vvEPA
         United States
         Environmental Protection
         Agency
           Industrial Environmental Research
           Laboratory
           Research Triangle Park NC 27711
EPA-600/2-78-189
August 197£
         Research and Development
Development and
Demonstration of
Concepts for
Improving
Coke-oven
Door Seals:
Interim Report

-------
                    RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental Protec-
tion Agency, have been grouped into nine series. These nine broad categories were
established to facilitate further  development  and application of environmental tech-
nology. Elimination of traditional grouping was consciously planned to foster technology
transfer and a maximum interface in related fields. The nine series are:

          1. Environmental Health Effects Research
          2. Environmental Protection Technology
          3. Ecological Research
          4. Environmental Monitoring
          5. Socioeconomic Environmental Studies
          6. Scientific and Technical Assessment Reports (STAR)
          7. Interagency Energy-Environment Research and Development
          8. "Special" Reports
          9. Miscellaneous Reports

This report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY
series. This series describes research performed to develop and demonstrate instrumen-
tation, equipment, and methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the new or improved tech-
nology required for the control and treatment of pollution sources to meet environ mental
quality standards.
                             REVIEW NOTICE


          This report has been reviewed by the U.S. Environmental
          Protection Agency, and approved for publication.  Approval
          does not signify that the contents necessarily reflect the
          views and policy of the Agency, nor does mention of trade
          names or commercial products constitute endorsement or
          recommendation for use.


          This document is available to the public through the National Technical Informa
          tion Service, Springfield, Virginia 22161.

-------
                                  EPA-600/2-78-189
                                       August 1978
Development and  Demonstration
     of Concepts for  Improving
        Coke-oven  Door Seals:
              Interim Report
                       by

              A.O. Hoffman, AT! Hopper, and R.L Paul
                 Battelle Columbus Laboratories
                    505 King Avenue
                   Columbus, Ohio 43201
                  Contract No, 68-02-2173
                 Program Element No. IAB604C
               EPA Project Officer: Robert C. McCrillis

             Industrial Environmental Research Laboratory
               Office of Energy, Minerals, and Industry
                Research Triangle Park, NC 27711
                     Prepared for

             U.S. ENVIRONMENTAL PROTECTION AGENCY
               Office of Research and Development
                   Washington, DC 20460

-------
                              ABSTRACT
    The report gives pre-engineering analyses, evaluations, and  recommen-
dations  in an ongoing research project dealing with the development of a
retrofittable concept for minimizing emissions from door seals on coke ovens. It
includes evaluations drawn from tasks dealing with mathematical and physical
modeling,  and from  a task dealing with  field-data  collection and  field
experiments. Based on these results, the recommended metal-to-metal sealing
system includes:  a simplified-shape seal, a  new  improved  procedure for
mounting and adjusting seals, and  high-strength heat-resistant materials. The
recommendations were approved by the sponsors, and engineering tasks are in
progress. It is recommended that  as new door jambs are installed, they  be
allowed to assume their natural curvature resulting from operating tempera-
ture gradients, while simultaneously restricted to prevent hourglassing, and that
they be ferfitic ductile iron castings. Limited experiments with luting com-
pounds were encouraging. If developed further, this might be an attractive
alternative, particularly for older batteries.

-------
                               TABLE OF CONTENTS

                                                                            Page

EXECUTIVE SUMMARY	   vii

                                    CHAPTER I

INTRODUCTION 	   1-1
    Background and Antecedents
    Project Objectives 	
    Organization of the Project
                                                                          -1
                                                                          -3
                                                                          -4
Research Staff 	    -4
                                                                          -7
                                                                          -8
    Acknowledgments 	
    Project Manager's Comments and Qualifying Statements  	

                                   CHAPTER II

ANALYSIS OF METAL-SEAL SYSTEMS AND RETROFITTABLE
  METAL-SEAL RECOMMENDATION 	  11-1

    Comments on Standard Designs for Seals  	  11-1
    Input From Tasks 1 and 2 (Mathematical and Physical Modeling)  	  II-1
    Input From Task 3 (Field Data Collection)  	  11-9
        S-Shaped Seals  	  11-9
        Fixed-Edge Seals 	  11-9
        Jamb/Door-Frame Profile Relationships	 11-10
        Damage/Changes Observed in Operating S-Shaped Seals	 11-15
    Input From Other Sources  	 11-18
    General Requirements for Upgraded Metal-Seal Systems	 11-19
    Technical Conclusions and Recommendations (Input to the Design Task)	 II-20
    The Recommended Design of Retrofittable Seals 	 II-22
        Description of the Recommended Seal  	 II-22
        Comments on Discarded Approaches	 11-30
        Rationale for the Shape of the Recommended Seal (The Corner Problem) .... II-33
        Critique/Limitations of the Recommended Seal	 II-36
    Material Considerations for Upgraded Metal Seals 	 II-37
        Comments on Present Seal Materials  	 II-37
        Material Selection Criteria 	 II-38
        Material Recommendation   	 M-41
    Research Outline for Task 5	 11-41
                                   CHAPTER III

JAMB ANALYSES: RECOMMENDATION OF A RETROFITTABLE
  JAMB DESIGN AND JAMB MATERIAL 	   111-1
                                       in

-------
                               TABLE OF CONTENTS
                                   (Continued)
                                                                              Page
    Introduction and General Statement of the Problem  	  111-1
    Data and Insights Obtained in Task 3 (Field Data Collection) 	  IH-3
       Long-Term Temperature Data on Jambs	  '"-3
       Jamb Contour Measurements  	  IH-4
       Internal Damage to Gray Iron Jambs 	  III-5
       The Hourglassing Problem	  HI-7
    Conclusions (and Discussion) Derived from the
      Analytical Effort Dealing With Jambs	  III-8
       Background Comments  	  HI-8
       Analyses of Jamb Distortions 	  III-9
       Summary of Recommendations for Retrofitted Jambs	 111-15
    Jamb Material Selection/Recommendation	 III-16
       The Progression of Approaches Used in Material
          Evaluation/Selection  	 111-16
       Basic Information and Evaluation of Jamb Materials  	 111-17
       Summary  	 111-25
       Concluding Comments	 111-26
    Design Considerations for Retrofittable Jambs  	 111-26

                                   CHAPTER IV

AN ANALYSIS OF THE PROSPECTS FOR SEALANT SEALING
  (INBOARD LUTING) OF COKE-OVEN DOORS  	  IV-1

    Background and Definitions  	  IV-1
    Results of the Phase I Study  	  IV-3
    Criteria Development  	.„	  IV-5
    Summary of Laboratory Results	  	  IV-6
    Conclusions of the First Phase of the Laboratory Work	 IV-11
    Summary of Field  Evaluations	 IV-11
    Summary of the Follow-On Laboratory Tests	 IV-15
       Variations on  Water Content in Sealants	 IV-15
       Effect of Water Content on Sealant Shrinkage 	 IV-18
       Preliminary Evaluation of the Possibility of Cracking
          Cast-Iron Jambs as a Result of Using Wet Sealants  	 IV-19
    Overall Summary of the Laboratory Research and Field Tests	 IV-21
    Feasibility Analysis	 IV-22
       Cost of Sealant Approach Including Additional Benchmen	 IV-25
       Conclusions	 IV-28
    Recommendations	 IV-28
                                        IV

-------
                                   LIST OF TABLES

Table 11-1.  Comparative Data on Candidate Alloys for Manufacture
           of Coke-Oven Seals  	   11-42

Table 111-1. Comparison of the Amount of Hourglassing of Two New
           6-Meter Jambs Operated About the Same Length of Time  	  II1-7
                                  LIST OF FIGURES

Figure 1-1.   Outline of Tasks in the Basic and Optional Program and
            Schedule of Summaries and Reports 	    1-5

Figure 1-2.   Outline of Activities in the EPA/AISI Research Program on
            Improved Systems for Sealing Coke-Oven End Closures 	    1-6

Figure 11-1.  Cross Section of a Typical Coke-Oven Door With a Fixed-Edge Seal  	   II-2

Figure II-2.  Cross Section of a Typical Coke-Oven Door With an S-Shaped Seal	   II-2

Figure II-3.  Cross Section of a Typical "Knock-Type" Seal	   II-3

Figure 11-4.  Side View of Adjustment Equipment on Fixed-Edge Seals  	   II-3

Figure II-5.  Distribution of Sealing Pressure  	   II-5

Figure II-6.  An Exaggerated Representation of the Contact Pressure Distribution
            Between Two Clamped Rules 	   11-5

Figure II-7.  Side View of S-Shaped Seal System  	   II-7

Figure II-8.  Examples of Variations That Exist in Terms of the Relative
            Contours of Jambs and Door Frames	  11-13

Figure II-9.  Tool Designed for Measuring the Horizontal Distance Between the
            Inboard Edge of any Door Frame and the Sealing Surface
            of the Accompanying Jamb  	  11-14

Figure 11-10. Sketch Showing Operation of the Measuring Tool  	  11-15

Figure 11-11. Photograph of a Discarded S-Shaped Seal Showing Indentations at the
            Plunger Contact Points and the Resulting Waviness in the  Seal Edge  ......  11-16

Figure 11-12. Recommended Retrofittable Seal	  II-23

Figure 11-13. Typical Koppers Seal Area With Seal Removed	  II-24

Figure 11-14. Typical Wilputte Seal Area With Seal Removed  		  II-25

Figure II-15 Section of Recommended Seal; Typical 4-Meter Koppers Oven 	  11-27

-------
                                  LIST OF FIGURES
                                    (Continued)
                                                                                Page

Figure 11-16. Section of Recommended Seal; Typical 4-Meter Wilputte Oven 	 H-28

Figure II-17. Concept of Retrofittable Mechanical Adjustment for Seal	 11-30

Figure 11-18. Alternate Concept for Wilputte Door  	 "-31

Figure 11-19. Concept Using Metal Bellows	 "-32

Figure II-20. An Example of a Sloping Seal Design	 11-34

Figure 11-21. An Example of a Parallel Seal Design	 11-34

Figure II-22. Sketches Showing Seal Actions at Corner of Doors 	 II-35

Figure 111-1. Six Examples of the Horizontal Cross Section of Existing Jambs 	  III-2

Figure III-2. Side View of the Top of a Jamb With a Proposed Packing Retainer	 III-15

Figure III-3. Typical Stress-Strain Curves in Tension for Various Cast Irons and Steels .. 111-22
                                                                               t
Figure 111-4. A Typical Stress-Strain Curve for Gray Cast Iron Showing the Amount
            of Permanent Strain That Develops During the Early Stages of
            Putting This Material in Load-Carrying Service 	 111-22

Figure 111-5. Laboratory Arrangement for Subjecting Metal Samples to
            Thermal Stress	 111-24

Figure IV-1. Luted Door Design in Use About 1946 	  IV-2

Figure IV-2. Luted Door Design in Use About 1929 	  IV-2

Figure IV-3. An Example of a Sealant-Sealing Concept Presented in the First
            Project (Study of Concepts)  	,	  IV-4

Figure IV-4. Sealant Applied Between the Door and Jamb Surfaces (Concept
            Taken From the Phase I Report) 	  IV-7

Figure IV-5. Simulated Jamb Channel (Door Edge Embedded in the Sealant) 	  IV-7

Figure IV-6. Photograph of a Laboratory Arrangement Used to Simulate
            the Jamb/Seal Relationship at Coke Ovens 	  IV-8

Figure IV-7. Interlake  Incorporated Drawing of a Proprietary Door Seal,
            i.e., a Modified Fixed-Edge Seal 	 IV-13

Figure IV-8. Illustration of the Concept of "Inboard Luting", i.e., the Application
            of Luting Material in the Gas Passage of Existing  Equipment	 IV-15

Figure IV-9. Existing Seals can be Equipped With a "Filler" Section to Minimize
            the Amount of Material Used in Inboard Luting   	 IV-17
                                          VI

-------
                           EXECUTIVE SUMMARY
ORIGIN
    In  1974 the Industrial Environmental Research Laboratory of EPA and  the
American Iron  and Steel Institute (AISI) reached an agreement to  fund  a program for
a phased technical study to  decrease/eliminate visible emissions from coke-oven door
seals on existing ovens. The phases  or projects of the program were:
     I. Study of Concepts for  Minimizing Emissions from Coke-Oven Door Seals
     II. Development of Concepts for Improving Coke-Oven  Door Seals
    III. Demonstration of Concepts for Improving Coke-Oven Door Seals
    In  competitive bidding, Battelle-Columbus  Laboratories was awarded a  jointly
funded research contract dealing with Phase I. This project  was completed in July of
1975.*  The conclusions of the  Phase I project were:
    • Of all of the concepts considered, upgraded metal seals have  the best
      potential  for meeting all  of the emission control and retrofit  criteria
    • The entire end closure should  be analyzed quantitatively because the
      dimensional  instability  of the jamb and other components is part of the
      leakage problem
    • Variations of luted seals  should be evaluated in a laboratory arrangement
      to  obtain information on  their potential for successful  development, accep-
      tance, and implementation.

THIS PROJECT
    In August  of 1976, Battelle-Columbus was awarded  research contracts  (with EPA
and AISI) dealing  with  Phase II of the program, i.e., the research/development aspects
of the conclusions and recommendations of the concepts study.
    The  specific objective of this project was to "innovate and to develop at least one
new system that will be proven in the field  to be retrofittable to existing coke ovens,
mechanically and  physically  suitable  for commercial use in  steel plants, and  highly
The report on this project was entitled "Study of Concepts for Minimizing Emissions from Coke-Oven
 Door Seals". It may be obtained from the National Technical Information Service, 5285 Port Royal Road,
 Springfield,  Virginia 22161. The report number is PB 245580.
                                        VII

-------
effective in containing and controlling emissions from the ends of ovens".  The effort
was broadened from improved seals alone to research consideration of all  com-
ponents of end closures, i.e.,  "from the bricks out".

THIS REPORT
    This particular development project was  organized into six interrelated  tasks. The
first three tasks were conducted concurrently and were entitled:
    1. Mathematical Modeling and Analysis of  Systems
    2. Physical Modeling and  Laboratory Experimentation
    3. Field Data Collection and Field Experimentation
    These  first tasks were  reported  to  the  sponsors  in the form  of  working paper
summaries.
    These were followed by Task 4 which  is entitled "Analysis, Evaluations, and
Recommendations". This report is  the summary of our evaluations and recommen-
dations resulting from  Task 4.
    While this report was being critiqued by the sponsors, work  was in progress on
Task 5  ("Full-Scale Unit Design and Testing  of Parts"). Contractual arrangements
allow the sponsors to exercise an  option of funding Phase III  (the demonstration
phase) without any lost time between  projects. For this possibility, Task 6 ("Decisions
Regarding Full-Scale Installations") is being completed concurrently with Task 5.

MAJOR CONCLUSIONS AND RECOMMENDATIONS
OF THIS PHASE II PROJECT
1.  Battelle's recommended, retrofittable door-seal system has the following  elements:
     • A preferred new design of seal (page 11-23)*
     • Use of  high-strength, heat-resistant  materials (page 11-37)
     • Special considerations regarding seal adjustment/attachment  (page 11-29).
      It is judged that  the recommended door-seal system will be applicable to 4 to
  6-meter batteries.
"The numbers in parentheses refer to page numbers in this report.
                                       VIII

-------
    The  recommended design of metal-to-metal seal is shown in cross section as
follows:
                    Sealing
 Packing Material ^  Ring-,
     In our judgment, this simple design, fabricated from high-performance material
 will:
   a. Flexibly absorb the latching force without  recourse to additional localized
     or point-applied forces. Point-loading of the seal has been eliminated to
     help equalize the  distribution of loading along the seal.
   b. Minimize stress-concentration problems at  the corners of the seals.
   c. Have an elastic seal-edge flexibility (allowable displacement)  2  to  3 times
     higher than existing S-shaped seals. Inconel X-750 and other alloys will
     be tested in Task 5 for this application.
   d. Be adaptable as a single design to both Koppers and Wilputte designs for
     doors.
   e. Have a longer good-performance life  than  standard seals. The use of
     harder and  tougher materials will lower the sensitivity to physical damage.
                                        IX

-------
    By a special procedure for attachment and adjustment, the  recommended  seal
can (we believe) be made to:
  • Contact  (without gaps) the seal-mating surface of perhaps 90+ percent of
    existing jambs  regardless of their present contours.*
  • Have a relatively uniform deflection along the entire length  of the seal.
    This would  result in (a) more uniform distribution of the latch forces along
    the seal  edge,  and  (b) lowering the possibility of overstressing the seal.
    Both results would  be favorable in terms of  long-life, improved emission
    control.
    The recommended installation/adjustment procedure  is as follows:
  a. At  an  operating end closure,  use a  tool developed during this  project
     (page 11-14) to  measure the variation in the horizontal distance between
     the jamb sealing surfaces and the backsides  of  the  door  frame.  This  is
     the space that the seal  has to accommodate (seal)  effectively.
  b. Install the recommended seal by using spacers (where required) un-
     derneath and along the seal holder to form the entire seal  (including the
     seal edge)  to  the same  contour as that of the jamb seal-mating  surface.
     The  amount of spacing  is set (within workable tolerances)  by the
     measurements obtained  at  the operating  oven.
    This procedure may custom fit a door to only one jamb. However, it is ex-
pected  that the  same door and seal will fit several jambs on the same  battery
because of the tendency for jambs on  a battery to assume similar contours.*
    One setting of  the spacers is expected to give satisfactory performance for
some long period of time. Part  of this  expectation is  based on the increased flex-
ural tolerance designed into the  seal. Mechanical-screw adjustments were con-
sidered for moving  the entire seal to obtain conformity with the jamb, but design
of this variation was held in  abeyance  because of (a)  the high cost of  this varia-
tion, and  (b) the concern that  manual adjustment can result in overstressing (warp-
ing) of  the seal.
*The actual limitations (if any) are to be established in the demonstration phase.

-------
2.  There may be some existing jambs that are distorted to the point where they can-
   not be fully contacted (no gaps) by  any metal seal without developing distortions
   in  the seal  itself.  In these instances,  it is recommended  that new jambs  be  in-
   stalled. Our analysis/judgment indicates  that these jambs  and those  on all new
   batteries and rebuilds should:
     • Be installed so that they are in a state of low or zero thermal stress when
       in their operating-temperature range
     • Have a short web depth to lower both the stiffness of the jamb and the
       temperature gradient that occurs  during transient thermal flux conditions
     • Be castings made of ferritic ductile iron castings   <
     • Be locked in place  to prevent hourglassing  and thermal flexing.
       A low thermal stress state occurs when the jamb has been allowed to or en-
   couraged to bow inward (top and bottom out from the oven) to relieve the stress
   generated by the normal temperature differential across the jambs. In the thermally
   bowed "natural" condition, jambs will develop less stress  when they are heated  or
   cooled (rain storms) outside of the normal operating temperature range. Lower
   stress will translate to less distortion and, therefore, to longer jamb life. During this
   project, it was learned that one coke-plant builder adopted the concept of "natural-
   ly  bowed jambs" several years ago. Also, the technical personnel of one coke plant
   had come to the same conclusion regarding the advantages of a shorter web
   depth, and have machined their jambs accordingly.
3.  Our research efforts dealing  with various ways to  use sealants developed into an
   evaluation of what we have termed "inboard luting". In this instance, improved
   luting material (clay-based mixtures) is  mechanically placed into the gas passage
   of existing  metal seals while the  door is off the oven. Sealing occurs when the
   water-tempered luting material is compressed  by latching the door. Luting material
   would have to be applied for every coking cycle.
   It was concluded that:
     • If inboard luting could be fully developed and implemented it has the
       potential of being a sealing  method  that is both smoke-tight and gas-tight
       through the entire coal-charging and coke-making cycle. It has this poten-
       tial because coal tars and liquids seal the pores of the luting material as it
       dries.
                                       XI

-------
     • The inboard luting method  is not sensitive to the condition of the jambs
      or door seals. It could be used where end closures are in "bad" condition.
     • The cost  of material for inboard luting is low, but the overall cost is ex-
      pected to be appreciably higher than for the  retrofittable metal seal  if ad-
      ditional workers must be hired  to  operate the  sealant-application equip-
      ment.
      It is recommended that the  coke-producing industry (producers of blast fur-
   nace coke and foundry coke) be polled to  determine whether there is enough in-
   terest in the inboard-luting approach to underwrite a development and implementa-
   tion effort. Inboard luting may be attractive in situations where the battery age
   does not justify the expenditure for new end-closure components.  The possibility
   exists that inboard luting will eliminate the need for  manual cleaning  of jambs and
   doors. If this  proves to be  correct, then luting may  at  some  locations be sub-
   stituted for the cleaning operation  with no  increase in crew  size.
4.  The recommended metal-to-metal  sealing system should be evaluated in  a
   demonstration project  on operating coke-oven batteries. As part of this demonstra-
   tion, a number of closures incorporating the recommended system should be com-
   pared with a number of closures incorporating new-component standard  designs
   and possible competing designs.
   Comparisons should be made in terms of:
     • Emission-control effectiveness
     • Level of attention  required to  maintain an  acceptable seal
    • Life of the system  in terms  of  acceptable emission control
    • Overall costs,  including  original installation and repair and adjustment
      operations
    • Operator acceptance.
      The demonstration  project may be too short to obtain data on the very impor-
  tant cost-per-year basis, but it would be expected that some indications may
  develop  in a year's time. Continued evaluation by  coke-producing companies, after
  the end  of  the demonstration project, should develop  long-term cost information.
      As part of the technology transfer effort, there will  be a requirement for  con-
  siderable technical and evaluation  input by steel-company  personnel at the host
  plants.
                                       XII

-------
                                          1-1
                                    CHAPTER I

                                   INTRODUCTION
    This report summarizes  Task IV (Analysis, Evaluations,  and Recommendations) of a
research project entitled "Development and Demonstration of Concepts for Improving Coke
Oven Door Seals". This report is being issued after 20 months of effort in a 24-month develop-
ment project. The main objectives of releasing this report prior to the completion of the pro-
ject are to:
    (a)  Formally present our conclusions before completion of the project
    (b)  Present our recommendations dealing with additional effort on this present pro-
        ject and a following demonstration project that is under consideration
    (c)  Obtain permission from the Sponsors to proceed with the remaining Tasks in this
        project
    (d)  Make  it possible for the Sponsors to  make plans for going into a  follow-on
        demonstration project without interruption in the overall program.
    The work  reported  herein was sponsored by  the Industrial Environmental  Research
Laboratory of the  EPA and by the American Iron and Steel Institute (AISI). The opinions,
evaluations, judgments,  and  recommendations expressed are  strictly  those  of  the par-
ticipating Battelle-Columbus staff.

