson's Ratio. One form of equation used to evaluate thermal
stress is as follows:
(Coef. of Thermal Expansion) x
Thermal Stress = (Thermal Gradient) x (Mod, of Elasticity)
2(l-Poisson's Ratio)
In the preceding spring-deflection calculations, it was found, as an
approximation, that a 2. 8-mm (0. 11 l-inch)-thick strip of age-hardened
Incoloy Alloy 903 could be selected for a spring material to reach a
required 13-mm (0. 5-inch) deflection. Presently some of the existing
spring-type seals are using 3-mm (1/8-inch)-thick austenitic stainless
steel strip. A comparison of the mechanical and thermoelastic
VII-2 5
-------
properties of Alloy 903 and Type 304 stainless steel (as one example
of austenitic stainless steel) is as follows:
Alloy Designation
Incoloy Alloy 903, Type 304 Stainless Steel.
Age-Hardened Cold-Worked
Yield Strength (0.2<7o Offset)
at320C(600F);MPa(psi) 1034 (150,000) Up to 827 (120,000)"
Coefficient of Thermal Expansion;
10-6cm/cm/C(10-6in./in./F) 2.67 (4.0) 6.40 (9.6)
Tensile Modulus of Elasticity;
GPa(!06psi) 150 (21.8) 172 (25.0)
Poisson's Ratio 0.23 (0.23) 0.32 (0.32)
Thermal Conductivity;
Watt/mK(Btu/hr/Ft2/in./F) 1.56 (130) 1.53 (128)
Introducing these data into the thermal-stress equation and
solving indicates that under identical conditions of heat input and
mechanical restraints, the 3.2-mm (0. 125-inch)-thick stainless steel
will develop a 3.5 times higher internal stress than 2.8-mm (0. 111-
inch)-thick Alloy 903 strip. That is, Alloy 903 will develop only about
one-third of the thermal stress, and will be able to tolerate three
times more stress without failure.
The properties of Incolqy Alloy 903 are unusually favorable
for minimizing thermal stress, but it is known that many other alloys
that can be considered for this application develop significantly
lower thermal stress than austenitic types of stainless steels. In
addition, there are alloys that can tolerate a significantly higher
stress than austenitic stainless steels. The final selection of the
spring-seal material must include consideration of numerous factors
including creep and relaxation characteristics, fabrication costs,
and material availability. In addition, it is possible to lower the
thermal stress level by design considerations. For example, if the
spring-seal element is permitted some thermal-expansion movement,
particularly in the door-height direction, the thermal stresses would
' This is a handbook value for very severely cold-worked Type 304 stainless steel. It is estimated that
the actual yield strength of the austenitic stainless being used on existing seals is in the 275 to 414
MPa range (40, 000 to 60.000 psi). Alloy 903 would then have a 2.5 to 3.75 times higher yield
strength than the stainless steel in existing seals.
VII-2 6
-------
be lowered significantly. This, in turn, would permit consideration
of other classes of alloys.
Feasibility Summary Statements. The general approach of
developing upgraded and retrofitable metal-to-metal contact seals
that will minimize emissions is feasible. The reasonableness and
probability of success of the contact-seal concept are high. Battelle
does not, however, recommend that an attempt be made in only one
step to design an upgr-aded seal (or jamb) and then subject it to per-
formance and endurance trials. Instead, it is suggested that more
fundamental information and analysis are required before design
should begin. In addition, Battelle recommends a program con-
sisting of mathematical and physical modeling prior to designing a
full-scale test unit. For example, it will be necessary to evaluate
the stress considerations in the seals at the corners of doors. The
interest in the full-scale unit would be both in its sealing performance
and in its dimensional stability relevant to existing equipment. With
this approach, there should be no need to complete an endurance
check of the test equipment prior to starting the retrofit of existing
coke-oven doors. Early physical measurements completed on full-
scale test units will determine whether the goal of dimensional
stability has been achieved.
It should also be noted that Battelle is not recommending the
curved-end cantilever spring seal for coke-oven doors. This geom-
etry was only an example sele'cted for the purpose of completing
preliminary calculations. The entire contact-seal family of concepts
(with additions) should be evaluated prior to making a selection of
seal geometry.
Other Comments on Contact Seals. Various representatives
of the AISI Task Force have expressed concern about (a) the problem
of cleaning of some of the contact-seal concepts and (b) how resistant
such seals are to abuse and damage. Some of the aspects of cleaning
are discussed in Chapter IV. It is expected that the cleaning of
upgraded metal-to-metal seals will not be more difficult than
with the present seals. In some instances it may be necessary to
develop variations and improvements on the present cleaning methods
to accommodate some contact-seal concepts, but it is not foreseen
that this will present any particular problem. The important con-
siderations are to clean the tar from the bottom of the gas passage on
VII-2 7
-------
every cycle and to remove any coal that may have fallen into the gas
passage.
The ability to withstand abuse is more a function of the tough-
ness of the seal material than other considerations. Abuse (such as
striking the jamb hooks with the sealing edge) can be eliminated by
upgrading the door-spotting equipment and procedures. Laboratory
tests will indicate whether reinforcement is required in the areas
where workers use hand tools for cleaning the gas passage.
Luted Seals (Rating: Possibility of
Successful Development and
Performance)
In hindsight, Battelle's use of the term "luted seals" could
more appropriately be described as the use of sealants foamed in
place on the jamb or door before a door is mounted, and foaming or
other materials that are injected into the gas passage or into a
special channel after the door is mounted on the oven.
The only possible way to evaluate and/or bypass the uncertain-
ties of this general approach would be through a laboratory testing
program or through a combined laboratory and coke-plant testing
program. As previously noted, all Battelle evaluators have reserva-
tions on various aspects of foam sealing, and have reservations
particularly in terms of confidence that all of the specifications and
functional requirements eventually can be met.
Rigid and resilient foamed materials consist of a mass of gas
bubbles trapped in a rigid or resilient matrix. Many types of plastic
materials (and some inorganic materials) are commercially avail-
able in foam form. These materials can be produced in either an
open-pore or closed-pore form. Some of the polymer materials char
rather than melt when exposed to high temperatures. Phenolic foam
is one example. Generally, the manufacture of foams consists of
mixing of selected liquids (including a gas-forming agent) and inject-
ing the mixture into a mold. Depending on the type of foam, the mold
can be either hot or cold. With the immediate action of the gas-
producing agent, the material swells rapidly to fill the mold. Once
fully expanded, the foam is solidified by either internal chemical
reactions (polymerization, often exothermic in nature), by heat
VII-2 8
-------
curing, or by hydration (as for inorganic materials). In some in-
stances, the foam can be generated mechanically by injecting gas
from a separate source.
In concept form, it is possible to visualize the injection of a
foamable liquid into a hot gas passage after the door is on the oven.
Within a minute or two, the injected liquid could (in theory) foam and
fill the entire gas passage area with a closed-pore foam. Additionally,
the foam could fill much of the space between the door plug and the
oven walls. If this foam could be rapidly stabilized and would only
char when heated (rather than melt) the present flow of dust, tar, and
volatiles to the metal sealing area would be minimized. The unknowns
are many, including (a) the possible problems in removing the foam
materials after a completed cycle, (b) lack of information on which
foamable materials will resist the coal-tar chemicals and the tem-
peratures to be encountered, (c) whether there will be an outward
movement of gases to a leaking metal seal along the interface between
the foam and the coke-oven wall, and (d) how much foam would leak
out (before solidification) between the present gaps between the metal
seal and the jambs. The resistance of foamed materials to coal-tar
chemicals and coke-oven temperatures could be tested in operating
ovens. The possible leakage of gas past the foam sealant, and the
level of leakage of foam through gaps, could be determined in labora-
tory equipment.
The use of commercial foamable plastics can be expensive in
this application when it is taken into consideration that a door is
mounted about 500 times a year and there are over 25, 000 doors.
In the evaluation of this concept family, Battelle researchers
assumed that a low-cost foaming material would be developed for
this application, and that the use of the foam would not introduce any
additional environmental problems. As a possible example, it was
noted that asphalt has been used to produce a rigid closed-pore foam
(U. S. Patent 2, 901,469). If asphalt can be foamed, it is thought
that perhaps coal-tar pitch can be foamed. It is not known whether
a coal-tar pitch foam can be cheaply and rapidly solidified once it
has been foamed in place, or whether the stabilized foam would melt
(rather than char) upon being heated. In a related application, the
Porter Paint Company (Louisville, Kentucky) sells a coal-tar epoxy
as a coating material that solidifies after application.
In evaluating the potential for success of developing effective
foam sealants for (a) use between the door jamb and the mating surface
VII-2 9
-------
on the coke door and (b) use in the gas passage inside of doors,
Battelle's final consensus was to consider only sealants between
mating metal surfaces. The major negative judgment against gas
passage seals was that the effectiveness of the internal seal would
still depend upon the effectiveness of the existing metal seals. A
gap in the metal-to-metal sealing arrangement would encourage leak-
age of emissions past the foam sealant. This same reasoning was
also used to downgrade the rating of the hot-zone concept family. It
is for this reason that no discussion of hot-zone sealing is included in
this chapter.