                             Background and Antecedents
    In January of 1970, Battelle-Columbus issued a formal report on coke-plant emissions con-
trol to the National Air Pollution Control Administration.* In that report, Battelle researchers
recommended  that the solution to air-emissions problems, which all coke-oven operators
have  in common, can "best be  achieved" by group  action  and joint  contributions. This
recommendation was particularly appropriate for emission-control problems having complex
technical components and involving unknown and/or  conflicting factors and opinions. The
term "best be achieved" takes into consideration the favorable odds involved in an adequately
funded, objective, cooperative, and  well-publicized technical approach, as compared with
uncoordinated  empirical  approaches. In this regard, we  hasten to say that empirical  ap-
proaches can be successful provided that motivation is high and sufficient time is allocated to
the problem. This present project (and a preceding project) are examples of combined effort
towards a common goal.
•Evaluation of Process Alternatives to Improve Control of Air Pollution from Production of Coke". Prepared for the
 Division of Process Control Engineering, National Air Pollution Control Administration under Contract PH 22-68-
 65. Available from the National Technical Information Service as Document PB 189266.

-------
                                         1-2

    In January of 1974, the Battelle-Columbus Laboratories responded to an EPA Request for
 Proposal (EPA RFP No. DU 74-A039) dealing with a research project to study, innovate, and
 evaluate concepts for minimizing emissions from coke-oven door seals. It was understood that
 the research would be sponsored  and financed jointly by EPA and  the AISI  and that the
 research was to be monitored by EPA.
    After competitive bidding, the contract for this research project was awarded to Battelle-
 Columbus in June 1974. This  research project followed the plan outlined in the RFP and was
 entitled "Study of Concepts for Minimizing Emissions from Coke-Oven Door Seals". The ap-
 proved final report was  issued in July of 1975. Copies of this report are available from the
 National Technical Information Service (NTIS)  as  Document PB 245580.* This report is
 reference material for the present project.
    In January of 1976, Battelle-Columbus was asked by the EPA/AISI (EPA RFP No. DU-76-
 A103) to present a proposal and a series of recommendations dealing with "Development and
 Demonstration of Concepts for Improving Coke-Oven Door Seals". This proposal was to be
 based  on the conclusions and recommendations of the July 1975 Concepts Study. As given
 above, the title of the possible new project was somewhat of an understatement because the
 Sponsors were interested in  research and  development of all aspects of  end closures that
 relate directly or indirectly to emission control. Battelle-Columbus presented a proposal that
 was responsive to the following points:
    (a) Follows through and develops further the conclusions and recommendations of
        the Concepts report.
    (b) Schedules arrangements to present our results and recommendations for the
        Phase III demonstration prior to the end of the project (as in this report) for criti-
        que by the EPA, members of the AISI and their research organizations, and other
        interested organizations. This critique period allows time for evaluation  by the
        Sponsors as to whether the follow-on demonstration project should be funded,
        and allows continuation of the overall program without interruptions between
        major phases.
    (c)  Takes into consideration and makes use of the end-closure and emission-control
        data to be supplied by several large steel company research laboratories. Battelle
        researchers were also to obtain measurement data, but were  to make indepen-
        dent conclusions and recommendations.
•The address of the NTIS is 5285 Port Royal Road, Springfield, Virginia 22161.

-------
                                          1-3
     (d)  Defers plant testing of recommended seals, jambs, doors, and other end-closure
         components until given technical and administrative clearance by EPA and AISI
         to do so.
    Our proposal was accepted and a contract was awarded in August of 1976. The develop-
ment work, including engineering drawings of our proposed designs, will be completed by
the end of August, 1978.

                                  Project Objectives
    The objectives as stated in the contracts with EPA and AISI are as follows:
          The general objective is to innovate and to develop at least one new system
      that will be proven in the field (in an optional follow-on project) to be retrofit-
      table to existing coke ovens, mechanically and physically suitable for commercial
      use in steel plants, and  highly effective in containing and controlling emissions
      from the ends of ovens.
          The primary objective is to develop a retrofittable metal-contact seal that (a)
      is more flexible in terms of conforming to large distortions of jambs, (b) is more
      heat resistant than existing seals, and (c) will conform to warped jambs without
      the need for periodic adjustments by coke-plant workers.
          The end-closure system also includes the jamb (along with the mounting of
      the jamb and  other considerations).  Because the warpage of jambs is the fun-
      damental cause of the emission-releasing gaps between the jamb and the seal, it
      is a second objective of this project to develop a more dimensionally stable jamb
      for both new coke-oven batteries and for replacement  of some jambs at existing
      batteries.
           It  is a third objective to develop a bank of quantitative information that
      describes the conditions and variables that affect the performance, effectiveness,
      and life of coke-oven end  closures. Analysis of the technical results and the
      quantitative information obtained should result not only in an effective solution
      to the problem, but also pinpoint why it is effective and  what needs to be done in
      the field to maintain effectiveness.
          A fourth objective is to explore the potential of systems using sealants. In
      Contract No. 68-02-1439, "A Study of Concepts for Minimizing Emissions From
      Coke Oven Seals", a number of unknowns surfaced in the subjective evaluations
      of conceptualized sealant systems. Sealing via the use of sealants is included for
      study in a laboratory research program to evaluate the potential applicability to
      the general objective.

-------
                                          1-4

                              Organization of the Project
    The overall development project was divided into six tasks as shown in Figures 1-1 and I-2.
This report, for example, is the Interim Report listed as Task 4 in the following figures. The Op-
tional Tasks shown are part of a possible follow-on demonstration project, with implementa-
tion at the option of the Sponsors.
    Our approach was to carry out the first three interrelated tasks concurrently. The titles of
these tasks are as follows:
    Task 1: Mathematical Modeling  and Analysis of End-Closure Systems (Seals, Jambs,
            and Doors)
    Task 2: Physical Modeling and Laboratory Experimentation
    Task 3: Field Data Collection and Field Experimentation
    As a general statement, the data collected in Task 3 were used as inputs to Task 1 and Task
2. The analytical results of Task 1 were verified in Task 2.
    Tasks 1, 2, and 3 were conducted on an accelerated basis over a period of 16 months. The
completion of Tasks 1, 2, and 3 was followed by Task 4 which is entitled "Analysis, Evaluation,
and Recommendations". This is a 3-month task which  is summarized in this interim report. The
recommendations of Task 4 will be implemented  in Task 5, which is entitled "Full-Scale Unit
Design and Testing of Parts". "Testing of parts" anticipated a need to study the sealing (and
bending)   action  of  sections   (and corners)  of  recommended,   retrofittable   seal
design(s)/materials.
    The last task in this project is Task 6: "Decisions Regarding Full-Scale Installations". This 2-
month task represents a period of interaction with  the representatives of the Sponsors in-
volving decisions required to initiate or define the Optional Scope of Work (a demonstration
project).
    The general timing of the project, and the timing of the tasks within the project, are
shown in Figure 1-2. Also shown is the timing of the Optional Project. Overall, the timing of the
basic project has been scheduled  as short as possible so as to find, as rapidly as possible, an
acceptable solution to an air-pollution problem.

                                    Research Staff
    This research is being conducted by coordinating the efforts of researchers within three
Departments at Battelle:
        (1) Metallurgy Department
        (2) Structures and Mechanics Research Department
        (3) Engineering and Manufacturing Technology Department.

-------
      BASIC PROJECT
                                         1-5
TASK 1
                                    TASK 3
                                         TASK 2
MATHEMATICAL MODELING
AND ANALYSIS OF
END-CLOSURE SYSTEMS
(JAMBS. SEALS, DOORS)
>^

FIELD DATA
COLLECTION AND
FIELD
EXPERIMENTATION
^**

PHYSICAL MODELING
AND LABORATORY
EXPERIMENTATION
AND TESTING
                                       MATERIALS
                                       TECHNOLOGY
                        	^	ANALYSIS OF RESULTS
                                   TASK
                                     RECOMMENDATIONS
                                     (INTERIM REPORT)
                               TASK 5
                                  DESIGN AND COMPONENT
                                TESTING OF FULL-SCALE PARTS
                                             ~l
                               TASK 6
               \      I
                   i
                                   DECISIONS REGARDING
                                 FULL-SCALE INSTALLATIONS
                              OPTIONAL TASK 1
                \    I
                 \   I
                                 FABRICATION, INSTALLATION
                                       AND REWORK
OPTIONAL TASK 2
i
                           DETAILED PLANNING FOR FIELD EVALUATION

                                   FIELD EVALUATION
                             OPTIONAL TASK 3
                   I
                                ANALYSIS AND PREPARATIONS
                               OF MANUALS AND FINAL REPORT
                                                  OPTIONAL
                                                   TASKS
      FIGURE 1-1. OUTLINE OF TASKS IN THE BASIC AND OPTIONAL PROGRAM AND
                 SCHEDULE OF SUMMARIES AND REPORTS

-------
                                                  DEVELOPMENT AND DEMONSTRATION OF CONCEPTS FOR IMPROVING COKE OVEN DOOR SEALS

                                                                              MONTHS
BASIC PROGRAM     TASKS              _1  2  3  4 5  6 7  8  9 10 11 12 1  14 15 16  17 18 19 20 21 22  23 24 25 26 27 28 29 30 31 32 33  34 35 36 37 38 39 40
      1.  MATHEMATICAL MODELING
         AND ANALYSIS OF SYSTEMS
         (JAMBS, SEALS, DOORS)


      2.  PHYSICAL MODELING AND
         LABORATORY EXPERIMENTATION
      3.  FIELD DATA COLLECTION AND
         FIELD EXPERIMENTATION
      4. ANALYSIS, EVALUATIONS,
         AND RECOMMENDATIONS
      5.  FULL-SCALE UNIT DESIGN
         AND TESTING OF PARTS
       6. DECISIONS REGARDING
         FULL-SCALE INSTALLATIONS
           OPTIONAL PROGRAM
       1. FABRICATION, INSTALLATION
         AND REWORK
       2. PLANNING AND COMPLETION OF
         FIELD EVALUATIONS
       3. ANALYSIS AND PREPARATION OF
         MANUALS AND FINAL REPORT
         SUMMARY SCHEDULE
         REPORT SCHEDULE
 '  II  I  I   I  I  I   I  I   I  I  I   II  T
               .- TASK SUMMARY
	r
                        - TASK SUMMARY
                                                                  . TASK SUMMARY
                                                                                 .INTERIM REPORT I
                                                                                           -DESIGN SUMMARY
                                                                                                 INTERIM REPORT II*
                                                                  OPTIONAL
                                                                                                               DRAFT FINAL REPORT,
                                                                                                                AND MANUALS
           XX
                                         X
_RNA_LJ*EPORT

      X
                              FIGURE 1-2. OUTLINE OF ACTIVITIES IN THE EPA/AISI RESEARCH PROGRAM ON
                                           IMPROVED SYSTEMS FOR SEALING COKE-OVEN END CLOSURES
                              "This report becomes a Final Report if the Sponsors do not elect to fund the Optional Program.

-------
                                          1-7
                    j
      •>
   The key or lead personnel within the three departments are as follows:
        A. O. Hoffman, Project Manager, Metallurgy Department
        H. W. Lownie, Technical Advisor, Metallurgy Department
        A. T. Hopper, Leader of the analytical effort, Structures and Mechanics Department
        R. L Paul, Leader of the design effort, Engineering and Manufacturing Technology.
   The Sponsorship personnel most directly concerned with this study are:
        Mr. Robert C. McCrillis, Project Officer for EPA
        Mr. John G. Munson Jr., Project Officer for AISI (now retired)
        Mr. Richard G. Phelps, Project Officer for AISI
        Mr. Calvin Cooley, AISI Headquarters, Washington, D.C.
        Mr. George C. Bennett, Contract Administrator for EPA
        Mr. M. P. Huneycutt, Contracting Officer for EPA.

                                  Acknowledgments
    Having the support of the AISI on this project was advantageous to Battelle researchers
working on this project. This was particularly evident during our plant visits and field-test
periods. During this and the prior project, Battelle researchers could have probably visited and
worked at any coke plant in North America.  For the support of the AISI and the input of many
coke-plant personnel, we wish to give acknowledgment.
    We give  special  acknowledgment to (a) the assistance given  to us at each of the coke
plants where we elected to obtain detailed measurements, and (b) to the United States Steel
Corporation  research organization for furnishing their internal technical report entitled
"Coke Oven-Door System Technology: Field Data From Clairton  Works" (December 1976).
We are fortunate in  being able to quote from  this report and in this manner are able to
acknowledge the technical contribution  of  the United States  Steel Corporation to this AISI
project. While several steel companies gave us permission to examine their internal reports,
the formal and detailed USS report had a significant impact on  our insights into the problem.*
    In our  field work there were  many examples  of concern for  our safety and well being.
These are particularly appreciated. In one instance, an Assistant Plant Superintendent came to
us where we were working on the  bench, handed us his personal snow shovel, and ejected us
from the plant. His objective was to minimize explanation time and get us out before the bliz-
zard hit. His action saved us from being snow bound with him  for several days or longer. Dur-
ing our "retreat" we had to use his snow shovel even before we had driven to the plant limits.
"The United States Steel Corporation has asked that requests for this USS report be directed to them.

-------
                                        1-8

    We would have liked to give acknowledgment to all individuals who gave the project
team assistance on this project but ran into a problem with "where do we stop the list?"

                Project Manager's Comments and Qualifying Statements
   (1)  Two important considerations in our research project are (a)  maximum retrofit-
       tability and (b) minimum visible emission from  the end-closure system. The
       retrofittability aspect limited the range of modifications that could be considered.
       Our recommendations, therefore, may or may not be appropriate for the taller
       ovens of the future. It is hoped, however, that our effort will be helpful in some
       aspects of new  designs for end closures.
           While our  Proposal and Work Plan listed our major objective as "develop-
       ment of at least one new system—that will be highly effective in containing and
       controlling emissions from the ends of ovens"  (no standard given), our target is
       complete elimination of emissions—including the first hour of any coking cycle—
       over a long period of operational time. Whether or not this can be achieved with
       the recommendations included in this report remains to be established.
    (2) Some of the comments  and judgments in this report could be construed  as be-
       ing critical of existing door-sealing designs and of the approaches of the builders
       of  coke batteries. In addition, discussion of any aspect of coke-oven design is a
       sensitive subject for one commercial interest or another. It should be appreciated
       that we are evaluating designs developed in the early 1940's (or earlier) when
       there was a different set of performance standards and competitive restraints.
       Our attention is directed towards the good and not-so-good features of standard
       designs as referenced to a new set of performance standards. In instances  where
    ,   we have judged that some feature of a design is desirable and should be retained,
       we have given credit to the builder.
           Our discussions with builders have been limited, possibly because they are
       actively developing their own new and  retrofittable designs. If this is the case, this
       bodes well for the development of several acceptable upgraded systems.
   (3)  Almost all of the older coke batteries in  North America (all shorter than the
       newer 6-meter  batteries)  have been built  by  either the Koppers or Wilputte
       engineering organizations. The major thrust of our research  effort has been to
       develop upgraded, retrofittable metal seals for the replacement of existing seals
       on  these "brands" of batteries. Within each brand  of battery there are variations

-------
                                    1-9
   in size, age, and design. Our development of engineering drawings (Task 5 is now
   in progress) will, however, be based on "typical" end closures of these two com-
   panies.  Koppers and  Wilputte  furnished Battelle-Columbus with end-closure
   drawings that they consider to be typical.  For this cooperation  and for these
   drawings we thank these two organizations. It is expected that the  manufacturers
   of upgraded  seals  will be able  to modify the recommended seal design to fit
   variations in the existing sizes and designs of end-closure elements.
(4) Because of the large amount of work that has been done and the large amount of
   data obtained, our  Sponsors have requested that this interim report be presented
   concisely. Therefore,  we have  not included the  working paper  summaries of
   Tasks 1, 2, and 3 in this  report.  Where necessary  to  our conclusions and
   recommendations, we have included extracts from these working papers. Copies
   of these working papers  have been sent to our Sponsors' representatives and
   copies can also be  made available to those with specific interest in details and in
   mathematical analyses. We suggest that the  EPA/AISI  report entitled "Study of
   Concepts For Minimizing Emissions From Coke-Oven Door Seals"  be considered
   as reference material for this report.
(5) Measurements during this study were made in terms of customary U.S. units.
   However, measurements of base units (length, mass, time, and temperature) are
   reported in SI units, in general agreement with conversion practices as given in
   ASTM Designation  E 380-72; "Standard Metric Practice Guide; A Guide to the Use
   of SI". To  enhance readability and rapid comprehension, customary U.S. units
   often are also given, and customary U.S. units are used for derived units (such as
   energy, force, pressure, and power).

-------
                                        11-1
                                   CHAPTER II

                       ANALYSIS OF METAL-SEAL SYSTEMS AND
                  RETROFITTABLE METAL-SEAL RECOMMENDATION
    The conclusion of the 1975 EPA/AISI report ("Study of Concepts For Minimizing Emissions
From Coke-Oven Door Seals") was that, of all the concepts considered, Battelle researchers
rated upgraded metal-to-metal seals as the best retrofittable approach to minimizing the emis-
sion problems from coke-oven door seals. Only upgraded metal seals were given a rating in-
dicating a 90 to 100 percent probability of successful development and successful performance
within the criteria specified jointly by the EPA and the AISI. This chapter presents a summary of
new work completed to advance the state of the art for upgraded, retrofittable metal-to-metal
seals.

                        Comments on Standard Designs for Seals
    Of the coke-oven batteries that have  metal-to-metal seals, about 25 percent have the
fixed-edge design shown in Figure 11-1. About 70 percent have the S-shaped design shown in
Figure II-2. Our research and development target was to evaluate the technical strengths and
limitations of these designs, and either to upgrade one or both of these designs to obtain im-
proved performance, or to recommend new designs and materials. As a basis for evaluation
and subsequent conclusions and recommendations, the inputs from Tasks 1,  2, and 3 are
presented in the following paragraphs.
                               Input From Tasks 1 and 2
                         (Mathematical and Physical Modeling)
    All of the end-closure components play some role in sealing the enas of coke ovens. Most
of these  components, however, play a minor role compared with (a) the conformity of the
jamb-seal surface and the seal edge, and (b) latch force. Seals and latch forces are discussed in
this section of the report. Jambs and other related aspects of end closures are covered in other
chapters of this report.
    Fixed-edge seals and S-shaped seals were considered in this study. These will be shown in
more detail in the following discussion. In addition, there is a variation of the fixed-edge seal
often called the "knock-type" seal. One example of this variation is shown in Figure li-3. With
the popular fixed-edge seal (Figure 11-4), adjustments to the seal are made by backup screws
which are rigidly attached to the body of the door. In the knock seal, coarse adjustments are
made with eccentric cams spaced around the periphery of the seal, while fine adjustments are
made by  "knocking" the seal from behind with a hammer.