It was judged, however, that there was considerable potential
for a successful development for those concepts that placed a suitable
sealant (foamed or unfoamed) between the jamb and a retrofitted
mating plate on the door before the door is mounted. The final eval-
uation of this approach is the one shown in the foregoing summary
Table VII-6. In this table reservations are noted as to whether an
approach can be developed to accommodate jambs that are severely
bowed. Also, it is not known whether it would be possible to avoid
new cleaning problems, and whether the procedure would be truly
dependable and repeatable.
It is recommended that the concept of placing a sealant be-
tween the jamb and a fixed door plate (as shown on pages VI-11 to -14)
should be studied in an experimental research program. It should be
noted, however, that in most instances Battelle researchers assumed
that a steel fixture would be mounted on warped jambs to (a) present
a straight surface to the door plate and (b) hold the sealant material.
Further consideration of this approach, however, goes full circle into
considering a fixture on the jamb to present a straight surface to the
existing seals and again introduces concerns as to whether an overlay
plate would promote or retard continued jamb warpage. Therefore,
it is judged that any research program on sealants should be directed
to being effective with existing jambs. This is the major reason for
the interest in swelling sealants - to close gaps.
Considering the fact that any sealant held between the jamb and
the mating plate on a door will be heated in air, it is to be expected
that the sealant will harden. With most organic foam materials, it
is probable that upon being heated the material will release emissions
into the atmosphere. To bypass this possible environmental problem,
it is necessary to consider inorganic materials for this application.
(The use of filled phenolic foam may be an exception to this judgment.)
VII-30
-------
The fact that the sealant will harden during one cycle is not considered
detrimental, provided the material remains under compression and
does not crack, deteriorate, or adhere to or corrode the metal parts.
Possible sealants for this application are foam concrete, gypsum
foams, and some foam-clay refractories using a mixture of clay and
gypsum. In some instances, these materials are available commer-
cially*, and they have been sprayed on vertical surfaces. However,
these are all water-based. This introduces concern for thermal
shocking the gray iron jambs. In this regard, it should be noted that
wet mud slurries are-presently being hand plastered against the out-
side of gray iron jambs and door frames on some older ovens that
are in operation. Here again, it is noted that the drop in the surface
temperature of a jamb on meeting a wet sealant can be measured
and the thermal stresses can be calculated.
The chuck doors on the pusher side of batteries might serve
well as experimental stations for the evaluation of sealants.
Resilient Seals (Rating: Unlikely to Result
in Potential Successful Development and
Performance)
This family of concepts is evaluated as unlikely in terms of
potential successful development. Shortcomings appear in meeting
the specifications on (a) acceptable tolerance for normal operating
temperatures (water cooling not considered), (b) gap-closing ability,
and (c) conformity to standard cleaning methods.
At 320 C (600 F) and higher, polymerized materials lose weight
slowly, become soft or brittle, and generally change in characteris-
tics with time. The advantages of "high-temperature plastics" are
mainly in low rates of deterioration at high temperatures when com-
pared with conventional plastics, but the high-temperature plastics
still deteriorate.
In the search for suitable nonmetallic sealants and materials,
Battelle considered many commercial elastomeric (rubber-like)
materials and reinforced high-temperature resins. The interest in
the reinforced resins at one time was in the possibility of using
* Rivkind. L. E.. "Improved Technology for Rigid Inorganic Foams", Journal of Cellular Plastics.
July. 1961. p. 329.
VII-31
-------
this fairly flexible type of material as a substitute for the relatively
stiff steel beam used to hold the metal sealing edge on the existing
fixed-edge design of seal. As one example, Monsanto was contacted
for information on their Skybond 703 product, a high-heat-resistant
resin which when reinforced with fiber glass has "excellent strength
retention up to temperatures of 350 C (650 F)". Discussions indicated
that the supplier did not expect retention of properties at 320 C
(600 F) for 2000 hours (about 3 months) or even less time. This
same response was obtained from other companies and from research
laboratories developing high-temperature plastics for the aerospace
industry.
The present operating temperatures of jambs (and metal seals)
are above the practical limits of elastomeric and resin materials,
if these materials are to be pressed against heated jambs for long
periods of time. This introduces consideration of lowering jamb
temperatures via either jamb insulation and/or water-cooling, in
one configuration or another. Water-cooling should be considered
for the jambs of new coke-oven battery installations, but present a
difficult and expensive problem when considered for retrofit applica-
tions. Battelle rates retrofitable water-cooling for jambs on existing
ovens as an approach of "last resort"; i.e., the approach to use if
and when other approaches are proven to be insufficiently effective.
Some coke-oven plants have tested Teflon as a potential sealing
material; i.e., as a secondary sealing system positioned either in-
board or outboard of the existing metal seal. Teflon-coated asbestos
has been mounted alongside the upstanding edge of the existing seals
in such a manner that the Teflon material would contact the jamb
prior to the contact of the metal seal. This is in effect an attempt
to (a) increase the gap-closing capability of the overall sealing system
and (b) obtain a better seal because of the ability of compressible
materials to deform and fill the asperities of the mating surfaces.
From the viewpoint of chemical inertness, Teflon is the best choice
to resist chemical attack. It is rated as being "stable" up to 260 C
(500 F) which is, however, a lower temperature than jambs reach at
the top of the oven. Also, the pressing of Teflon against a heated
jamb acts to insulate the covered portion of the jamb and further
increases the contacting surface temperature of the Teflon. The most
ambitious effort in this regard, and on which information is available,
was to equip all the doors on a battery with outboard seals of Teflon-
coated, asbestos-cloth tubing having Monel fibers in the tubing to
provide resiliency. The early results of this test program were
VII-3 Z
-------
encouraging in terms of emission control, but, overall, the results
were not successful because of the deterioration of the expensive
tubing and because it was not possible to clean the tubing by conven-
tional scraping methods.
If effective resilient-seal materials were available, considera-
tion would be given to either mounting the material as a guard seal
(on the existing metal seals) or developing a single seal based only on
the resilient material. It is doubted that a resilient material by
itself, can close.the major inward bowing on some jambs. In this
instance, it would be necessary to design a more flexible metallic
holder for the resilient material. Seal retainers or holders could
be designed to provide the extra flexibility required, but, then, the
resilient material would not be required.
Inflatable Seals (Rating: No
Possible Application)
This concept family has appeal because (a) it would be a gas-
tight seal and (b) the inflatable rubber-type tubes are not expensive.
However, as indicated in Table VII-5, the drawbacks of this approach
are many if water cooling and inert-gas back-pressurizing are not
included as part of the system.
The original interest in feeding of inert gas into the gas passage
between an inflatable seal and1 the standard metal-to-metal seal was
to prevent coal-tar volatiles from reaching the inflated tubes. Dis-
cussions with steel-plant technical personnel working in the by-
products departments of coke plants have indicated that all elastomers
are attacked by coal-tar volatiles in one way or another. Research
is in progress in this field by material manufacturers, but discussions
with these companies indicated that they will not recommend even the
most volatile-resistant product (trade name Chemigum) for this
application. It had been suggested that the inflatable-tube material
be coated with Teflon, but no tube manufacturer indicated that this
could be done.
To avoid the need for disconnecting and connecting the feed line
for the inflating gas every time the coke door is opened and closed,
the most logical place to install an inflatable tube seal is on some
part of the door jamb. This, however, presents a problem in pro-
tecting the bottom portion of the tube from the hot coke during and
following coke pushing.
VII-3 3
-------
Battelle researchers do not foresee a ay possible application of
inflatable seals on coke-oven doors.
Noncontact Seals (Rating: No
Possible Application)
In general, the noncontact family of sealing concepts depends
on the injection of pressurized gases into the seal area to prevent
the outward flow of emissions. Labyrinth seals, in which intimate
contact between'mating members is not essential, were considered.
The pressure of gas injected within the labyrinth would exceed the
gas pressure in the oven. Thus, the sealing gas would leak into the
oven, as well as out into the atmosphere. It was concluded that this
approach would be feasible only for jambs that had only a minor
degree of warpage; i.e., for jambs that can already be sealed by
existing designs. This family of concepts was given a rating of "not
possible" primarily because it does not have increased gap-closure
capability over and above the existing systems.
Any approach that injects back-pressurizing gas into the seal
area of coke-oven doors will release some of this same gas into the
atmosphere through gaps in the primary or existing sealing system.
Where the gaps in the primary system are large, even the idea of
using an inert gas (such as nitrogen) as a "sealant" does not appear
to be practical^ However, there is an interest at Battelle in con-
sidering coupling back-pressurizing with tight primary seals. The
interest in this instance is in finding out whether this approach
minimizes the cleaning effort presently required on door plugs,
jambs, and seals. It is understood that some experimental work
of this nature has been done in Europe, but details are not available.