-------
                                      1-2
    ENLARGEMENT OF THE
    SEAL DESIGN

        Jamb
    Adjustment Screw

  FIGURE 11-1. CROSS SECTION OF A TYPICAL COKE-OVEN DOOR WITH A FIXED-EDGE SEAL
                                                   Seal Ring
                                                                    S-Shaped
                                                                    Seal
                                                    Spring
                                                    Backup and
                                                    Adjustment Arrangement
FIGURE 11-2.  CROSS SECTION OF A TYPICAL COKE-OVEN DOOR WITH AN S-SHAPED SEAL

-------
                               11-3
                 jamb
      FIGURE 11-3.  CROSS SECTION OF A TYPICAL "KNOCK-TYPE" SEAL
                                     JAMB
            ADJUSTMENT
            SCREW
FIGURE 11-4. SIDE VIEW OF ADJUSTMENT EQUIPMENT ON FIXED-EDGE SEALS

-------
                                         11-4

     Both of the fixed-edge seals have features in common. The latch force is generated by a
 door-mounted spring  system. Because  the latch forces  "clamp" the jamb  and the door
 together, this system is self-balanced. The total contact force between the seal and the jamb
 equals the latch force. Another feature in common is the rigidity of the seal. Rigidity in this
 context does not refer to the resistance of the seal edge to bending (although this is present in
 these designs), but rather to the fact that these seals lack capability for automatic adjustment to
 small changes in the profiles of the jamb or  door. This feature of the  rigid sealing system
 manifests itself in the development of leaks and in a nonuniform sealing pressure around the
 seal, i.e., the contact pressure between the seal and the jamb is different at one place on the
 seal than another. This is depicted in Figure 11-5 where the sealing pressure is shown  to be
 significantly higher under the latches than between the latches. Once contact has been  made
 between the seal and the jamb at the latching levels, increasing the latch pressure will increase
 the sealing pressure only under the latches. Thus if the seal is leaking elsewhere, only manual
 adjustments near the leak can cure the situation.
     The  two points made so far—uneven sealing pressure and low effectiveness of increase in
 the latch force—are easily  envisioned by taking two yard sticks and squeezing them together
 at two locations. This is depicted in Figure 11-6. The pressure between the edges of the rules
 will clearly be higher at the locations where the rules are being squeezed together than say be-
 tween the clamps. Moreover, squeezing harder will make little or no change  in the contact
 pressure at locations between the "latch levels".
     Because  of the preceding two points, and because the rigid seals cannot of themselves
 assume a different contour, rigid seals are susceptible to development of leaks  if a  slight
 change in the contour of the jamb or door arises because, for example, of changes in the ther-
 mal gradients. Such leaks can be stopped only by adjusting the seal manually with the available
 devices.  Adjustments to the seal  while the door is in place can have several possible conse-
 quences, each of which needs to  be considered.
    Recall from preceding  discussions that the latch force is distributed around the seal/jamb
 interface as a contact pressure which we have referred to as the sealing-pressure distribution.
The  total latch force equals the  sum of  the contact force around the seal perimeter. Any
 manual  adjustment to the seal  is  made  with  the intent of changing the  seal pressure
somewhere around the seal, most probably to increase it from zero where a leak is occurring.
Thus, any adjustment will either increase, leave unchanged, or decrease the total latch force
and will  redistribute the seal pressure. Because intuition says  that if there is a leak there is a
gap, and, if the seal edge needs to be advanced to close the gap, most adjustments will result in

-------
                                 1-5
              CONTACT PRESSURE
              BETWEEN SEAL
              AND JAMB
             Total X
             Latch
             Force
            (Upper)
            Total
            Latch
            Force
           (Lower)
            FIGURE 11-5. DISTRIBUTION OF SEALING PRESSURE
              Hand Grip
                Forces
                                        Paper Is Lightly
                                           Gripped
FIGURE 11-6. AN EXAGGERATED REPRESENTATION OF THE CONTACT PRESSURE
           DISTRIBUTION BETWEEN TWO CLAMPED RULES

-------
                                         11-6

 increasing the latch force. Note that it is not inconceivable that a gap could be closed by
 loosening the appropriate adjustments but to know how to do this requires skill and ex-
 perience.
     If the latch force is increased by adjustments made to the seal while the door is in place, in
 the fixed-edge systems this implies that the main latch springs are compressed further and the
 door main body is a bit further from the jamb. This design can reproduce this increased latch
 force on the next placement  of the door  provided the springs are not fully compressed
 through relaxation or through damage from  fires.
     The fixed-edge designs have two distinct positive features which need to be emphasized.
 One of these is the point just made that the springed latching force has the ability to increase
 and decrease as fine adjustments are made to the seal and, more importantly, the ability to
 reproduce this  new sealing force automatically. The other advantageous feature is the way in
 which the  entire seal element is adjusted by moving it back and forth. These features are
 recognized as important and will be discussed again relative to our recommendations.
     The S-shaped seal  system shown in Figure 11-7 has a more flexible sealing element and is
 backed up by springed plungers spaced about every 20 to 25 cm (8 to 10 inches) around the
 seal. These devices give adjustability but with the retention of flexibility at the seal edge. The
 latch force in this system is usually applied by a screw mechanism that is motor driven and cuts
 off by means of a  current limiter to the drive motor. The motor and drive are a part  of the
 door-handling equipment. One other feature of this system which does not show in the figure
 is the presence  of stops to keep the door from advancing so far as to overstress the seal. These
 stops are at the corners of the doors on 4-meter ovens and  at the corners and two places in
 between on the taller 6-meter ovens.
     The important differences between this system and the rigid seal systems are in the latch-
 force devices and in the seal design. Because of the way the latch force is generated on the S-
 shaped  seal system, it will not respond to adjustments in the manner described previously for
 the rigid seals. Here, with the door in place, adjustments to the seal will again most probably
 increase the latch force above what was present on placing the door. Assuming the latch force
 to be increased in  this way, when the door is removed for pushing and then replaced, the
 door-handling machine will essentially reproduce the  previous  lower latch  force.  This will
 cause still another distribution of latching force. If leaks occur, adjustments will increase the
 latch force  once more to a level which cannot be reproduced by the drive motor when the
door is placed in the next cycle.
    This difficulty with the more flexible S-shaped seal system can be overcome by  incor-
porating a spring-latch system along with the spring seal.

-------
                                          1-7
                          Sealing Ring
Jamb
                  FIGURE 11-7. SIDE VIEW OF S-SHAPED SEAL SYSTEM

    There are advantages of the flexible seal which are worth noting. Flexibility is used in this
discussion to mean the ability to flex or to bend in a spring-type fashion. Thus, the flexible seal
can automatically adjust  to small changes in the jamb or door profile without losing sealing
pressure. Flexibility also means that the sealing pressure at any point can be about the same as
at any other point—that is the sealing pressure can, in theory, be relatively uniform around the
door. A final advantage which can be expected but which  is not quite so obvious is that, for a
given latch force, the flexible seal is capable of  sealing a greater mismatch between the con-
tours of the seal and the jamb. While this was expected, the Task I  analysis indicated that the
difference is more than minor.
    Up until this point in our discussion, the topics have been limited to features which were
particular to either the rigid or to the flexible  seals. Certain features which they have in com-
mon are also worthy of mention. One way a seal edge can be permanently damaged is by sub-
jecting it to  a force which develops stresses in excess of its yield strength—a property of the
material. Both flexible and rigid seals are easily damaged if they are banged against the jamb or
against the latch hooks by the door-handling machinery. But there are other, and more-subtle
ways of damaging a seal by overstressing. Both of the sealing-system designs have a common
feature that  is undesirable and which the research team has termed "point loading". Referring

-------
                                         11-8

 to the drawings of the various designs, it is clear that each seal is affected from the rear by
 cams, screws, or spring-loaded plungers. Each of these devices applies a force at a small area at
 the back of the seal. The difficulty with point loading is that it causes a stress concentration,
 particularly on the more flexible S-shaped seal. Overstressing of the seal at such a point can
 result in damage by yielding or in dimensional changes resulting from creep and relaxation. In
 addition, point loading does not necessarily guarantee distribution of the load between the
 points of force application. In this regard, the situation is somewhat similar to that described in
 the previous discussion concerning latch-force distribution with fixed-edge seals. It is for these
 reasons that it was recommended to the design researchers that consideration should be given
 to the elimination of point loading on the rear of seals. This recommendation was considered
 to be particularly important where the seal system is to have a high degree of flexibility.
    A more difficult similarity to discuss is the yielding of the seal member through bending.
 The rigid seals are bent as a unit. The entire rigid seal is bent or deflected to attempt to con-
 form to the jamb contour. Flexible seals on the other hand  are bent in a rather complicated
 way. The rear of the seal is attached to the body of the door  and thus has the same contour as
 the door. The sealing edge, on the other hand, is loaded  by  plungers and bends in a way that
 may be contrary to the rear of the seal. Thus the seal is being asked to contort in ways which
 may cause it to yield permanently. The analysis in the Task I effort also showed that the amount
 of forward or backward  adjustment of the spring seal that can occur before it yields depends
 upon whether only the leading edge of the seal is bent to conform to the door profile or
 whether the entire seal is customized by initially shimming it to match the jamb profile.
    The physical-modeling experiments in Task 2 were concerned mainly with  a study of
 thermal-stress variations in jambs and doors. Rather than attempting to evaluate new seal
 designs on the heated quarter-scale model, the work  plan calls for evaluation of full-size sec-
 tions of seals in a separate task (Task 5). This work is now in progress.
    During physical-model experiments, however, it was learned that even a small fire bur-
 ning at a gap between a seal and jamb can heat the seal material to about 810  K (1000 F).
 Because of the thermal mass of the jamb and door, these components do not rise rapidly in
 temperature when near small fires. It was judged that localized heating by fire of any seal
 design/material under conditions of constraint and seal-bending stress can be detrimental to
 the seal.
    When the physical model was cycled in temperature to mimic operational conditions,
sealing between  the seal and the jamb was not lost at any  time. Although the temperature
levels  were changed, the temperature gradients remained almost constant. The gradients
remained near constant because the temperature changes occurred slowly. Tests were run in

-------
                                          H-9

which the end-closure component gradients were increased rapidly by spraying water against
the equipment in a simulation of a rain storm. There was a decrease in latch loads during the
water-cooling period. This decrease had to be accompanied by a change in the distribution of
sealing pressure and a change in the profile of the door or jamb, or both.

                         Input From Task 3 (Field Data Collection)
    When obtaining measurements of jamb profiles at 4-meter and 6-meter batteries having
both  S-shaped and fixed-edge seals, the performance of seals was observed and numerous dis-
cussions were held with operational personnel about their experiences (and improvement ap-
proaches) with the performance of the two types of seals. Information that relates to our ob-
jective of upgraded metal seals is as follows:

S-Shaped Seals
    Members of the project team have witnessed 100 percent emission-control performance
with new and well-adjusted S-shaped seals in operation against relatively straight and relatively
clean existing jambs. Performance of "100 percent" means no visible emissions at any time, in-
cluding the period during and just after charging of coal to  the oven. "Well-adjusted" means
that considerable skilled effort was expended to adjust the  plunger-spring pressures that are
acting as a point loading against the back of the seal edges. This observation relating to seal
performance agrees with the analysis summary of Task 1 to the effect that the S-shaped seal is
fundamentally a good approach. This observation also confirms that, in general, the metal-to-
metal seal approach is sound.
    However, it was observed that with time (often less than 6 months), the emission-control
performance of these standard seals begins to deteriorate.* Deterioration of standard S-seal
performance also occurs at 6-meter batteries where the jambs are mechanically cleaned, the
jambs are not warped, and the seal edges have not been nicked. Reasons for this deterioration
will be discussed in a later section.

Fixed-Edge Seals
    Members of the project team have never witnessed 100 percent emission-control perfor-
mance with fixed-edge seals.  Operational personnel were  asked in several instances to
attempt to reach 100 percent control by adjusting the seals working against both well-cleaned
*At project review meetings, U.S. Steel Research personnel have reported that in 20 months of coke-plant opera-
 tion, their new seal design has evidenced no sign of performance deterioration. Information on the U.S. Steel seal
 design should be obtained from them.

-------
                                         11-10

 jambs and uncleaned jambs. It is appreciated that it takes skill and experience to adjust a fixed-
 edge seal, because the seal itself acts as a beam so that forcing it inward at one location can
 result in a rocking action causing a force on the beam acting outward at another nearby loca-
 tion. Even with diligent and experienced effort, 100  percent emission control  was never
 achieved. Leakage is particularly severe when attempting to adjust the fixed-edge seal in con-
 tact with a clean jamb. This is a negative observation on fixed seals, but it should be noted that
 there is at least one battery having fixed-edge seals that has a low percentage of leaking doors.
 It has been stated that, to reach a standard of less than 10 percent leaking ovens on a battery, it
 is necessary to lower the average leaking time per oven (both doors) below 1 hour. There is at
 least one  fixed-edge-seal battery where the oven leaking time is rather uniformly between 15
 and 20 minutes, indicating about 3 percent door-leakage rating for the battery. Our investiga-
 tion at one of these "rapid-sealing batteries" indicates that (a) the jambs are only mildly bow-
 ed, and (b) the door and jambs are bowed in the same direction, i.e., the door and jamb were
 almost congruent. It was judged that the jamb deposits in this instance helped fill  gaps. The
 amount of leakage and the duration of leakage was always high when we asked operators to
 scrape the jambs to bare metal.
     Task 1 results indicated that with fixed-edge seals it is not possible to obtain a transfer of
 the door-latching force to  a well-distributed seal/jamb contact pressure. Considering that our
 research goal is long-life, 100 percent emission-control performance, Battelle researchers lost
 interest in the fixed-edge  seal approach except for (a) the concept of attempting to achieve
 jamb and  seal-edge congruency by moving the entire seal (as opposed to  bending part of the
 seal in the S-shaped seal approach) and (b) the possibility that a flexible seal element could be
 mounted  inboard of the existing sealing plate.

 Jamb/Door-Frame Profile Relationships
     Our "Concepts Report" stated rather positively that "the door-mounted seals often
 received the blame for coke-oven  emissions, but the  fundamental cause of the  emission-
 releasing gaps is the pronounced degree of warpage that has occurred in differing degrees on
 most (if not all)  of the 25,000 or more cast-iron jambs in operation". This is what we believed in
 1975. In 1978 we say  that the 1975 statement was somewhat of an oversimplification—there are
 other factors that are pertinent. It is the purpose of this section of the report to indicate our
 "shift" in stand and  the relationship of this more complete understanding to our recommen-
dations.
    Based on our past labeling of the "warped" jambs as the culprit in emission problems, in
this development project we initiated a subtask of Task 3 entitled "Development of a Practical

-------
                                        11-11

Profilometer" (for measuring jamb contours). This profilometer development, coupled with
the measuring of jambs in various plants, consumed a considerable amount of effort allotted
to Task 3.
    With time and experience we concluded that:
    (a)  Operating jambs, as a generalization, are not so severely warped as we had ex-
        pected nor so warped as indicated by some steel-company data that were in-
        cluded in the Concepts Report.
    (b)  Our early views considered jamb warpage to be any displacement outside of the
        plane between the bottom and the top of the jamb, i.e., any deviation from flat.
        This viewpoint had to be modified when it became apparent that (a) if a jamb
        were allowed to take its "natural" thermal-gradient-induced inward bow, such
        bowing would be beneficial in lowering the thermal-stress level of the jamb, and
        (b) some jambs had rather smooth inward bows coupled with door frames hav-
        ing about  the same inward  bows.* At the present time we are not concerned
        with any rather abrupt profile changes that may span only a short distance along
        the jamb. It  may prove difficult for a seal to conform to these  short-pitch
        variations. However, based on our measurements and observations, we do not
        expect to find many jambs having short-pitch problems. Quantification of what
        constitutes "out-of-bounds" warpage is part of Task 5, which has been started.
    (c)  The most important measurement, with  regard to retrofittable seals, was found
        not to be that of the jamb profile itself but rather the measurement of the varia-
        tion in the horizontal distance between the fastening point of the seal on the
        door frame and the seal-mating surface on the jamb. This distance is discussed in
        the following paragraph.
    Measurements were taken  (from a vertical  reference plane) on the end closure of a 4-
 year-old, 6-meter battery to obtain the relative profiles of the jamb sides and the outboard side
 of the flat, steel-door frame. The thickness of the cross section of the steel door frame was con-
 stant along the length of the door and the measurements were used to calculate the variation
 in horizontal distance between the jamb and the line on the door where the S-shaped seal was
 fastened to the door. These data could be  presented as profile relationships, but it is more in-
 formative to adjust the data to simulate the first contact of the seal and jamb. These adjusted
 data are as follows:
*An inward bowed jamb or door has its ends moved away from the oven.

-------
                                         11-12
                 Vertical Distance from
                                             Distance Between Seal
                                           Edge and Jamb When First
                                             Contact is Made, mm
                  Bottom of Jamb, m         Left Side       Right Side
                         6.6                   5.1            5.7
                         5.7      ,  „ .         2.0            2.8
                         4 8-Latch Point       Q Q            0 2
                         3.0                   0.4            0.0
                         1.5   .   .  _ .  t       3.6            0.1
                         0 9-Latch Point       ^            Q5
                         0.1                   5.0            3.9
In this case, both the jamb and the door had an inward bow with the latched door having a
greater inward  bow (smaller radius of curvature) than the jamb by about 5 mm (about 0.2
inch). The measured curvature of the door was more pronounced above and below the latch
points.
    It was concluded that, unless the upper and lower edges of the seal had been pushed
forward (inward) by increasing the backup  spring pressure in these areas, upon latching the S-
seal portion between the latches had to bend back about 5 mm (about 0.2 inch) before the seal
edge made first contact with the top and bottom cross pieces.
    Relative to  other existing jamb and door profile relationships, the above example could
be labeled a "good" congruency. There are instances in which the jamb and the door body
bow towards each other. A sketch of this relationship and other variations that exist in terms of
relative contours are shown in Figure II-8.
    Most existing  coke-oven door frames are rough castings. This roughness presents  a
problem in terms of obtaining accurate and credible door-profile measurements. When this
difficulty was encountered, the design researchers were asked to develop a tool to directly
measure the distance from the jamb seal-mating surface to the inboard fastening point of the
S-shaped seal on the companion door. The first tool that was designed  and used is shown in
Figure II-9. The way in which this tool was used is shown in Figure 11-10. It is expected that this
tool can be improved for use in a follow-on demonstration project. Because of the difficulty of
taking readings near the top of a door, there is a pronounced tendency to take them fast, and
thereby to penalize accuracy. A modification of this tool could include electronic sensing
equipment, perhaps with a digital readout.