Battelle researchers suggest that a study should be completed in
which (a) the back-pressurizing gas is injected behind the jamb (not
through the door), and (b) the gas passage on the door and the space
between the plug and the door are restricted at the top of the oven to
force a portion of the back-pressurizing gas into the coal. As noted
in Chapter IV, any procedure that eliminates the deposition of tar on
the jamb will result in outward gas leakage past a metal-to-metal
seal. It is expected that any back-pressurizing system will not
eliminate deposition of all tars on jambs. However, if it decreases
the outward flow of tars significantly, it should decrease the cleaning
problem. From this viewpoint, Battelle researchers consider this
approach worth testing.
VII-34
-------
CHAPTER VIII
DEVELOPMENT OF AN EMISSION EVALUATION METHOD
To aid in developing improved coke-oven end seal, it is desirable
that a method be available to demonstrate improvement and to measure
the degree of improvement. Therefore, as part of the overall research
program, a task was established with the objective of developing a
method for quantitatively measuring door-seal particulate emissions.
Comments on the Measuring Method
The emphasis in this phase of the research was on weighing the
particulate emissions collected over time on a 3. 15 x 3. 94 cm
(8 x 10 inch) glass-fiber filter. The filters were obtained from Mine
Safety Appliance Company and had the designation 935-BJH J-1453
(46-26-Natural). These filters are listed as having a collection effi-
ciency of over 99 percent for particle sizes of 0. 3 micron and larger.
It was decided that performance of coke-oven door seals could
be evaluated best through a comparison of the quantity of particulate
matter that escapes from around the door during the first few hours
of the coking cycle. In most instances, ovens tend to seal themselves
within the first hour or two of the coking cycle even those ovens that
are considered to be leaking badly early in the cycle.
After a review of literature dealing with tests that have been run
on coke ovens, and consideration of various approaches suggested by
Battelle staff members, the idea of a sealed hood was selected. Al-
though this method has some inherent disadvantages, it does lend
itself to a quantitative measurement of coke-oven door leakage.
VIU-1
-------
The initial basic idea was to seal off an oven door by placing a
cover or hood over the backstays and sealing between the buckstays
for the total height of the oven doors, except for a controlled air inlet
at the bottom of the hood and an exit at the top. A mete red supply of
clean air would be drawn into this space between the buckstays, and
would sweep up over the door and out through a duct connected to the
blower. A high-volume sampler could then be attached to the duct at
the top to draw off a known amount of gas which would be passed
through the sampler's filter to collect the contained particulate matter
for weighing.
Initially, consideration was given to drawing air into the bottom
of the hood through a charcoal filter. As air was drawn into the space
beneath the hood, particulates and other matter from adjoining ovens
would be filtered out by the charcoal. Thus, unpolluted air would be
mixed with the gases leaking from around the door. Because of lack
of clearance between the buckstays and the coke guides or pusher
machines, it was necessary to abandon this idea. Instead, the bottom
of the hood was completely sealed off between the buckstays. Air
from a filtered compressed-air line was then piped into the space be-
tween the buckstays for release near the bottom of the door.
Discussion of Equipment and Procedure
So that the hood could be1 readily transported and put in place, it
was made of aluminum sheets about 1. 1 meters wide by 0. 9 meter high
(42 x 35 inches). These sheets were fastened together with hinges;
five were the full 0. 9 meter (35 inches) high and one was about
0. 4 meter (16. 5 inches) high. The hood could be carried by one man
as it weighed only about 18 kilograms (40 pounds).
Figure VIII-1 shows a man holding the hood prior to installing it
on the buckstays of the simulated coke-oven end built in the laboratory.
The shiny surface is the aluminum backing of glass-fiber insulation
placed on the outside surface of the hood to minimize heat loss and
thereby avoid excessive condensation on the inside surface.* This
simulated coke-oven end was used to assemble, evaluate, and try out
the sampling equipment before taking it to an operating coke-oven bat-
tery. The simulated oven end was built of plywood, with metal strips
attached to the simulated buckstays for attaching the hood magnetically.
During the field testing of this equipment it was found that the insulation of the hood was not required
or desirable.
VIII-2
-------
Figure VIII-2 shows the hood being pulled up into place on the
buckstays. Figure VIII-3 shows the hood with three of its five full
sections in place. On each side of each section, strips of magnetic
tape were fastened to hold the section in place during installation.
These magnetic strips, which show along the section edges in
Figure VIII-1, hold the hood sections to the buckstays rather well
until they encounter high temperatures. With high temperatures, the
1.6-mm (16-gage) metal used in the hood sections tends to warp and
pop the metal back in spite of the magnetic strip.
/
To hold the hood more firmly in place during operation, and to
avoid leakage of gases around the edges of the hood, special clamping
devices were fashioned to attach to the outside flange edge of the buck-
stay. These clamps have angle-iron extensions welded to them so that
when the clamp is applied to one side of the hood section, the extension
presses firmly against the entire height of the section edge to hold it
tightly in place. Figure VIII-4 shows one of these clamps in open
position, not yet applied to the hood section. Figure VIII-5 is a view
of the clamp in closed or applied position. Figure VIII-6 is an over-
all view of a clamp in closed position against one of the section edges.
Figure VIII-7 is a view of the top plate which fits between the
buckstays and provides the top enclosure for the space enclosed by the
hood. This view shows the pressure-tap tube (between the man's
hands) used to monitor pressures within the space around the coke-
oven door and between the buckstays. The 10.2-cm (4-inch) gas-
exit port is shown. The Length adjustment which enables the plate to
be put into place past protruding objects can be seen on the left end.
Railing hooks attached to the top of the plate by stiff coil springs serve
to support the plate over its span during sampling runs.
Figure VIII-8 shows the under side of the top plate as viewed
from a point looking up between the buckstays in the direction that
must be taken by gases escaping from around the coke-oven door and
by the introduced air which sweeps upward over the door.
The pipe that carries air to the bottom of the hood cavity is
visible in the upper righthand corner of Figure VIII-8. This same air
is shown in Figure VIII-9 before installation. Here it is shown attached
to the compressed-air supply line which contains a pressure gage,
pressure regulator, and a filter.
VIII-3
-------
FIGURE VIII-1.
COKE-OVEN HOOD HELD
BY RESEARCHER (D. HUPP)
IN FOLDED POSITION
PRIOR TO MOUNTING ON
BUCKSTAYS OR SIMULATED
COKE OVEN. MAGNETIC
STRIPS SHOW ON BACK OF
HOOD SECTION.
FIGURE VUI-2.
HINGED HOOD
BEING PULLED
INTO PLACE ON
BUCKSTAYS
-------
I
01
FIGURE VIII-3.
HOOD PULLED UP TO
COVER OVER HALF
OF CAVITY BETWEEN
BUCKSTAYS
FIGURE Vin-4. HOOD CLAMP
ATTACHED TO
BUCKSTAY FLANGE
AND FIXED IN OPEN
POSITION
-------
0
I
o
FIGURE VIII-5.
HOOD CLAMP ATTACHED
TO BUCKSTAY FLANGE
AND FIXED IN CLOSED
POSITION
FIGURE VIII-6.
OVERALL VIEW OF HOOD
CLAMP IN CLOSED
POSITION
-------
FIGURE VIII-7.
VIEW OF HOOD-TOP COVER PLATE SHOWING
PRESSURE TAP, GAS-EXIT PORT, LENGTH-
ADJUSTMENT FEATURE, AND PLATE-
SUPPORT HOOKS
VIII-7
-------
FIGURE Vin-8.
HOOD-TOP COVER AS VIEWED LOOKING
UPWARD FROM BETWEEN BUCKSTAYS
FIGURE VIII-9.
HOOD AIR SUPPLY BEFORE INSTALLATION.
SHOWN ATTACHED TO COMPRESSED-AIR
SUPPLY LINE CONTAINING PRESSURE GAGE,
PRESSURE REGULATOR, AND FILTER.
VIII-8
-------
During emission-measurement runs, sampling equipment was
located on top of the coke-oven battery at the charging-port level.
Figure VIII-10 is an overall view of the sampling system. It extended
from the top of the hood top-cover plate to the main blower which
exhausts gases from within the hood. From left to right, Fig-
ure VIII-10 shows the 10.2-cm (4-inch) flexible line that connects
the top plate to a 10.2-cm metal duct. A 5. 1-cm (2-inch) flexible
line carries a gas stream withdrawn from the hood exhaust stream
and carries it to high-volume (Hi-Vol) samplers. (This equipment
is manufactured by General Metals Works, Cleves, Ohio.) This
5. 1-cm flexible line can be switched from one sampler to the other in
about 10 seconds. Downstream of the point where the air sample is
withdrawn for the Hi-Vol sampler, an orifice is located in the 10. 2-
cm duct so that measurement can be made of the volume of gas that
flows past (bypasses) the Hi-Vol samplers. Downstream from the
orifice, the duct includes a damper that can be adjusted to control the
flow coming from the space within the hood. Immediately downstream
from the damper there is in the duct a "T" which can serve as a
balance in controlling flow. The leg of the "T" contains a damper
which can vary the amount of air pulled to the blower from outside.