-------
                                1-13
                                       Away from Oven
     Jamb
 Door
 Jamb
     Seal
 Example A. Horizontal distances
           between jamb and door
           are all the same.
                                                             Door
                 Seal
             Example B. Door has less curvature
                       than jamb. Jamb corners are
                       closer to door than at the
                       middle of jamb/door.
    Jamb
    Seal
Door
Jamb
Door
Example C. Door is curved inward and
          is facing either a straight
          jamb or distance between
          jamb and door is greater
          at the corners.
                Seal
             Example D. Jamb and door bow towards
                       each other.
    FIGURE II-8. EXAMPLES OF VARIATIONS THAT EXIST IN TERMS OF THE
               RELATIVE CONTOURS OF JAMBS AND DOOR FRAMES

-------
                                11-14
                 • Lock Ring
                      Scale
                    Rotatable, Spring-Loaded Contact Arm
                                                                Saddle
FIGURE 11-9. TOOL DESIGNED FOR MEASURING THE HORIZONTAL DISTANCE
           BETWEEN THE INBOARD EDGE OF ANY DOOR FRAME AND
           THE SEALING SURFACE OF THE ACCOMPANYING JAMB

-------
                                          1-15
                       Door Frame
        FIGURE 11-10. SKETCH SHOWING OPERATION OF THE MEASURING TOOL
                     Distance being measured is being the seal surface on the
                     jamb and the inboard edge of the door frame.
 Damage/Changes Observed in Operating S-Shaped Seals
    This section of the report deals with the observed or measured damage and dimensional
 changes that often take place in standard S-shaped seals.
    Coke-plant superintendents often indicate that they think that their 304 stainless steel S-
 shaped seals are "too soft" and "damage too easily". Relative to other materials, 304 stainless
 steel does have these disadvantages. A comparison of 304 stainless with other materials is in-
 cluded in a  following section of this chapter. Many examples of nicks and bends on seals can
 be seen on  visiting or working at coke plants. It was judged that 304 stainless steel should be
 replaced with a harder and tougher material if for no other reason than to minimize dents and
 bending. With regard to this type of physical damage, Battelle researchers did not include any
 effort  in this project to devise methods  for shielding seals  from bumping damage. One ap-
 proach would be the retrofitting of additional bumpers and  guides. There are indications that
 the research organizations of steel companies have done research on this subject, and that the
 results may  become available. The development of suitable door guides and door stops is an
 important consideration in long-life seal  performance. If required, this development may be
 part of a follow-on demonstration project.
    Another type of visible damage is caused by the spring-plunger back-up loading of the S-
shaped seal. Calculations made in Task 1  indicated that, on  the average, about 70 percent of
the latching  load is absorbed by the spring-loaded plungers pressing against the back of the

-------
                                        11-16
seal  edge.  Under conditions where the seal  and plungers are forced  backward in  a large
deflection, damage to the back of the standard seal can occur as shown in Figure 11-11. Shown
is a photograph of a portion of a discarded S-shaped seal. Indicated are the depressions made
in the back of the seal edge by the plungers and the resulting waviness of the seal edge. Accor-
ding to coke-plant superintendents this type of damage is not rare.
    According to calculations made in Task 1,  the use of point loading (as with plungers in the
S-shaped seal system) does not distribute the latching force very well between the plungers. It
is also indicated in  Figure 11-11 that, under some conditions, point loading of the back of the
seal strip can result in (a) physical damage to the seal, and (b) the opening of gaps between the
seal edge and the jamb.
        Indentation from '
         Backup Plunger
Waviness of
Edge of Seal
      FIGURE 11-11. PHOTOGRAPH OF A DISCARDED S-SHAPED SEAL SHOWING
                   INDENTATIONS AT THE PLUNGER CONTACT POINTS
                   AND THE RESULTING WAVINESS IN THE SEAL EDGE
    Creep and  Relaxation. There are indications that the S-shaped seal rather rapidly goes
through dimensional changes during service. These changes occur by the action of yielding as
creep and/or stress relaxation with the result that the seal edge assumes or attempts to assume
the same profile as the mating jamb. U.S. Steel Research data indicate that this accommodation
of the seal can take place in four coking cycles or less. This accommodation appears to be ad-
vantageous in terms of attaining congruency between seal edge and jamb profile, but gaps
between the seal edge and the jamb could develop due to nonuniform plastic deformation.
Battelle researchers believe that the stress level and temperature level vary along the standard
S-seal.  Therefore, the degree of plastic deformation  could vary with the result that gaps can
develop between the seal edge and the jamb.

-------
                                         11-17

    It would be expected that the gap-forming problem caused by deformation would be
more severe at coke plants where the seals are at a higher temperature than at other batteries.
In general, this has been observed to be correct. It is recommended that future batteries
should be designed to have significantly lower operating temperatures on all elements of the
end  closures. However, in terms of retrofit, one approach to be considered is the use of
materials more resistant to plastic deformation.

    Changes in Jamb and Door Profiles. The degree to which operating jambs and doors
change their profiles during temperature gradient changes has a bearing on the amount of
flexure that will be required in upgraded seals. The results of our measurements, with com-
ments, are included in Chapter III of this report. Chapter III deals with retrofittable jambs.
    For this seal chapter of the report, it is appropriate to state that our greatest concern with
changes in relative profile relates to the changes that occur during rain storms that rapidly cool
the door and the jamb. This problem is greater in the taller ovens because the degree of bow-
ing is a function of the square of the height (all other factors being equal).

    Seal Adjustments and "Life on the Bench". To a limited degree, both the fixed-edge seal
and the S-shaped seal are adjustable. As noted in the introduction to this section of the report,
if a considerable amount of "tuning" of a new S-shaped seal is performed by skilled personnel,
a new S-shaped seal can be made to be 100 percent emission free—for a time, and if the door-
to-jamb spacing is within the limits of the adjustability of the seal. This can be done, but it is
not being done, or it is not being done often. There are probably more reasons for why it is not
being done than we appreciate, but, based upon our  experience, it is not being done mainly
because there is really not enough time between equipment moves to do the job efficiently or
effectively. Adjusting seals or making repairs (patching) on oven ends is somewhat akin to
doing ceiling repairs in an operating subway. You have to know the schedule and estimate the
time you have to work. With this information in hand, the next decision to make is whether to
attempt to do the job working from the ground (the bench) or do it right and get up to the job.
Given that the "work" is at the top of the jamb, the next step is to move the scaffolding equip-
ment  into place (team operation) and  start  climbing. Once in  place, whether doing
measurements or repairs, the feeling persists  of "get the job done fast and get the heck down
from there". This feeling exists because it is not unknown for a machine operator to "wipe
out" a  scaffold arrangement on the bench. In moving from working on a 4-meter battery to
working on a 6-meter battery, the difference is somewhat akin to working on a step ladder as
compared with working at the top of a long extension  ladder. In summary, the equipment and

-------
                                        11-18

the tools and the time are just not available for the job that has to be done. For the taller
batteries of the future, we suggest that the designers first personally work for a brief period at
the top of jambs and then perhaps equip the battery with powered monorail cabs for single-
man repairs.
    As a result of our own experiences taking measurements on jambs and doors, it  was
suggested to our designers that "to the degree possible, attempt to eliminate the need for
above-the-bench adjustments". Based upon our discussions with coke-plant supervisors, this
approach has to be demonstrated  before industry acceptance is obtained.
                             Input From Other Sources
    Prior to the start of this project, it was understood that steel companies having data on
end-closure measurements and  seal performance would share these data with Battelle
researchers. One of the objectives of this sharing was to minimize the duplication of efforts. In
the introduction to this report, special acknowledgment was given to the Research arm of the
United States Steel Corporation for forwarding to Battelle a detailed  report entitled "Coke
Oven-Door System Technology; Field Data From Clairton Works".
    Some of the data presented in this USS report were new to the project team and some of
the data confirmed  our own measurements. Some of the more important information  ex-
tracted from this report is as follows:
    1.  Most of the measured gaps between S-shaped seals and the jambs were in the
       range of 0.4 to 0.8 mm (0.015 to 0.030 inch). For this size of gap, sealing time (filling
       with tar) exceeds 1 hour.
    2. The temperatures of seals and their temperature profiles varied from cycle  to
      cycle.
    3.  USS researchers recorded  an unusual temperature excursion during the  period
      when they were monitoring seal temperatures. With a door on an empty oven for
      about 3 hours, the maximum seal temperature reached about 660 K (725 F). At this
      point, the aspiration steam was turned on for charging. The seal temperature
      rapidly rose to 710 K (820 F), but began to fall rapidly as the oven was charged. In
      subsequent tests, this temperature excursion was not duplicated.
    4. Periodic measurement of the vertical contour of the seal edge interfacing with the
      jamb (taken with the doors  off the oven) indicated that the seal edge rather rapidly
      developed  a  permanent "set" in conformation  with the jamb contour. These
      measurements were taken after four coking cycles. It was stated, however, that this
      yielding of the seal edge to assume the contour of the jamb did not necessarily, by
      itself, result in satisfactory sealing.

-------
                                         11-19

    The above extracts are not direct quotes from the USS report. The information on the
yielding of the S-shaped seals was particularly important to the project team.

                 General Requirements for Upgraded Metal-Seal Systems
    This present development project  is an outgrowth of a preceding project. As in the
preceding project,  the Sponsors specified that they wanted an analytical approach as con-
trasted to an empirical approach.
    The original scope of work for the Concepts Project, as stated by the joint contract with
EPA and AISI, specifies that the techniques and technical advances developed should meet, to
the greatest extent possible, the following functional criteria:
    (a)  Capable of being retrofitted to current and contemplated slot-type coke ovens,
        encompassing all oven heights and construction types
    (b) Compatible with existing door-handling and oven-end working machinery
    (c)  Operability and reliability commensurate with present coke-oven practice
    (d) No creation of additional or different environmental problems
    (e) No adverse effect on product quality.
    During the Concepts Study, evaluation  criteria were  developed by  joint efforts of
 EPA/AISI/BCL. In this instance each criterion was given a weighting factor to indicate a relative
 degree of importance. These evaluation criteria (with comments) and the weighting factors are
 as follows:
                                                                 Weighting Factor
                   Criteria Listing                                  (Scale of 1 to 10)
      1. Relative Effectiveness to Lower Emissions (Stop Smoke)      (highest rating) 10
      2. Low Operating and Maintenance Cost                                   10
        (Note that the above criteria are rated higher than
        initial cost or development cost)
      3. Capability of Retrofit                                                    9
      4. Relative Life (Longer than the standard seals)                               8
      5. Avoids New Safety or Environmental Problems                             8
      6. Avoids Increasing Length of Door-Handling Cycle                          7
      7. Equipment Less Sensitive to Damage, Error, Abuse
        or Nonstandard Operating Conditions                                    6
      8. Relative Cost Installed                                                   6
      9. Avoids Affecting Coke Quality                                           6
    10. Availability of Expendable Materials or Components                        4

-------
                                       11-20

    11. Does It Decrease Operation Complexity and/or
       Minimize Operator Options?                                           4
    12. Relative Maintainability                                                4
    13 Cleanability                                                          2
    14. Operator Skill Requirements                                           2
    15. Cost for Development                                                 1
    In addition to these jointly developed evaluation criteria, in the Concepts Report Battelle
researchers added additional technical specifications in question form as follows:
    1. Will it tolerate a 480 to 590 K (400 to 600 F) operating temperature?
    2. Will it tolerate a temperature excursion to 700 K (800 F) for short periods without
      destruction?
    3. Will it have increased gap-closure capability?
    4. Automatic gap-closure capability? (Can the need for  manual adjustments be
      avoided?)
    5. Resistant to corrosion and chemical attack?
    6. Total-failure proof?
    7. Avoids new cleaning problems?
    After the  presentation of our recommended seal design/material, the new seal will be
critiqued using the above criteria as a frame of reference.

                    Technical Conclusions and Recommendations
                             (Input to the Design Task)
    The summaries of Tasks 1, 2, and 3 with reference to seals have been presented in non-
mathematical terms in foregoing sections of this Task 4 report. These first three tasks provided
the source material for the list of conclusions and recommendations that were developed in
the early part of Task 4 (Analysis, Evaluations, and Recommendations). These conclusions and
recommendations lead to the preliminary designs presented later in this chapter. The con-
clusions and recommendations were:
    1. FOR MAXIMUM VISIBLE EMISSION-CONTROL EFFECTIVENESS, CONVERT ALL
      FIXED-EDGE  AND  S-SHAPED SEALS  TO  UPGRADED  SPRING-TYPE  SEALS.
      SPRING-TYPE  SEALS HAVE INHERENT ADVANTAGES  IN TERMS OF (A)  IM-
      PROVED CONTROL OF EMISSIONS DURING THE EARLY PART OF THE COKING
      CYCLE,  (B) ACCOMMODATING SMALL  CHANGES  IN  DOOR AND JAMB
      PROFILES EITHER DURING IN-CYCLE VARIATIONS, OR TEMPERATURE EXCUR-
      SIONS   CAUSING   HIGH  THERMAL GRADIENTS IN   THE  END-CLOSURE
      COMPONENTS.

-------
                                   11-21

        Most fixed-edge-seal end closures need the retrofit of a spring or flexible type
    of seal to meet performance standards. There may be some batteries with fixed-
    edge seals that, as they stand, (or with an increased level of manual adjustments)
    will meet some accepted standard of performance. Where standards are being
    met, retrofit to spring-type seals will depend upon a costs/benefit evaluation.
2. THE NEW SPRING-TYPE SEALS SHOULD HAVE NO POINT LOADING OF ANY
  KIND. THIS MEANS A SHIFT AWAY FROM SPRING-LOADED CONTACT POINTS
  TO HAVING THE SEAL DESIGN/MATERIAL ABSORB ALL OR ALMOST ALL OF
  THE LATCHING FORCE.
      Stop-bolts may be used to limit the maximum deflection of the seal, depend-
  ing on the particular combination of seal and material recommended.
3. DEVELOP A PREFERRED SEAL-ADJUSTMENT METHOD THAT MOVES THE ENTIRE
  SEAL ELEMENT IN OR OUT RATHER THAN BENDING ONE PART OF THE SEAL TO
  MAKE A CLOSURE. TO BE CONSIDERED ARE THE RELATIVE MERITS AND COST
  OF MECHANICAL ADJUSTMENTS VERSUS PERIODIC SHIMMING OR SPACING
  ADJUSTMENTS.
      This will (a) help to distribute the latching force more evenly in terms of seal/
  jamb contact pressure  and  will also (b) tend to average out the stress level at
  various locations on a seal by minimizing the high-pressure points and "filling" in
  the low-pressure (or no contact) points.
4. CONSIDER  THE USE OF HIGH-PERFORMANCE, HIGH-TEMPERATURE ALLOYS
  FOR SEAL CONSTRUCTION. THIS SELECTION IS TO BE AIMED AT:
  A.  ALLOWING AN  INCREASED  AMOUNT  OF  SEAL  DEFLECTION WHILE
      REMAINING WELL WITHIN THE MATERIAL'S ELASTIC LIMIT AND AVOIDING
      CREEP AND RELAXATION PROBLEMS
  B.  UPGRADING THE  SEAL  MATERIAL IN  TERMS  OF INCREASING THE
      RESISTANCE TO  DAMAGE AND ABUSE  AND  RESISTANCE TO THERMAL
      STRESS.
      It is expected  that  most high-temperature  alloys will  have  acceptable
  resistance to coke-oven corrosion conditions.
5. THE CONCEPT OF RETROFITTING A FLEXIBLE SEAL ELEMENT IN FRONT OF THE
  DIAPHRAGM  OF THE  FIXED-EDGE SEAL SHOULD BE COMPARED WITH THE
  POSSIBILITIES OF INSTALLING ONE DESIGN AND SIZE OF FLEXIBLE SEAL FOR
  BOTH THE EXISTING  FIXED-EDGE SEAL AND THE S-SHAPED SEAL.

-------
                                       11-22

    6. SPECIAL CONSIDERATION AND DESIGN ATTENTION SHOULD BE GIVEN TO
      THE STRESS DISTRIBUTION THAT DEVELOPS ON FLEXING THE CORNERS OF
      SEAL DESIGNS.
    7. CONSIDER INCREASING THE SEAL/JAMB CONTACT PRESSURE BY NARROW-
      ING THE CONTACTING WIDTH OF THE SEAL EDGE. THE CONCEPT OF HAVING
      A REPLACEABLE SEAL EDGE  SHOULD BE CONSIDERED FOR RETROFITTING
      SEALS TO BE PLACED IN OPERATION AGAINST HOURGLASSED JAMBS.
    8. COSTS AND IMPLEMENTATION SPEED SHOULD BE CONSIDERED IN SELECTING
      A RETROFIT DESIGN. IT IS CONSIDERED PROBABLE THAT THE LEAST-COST AP-
      PROACH WILL  BE TO FIT A NEW SEAL  INTO THE SPACE THAT IS PRESENTLY
      AVAILABLE.

                   The Recommended Design of Retrofittable Seals
    During the first  three tasks of this project, various design concepts and the listing of
 recommendations were developed  concurrently.  Following the final  listing  of  desired
 specifications in the  early stages of Task 4,  various concepts were discarded until only one
 remained. This remaining design which became our recommendation is presented in this sec-
 tion. A discussion of material selection is included in a following section along with some dis-
 cussion of the designs that were discarded.  It should be appreciated that the recommended
 design/material is the result of an analytical and pre-engineering effort—the specifics have yet
 to be decided upon and laboratory tested in Task 5. Upon hearing an oral presentation of our
 recommendations, the Sponsors gave permission for the project team to proceed  with Task 5.
 This design and laboratory testing effort is in progress.

 Description of the Recommended Seal
    The retrofittable seal recommended by Battelle is illustrated in Figure 11-12  as it would
 appear as a replacement for an S-shaped seal. The seal ring is somewhat similar to that of the S-
 shaped seal, but is different in several features. The seal lip is 9.5 mm (V8-inch) high instead of
 the approximately 25 mm (1-inch) lip on the S-shaped seal. Two  bends are required to form
 the new seal rather than the three required for the S-shaped seal. The seal ring may be backed
 up by optional leaf springs as determined by the engineering effort in Task 5.
    The seal ring and flat leaf springs are secured to a mounting angle which is a standard
 structural shape. The  mounting angle, in turn, is mounted on the door frame using studs at the
 same locations used for mounting the conventional S-shaped seal. A spacer (whose function
will be described later) is placed between the mounting angle and the door frame.

-------
                                         11-23
                      Mounting Angle /- Sealing Ring
                FIGURE 11-12. RECOMMENDED RETROFITTABLE SEAL
    Conventional gaskets and sealers are used between metal parts in the same manner as for
present door assemblies.
    No changes are made to the jamb, the door frame, or the lining, or to their  spatial
relationships.
    This same seal with optional leaf springs  may also be used as a replacement for a typical
fixed-edge seal. However, the mounting method will be different and some modification of
liner and liner retainers will be required.
    The space available for retrofitting seals in a typical 4-meter Koppers and Wilputte battery
is shown in Figures 11-13 and 11-14. These dimensions are based on reference drawings con-
sidered by the designers to be typical of each of the conventional systems in the 4-meter size
range.  Existing batteries may have dimensions different from the typical designs, because
operators request changes at the time of construction or make modifications in later years.

-------
                        11-24
                                                Jamb
FIGURE 11-13. TYPICAL KOPPERS SEAL AREA WITH SEAL REMOVED

-------
                                  11-25
Jamb
             FIGURE 11-14. TYPICAL WILPUTTE SEAL AREA WITH SEAL REMOVED

-------
                                        11-26

    Figures 11-15 and 11-16 are sections of the recommended seal as applied to typical 4-meter
Koppers and Wilputte end closures. The seal ring and optional leaf springs are the same in
both cases. The mounting components are different and will be discussed later.
    Material for the seal ring as shown is 2.5 mm (0.10-in.) thick. Another thickness could be
selected, as appropriate. The thickness or width of the optional leaf springs is to be deter-
mined for each application. The seal lip is 9.5 mm (0.375-inch) wide and is much more flexible
along the jamb than the S-shaped seal. This increased flexibility allows the seal ring to conform
more easily  to a bowed jamb. Across the seal  ring, which must be made of  high-strength
material, the seal can be deflected 7.6 mm (0.30 inch) before exceeding the elastic limit of the
suggested materials. Addition of leaf springs increases the transverse stiffness of the seal ring to
absorb latching forces, and has only a minor effect on the deflection capability of the seal ring.
    As shown and described, the  recommended seal  has a greater flexibility  than conven-
tional S-shaped seals.
    Mounting features for the recommended seal are different for the Koppers and Wilputte
applications. Figure II-15 shows the seal mounted on a typical 4-meter Koppers battery. The
seal is secured to a structural angle as shown. A fixed dimension results when the seal ring is
secured to the angle. This  dimension is identified  in the  illustration.  Another dimension
(between the door frame and the jamb surface) varies from top to bottom of the jamb depend-
ing on the warpage and relative bowing of the jamb and door frame. This variable dimension
must be measured at selected points along the jamb by  using a tool similar to one described in
the summary of Task 3.  The difference between  the fixed dimension and the variable dimen-
sion determines the thickness of the spacer at that location.
    Spacers bow the seal ring in such a manner that the seal edge (and the entire seal) assume
the same profile as the jamb surface. More uniform sealing pressure is achieved with more
uniform deflection all around  the seal. If this minimum deflection  is in the range of one-
quarter to one-half the total deflection capability of the seal, the  balance of the deflection
ability is available for jamb and door changes during a coking cycle or over some period of
time.
    Several methods of installing spacers may be used.  One is to use metal blocks of required
thicknesses at suitable spacings. A suitable nonshrinking cement would be used to fill the gaps
between blocks. Another method would use metal bars of appropriate length and thickness
stacked to attain the desired  space between  seal mounting  angle and door frame. This
method,  also, would require a  suitable cement, but the gaps would be smaller than in  the
previous method.

-------
                                   11-27
                                             jamb
            Door Frame
                                       Scale:
                                           I
                                           0
   1
Inches
 I	
                                               Centimeters
FIGURE 11-15. SECTION OF RECOMMENDED SEAL; TYPICAL 4-METER KOPPERS OVEN

-------
                                     1-28
         Liner
                                                              Variable
                       Door Frame
                                     Scale:
                                         I
                                                    I
                                                    1
                                                  Inches
                                         02
                                             Centimeters
FIGURE 11-16. SECTION OF RECOMMENDED SEAL; TYPICAL 4-METER WILPUTTE OVEN

-------
                                         11-29

    The use of spacers requires the following steps:
    • Determine the spacer requirements from measurements taken with the door in
      place on the jamb. These measurements can be obtained rapidly and accurately
      with simple but specialized measuring equipment.
    • Take door to repair facility.
    • Repair door as required.
    • Install recommended seal using spacers and cement as required.
    • Set corner stops to avoid overstressing the seal corners while the door is being
      heated during the first several cycles.*
    • Place door in service.
    Another means for adjusting the seal ring to conform to the jamb uses mechanical screw-
 type adjustments. Figure  11-17 illustrates a possible means of retrofitting a mechanical adjust-
 ment to typical Koppers doors. This concept, which has yet to be engineered, requires no
 modification of the existing door frame. Other means can be visualized, but would require ex-
 pensive modification of existing door frames or design of new frames. In any event, a large
 number of small parts would be required.
    Use of a mechanical adjustment system has features such as these:
    • Mechanical screw adjustments made with the door in place eliminate the need for
      measurements to determine spacer thickness.
    • Subsequent small adjustments are possible.
    • Skill and judgment are required to make seal adjustments by this method.
    • Misadjustments may be made by uninformed and unauthorized personnel.
    • Sealing behind the mounting angle could be a problem.
    Adjustment of the seal ring to conform to the jamb may be accomplished by either of the
 two methods described, i.e., spacers or screws. Considering the increased degree of flexibility
 to be expected from the recommended seal, it is believed that one setting of spacers may give
 an extended period of emission-free service (to be  proven in a demonstration project). Our
 negative attitude toward mechanical adjustment screws is  based on (a) the added cost and
 time required to retrofit adjustment screws,  (b) the concern that screws will be overtightened
 with the result that the seals will be permanently  distorted, and (c) the conclusion of the
 analysis work that tightening the screws with the door in place sets up an increased latching
 force (where screw latches are used). This latch force cannot be duplicated the next time the
 door is mounted on the jamb if the latch force is generated by a screw mechanism.

*U.S. Steel Research data indicate that it takes two or more coking cycles before a door achieves its equilibrium
 thermal bow.

-------
                                        11-30
                 Gasket
                                                                \
                  Continuous Plate
                  Secured to Seal-
                  Assembly
                        Swivel
                       Adjustment Screw
               FIGURE 11-17. CONCEPT OF RETROFITTABLE MECHANICAL
                            ADJUSTMENT FOR SEAL
Comments on Discarded Approaches
    One of the conclusions of Task 4 was that the concept of retrofitting a flexible seal in front
of the diaphragm of the fixed-edge seal should be considered. This concept is illustrated in
Figure 1I-18. As previously noted, there is only a small space for mounting a seal as a replace-
ment for the fixed-edge design. Therefore, the spring element must be smaller than the
replacement spring element for the S-shaped seal cavity. Calculations indicate that it is not

-------
                                 11-31
Seal
Diaphragm
Existing Clip
for Present
Adjusting Screws
                                           New Push-Pull
                                           Adjustment
        FIGURE 11-18. ALTERNATE CONCEPT FOR WILPUTTE DOOR

-------
                                         11-32
possible with the smaller spring element to combine a high degree of flexibility with the ability
to absorb the entire latch loading. Also, to prevent overstressing of the seal element it would
probably be necessary to  mount a stop bolt or stud beside each adjustment point on the
diaphragm. The next factor to be considered is relative costs of retrofit. It is believed that the
retrofitting of the recommended seal will be about equal  or less than the cost of retrofitting
concept shown in Figure 11-8. Our recommendation is to retrofit fixed-edge seals with the ap-
proach illustrated in Figure 11-16.
    One concept that received attention was the thought of forcing the seal edge against the
jamb  by  means of an array of pneumatically inflated, metallic bellows. This concept  is il-
lustrated in Figure 11-19. The expansion of the bellows would increase the latch pressure and
would also rather uniformly distribute the pneumatic force around the periphery of the seal.
At least in theory, the portion of the seal element other than the upstanding  edge could be
thin and highly flexible. A major factor that works against this concept is the probable short life
of the bellows units. Metal bellows are seriously affected  by fatigue and  high temperatures.
Fatigue life is closely related to the ratio of stroke/length. A low ratio of stroke/length ratio is
                     FIGURE 11-19. CONCEPT USING METAL BELLOWS

-------
                                          11-33

desirable to obtain a long life. In this instance, a stroke of 8 mm (0.30 inch) or longer would be
required. With this requirement, a long and, therefore, more expensive bellows would be re-
quired to provide a long service life.
Rationale for the Shape of the
Recommended Seal (The Corner Problem)
    One of the conclusions developed during this project was that special attention should be
given to the stresses that develop when flexing the seals at the four corners of any door. The
level of stresses at the corners relates to the shape of the seal. Figures 11-20 and 11-21 show two
designs that are used as examples to illustrate that the orientation of the seal relative to  the
jamb face  is  an  extremely  important consideration in  the selection  of  the seal  design.
Figure 11-20 shows a sloping seal, and Figure 11-21 shows a seal in which the major portion of
the seal is parallel to the oven face.
    Figure II-22 illustrates the action  of several seals at the  corner as the door is latched in
place. On this figure, View (a) is the upper right hand corner of a typical standard seal in place
on an oven. The deflected edge of the sloping View (b) and parallel seal shapes View (c)  are
also shown to illustrate the dimensional changes the seals must make as they are deflected. In
View (b), a given deflection requires a relatively large change in the radius, in this case an in-
crease. Such a change requires a corresponding change in the length of the seal edge around
the corner. The seal material resists this change, and thereby stiffens the seal at  the corner to
the point of immobility or triggers self-destruction.
    If the seal is parallel to the oven face as in View (c), the same deflection as in View (b)
results in little change in radius, in this case a practically negligible reduction. Appreciable
change in the length of the seal edge around the corner does not occur, and hence seal flex-
ibility does not change at the corner.
    The action of the standard S-shaped seal is illustrated in View (d). Along the sides of  the
seal ring, deflection is distributed over the portions parallel and perpendicular, respectively, to
the jamb face; and also at the two  bends adjacent to the vertical portion. At a corner, the  top
(or  bottom) of the seal ring is joined to a side of the ring. The vertical portions of the seal  are
welded together to form a box-like corner thus stiffening each vertical portion and preventing
deflection.  The horizontal portion of the S-shaped seal is the only portion that  can deflect at
the corner. The result is a stiff seal with less deflection capability than along the  side.
    The geometric considerations discussed above  resulted in the selection of the parallel seal
approach.

-------
                                    11-34
               r
                                          FIGURE 11-20. AN EXAMPLE OF A SLOPING
                                                     SEAL DESIGN
FIGURE 11-21.  AN EXAMPLE OF A PARALLEL
            SEAL DESIGN               h-
                                     i

-------
                                      11-35
       r	 Weil
       \   of SI
Deflected Edge
of Sloping Seal
                                    Section
                                    at Corner
Deflected
Edge of
Parallel Seal
  (a) Corner of Typical Seal (Looking
     Toward Oven)
                                                     Change in
                                                     Radius
                                                                     Deflection
                                        (b) Section of Sloping
                                            Seal at Corner
      Change in
      Radius
      r
                 Deflection
                                                         Deflection at Corner
                                            Deflection Along Side—'
     (c) Section of Parallel
         Seal at Corner
                                   (d) Action of S-Shaped Seal
                                       at Corner and Along Side
    FIGURE 11-22. SKETCHES SHOWING SEAL ACTIONS AT CORNER OF DOORS

-------
                                         11-36

Critique/Limitations of the Recommended Seal
    The purpose of issuing this summary of Task 4 before the completion of the project was to
obtain the critique of our Sponsors (and others) prior to committing a large level of effort into
engineering, design, and laboratory testing.  As  previously described, there are functional
criteria, evaluation criteria, technical criteria,  and finally the recommendations that resulted
from the mathematical analyses completed during this project. This portion of this report deals
with a comparison of our recommended seal design with the criteria listed beginning on page
11-19 of this report. In addition we are including some comments on possible limitations of our
recommended design.
    With regard to functional  criteria,  the  recommended  seal will meet all of the  re-
quirements except that:
    (a) It will not be readily retrofittable to the relatively small number of doors having
        knock-type seals. Retrofit  in these  cases would  probably require a major
        modification of the door frames. The principles that are needed for steel com-
        panies to engineer these modifications are included  in this report and in the
        forthcoming Task 5 report.
    (b) The design may or may not be applicable to ovens taller than 6 meters. One of
        our   reservations deals  with the fact  that  we  do  not know the  latching
        arrangements that will be used for the super-tall batteries of the future.
    It is expected that all of the evaluation criteria (pages 11-19, 11-20) will be satisfied  by the
recommended design. Sealing effectiveness is expected. The life of the recommended seal is
expected to be appreciably extended over that of present standard S-shaped seals. Our defini-
tion of seal "life" is the length of time the seal gives acceptable emission-control performance.
It is expected that the installed cost for the recommended seal will be higher than that of the
existing seals. However, this cost criterion was given a weighting factor of only 6 by the AISI
Task Force in terms of the  relative importance. Estimation of the  fabrication cost of the
recommended seal is presently being done, but we are suggesting that the seal be evaluated
on the basis of overall cost effectiveness. Overall cost effectiveness is defined as:
                      (Original Installed Cost + Maintenance Cost)
                                    Effective Life
The measure of  effective life will,  of course,  depend upon the emission standards that are
used. Field experience is required before any evaluation can be made of maintenance cost.
    Battelle  researchers do have some reservations about  the level of emission control that
will be achieved with the recommended seals at batteries that do  not have mechanical jamb-
cleaning  equipment.  There  are  technical  variables involved  including  temperature,

-------
                                         11-37

temperature variations, type of tar, smoothness of the jamb surface, and other considerations.
These variables have not been studied (particularly "type" of tar) and there are unknowns.
Our concern is in sealing against hard carbon deposits on jambs that have been nicked in
manual cleaning operations. The formation of hard carbon on jambs is far from uniform from
plant to plant. Some plants, for example, have little or no carbon formation on jambs.
    We believe that the technical criteria developed in the Concept Project (page 11-20) will
be fulfilled with the recommended design, inasmuch as it is expected to remain elastic at
temperatures to 700 K (800 F) and does have increased deflection capability. During the Con-
cepts Project, we were thinking in terms of a seal that when mounted on the door would have
a wider range of flexibility in terms of "gap-closure capability", i.e., seal-edge bending. In this
approach, the seal/jamb contact pressure would be low where the seal was just touching but
the seal was not deflected.  On the other hand, the seal/ jamb pressure would be very high
where the seal was deflected a large amount. This approach was abandoned during this pro-
ject with the.objectives of (a) obtaining more uniform seal/jamb contact pressure while main-
taining increased deflection capability, and (b) decreasing the stress levels in the seal.
     It  is judged that all of the criteria/specifications developed during this project will be met.
We do not foresee any new cleaning problems associated with  the recommended design.
However, a demonstration project is required  to prove or disprove these judgments.

                    Material Considerations  for Upgraded Metal Seals

    During the course of this project, it became apparent that with the seal criteria being
developed there would be a need for higher-performance materials. This section of the report
deals with the criteria developed to aid in material selection. Also included are other factors
that led to our recommendation that Inconel X-750 be tested in Task 5. It is expected that the
evaluation of materials  for seals will continue in more detail in Task 5. Our interest in this
regard is related to the possible use of materials that have a  lower price per pound than X-750.
Because our search for materials for recommendation is not complete, this summary is
preliminary in nature.

Comments on Present Seal Materials

     Type 304 stainless steel is used presently for both the fixed-edge and S-shaped seals. This
material has excellent resistance to the corrosive vapors that are emitted by some coals, and it

-------
                                          11-38

has good formability. Where 304 is used to fabricate a complex shape (such as the S-shaped
seal), the starting material is usually annealed strip or sheet. In this annealed condition the
yield strength is only about 210 MPa (30,000 psi), which decreases to about 140 MPa (20,000 psi)
at 700 K (800 F). Because 304 stainless steel cannot be  heat treated to increase hardness and
strength, the hardness of 304 stainless S-shaped seals is only about 160 to 200 Brinell.
    The coefficient of thermal expansion of 304 stainless steel to 700 K (800 F) is a very high 17.8
X 10"" m/m/K (9.9 X 10~" in./in./F). This expansion rate is about 50  percent higher than that of
the gray cast iron door frames to which the material is bolted. This  difference in expansion
coefficients presents a thermal stress situation when the seal and the door frame are heated in
service. The problem becomes amplified when the seal  reaches a higher temperature than the
door frame.
     The reported  maximum allowable temperature for 304 stainless steel when used as a
spring material is in the range of 530 to 560 K (500 to 550 F).* Above this temperature range, un-
der  loaded (stressed) conditions, the creep/relaxation rate  of this stainless is high, and the
original spring shape and/or its ability to carry the load (resist the stress) are lost. They are lost
either slowly or rapidly, depending on conditions.
     At some plants, carbon  steel and ASTM A588 low-alloy steel are used as materials for S-
shaped seals. ASTM A588 is  also known as Ni-Cu-Ti because the steel contains small amounts
of nickel, copper,  and titanium.  Ni-Cu-Ti has a yield strength of 350 MPa (50,000 psi) at room
temperature. It is sold  in the form of lightweight channel which is an excellent starting form
for the fabrication of S-shaped seals. At some plants, this material is  rapidly corroded in ser-
vice. We  have no information on Ni-Cu-Ti's high-temperature 533+ K (over 500 F) strength or
on its creep and relaxation characteristics, but it is expected that this alloy does not have the
deformation-resisting strengths required for the recommended seal design.
Material Selection Criteria
    Over the  time period  for  Tasks  1,  2,  and  3;  a  list  of  criteria  evolved for the
selection/evaluation of materials for upgraded metal seals. The final criteria list is as follows:
    1.  Material should have  a high modulus of resiliency at 700 K (800 F).
           The modulus of resiliency is  the area under the elastic portion  of the stress-
       strain  curve.  This modulus  is a  measure of the ability of a material to absorb
       energy elastically. The equation for this modulus is: (Yield Strength)/(2X Modulus
       of  Elasticity). In  one version of the recommended design, the seal material (with-
       out backup leaves)  could  be called upon to absorb the entire latch force. An im-
       portant consideration, then, is high yield strength at operating temperature, which
*Siegel, Martin }., "High-Temperature Springs", Machine Design, March 30,1967.

-------
                                          11-39

      can also be an important factor in allowing an increase in amount of elastic deflec-
      tion of the seal.*
           Most seals normally operate in the range of 480 to 590 K (400 to 600 F), with the
      highest temperatures at the top of the door. Test  work in Task 2 (the physical
      model) indicated that even a small fire at the seal would heat the seal to near 810 K
      (1000 F), although the temperature rise of the more massive door frame and jamb
      (with a small fire) was small. USS Research has reported on one example of seal
      temperatures climbing to 700 K (800 F) during the  steam-aspiration  period just
      before charging of coal to an oven. We selected 700 K (800 F) as the temperature at
      which the mechanical and physical properties of materials should be compared.
    2. Materials should have a low relaxation and creep rate at 700 K (800 F).
           Relaxation (loss of load carrying ability) is an important consideration in selec-
      ting materials for  springs, but  the amount of data on relaxation rate is rather
      meager except for some alloys.  A literature search was completed looking for the
      recommended maximum continuous operating temperatures for various materials
      when used as springs. It was assumed that the technical personnel setting these
      upper-limit temperatures had experienced creep (dimensional change) or relaxa-
      tion (loss  of  load) under  heavy spring loads at temperatures higher than those
      specified. Some of the data taken from the literature are as follows: •
                   Material             Maximum Continuous Service Temperature
     Carbon Steel                               394 K               250 F(a)
     High-Carbon Steel                          422 K               300 F(h)
     SAE 6150 Steel (Low Cr + Mo)                450 K               350 Flh)
     304 Stainless                              533-560 K          500-550(c>
     17-7 PH, Age Hardened                    588-616 K         600-650 F(a'h)
     Inconel600                                 588 K               600 F(hl
                                                                   «-,— — fM
     A28, Age Hardened
     Inconel X-750, Age Hardened              755-866 K          900-1100*
     Inconel 718
                          h)
977 K              1300 F(d)
     (a)  Carson, Robert W., "Flat Spring Materials", Product Engineering, March 1,1965.
     (b)  Crooks, R. D., and Johnson, W. R., "Flat Springs Above 500 F", Product Engineering, February 18,
        1Qfi^
     (c)  Siegel, Martin ]., "High-Temperature Springs", Machine Design, Marcr< 20,1967.
     (d)  Moeller, R. H., "Coil Springs Above 400 F", Product Engineering, October 25,1965.

bother possible versions of the recommended design backup ^™™*™
 material, and/or some of the latch force could be shifted to the stop bolts. See page

-------
                                     11-40
3.  Material should be metallographically stable at maximum temperature.
       With sufficient time and temperature, some change in metallurgical structure
   can be expected for almost any alloy. Some alloys change structure (and proper-
   ties) rather rapidly above some temperature, whereas other alloys react  more
   slowly and can tolerate occasional excursions through "maximum" temperatures.
   For demonstration work, it is considered best to start with stable materials. Other
   less-stable materials can be evaluated once  a  performance standard has  been
   established.
4.  Material should have a thermal expansion coefficient about that of gray cast iron
   and steel.
       As previously indicated, we have reservations relative to the high coefficient
   of expansion of 304 stainless steel (when this material is bolted down on a gray cast
   iron frame). Battelle researchers believe that they have seen indications of thermal
   buckling of S-shaped seals in operation. If the door frame and the attached seal
   expand about the same distance in coming to operating temperature, the level of
   thermal stresses will be favorably low.
5.  Material should have a high thermal conductivity.
       The thermal stress level to be developed in  a seal is a function of several fac-
   tors, including the thermal conductivity of  the  material. The higher the thermal
   conductivity, the lower the stress level, all other factors being  equal. Seals will con-
   tinue to be cooled rapidly by  rain, and seals will occasionally be heated on one
   side by fires. Under these conditions, higher conductivity material is favored.
6.  Material should be available in strip form now.
       The availability of high-performance material in strip  form is in many in-
   stances a problem. For some materials, it  may be necessary to place a special mill
   order.  Battelle researchers need  readily  available  material  for  small-order
   purchases for test work in Task 5 and for possible demonstration work.
7.  Material should be easily formed, weldable  with no problems, and heat treatable
   with no problems.
       Most of the high-performance materials can be formed (but not as easily as
   304 stainless) by working with  annealed material which is then aged to develop
   high strength and high hardness. We have been told that welding and heat treat-
   ment with some materials represent "no problems", but we are concerned about
   (a) the complicated aging step required for  some alloys, and (b) the possibility of

-------
                                         11-41

      warpage during structural transformations and upon cooling a rangy, elongated
      shape such as coke-oven seals. Evaluation efforts in this direction are in progress.
    8. Material should be adequately corrosion resistant in coke-plant service.
           A summary of data on various alloys  as they relate to the  above criteria is
      shown in Table 11-1.

Material Recommendation

    At this time, our recommendation for a first choice of material for Task 5 testing is Inconel
X-750. This is the material normally used for long-time service over 590 K (600 F).  It is available
in various thicknesses from two suppliers. Information is being obtained on availability of long
sheets or strips.
    The testing of Inconel X-750 will be followed  by testing of various types of precipitation-
hardening stainless steels. It is possible that  several alloys will be  selected for  comparative
testing in  any demonstration project.  Discussions  are in progress with technical  personnel of
alloy suppliers.