The blower appears at the extreme right of Figure VIII-10.
Figure VIII-11 is a closeup view of the Hi-Vol samplers and
metering equipment. On the front of the Hi-Vol on the right is an
adsorber column which is attached during tests by tubing to the Hi-
Vol sampler downstream from the filter. A sample of about 0. 021 m^
(0.75 ft^) per minute is drawn through the adsorber by the pump shown
to the left of the gas meter in Figures VIII-10 and VIII-11. Tubing
during tests is arranged so that flow from the adsorber goes to the
pump and then through the gas meter.
The open door on the Hi-Vol reveals the sampler's flow recorder.
The inclined manometer at the top of the instrument board to the right
of the Hi-Vol connects by tubing with the pressure tap on the hood-top
cover plate shown in Figure VIII-7. This manometer is used in
monitoring static pressure beneath the hood. The manometer on the
left side of the board records the pressure of the flow transducer on
the Hi-Vol, permitting the sample flow to the sampler filter to be
kept constant. The manometer on the right side of the board measures
the pressure differential across the flow orifice, thus monitoring the
flow in the duct downstream from the flow line to the Hi-Vol.
VIII-9
-------
FIGURE VIII-10.
OVERALL VIEW OF SAMPLING EQUIPMENT
USED IN TAKING SAMPLES AND MAKING
MEASUREMENTS
FIGURE VIII-11.
CLOSEUP OF HI-VOL SAMPLER AND ITS FLOW
RECORDER, ADSORBER COLUMN WITH FLOW
METER AND PUMP, AND THE MANOMETER
BOARD
VIII-10
-------
A critical-flow orifice in the compressed-air line enabled the
test team to put a measured quantity of air down into the hood cavity
between the buckstays. This air plus gases leaking from around the
coke-oven door constituted the total gas flow from the hood into the
10.2-cm (4-inch) duct. A known volume of gas was drawn off to the
Hi-Vol sampler, and the remaining gas volume that flowed through
the orifice in the 10.2-cm line accounted for total gas flow. Thus,
orifice-flow volume plus Ni-Vol volume minus compressed-air
volume equaled volume of leakage from around the coke-oven door.
Figure VIII-12 is a view of the Hi-Vol sampler with the inlet
cap and filter holder tipped back showing the coarse mat and regular
glass-fiber filters used in preliminary laboratory and coke-oven-
battery test runs.
FIGURE VIII-12.
HI-VOL SAMPLER WITH COARSE (UPPER)
AND REGULAR (LOWER) FILTERS BEING
INSTALLED
VIII-11
-------
Laboratory Test Runs
To try out the sampling hood in the laboratory, a mockup of a
coke oven and its buckstays was built to full scale. The mockup was
built of plywood except for metal braces and metal strips on the buck-
stays to simulate the metal of real buckstays. Figures VIII-1 through
VIII-8 show parts of the mockup with Figure VIII-8 revealing the metal
braces used to strengthen the wooden buckstays. Figure V1II-2 shows
the metal buckstay strips clearly.
As a rough check on the system, a 3-minute smoke bomb was set
off on the floor inside the hood. This bomb released more particulate
matter than a regular filter in the Hi-Vol could handle. The passages
of the filter were plugged in about 1 minute, as evidenced by the flow
through the filter dropping to zero.
Although the concentration of particulates provided by a smoke
bomb may be greater than would be encountered ordinarily with leak-
age from a coke-oven door, provision was made to handle such a
concentration. This was accomplished by adding a second Hi-Vol
sampler with provision for switching the gas line rapidly from one
sampler to the other. The switch can be carried out within 10 seconds.
The second approach to handling a very high concentration of
particulate matter through the Hi-Vol sampler was to add a coarse
glass-fiber mat ahead of the normal filter (which has an efficiency of
over 99 percent for 0. 3-micron particle sizes and larger). With the
coarse filter in operation ahead of the regular filter, another smoke
bomb test was run. With the two filters in use, the Hi-Vol sampler
operated throughout the full 3-minute life of the bomb without any
measurable decrease in flow through the sampler.
Coke-Plant Test Run
It was recognized that although the system performed satisfactorily
in the laboratory, operation in the field was the purpose of the system.
Therefore, a single trial run was performed at an operating battery.