                              Research Outline for Task 5
    The primary objective of Task 5 will be the design of a full-scale sealing system for typical
4-meter  Koppers  and Wilputte  end closures. The  drawings and supporting documents
resulting from the design effort will be prepared to Level 2, Production Prototype and Limited
Production, defined in MIL-D-1000A, 15  October 1975.  End-closure components made to
these drawings will fit existing end closures matching the drawings furnished by  Koppers and
Wilputte Companies as typical of 4-meter ovens. End closures which  do not  match these
typical  drawings will require that the battery operator and the vendor be  responsible for
modifying the drawings accordingly. Variations should not adversely affect the  intent of the
design in any way.
    Various tests will be conducted of  portions of full-scale seals made of recommended
materials. The following are typical of the expected tests:
    • Relaxation Test. This test will be directed at verification of the high resistance to
     relaxation claimed by suppliers of some candidate materials. A short length of full-
     size seal will be clamped in a fixture and deflected  a predetermined amount. Fix-
     ture and specimen will be placed in an oven at operating temperature for 16 hours
     (one coking cycle). The  specimen  will then be removed from  the fixture and
     checked for relaxation and creep. This test may be repeated several times (some at

-------
            TABLE 11-1. COMPARATIVE DATA ON CANDIDATE ALLOYS FOR MANUFACTURE OF COKE-OVEN SEALS
Maximum Continuous
Service Temperature 0.2% Offset Coefficient of Thermal
With Low Creep or Index of Relative Yield Strength Expansion(b) Thermal Conductivity
Relaxation Under Load Modulus of Resilience at 700 K (866 F) 293-700 K (68-800 F) at 700 F (800 F)
Material
304 Stainless
(present material)
Ni-Cu-Ti
(present material)
1 7-7 PH, Age Hardened
A286, Age Hardened
Inconel X-750,
Age Hardened
Material
304 Stainless
Ni-Cu-Ti
1 7-7 PH, Hardened
A286, Hardened
Inconel X-750
Hardened
K F at 700 K (800 ?)(*)
533-561 500-550 1
(reference level)
350 450 4 est
588-61 6 600-650 23
700 800 1 2
755 900 1 6
Basic Price
Material per Pound of
Availability Strip Material, $/lb
Available now $1.00/lb
Yes, as channel 0.40
Yes — various sizes 1 .25
Some sizes 3.85
Yes, various sizes 5.00
but length may be
a problem
MPa
124
172
896
620
965
Corrosion
Resistance
High
Low
High
High
High
Ksi 10-6m/m/K 10-6 in./in./F w/m/K
18 17.8 9.9 20
25 13.8 7.7 est 32
130 12 6.6 21
90 17 9.6 20
140 14 7.8 17
Hardness Comments on Forming
(Brinell) and Treating
160-200 Easily formed. Easily welded.
No heat treatment possible.
-
400 Formable. Weldable.
Complex heat treatment.
Ditto
290 Formable. Weldable.
Complex heat treatment.
Long-time heat treatment
required.
Btu/hr/ft2/in./F
144
220 est
150
140
120
Stability
Comments
-
-
Not stable
over 600 F
Stable to 81 OK
(1000F)
Stable to 81 OK
(1 000 F) and
higher
(a) 0.2% offset tensile "strength (-it 800 F) was used in calculation (instead of yield strength) because ot" the availability of the offset data.
(b) dr.iv «..isi iron expansion ,»t 700 K (800 I } is about I 2.2 lo I 2.6 \ 10'6 m/m/K (6.8 to 7.0 \ 10*6 in./in./F).

-------
                                         11-43

     elevated temperature) to span the initial high relaxation rate expected before the
     material settles into a long-term low relaxation rate.
   • Spacing Test. This test will determine the effect of spacing or shimming (bowing)
     upon the deflection capability of the seal and will provide a measure of the max-
     imum jamb warpage or bowing  which can be handled by the seal. A length of
     about 1.2 m (4 feet) of full-size seal will be mounted ina fixture in a shimmed condi-
     tion. Shimming will induce a bow of 1 mm (0.04 inch) in 1.2  m (4 feet) which is
     equivalent in terms of seal styles to a bow of 13 mm (0.5 inch) in 4.3 m (14 feet). The
     shimmed specimen  will  be forced  against a simulated  jamb seal  surface of
     matching bow, deflected the design  maximum, released, and checked to deter-
     mine if the elastic limit has been exceeded.
          If, in the test described above, the elastic limit has not been exceeded, the seal
     will be reshimmed  to increase the bow some  discrete amount. This step can be
     repeated increasing the bow each time until the elastic limit of the seal is exceeded
     and permanent deformation of the seal can be detected.
   • Corner Test.  Battelle does not  expect  any severe  problems with the corner.
     However, the length of seal used in the spacing test can have a fabricated corner
     attached to one end so that it deflects each time the spacing test is conducted. This
     arrangement will provide  visual  evidence that the corner  does perform as ex-
     pected. This arrangement can also be used to determine if the optional  back-up
     leaf springs must be carried around the corner, as must the seal ring, or if they can
     be terminated  at the corner at the point where the seal edge begins its curve.

   The output of Task 5 is to include recommendations dealing with the following questions:
   (a)  Are backup spring leaves required?
   (b) Will stop bolts be required to limit the loading (deflection)  on the spring seal?
   (c)  Is it possible  to design a  single, leaf-spring seal that can absorb the entire latch
       loading with  stop bolts only being used to protect the seal during the mounting
       of the door?

   It is  probable that the final answer for some of these  questions will depend upon the
results of any comparative testing tasks recommended for the demonstration project.

-------
                                        111-1

                                   CHAPTER III

             JAMB ANALYSES: RECOMMENDATION OF A RETROFITTABLE
                         JAMB DESIGN AND JAMB MATERIAL

                  Introduction and General Statement of the Problem
    In the context of this project, production of coke  is a high-temperature, gas-producing
process that is conducted in brick structures that are difficult to keep gas and vapor tight. The
particular emphasis of this project is to develop upgraded and retrofittable end-closure com-
ponents that will give long life and significantly improved emission-control performance.
    The emphasis in this chapter of the report is on  jambs. This  major component of all
modern end-closure systems  exists in the field in ten or more cross-sectional "shapes". Six of
these shapes are illustrated in Figure 111-1. Although jambs are massive, they appear to be
vulnerable to problems. Our interest in this project was to find out: Why do  they warp or dis-
tort? How do jambs interact with buckstays and doors? What can be done to  design and make
jambs that are both retrofittable and more stable?
    In the preceding Concepts Study, evidence of the jamb warpage was reported. It was one
of the major conclusions of that study that "the door-mounted seals often receive the blame
for coke-door  emissions, but the fundamental cause of the emission-releasing gaps is the
pronounced degree of warpage that has occurred in differing degrees on most (if not all) of
the 25,000 or more cast-iron  jambs in operation". Another related conclusion was that "ex-
isting seals were not designed to conform to more than a minor amount of jamb bowing and
warpage". These represented generalized  conclusions supported by observations and some
measurements, but unsupported by any analytical effort or information useful for the design
of upgraded jambs.
    In the present project, our "jamb problem" was to:
    (a)  Develop an  understanding (to the degree possible) about the fundamental
        reasons why and how designs and materials of existing jambs present problems
        in terms of their contribution to end-closure emission control. The recommend-
        ed  approach was an analytical effort supported by field measurements and by
        operation of a quarter-scale, heated physical model.
    (b) Develop recommendations for a design of retrofittable jamb that would  be more
       stable, both in the short and long term, and would perform in an improved way
       to give better sealing  at both the back and the front sides of the jamb.

-------
                              1-2
                    A
                   0
                    I
3
I	I
Inches
  6
i  I  i
                                           12
                         76
     152
     mm
                                     228
                                           304
FIGURE 111-1. SIX EXAMPLES OF THE HORIZONTAL CROSS SECTION OF
            EXISTING JAMBS

   The oven bricks are toward the top of the figure. The longer portion
   of the cross-section (shown vertically in five of the examples) is
   the jamb web or flange.

-------
                                         111-3

               Data and Insights Obtained in Task 3 (Field Data Collection)
    A detailed presentation of the data and information gathered in Task 3 is available to the
Sponsors in a working paper summary that has been sent to them. This portion of the report
discusses the highlights of Task 3 and serves as background material for the following sections
dealing with analytical results and conclusions.

Long-Term Temperature Data on Jambs

    At the start of this project it was reasoned that the warpage and distortions of existing
coke-oven jambs may  have been the result of past overstressing (thermal stressing)  of this
equipment under conditions that may well have only existed in the past. No coke plant keeps
records of fires or records of any temperature data that can be related to dimensional changes
in the end-closure components. To develop data for analysis, the field-team personnel wanted
to start with new jambs and to obtain temperature and physical measurement data over a long
period of time. This had the strong endorsement of coke-plant superintendents who believe
that some significant portion of the damage to jambs occurs during the initial battery-heating
operation. It was, therefore, decided to obtain temperature and strain-gauge data for new
jambs  on a  new coke battery  about to  be heated, and  on new  jambs  being installed as
replacements on operating batteries.
    At a new 6-meter battery, 39 thermocouples were installed in, on, and behind an un-
heated coke-side jamb prior to the installation  of the jamb.  In addition, calibrated strain
gauges were installed on the latch hooks, jamb clips, and tie rods. Special data-acquisition
equipment was also installed.
    Unfortunately this data-collection installation was inadvertently destroyed  by  malfunc-
tioning equipment during a pushing operation after only about 50 low-speed coking cycles. It
was not possible to replace this sensing equipment, but it is hoped that some coke plant (with
the possible support of the AIS1) will attempt to  obtain long-term data of this type. Battelle
researchers consider this suggestion to be important, inasmuch as taller and taller coke ovens
are being considered by steel companies and coke-plant builders.
    During the short life of our installations, it was learned that:
    (a)  The temperature differential across "square" jambs (during coking cycles) was
        low in comparison to jambs having an exposed web or flange. The square jambs
        were partly embedded in the end bricks of ovens, so that only part of one face is
        exposed to convection and radiation cooling. Temperature gradients of only 1.3
        K/cm (6 degrees F/inch) were measured on the square jambs, and gradients of
        3.3 to 4.4 K/cm (15 to 20 degrees F/inch) and higher were measured on flanged
        jambs.

-------
                                        111-4

    (b)  During the early stages of preheating a battery, the top of jambs can be heated to
        as high as 780 K (950 F) if hot gases can leak out past the back of the jamb. This
        high temperature was measured during the  period when the oven was being
        heated by firing through the door.
    (c)  Directing a stream of water at an end closure (simulating a heavy, slanting rain
        storm) can increase the temperature differential across a square jamb to 110 K
        (200 F) in less than 3 minutes. It took 3 minutes before the backside temperature
        began to drop. The longest time that water was used to extinguish any door fire
        was measured as less than 5 seconds.

Jamb Contour Measurements
    At the beginning of this project, Battelle researchers were especially interested in measur-
ing the vertical profile or contour of jambs. Our experience coupled with  the evaluation of
plant data indicated that many errors are possible in using measurement methods that depend
on personnel taking  readings while working  on scaffolds or hoisting  equipment. Our
preferred measurement method for  jam  profiles was developed  for shorter ovens using a
relatively  simple  wire-referenced  straight edge. We prefer this over  taut-wire methods
because the  tool users only have to move the horizontal scales into contact with the jamb and
then clamp the scales. The actual readings are taken at the bench level or off the bench after
the tool has been lowered. The repeatability of this method has been established.
    As noted  in Chapter II,  late in  this  project our  interest shifted from measuring jamb
profiles to measuring the horizontal spacing between the jamb seal-mating surface and the in-
board side of the door frame (at attachment point for S-shaped seals). However, some jamb-
contour measurements  were  taken at five batteries during this project. While this is only a
small fraction of the batteries  in operation, the results of our measurements (after the learning
period) indicated that the degree of warpage  to  be expected  in the field will not be  so
pronounced as we had expected from the plant data given to us during the Concepts Project.
To elaborate further, most existing jambs were installed to remain straight. However, the ther-
mal gradients across the jambs represent a force that is driving the jambs to take an inward-
bowed shape (top and bottom of the jamb are further out from the battery than the center).
Bowing was  once considered  warpage by  the field team, but inward bowing later came to  be
considered to be natural. It was judged that bowing  is  only a problem if the seal cannot effec-
tively contact the jamb. With this judgment, our  research emphasis concentrated on ways and
means for getting seals to contact bowed jambs. Fortunately, all but a few door frames take the
approximate congruent  bow  of the  inward  bowed  jambs.  In a  manner of speaking,  we

-------
                                         111-5

now consider jamb warpage as jamb profiles that an upgraded and more flexible retrofittable
seal cannot contact and continue to hold in contact. We have been told that (a) there are jamb
sides that have both an inward and outward bow on the same side (an S-shaped profile) but (b)
these profiles are rare.  In Task 5 the limits of adaptability of the recommended upgraded seals
will  be quantified. It remains to be seen whether shapes as unusual or as rare as S-shaped
jambs can or cannot be accommodated by upgraded seals. It is expected that the percentage
of jambs that will have to be replaced to achieve improved emission control on end closures
may be small.
     Movement or flexing of jambs  during changes in thermal gradients (during coking cycles,
during rainstorms, and during temperature excursions) remains an important consideration.
The field team was not able to detect any amount of flexure nor were we able to confirm one
steel company's data that  there is jamb flexure during the period  from "just before the
charge" to "just after the charge". The fact that we could not detect flexure probably means
only that we were not measuring  before and after a significant gradient change or that we
were taking measurements where the jamb was tightly restrained.* The force for jamb flexure
is always present during a gradient  change. If the jamb is not restrained, it must flex or change
profile during a gradient change. Our approach to this possible problem was to place jamb-
fastening methods and increased seal flexibility high on the research priority list.

Internal Damage to Gray Iron Jambs
     Changes in microstructure of gray cast iron can be generally interpreted to indicate the
range of temperatures to which the cast iron had been heated. However, the time and
temperature relationships cannot be determined by this technique, because high temperature
for  a short time results in structural changes about the same  as longer times at lower
temperatures.
     Examination of the microstructure changes in discarded jambs should have been  a con-
tinuing task throughout this project.  It was handicapped, however, by the low rate at which
jambs are being replaced and the difficulty in cutting out cross-section samples.
    From the examinations that were completed, the indications are that:
    (a) The top section of the jamb (from the coal line up) continues to be the location
        showing  signs of  past  overheating.  Overheating  is defined as  heating  to
        temperatures  well  over  700 to  760 K  (800 to  900 F). These  and higher
        temperatures result in weakening of gray iron and expansion of the gray iron
        due to graphitization and internal oxidation.
'Jamb restraint is a variable at every end closure. No method was developed to measure this variable.

-------
                                         111-6

    (b)  signs of overheating of jambs show up first on the backside of the jambs, i.e., on
        the side of jambs facing the oven bricks. It was concluded that the "inside" fires
        referred to by coke-plant superintendents did exist, at least in the past. Whether
        "inside" fires have been eliminated is questionable because in some field trips
        batteries were observed to  be "breathing". In these instances, fires at  chuck
        doors would go  out,  restart,  go out, and restart at regular  intervals. This,
        however, may have been due to  a  repairable fault in the governor-house
        controls.
    In the last part of this particular investigation, we attempted to answer the questions:
    (1) Are there long-service jambs in  existence that show no evidence of microstruc-
       tural changes?
    (2) If the answer to Question (1) is affirmative, is there less warpage of these jambs?
    (3) If the answer to Question (2) is affirmative, is the door-sealing performance better
       than average?
    Our search for  such a battery took into consideration (a) battery age, (b) our  general
rating of the sealing performance, (c) our judgment as to whether the battery had been well
maintained, and (d)  whether or  not it was a single battery or one of a string of batteries. This
last consideration centers on the judgment  often expressed by coke-plant superintendents
that "it is easier to take care of  and operate a single battery than a string of batteries". Our
search resulted in the selection  of a single 24-year-old battery having fixed-edge seals. This
battery  was just starting to replace coke-side jambs because  of  an occasional cracking
problem.
    Metallographic examination of samples taken from the next jamb that was replaced at the
above battery indicated (a) no change in the microstructure of the of the gray cast iron after 24
years, and (b) a trace of a spheroidized  (overheaded) zone on the backside of the jamb (top
cross piece). It was inferred that coke batteries can be operated in a manner that does not
result in heating of jambs to high  temperatures—at least  not for  any long-time  heating
period(s). At this plant, the coke-side jambs  have a slight  outward bow [maximum 6 mm (1/4
inch) on those measured] and the steel door frames are unusual inasmuch as they also have an
outward bow; i.e., the jambs and  doors bow in the same direction. The end closures at this
battery all release emissions when coal is being charged to the ovens, but each end closure is
emission-free in less than 20 minutes. The door-sealing performance at this battery was, in our
opinion, well above average.
    In general, the work practices were not much different at this plant than at other plants,
with the possible exception that  the workers stated that they concentrated on "patching

-------
                                            111-7

behind the jamb". In addition, they felt that hand troweling of refractory paste into the crack
between the jamb and the bricks (and building up spaded areas on bricks) was much preferred
to spray patching.  In  their operation they placed a heat shield in the open empty oven,
climbed to the work site, and  hand-pressed refractory paste into cracks using trowels.
    Later in  this project, Battelle researchers came to appreciate the difficulties involved in
patching behind the jamb, particularly on 6-meter ovens. Another lesson learned was the large
amount of patching required-at some plants it is a continuing program. Battelle researchers
gave some consideration to improved patching methods and/or improved placement tools.
These considerations are presented in a later portion of this chapter.

The Hourglassing Problem*
    For Task 3, a special measuring tool was developed to follow the hourglassing rate of new
jambs.  The  data obtained using this tool  were  as shown  in  Table 111-1.  Our periodic
 TABLE 111-1. COMPARISON OF THE AMOUNT OF HOURGLASSING OF TWO NEW 6-METER
             JAMBS OPERATED ABOUT THE SAME LENGTH OF TIME(a)

                                              Total Dimensional Changes from Original
                                             Measurements (Ambient Temperature) to Last
                                             Measurement at Operating Temperatures(°)
         Jamb Cross-Section:
         Jamb Material:
         Jamb History:
         Total Service Time at Last Measurement:
          Approximate Reference Location
              Above Bottom of Jamb
"Square"          "Square"
Ductile Iron        Unalloyed Cray Iron
Replacement Jamb  New Jamb on a New Battery
About 8 months     6 months
Meters
6.25
56
4.9
44
40
34
27
21
1 4
0.2
Feet
20.5
18.5
16.0
14.5
13.0
11.0
9.0
7.0
45
0.6
Millimeters
+3.0
+1.5
-0.25
-0.75
-2.3
-3.3
-3.3
-2.5
-1.0
+3.0
Inches
+0.12
+0.06
-0.01
-0.03
-0.09
-0.13
-0.13
-0.10
-0.04
+0.11
Millimeters
+1.8
+1.3
-1.0
-3.0
-2.0
-3.0
-4.0
-3.5
-3.8
+1.3
Inches
+0.07
+0.05
-0.04
-0.12
-0.08
-0.12
-0.16
-0.14
-0.15
+0.05
        (a) In the design of both of these end closures, there was no provision for locking the jambs to attempt to
           prevent hourglassing; i.e., the sides of the jambs were free to move toward each other.
        (b) Measurements were between trammel marks indented in the Jamb pnor to ,amb ,n.»llat onjhe mdicated
           distance changes are between the initial ambient temperature read.ngs and measurements made on operat-
           ing jambs after 6 and 8 months of operation.
'Hourglassing has occurred when the midpoints of the two long vertical sides a.r,Jamb> have distorted towards each
 other so that these sides are closer to each other at their m.dpomts than at the.r ends.

-------
                                         111-8

measurements indicated that almost all of the hourglassing dimensional changes took place
very early in the operational period of these jambs, i.e., there are indications that these jambs
are now stabilizing. If this is happening, these particular jambs will not have an hourglassing
problem.
    In any demonstration project in which retrofittable jambs are tested, the changes in jamb
dimensions will be compared with the data shown in Table 111-1.

                     Conclusions (and Discussion) Derived from the
                         Analytical Effort Dealing With Jambs
    The detailed mathematical analyses and data summaries developed in Tasks 1 and 2 are
available to the Sponsors in working paper summaries of these tasks. These working  papers
were presented  without conclusions.  This portion  of  this report deals with  a descriptive
presentation of the conclusions of the analytical work.