* Under a separate contract, EPA funded additional testing of this equipment. The results of this
program are not available at this time.
vm-12
-------
The objective of the trial run was to operate the collection sys-
tem on a producing coke oven to determine if major modification was
required before making test runs to evaluate new oven-door sealing
concepts.
Permission for a trial run was granted by Empire-Detroit
Steel Division, Detroit Steel Corporation, Portsmouth, Ohio. The
test site was a high-bench Koppers coke-oven battery that was built
in 1964. Door leakage from these ovens is generally light compared
with other ovens in the industry. Figure VIII-13 is a view of Ovens 1
through 3 from'the coke side of the battery. Figure VIII-14 is a view
looking from Oven 1 toward the coke received car on the coke side.
It illustrates how the buckstays on this oven extend down to the bench
level.
Specific items that were listed for checking were enumerated as
follows:
(1) Whether or not the filter system could be operated for
extended periods without particulate emissions over-
loading the filters.
(2) Operation of adsorption columns with samples taken after
the high-volume filters. These samples were wanted for
quantitative and qualitative analysis of the emission
gases.
(3) General functioning of the hood and sampling system
when mounted on the buckstays of an operating coke
oven.
(4) Temperature of the hood surface to determine if insula-
tion was needed on the outside surface. Concern had
been expressed regarding the possibility of condensation
on the inside surface of the hood.
It was found that the coarse and regular filters handled with ease
the particulate emissions encountered. In this test, the coarse filter
probably would not have been needed over the 2. 5-hour trial run.
The adsorption column performed well throughout the run. Flow
was held consistently at about 0. 018 m^ (0.65 ft^) per minute and a total
sample of 2.45 m^ (86. 5 ft^) was passed through the adsorption column.
VIII-13
-------
FIGURE VIU-13.
KOFFERS HIGH-BENCH
COKE-OVEN BATTERY
FROM COKE SIDE OF
BATTERY
[ -
FIGURE VIII-14.
OVEN DOORS AND
BUCKSTAYS ON COKE
SIDE OF BATTERY
-------
It was found that the hood covering the backstays might cause
abnormal heating of the oven door. At 3 hours and 45 minutes into
the coking cycle, the door-surface temperature was about 65 C
(150 F) above the temperature of another oven that was charged
15 to 20 minutes after the test oven. Temperature measurements
on the coke-end doors taken 10 to 15 minutes after the hood was re-
moved were 220 C (430 F) for the test-door surface and 150 C (300 F)
for the surface of a door used in comparison. Similar measurements
made on the pusher end of these same ovens gave 150 C (300 F) for the
test oven and 170 C (340 F) for the comparison oven. Thus, there is
variability in temperatures recorded on door surfaces from oven to
oven.
Temperatures of 80 C (175 F) to 85 C (185 F) were recorded on
the outside surface of the hood. This led to the conclusion that for
future use the insulation should be removed from the outside surface
of the hood.
In general, the test at the coke-oven battery showed that this
system for collecting and measuring emissions of particulates can
be used to evaluate and compare new methods for sealing coke-oven
doors. The approach and equipment appear satisfactory for about
the first 2 hours of coking operation. For longer periods of measure-
ment, intermittent application of the hood or other means of prevent-
ing significant increases in door temperature should be considered.
VIII-15
-------
TECHNICAL REPORT DATA
(rii me read Iniiructions on the ret ene before completing)
1. REPORT NO
EPA-650/2-75-064
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Study of Concepts for Minimizing Emissions from
Coke-Oven Door Seals
5. REPORT DATE
July 1975
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
H.W. Lownie Jr. andA.O. Hoffman
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
10. PROGRAM ELEMENT NO.
1AB015; ROAP 21AQR-012
11. CONTRACT/GRANT NO.
68-02-1439
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Control Systems Laboratory
Research Triangle Park, NC 27711
T
ERIOD COVERED
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
The report gives results of a study aimed at minimizing emissions from coke-oven
door seals. It identifies problems associated with the sealing of slot-type coke oven
end closures, and quantifies them to a limited degree by test results presented in the
report. It analyzes coke-oven door sealing systems--those which have been devel-
oped in the past, as well as those currently in usewith respect to individual
strengths and weaknesses. It develops and critically analyzes concepts to improve
the seal design, and recommends the development of the two most favorable
concepts.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution
Coking
Doors
Seals (Stoppers)
Air Pollution Control
Stationary Sources
13B
13 H
13M
11A
18 DISTRIBUTION STATEMbNT
Unlimited
19 SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
235
20. SECURITY CLASS (Thu page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
VIII-17
------- |