Background Comments
    There are certain obvious reasons for the existence of jambs on coke ovens. The massive
metal jambs in use on most batteries serve at least four functions as follows:
    (1) Support for doors
    (2) Provision of frameworks against which the latching forces can react
    (3) Protection of the refractory-oven ends from the  forces and loads involved in the
       pushing operations
    (4) Provision of surfaces for the door seals to contact.
    The Battelle team recognizes that jambs should serve these functions, but has reservations
about the massiveness of the jambs in use on some ovens. There are also reservations about
the way that some jambs are attached to ovens so as to interact with other components in un-
necessary and possibly harmful ways.
    If it is assumed that the above functions are the only functions a jamb needs to perform,
then it is clear that the cross section of the jamb needs to be only strong enough to support the
weight of the door and to react in an appropriate way with the seal to give a relatively uniform
pressure around the seal. Because operators noted jamb warping and bending in the past, it is

-------
                                         111-9

understandable that  intuition would lead to the development of more massive and stiffer
jambs. Intuition, however, can lead to erroneous conclusions. This is particularly true when it
comes to thermal action.
    Coke-oven jambs have a  higher temperature on the back side against the bricks than on
the exposed side. There is a heat flow through the jambs with an accompanying temperature
gradient across the jamb. Because the hotter surface expands more than the cooler surface,
there is an inclination for the jamb to change its shape. With the back side being hotter, the
top and bottom of the jamb attempt to move outward away from the oven (provided that the
center portion of the jamb cannot move inward). The above statements are true regardless of
shape and dimensions of the cross section. The pertinence of the above discussion lies in the
fact that the force  needed to  hold a jamb in place (i.e., to restrain flexing) depends upon the
stiffness of the jamb.  Stiffness  refers to the resistance of the jamb to bending. As a generaliza-
tion, the thicker a  section in the direction of bending, the more difficult it is to bend and the
more difficult it is to  restrain it from thermal bending.
    From  this conclusion came the recommendation that the jamb cross section be  no more
massive than is necessary to perform its functions. Another favorable factor for shallower jamb
sections (in  the direction of bending) is that the overall temperature differential that results
from  sudden heating and  cooling of jambs will be smaller, and therefore, the stress that
develops during transient heat flows will be lower. No single numerical value can be placed on
these design parameters, because they depend on the oven height, door weight, seal type, etc.
    Of the four identifiable functions of a jamb, the most demanding is that of providing a sur-
face for the door seal to contact. While this sounds like a simple passive task, it carries with it
the implication of stability. That is, the sealing  surface of the jamb should not move around
significantly during a coking cycle or during any other operational sequence which it might
experience

Analyses of Jamb Distortions
    One of the approaches of the analytical team was to develop a numerical analysis of the
factors controlling two types of permanent warpage of jambs. These distortions are  (1)
hourglassing and (2) the in-or-out displacements of the jamb sides relative to a natural or
desired shape. Because almost all end-closure designs call for holding the jamb straight on the
battery, in most instances jamb warpage was considered to be the bulges that deviate from the
desired straightness.
    The analyses performed on the jamb indicate that the mechanism for hourglassing is most
probably creep accompanied with  relaxation.  In most jamb-attachment schemes,  the jamb

-------
                                         lil-10

sides are free to move in the hourglassing direction. If the brickwork is rigid enough to restrain
an outward bending of the jamb (in the anti-hourglassing direction) then the analysis showed
that should creep or relaxation or both occur on either the inner or outer surface of the jamb,
the result would be the  classic hourglassing pattern. Basically, this occurs because the jamb
must move, and it is free to move in only the one direction. A somewhat detailed analysis of
this phenomenon led to the conclusion that the force required to restrain the jamb from
hourglassing  would  be a relatively modest several  hundred  pounds.  Therefore,  it  is
recommended that the jamb sides be mechanically restrained (say, to the buckstay) to prevent
hourglassing.
    With regard to out-of-plane jamb warpage, no one mechanism has surfaced as the cause
of this distortion. When this project started, several possible causes of jamb bowing were
suggested. These included  (1) brickwork expansion,  (2) carbon buildup behind the jambs, (3)
fires causing yielding of the material, and (4) structural changes within the metal of the jamb.
To this list could also be added failure to adjust the jamb clips, bowing of the buckstays, and
improperly adjusted latching forces. A few words  can be said about each of these possibilities
as they relate to the work on this project.

    Brickwork Expansion.  It is well known that brickwork  in coke batteries  expands with
time. This growth can cause a problem relative to the stability of the end closure. Our analysis
indicates that this expansion cannot be stopped by increasing the level of constraint. Any
attempt to stop such movement would either be futile or would result in the permanent dis-
tortion of the jamb, the  buckstays, the tie rods, or some combination of these components.
The safest course of action is to make adjustments to the end closure to allow it to "ride with
the expansion".
    Coke-plant superintendents indicate that they attempt to operate batteries as steadily as
possible; i.e., they attempt to minimize adjustment  to the heating rate. Adjustments are re-
quired, however, and those that we know about concern changes in production rate, produc-
tion interruptions, and slow downs during air pollution alerts. It is possible that unavoidable
adjustments in the heating rate may have something  in common with a problem that occurs at
metal grain silos. Cyclic  thermal expansion  of the silo causes the grain to settle a little more
each cycle, and then when the silo cools it  becomes more and more highly stressed. It is not
unreasonable to imagine a battery of coke ovens "breathing" in this fashion so that coal, coke
dust, and tar can gradually work their way into small cracks during a "low-heat phase", and not
allow them to close up on the next "high-heat phase".

-------
                                         111-11

    Carbon Buildup Behind Jambs.  The field team working on this project has reported that
in some instances when jambs are replaced, the old jamb has a rather thick [up to 13 mm (1/2
inch)] deposit of hard carbonaceous material on the side of the jamb closest to the bricks.
Some coke-plant personnel believe that the carbon "grows"  and exerts pressure against the
jamb. We have no  explanation for a mechanism that causes swelling or expansion of existing
carbon. However,  the fact that the  carbon layer gets thicker with time cannot  be  ignored,
because operators  have reported displacement of massive equipment (such as gas-main sup-
ports) which was alleviated by removing carbon buildup in connecting joints.
    Our speculation is that jambs that have a rather thick back-side layer of carbon have been
"breathing" or going through  a ratcheting operation. Under  conditions where the  jamb-
restraining clips are not tight, the jamb can flex during a change in the  horizontal thermal
gradient. This would open a gap behind the jamb,  with subsequent  clogging of this gap with
tar. This tar converts to carbon by the loss of the lower-boiling-point constituents in the tar.
This residual material can prevent the jamb from returning to its original  position or profile.
The jamb is now under additional stress which can exert more force at other clips. If this force
results in jamb flexing, another gap is opened at another location, with the result that carbon is
deposited behind the jamb at the other location.  If the  stress in the jamb is relaxed over a
period of time, the jamb has a new permanent  profile (i.e., warped), so that carbon buildup
can continue provided the clips allow additional flexing of the jamb.
    Our field team has indicated that  (a) most superintendents are not happy  with the
strength of the jamb-restraining clips on the older ovens, and (b) some plants have "given up"
on the plan of  keeping jamb clips tight. It was concluded that weak or loose jamb fasteners
permit jamb flexing and results in carbon buildup  behind the jambs. This can contribute to
jamb warpage.  It is recommended that jamb-restraining connections should be upgraded on
retrofittable jambs to minimize movement of the jambs both inward  and outward.

    Intense Fires.  Fires can cause numerous problems on end closures, at least in theory. In
Task 3, metallographic  examinations of cross sections  of  discarded jambs showed that
sometime in the past the upper portion of some jambs may have reached  920 K (1200 F). Dis-
cussions with operators  revealed stories of jambs  in a "cherry red" condition  due to fires
"behind the jamb". Other  discussions suggested a general  feeling that jamb warpage and seal-
ing failure were, to a large  extent, the result of fires along the seal/jamb and jamb/brick inter-
faces where visible leakage occurs.
    It is reasonable to predict that should the jamb be subjected to thermal mistreatment, it
will deform permanently. Moreover, if deformation should  occur,  it  will  aggravate the
problem through more and larger fires due to increased volumes of combustible gases and

-------
                                        111-12

vapors leaking past the seal. The ability of most common structural steels and irons to recover
elastically from some load becomes less as the temperature of the material increases. Because
of observed fires and the generally elevated thermal environment of the jamb in a coke oven,
it is natural to blame this degradation in strength for much of the problem of sealing coke-
oven end closures. Several results obtained on the physical model and  in the analyses make
this ready explanation much less "ready".
    On the model, fires were considered as a possible cause of jamb warpage. However, the
jamb is a rather large thermal sink.  Fires occurring at the door  seal tend to (a) lower the
gradient through the jamb, i.e., lower the stress and (b) rapidly overheat the thin door-seal
material. To obtain the large thermal gradient required to build up a damaging stress in the
jamb (under restrained conditions), it is  necessary to have an intense fire behind the jamb. A
fire in this location requires (a) air passageways around the back of the jamb, and (b) a pressure
inside the oven  lower than the ambient pressure (including wind pressure effects).
    It is expected that an improved seal on jambs (backside and frontside) will minimize both
emissions and overheating problems.  Fires at the seal should be put out to prevent overstress-
ing of the door seal. Changes should be made  in the steam-aspiration controls that would
make it necessary to use extra human effort to keep steam aspiration in operation longer than
a required length of  time (e.g., steam  valves equipped with timers).
    In some  pushes, it is difficult to shove the hot coke out of the oven in one  motion.
"Stickers"  can "park" the hot coke (for a period)  directly alongside  the coke-side jamb.
Measurements made in the field show that the rate of heat rise on the heated jamb surface is
about 33 K per minute (60 F per minute). However, when these data were used in calculation,
the temperature difference that developed even  in 20 minutes was not large enough  to cause
stresses close to the yield strength. The simple fact is that coke cannot deliver heat to the jamb
fast enough to develop large thermal  differences across the jamb cross section.

    Sudden Cooling of Jamb.  To develop stresses  high enough to cause yielding  in a gray
cast iron jamb,  a temperature differential in excess of 110 K (200 F) must be realized through
the thickness of the jamb. As noted  above, back-side fires must  be very intense to develop
high temperature differentials. It seemed to the project team that if heat could not be supplied
to the jamb fast enough to warp it, then  perhaps heat could be removed at a rate fast enough
to do damage. Although some coke batteries have sheds over the coke side, most batteries are
completely exposed to rain. Tests were performed both on the model and in the field to
measure the magnitude of temperature differentials which might result from a hard, inward-
slanting, driving rain.  In  both test situations, the temperature differentials generated were

-------
                                         111-13

higher than 110 K (200 F), and if the jambs were totally restrained from thermal bowing they
would have yielded to some degree. The large differential is the result of lag in temperature
change across the width of the cross section of the jamb. The maximum differential occurs at
the time when the exposed surface has cooled and the back-side temperature just starts to fall
as a result of the increased heat extraction. This differential can be minimized by reducing the
cross  section  of the jambs. This recommendation is consistent with a previous recommenda-
tion dealing with the level of forces required to restrain thermally induced flexure of jambs.

    Jamb Restraint Versus Jamb Stresses.  For most end-closure designs, jamb restraint goes
hand  in hand with the development of thermal stresses in the jamb. What is wanted is firm
restraint  and  low  thermal  stress. Low  thermal stress can  be  achieved  in two ways. The
temperature gradient across the jamb can be reduced (as in the square cross-section jambs in-
stalled with only part of one face exposed to heat extraction), or the jamb can be installed so
that it is allowed to  take its thermally induced  inward bowing  once it is at operating
temperatures. A jamb that is secured in  position and has taken its "natural" heat curvature is
practically in a zero thermal  stress  state. With this condition, any variation in  temperature
gradients gives rise to stress which varies about zero, instead of around some  already high
value. A practice that should be abandoned is that of securely attaching ambient-temperature,
straight jambs to ovens  and  then allowing them to heat to  operating temperature while
preventing bowing. These jambs enter their operating life at a high level of thermal stress.
    Late in this project the field team reported that a 4-year-old, 6-meter Koppers battery was
in operation with the "allow the jamb to bow and then secure it" concept. We agree with this
approach, and recommend it for retrofitted jambs on older batteries. Also reported were the
facts that (a)  the jamb web in the above battery was very deep, and (b) back-side patching
appeared to be a problem with the "curved, stress-free" jambs. It will be recalled that we are
recommending minimum-depth flanges  on jambs.
    There are various possible methods of mounting a jamb so  that it is in a zero-stress condi-
tion when reaching  operating temperatures. Late in  project it was learned that Inland Steel
Company is using a method  that effectively achieves this result. As understood by Battelle
researchers, Inland mounts the jamb and uses shims and mechanical force to bend the cold
jamb  to the shape that will be stress free when the jamb will  reach operating temperature.
After  bending the jamb, the gap between the jamb and the bricks is grouted. After  this
grouting operation, the jamb is allowed to reach operating temperatures in  preparation for
service. In this instance, the jamb is stressed while cool, and as it heats the stress automatically
slacks off. It remains to be established whether this approach can be adapted to  other retrofit
situations.

-------
                                        111-14

    Jamb-Attachment Methods.  The way that the jamb is held in place can be important to
the effectiveness of sealing. Regardless of the particular design, nearly all jambs are held in
place by reacting in some way with the buckstays. Some designs have a plate behind the
buckstay, with the jamb attached to the plate by studs. In other designs, the jamb is either
directly bolted to the buckstay or is clamped by means of clips which react with the buckstay.
This last type of attachment results in what is sometimes referred to as a unilateral constraint.
That is, the clips tend to keep the jamb from moving outward, but offer no resistance to the
movement of the jamb towards the oven.
    Because of the way jambs are held to a buckstay, the jambs gain all or  a lot of the stiffness
of the buckstay. That is, the stiffness of the jamb is effectively that of  the jamb plus the
buckstay. The measure of stiffness of a structural component is the product of the second mo-
ment of area and the modulus of elasticity of the material. Jambs considered in this study
generally have a moment of area which varies from 60 to 700 in.4, while the  buckstays have
moments which varied from 900 to 2500 in.4. Now, if the jamb is held rigidly to the buckstays,
and if the buckstays are so much stiffer than the jamb, it becomes apparent that the jamb can
be effectively restrained from motion so long as the buckstay remains stable. Conversely, the
jamb can "be at the mercy" of the buckstay should the top of the buckstay move away from
the oven. Clips do not restrict any tendency for a jamb to hourglass.
    It is recommended that the jambs be firmly attached to the buckstay (with special connec-
tors) to prevent hourglassing and to minimize flexing of the jamb. With a locked-in-place
jamb, it appears that the difficulties of maintaining a patching seal behind the jamb would  be
minimized. However, if the top of the buckstay moves outward, the top of  the jamb will move
outward the same amount. This would  result in  heavy  leakage, carbon buildup, and other
problems. It is, therefore,  recommended that consideration be given to adding additional
sealing strips that (a) are attached to the oven brickwork rather than to the buckstays,  and (b)
permit a sliding action with no gap development in the event the top of the jamb should move
outward. A possible seal for this purpose is shown in Figure 111-2. This approach assumes that
carbon will not form in a  gap (near the top of a jamb) that does not connect with the at-
mosphere, i.e., it assumes that it takes a flow-through to deposit tars in gaps.We do not  know if
this assumption is correct.

-------
                                        111-15
     Angle Iron Packing Retainer

     Retainer Is Fastened to the
     Oven Bricks and Not to
     the Buckstays
       Packing Material
            Top Cross-Piece of Jamb
              FIGURE 111-2.  SIDE VIEW OF THE TOP OF A JAMB WITH A
                           PROPOSED PACKING RETAINER
Summary of Recommendations for Retrofitted Jambs
    (1) For retrofittable L-shaped jambs, lower the stiffness of the design by reducing the
       depth of the web section. This will make it easier to hold the jamb stable during
       temperature variations. This will also lower the temperature gradient developed
       during cooling of the jamb by rain.
    (2) Mount the jamb so that it takes its natural thermal bowing without restraint when
       it is at operating temperature. This will help lower the stress level reached during
       any subsequent temperature excursions.
    (3) In  attaching the jamb to fulfill Recommendation 2 (above), fasten the jamb
       securely to the buckstay.
    (4) In preparation for occasions when the top of the buckstay (and attached jamb)
       may briefly move outward away from the oven, design side seals that will allow
      jamb movement without opening flow-through gaps.
   (5) Only elastic materials should be used for jambs. Gray cast iron is not elastic until
      such time as it has already taken a permanent set.

-------
                                        111-16

    (6) Although small fires at the door seals do not result in damage to the jamb, they
       can damage the seal. The incidence of seal fires should be minimized by the in-
       stallation of upgraded  door seals, but such fires that do  develop should be ex-
       tinguished promptly.
    (7) Minimize  the length of time that any portion of an oven  is at  less  than  at-
       mospheric pressure.
    (8) For future coke batteries, serious design efforts should be directed to lowering of
       the operating temperatures of the end-closure components.

                       Jamb Material Selection/Recommendation
    This portion of the report deals with the rationale that led  to the  recommendation to use
annealed ferritic ductile iron for coke-oven jambs. The analytical effort on thissubtask passed
through various stages in response to the information and guidance received from the output
of Tasks 1, 2, and 3. Because consideration of materials was not included in detail in Progress
Reports, the reasoning that led to our recommendation is presented in some detail in the
following  paragraphs. The essence of the oral presentation  made to the Sponsors on this sub-
ject is included in the summary of this section.

The Progression of Approaches Used in
Material Evaluation/Selection
    Because truly warped jambs obviously have been overstressed, the first approach was to
search for materials which, for a minimum increase in cost, would have the highest ratio of
                         Stress Tolerance at High Temperatures
                        Stress Developed at High Temperatures
    While this search was in progress, it became appreciated that attempts to hold coke-oven
jambs "straight" (on ovens) contributes to  the warpage problem; i.e., stressing of jambs to
force them to remain straight against their natural inclination to assume a bowed shape (as dic-
tated by thermal gradient) was judged to be an approach to be abandoned in the retrofitting
of jambs. At this point it appeared possible that pearlitic gray  cast iron could continue to be
used for jamb  castings if heating above 640 K (700 F)  for prolonged periods of time could be
avoided. Above 640 K (700 F), unalloyed gray iron develops microstructural changes that can
cause warping.
    On further consideration, it was concluded that although securing of jambs on ovens only
after the thermal bowing of the jambs has occurred would  significantly lower the normal
operating  stress on the jamb, it would not eliminate stresses. Additional  significant stresses

-------
                                          111-17

could arise due to localized heating (e.g., door fires or fires behind the jamb) and also during
the sudden thermal gradient increases that occur during rainstorms. Further investigation in-
dicated that the stress tolerance and temperature resistance of gray iron could be significantly
increased by addition of small amounts of alloying agents during the foundry process. At this
point the researchers were discussing alloyed gray cast iron at progress review meetings with
the Sponsors.
    Later in the project, the analysts working on Task 1 (the mathematical model) and Task 2
(the physical model) were reporting that some of the highest levels of stress in jambs occur
during the early cooling period that takes place during a slanting rainstorm. This rapid cooling
causes a  high, short-term increase in stress in jambs. The duration of the high-stress period is
too short to result in significant  creep or relaxation (these are time  dependent), but the
analysts and metallurgists began to request consideration of materials that have a true elastic
range. Steel has an elastic range and ductile irons approach the elastic properties of steel. It is
reasonably well known that cast irons (plain or alloyed) containing flake graphite do not have a
true elastic range, and that they exhibit some initial plastic strain upon being loaded (either in
compression or tension). Stated another way, cast iron parts  upon being heavily stressed
undergo a small amount of permanent dimensional change. In the case of coke-oven jambs,
this  could lead to warpage. After the  first  stressing, the gray iron  will behave essentially
elastically, unless the stress level later is raised above the original stress.
    It was not appreciated how much this plastic strain (in the form of bowing) could be until
prestressing (prior to machining the jamb) was considered as a method for minimizing the
plastic component in gray irons prior to service. Depending on the strength of the cast iron
and the length and thickness of the jamb, heavy prestressing would result in a bowing of from
5 to 13 cm (2 to 5 inches). This was considered to be out of range of practical manufacturing
considerations and machining. These  calculations were followed by laboratory tests in which
water was sprayed against one side of samples of heated steel  and cast iron. These samples
were restrained from bending. Results confirmed the need for an elastic material as further in-
surance against the warping of retrofitted jambs. These results are covered in a later portion of
this part of the report. Further consideration of gray irons  was discontinued, and  the con-
sideration of elastic materials was started.

Basic Information and Evaluation of Jamb Materials
    At the present time, the materials being used in jamb castings or jamb fabrications are (a)
gray cast iron, (b) alloyed gray cast iron, (c) ferritic ductile iron, and (d) fabricated  (welded)
carbon steel. It is not  clear whether various builders of coke plants are making speof.c

-------
                                         111-18

recommendations to coke-oven operators about jamb materials, or whether the builders are
asking the operators to make the selection, judging from the comments of recent purchasers,
some German builders appear to be recommending ordinary gray cast iron, and the Ikio Com-
pany of Japan (supplier of jambs, seals, and doors) is recommending gray cast iron alloyed with
a low percentage of  chromium  to decrease the rate of microstructural changes in the con-
tained pearlite. Some buyers have mentioned that the Koppers Company is recommending
ferritic ductile iron, and operators of several older coke plants are replacing their cracked or
warped gray iron jambs with fabricated steel jambs.
    Late in this project visits were made to several relatively new 6-meter batteries including
several Koppers-built batteries, one of which was equipped with gray cast iron and one with
ferritic ductile iron. These jambs were originally installed to allow the jamb to take the cur-
vature dictated by the thermal gradient.  It became evident that the Koppers Company had
developed the "let the jamb curve to hold down the stress level" approach well ahead of our
analysis. Both the "curved" gray iron and  ductile iron jambs appeared to have largely avoided
dimensional distortions.  However, the fact that the "curved" gray cast iron jambs on a new
battery seemed to be dimensionally stable was not, in this  instance, taken as an indication that
curved gray iron jambs can be used in retrofit operations on older batteries. Our reservations
with regard to this field information  on curved gray iron jambs are based on the fact that the
battery having this arrangement  has unusually low temperatures on all of its end-closure com-
ponents. This favorable and unusual  condition is believed to be a function of the innovative
end-flue combustion control installed on  this battery; i.e., this battery is not typical in terms of
end-closure conditions.

    Thermal Stress Considerations. Given that metal is (a) being heated from one side, (b) in
steady-state heat transmission, and (c) restrained from bowing; the basic equation  for the
stress on the outer "fibers" of the material (compression on the hot side and balancing  tension
of the cooler side) is usually given as:
  Maximum   _  Elastic Modulus x Thermal Expansion xTemperature Difference  (hot to cold side)
    Stress     ~                           2
For our purposes the  equation can be restated as:
  Maximum   _  Elastic Modulus * Thermal Conductivity * Heat Flux x Thickness
    Stress     ~                           2
If a simplifying assumption is made  that  the candidate materials all have the same thickness
and are receiving the  same heat flux into the back of the jamb, then the above equation can be
used to rank the materials in  terms  of the relative maximum  thermal stress developed. Our
relative stress index is:

-------
                                              1-19
               Relative Stress Index  =  Stress Developed in Alternative Material
                                           Stress ueveioped in Gray Cast Iron
Using  the foregoing equations  and  assumptions, the ranking of  candidate materials is as
follows:

                 Relative Stress Level Developed in Candidate Jamb Materials
  Candidate Materials
 Class 20 Gray Iron
 Class 30 Gray Iron
 Alloy Gray Iron
   (0.8% Mo and 0.6% Cr)
 Ferritic Ductile lron(3)
 Carbon Steel
 Low-Alloy Steel
   (0.5% Mo)
Ranking Relative to
 Class 20 Gray Iron
 Ranking Relative to
Class 30 Gray IronO)
                                                                          Nominal Price,
                                                                         Dollar Per Pound
1.0
1.4
1.8

1.0
1.3
$0.40
$0.40
$0.45-$0.50
        2.8
        3.0
        3.8
        2.0
        2.2
        2.7
$0.60
$1.30
$1.50
 (1)  Class 30 gray iron is usually specified for jamb castings. In some instances of early and rapid warpage of jambs,
    reports of investigation by steel companies indicate that they had received Class 20 gray iron instead of Class 30
    gray iron. There was no information on whether the "rapidly warping" jambs had been subjected to fires, but
    the samples taken from the discarded jambs showed no indication of graphitization or internal oxidation.
    Class 20 gray iron exhibits a greater strain per unit of stress than Class 30 gray iron.
 (2)  Price does not include machining or any heat treatment. Shorter jambs carry a higher price per pound than
    taller jambs. Cost of fabricated steel jambs include the labor cost of fabrication.
 (3)  Only the ferritic grade of ductile iron was considered because pearlitic grades develop "growth" by
    graphitization above 800 F. This is another "safety factor" consideration.
 If the only required viewpoint in material selection would be to find a material with the lowest
 level of developed stress, then low-tensile-strength gray cast iron would be the obvious selec-
 tion. With the large amount of flake graphite in the low-strength iron, the modulus of elasticity
 is  low  and  the  thermal  conductivity  is  high.  There  are,  however,  stress-tolerance
 considerations.
     Stress Tolerance  Considerations.  There is available in  the literature a considerable
 amount  of information on the high-temperature mechanical  and physical properties  of the
 candidate materials. However, there is only a limited amount of stress-relaxation data; i.e.,
 measurement of the reduction in stress under constant strain. Also, there are few data on com-
 pression creep or comparative thermal straining tests in which the magnitude of the stresses
 set  up in the material depend to a large extent on the physical properties of the material; e.g.,
 the expansion coefficient, elastic modulus, and thermal conductivity.

-------
                                          111-20

    An ideal arrangement to obtain measured data as compared to calculated data would be
to develop a laboratory arrangement in which it would be possible to obtain material dimen-
sional changes (particularly with regard to bending) under conditions of:
    (a) Varying degrees of restraint
    (b) Varying degrees  of heat  input  (and  therefore  temperature)  to  one side of
        samples, and
    (c) Varying levels of thermal gradients, including water cooling of the unheated side.
    Consideration was given to this approach but extensive testing of this type was not includ-
ed  in our proposal and work plan. Therefore,  only limited testing was  done by quenching
restrained and heated samples. The results are presented later in this section.
    In our search for upgraded materials, discussions were held with technical personnel in
companies manufacturing diesel engines. These companies wish to keep  material costs as low
as possible, but they do have heat input to one side of gray iron (as in the cylinder heads) and
the parts are restrained. Engine parts are,  however, water-cooled on one side and they do not
reach the temperatures reached in warped jambs (as inferred from metallographic examina-
tion). It was judged that the papers of Baker and Pope* (1961) and R. Bertodo** (1970) were the
most appropriate for examination from the viewpoint of selection of a material for coke-oven
jambs. Baker and Pope heated cylinders of gray and ductile iron after locking them in place in
a special testing  machine having fixed end-blocks. The samples  were heated slowly and uni-
formly, and did not, therefore, reflect the effect of thermal conductivity of  the material. On
heating  the  bars,  the compressive  forces  were  measured as a function  of  time  and
temperature. During the slow heating and hold period  (thermal gradient was low), there was
plastic strain that translated into tensile stress after cooling to room temperature. In comparing
pearlitic  ductile iron, alloyed gray iron,  and unalloyed gray  iron (tensile strengths not
specified), the unalloyed gray iron developed the lowest compressive stress on heating but the
highest residual tensile stress on cooling.  In this work it was indicated that gray iron could be
alloyed (1.4% Ni and 0.4% Mo in this instance) to have thermal-strain properties superior to the
ductile iron and particularly superior to the unalloyed gray irons.
     Bertodo published an article  describing his evaluation of 166 heats of  plain and alloyed
 gray  cast irons for  use in diesel-engine components. At various temperatures and  loadings,
 compressive creep and relaxation tests were obtained on samples of all the materials. Also
 *Baker, S. G., Pope, J. A., "The Creep, Stress-Relaxation, and Thermal-Strain Properties of Several High-Grade Cast
  Irons", Report No. 363, The British Shipbuilding Research Association, 1961.
**Bertodo, R., "Grey Cast Iron for Thermal-Stress Applications", Journal of Strain Analysis, Vol. 5, No. 2,1970.

-------
                                          lil-21

measured were expansion coefficients,  the  elastic  modulus, thermal conductivities, and
fatigue data. Although not directly stated by Bertodo, it appears that he was working towards
increasing the strength and the creep and relaxation resistance of gray iron (by alloying) while
attempting to retain the maximum amount of the low-thermal-stress-generation properties of
gray iron (e.g., low modulus, and high thermal conductivity). One of his considerations was to
minimize the increase in hardness to retain  a more ductile matrix (to retain low-cycle fatigue
resistance). Bertodo developed alloyed pearlite gray irons [0.4% Mo + 0.4% Cr or 0.4% Mo +
1.3%(Ni + Cu) ] having a resistance to thermal stress 2.8 times that of ordinary cast iron. Both the
plain cast iron and the alloyed irons were in  the 276 MPa (40,000 psi) ultimate-tensile-strength
class.
    At this point in the project, consideration was being given to recommending alloyed gray
cast iron for retrofittable jambs taking advantage of the ability of molybdenum additions to in-
crease the creep strength and relaxation strength. Chromium additions were also considered
necessary to increase  the stability of the pearlite.

    Material Elasticity and Thermal Shock Resistance. As previously noted, near the comple-
tion of Task 1  (mathematical modeling) the analysts  were requesting consideration of more
nearly elastic materials for the jamb application. The stress-strain data of some of the important
candidate materials are shown in Figure 111-3. The gray irons (even if alloyed) have a stress-
strain plot curved almost from zero stress. On loading these materials (if they have not been
prestressed), a small amount of permanent deformation occurs as shown in Figure 111-4. Once
the peak stress occurs, further deformation  essentially stops and the gray iron then begins to
act almost  as if it were an elastic material.  In comparison, steel can be loaded to about its
proportional limit and will return to its original dimensions upon being unloaded. The nodular
or ductile irons, with their graphite in spheroidal form, approach the straight stress-strain line
of steel, at least to the 170 MPa (25,000) psi level of stress.
    In  Figure  111-4,  various  percentages  of  permanent strain have been converted to
equivalent  bowing in  a 4.3-m  (14-foot) jamb. Although the plastic component of gray  iron is
well known, in this work it became apparent to the researchers that a small percentage of
plastic strain in cast iron can translate to an appreciable amount of permanent bowing in a gray
iron jamb.
    Working from stress-strain data, calculation can be made (as shown in Figure III-4) of the
possible permanent distortions that will occur during the "shake down" effect in a laboratory
test. Tests were made on annealed gray cast iron and steel plates measuring 6 mm (1/4 inch)
thick by 38 mm (1.5 inches) wide by 14 cm  (5.5 inches) long. These plates and  fixtures were

-------
                                 1-22
               45
               40
               35
               30
               '25
               20
                10
	f
-    A/7
         -Low Carbon Steel
    /  ~—Pearlitic Nodular
    I  /
                                   Gray Iron-Class 30
                           Gray Iron-Class 20
                                           I
                                                    300
                                                    250
                                                    200
                                     E
                                     e
                                                    150
                                                    IOO
                                                    50
                 0   0.10  0.20  0.30  0.40  0.50  0.60  0.70  0.80
                               Strain, %
      FIGURE HI-3.  TYPICAL STRESS-STRAIN CURVES IN TENSION
                   FOR VARIOUS CAST IRONS AND STEELS
                                              A stress-strain curve in
                                              tension for a class 30
                                              gray iron in which the
                                              load was removed to show
                                              the permanent
                                              deformation*.
                           0.4         0.6
                    Strain — Percent
FIGURE 111-4. A TYPICAL STRESS-STRAIN CURVE FOR GRAY CAST IRON
             SHOWING THE AMOUNT OF PERMANENT STRAIN THAT
             DEVELOPS DURING THE EARLY STAGES OF PUTTING
             THIS MATERIAL IN LOAD-CARRYING SERVICE

           Permanent Strain Percentages Converted to Amount of
                 Bowing in a 14-Foot Jamb (6x6 In. Jamb)
                                  Equivalent Bowing
            Percent Strain

                0.01
                0.10
                0.20
            Millimeters

                2.5
               30.0
               58.0
Inches

 0.1
 1.2
 2.3

-------
                                          111-23
heated to 810 K (1000 F), and then were water quenched from one side as shown in Figure I.I-5
Typical results of several tests using different samples of gray iron and steel were as follows:
	 Measured Bowing
Cast
Millimeters
-0.25
-0.25
-0.25
-0.25
-0.25
Iron
Inches
-0.01 (i)
-0.01(2)
-0.01
-0.01
-0.01
— 	 — 	
Steel
None
None
None
None
None
                 Cycle Number
                       1
                       2
                       3
                       4
                       5
               (1) Negative value. The cast iron samples took a permanent
                  bow (as measured over the 5.5-inch length) in the convex
                  direction, i.e., outward at the middle of the plate toward the
                  the water cooling.
               (2) There was no increase in bowing after the first cycle.
    A bow of -0.25 mm (-0.01 inch) over a distance of 14 cm (5.5 inches) represents a perma-
 nent strain of about 0.2 percent (see Figure III-4). Depending on the length and thickness of a
 jamb, this could cause a bowing  of up to 5 cm (2 inches). This test is considered to be rather
 drastic. However, the point of interest was that although the stress level on the steel samples
 had to be 2 or more times higher than in the gray iron (during the cooling step), steel did not
 develop a permanent set and gray iron did.
    When gray iron jambs are replaced on operating coke ovens, in most instances the need
 for replacement is the result of cracks that develop after long periods of service. The replace-
 ment rate is about  5 to 10 times higher on the hotter coke side than on the pusher side. It is
 probable that this cracking is the result of tensile failure during thermal shocking of jambs dur-
 ing rain storms (and not due to water quenching of fires). Stress concentration and/or fatigue
 failure may also  play a role. However, a change of material for retrofitted jambs was  con-
 sidered to minimize/eliminate cracking of jambs.
    To be of most value, thermal-shock tests should rather closely approximate the operating
 conditions to be expected. A test procedure that  is akin to the  sudden cooling of the

-------
                                       111-24
               Heat Sink
             (Steel Plate)
                                                               Water
                                                               Spray
    Procedure:   1. Samples and heat sink are heated to 1100 F
                2. After heating, samples are clamped to heat sink
                3. Both samples are water sprayed on one side
                4. After cooling, samples are undamped for measurement of bowing
         FIGURE III-5. LABORATORY ARRANGEMENT FOR SUBJECTING METAL
                     SAMPLES TO THERMAL STRESS


restraining jambs was completed by Kattus and McPherson* in 1959. In their thermal-shocking
procedure, notched cylinders of materials were subjected to a bending moment and were
then heated evenly to 700 K (800 F) prior to water cooling to 340 K (150 F) in 15 seconds. The
functions of interest were the number of heating and cooling cycles to fracture as function of

bending load.

    Their rating in terms of thermal-shock resistance was as follows:
                                               Approximate Stress Level
                                                  to Obtain Cracking
                                                     in 100 Cycles
                Ranking

                1 (Best)  Steel

                2

                3
       Material
Ferritic Ductile Iron
Low-Alloy Gray Iron
  (0.6% Cr + 0.8% Mo)
Unalloyed Gray Iron
  [200 MPa at 700 K
  (30 ksi at 800 F) ]
MPa
700
350
300
Ksi
100
50
40
                                                   200
30
•Kattus and McPherson, "Properties of Cast Iron at Elevated Temperatures", ASTM STP No. 248,1959.

-------
                                           1-25
    These results indicate that ferritic ductile iron and alloyed gray iron have more resistance
than unalloyed gray iron to cracking under thermal-shock conditions, even though the ther-
mal stress is higher in the steel.

Summary
    Pearlitic gray cast iron has the advantages of:
        (1) Lowest cost
        (2) A high thermal conductivity and low modulus of elasticity, with the result
           that it (as a class) develops a relatively low thermal stress when operating un-
           der a thermal gradient.
    Pearlitic gray cast iron has the disadvantages of:
        (1) Susceptibility to damage (internally) by temperatures over about 700 K (800 F)
        (2) Low resistance to creep and relaxation under load and heat
        (3) Tendency  to crack in long-term coke-plant service
        (4) Inelasticity.

    Pearlitic gray iron  can be alloyed to minimize the first three disadvantages. Alloying will
improve:
        (1) Resistance to structural damage by high temperatures
        (2) Creep and relaxation strength
        (3) Resistance to thermal cracking.
    To a degree, upgrading of gray iron by alloying decreases both of the listed advantages.
Alloying:
        (1) Increases the cost
        (2) Increases the elastic modulus and decreases the thermal conductivity
    Because gray cast  irons contain graphite flakes, nothing can be done to make gray iron a
truly elastic material.
     It is our recommendation that:
         (1) Annealed, ferritic ductile iron should be tested in any follow-on demonstra-
            tion  project. Steel has some properties that are improvements over ductile
            iron, but these improvements  do not appear to justify the increased cost of
            steel jambs.
         (2) In retrofit situations, ferritic ductile iron jambs should be installed on ovens
            in such a way  that the jambs have  assumed an  unstressed "thermal cur-
           vature" prior to final bolting. This method is being used by the Koppers
           Company.

-------
                                         111-26

    It is our conclusion that annealed, ferritic ductile iron is the best compromise material for
jambs when taking into consideration:
        (a)  Cost
        (b)  The need for a material whose elasticity approaches that of steel
        (c)  The need to eliminate jamb failure by cracking.

Concluding Comments
    (1) Annealed ferritic ductile iron  is recommended in order to bypass the possibility
       of having residual casting stresses in the jamb casting. Ferritic ductile iron can be
       produced (1) as cast (without following the casting operation with a ferritizing
       anneal) or (2) by a ferritizing anneal. It is recommended that tests be completed
       to confirm that sufficient residual stresses do not remain in as-cast ferritic ductile
       iron to cause jamb warpage. In a demonstration project, the two approaches to
       producing ductile iron jambs could be compared.
    (2) The higher-strength ferritic ductile irons should be considered for jamb castings.
       If, during early trials, there is evidence of distortions that could be caused by
       creep/relaxation, it is recommended that the ferritic iron jambs in subsequent in-
       stallations should be alloyed to increase overall stress resistance.
    (3) The depth of the jamb web (or flange) should be decreased to minimize the high
       level of thermal gradients during transient heat input or heat-extraction periods.
       This consideration is covered in our recommendations on jamb design.
    (4) As a generalization, the seal-mating surface of retrofittable jambs should have a
       surface as smooth  as  possible within  the  limits of reasonable  cost. This
       recommendation can  be explored during  any  demonstration  project. The
       smoother the seal-mating surface of jambs, the easier it is to remove hardened tar
       (carbon) deposits.

                      Design Considerations for  Retrofittable jambs
    The analytical team has indicated their requirements and preferences for upgraded jamb
shape, performance, and jamb-attachment  methods.  However, in considering jamb retrofit,
certain compromises must be considered.
    While  the analytical team prefers a  compact  "rectangular" cross  section, economical
retrofit of existing jambs limits the nature and degree of modification that can be applied to
other end-closure components such as bricks, door frames, and latches. Modifications are also

-------
                                         111-27

limited by operating clearances of components and by existing door-handling machinery.
These limitations require that the jamb-seal surface and the jamb/brick surfaces remain un-
changed on retrofittable jambs and that only the jamb flange or web can be considered for
modification.
    The jamb flange provides strength to resist latching forces through the latch hooks that
are attached to the flange. The jamb must also absorb mounting restraints exerted through
fasteners of various designs. As specified, attempts will be made to shorten the flange width on
retrofittable jambs. There will be variations at  individual ovens depending on the original
design of the jamb.
    Laboratory evidence suggests that smoother jamb-sealing surfaces would permit easier
removal of  hardened-tar deposits. This effect should be further investigated in the optional
demonstration project. Many jambs have a roughness value of 125 microinchesfor the sealing
surface. Finer surface finished should be tried, provided capability is available to produce finer
finishes.
     During Task 5 (Design and Testing) consideration will be given to jamb design  con-
siderations that might make it easier to upgrade the procedure for grouting behind the jambs.
To be considered is the placement of grouting "ports" along the junction of the jamb sealing
surface and the jamb web.  Pressurized grouting  may  minimize  some of the  problems
associated with  back-side sealing. Some  of the information developed during the develop-
 ment of sealing with applied sealants may be useful in this experimentation.

-------
                                          IV-1

                                    CHAPTER IV

                AN ANALYSIS OF THE PROSPECTS FOR SEALANT SEALING
                      (INBOARD LUTING) OF COKE-OVEN DOORS
    Our definition of "prospects" for sealant sealing of coke-oven doors includes:
    (1) Potential for emission control
    (2) Potential for successful development
    (3) Possible degree of acceptance by the coke industry.

                              Background and Definitions
    Within the steel industry there are companies that routinely seal some 2000 coke-oven
 doors (on their  older batteries) by a procedure called luting. In addition, there are foundry
 coke  merchant plants that use this venerable sealing method. Luting consists of hand plaster-
 ing the outside of the jamb/door interface with a low-cost water-tempered clay/coke mixture.
 This sticky mixture is applied before the coal is charged to an oven. In most instances, there is
 some release of emissions through or past the luting material. Following the completion of the
 coking cycle, removal of the door breaks the dried luting material. Fresh luting material is then
 applied to the closed door prior to the charging of coal for the next cycle. Figures IV-1 and IV-
 2 show the cross sections of two existing types of luted doors.  Batteries that are designed to
 have  luted doors do not use pressurized latching of the doors to the jambs. There are latches
 but they serve only to secure the door in position.
    The  emission-control performance  of the luting operation is a variable depending on
 many  factors. There is, for example, an element of skill involved in the preparation and
 application of the luting material. In one survey conducted by a large steel company, it was
 reported internally  that "generally speaking, luted  doors have shown no  outstanding
 differences (in terms of emission control) than self-sealing (metal-to-metal) doors". One of the
 problems, for example, is that the mild  combustion explosion that occurs when  coal first
 enters an oven can jar and crack the luting material. In instances where these cracks and other
 points of emission leakage are repatched, the emission control can be good.
    In visits to luted-door coke plants in the Phase I project, Battelle researchers had noted
with interest that emissions do not come .through the luting material after it has dried and
become porous. This is a favorable consideration in terms of sealant development but, at the
beginning of this project, this observation was tempered by the knowledge that the internal
pressure  in luted ovens is relatively low because of the shorter heights (older batteries) and

-------
                                   IV-2
          FIGURE IV-1. LUTED DOOR DESIGN IN USE ABOUT 1946
 Refractory Plug Retainer
Luting Location
Roll-Formed Steel
       Buckstays
Gas Passage
         FIGURE IV-2. LUTED DOOR DESIGN IN USE ABOUT 1929

-------
                                         IV-3

 longer coking cycles (lower temperatures). Also, the luting material in most instances dried
 very slowly. Thus the porosity does not develop in the material until the internal pressure peak
 may have already subsided.
    Another element of background information is that a company has recently announced
 the intent to build a modern 5-meter battery (to produce foundry coke) that will be using the
 luting approach for sealing the doors.

                               Results of the Phase I Study
    In our Concepts Study, variations of the luting approach were considered and evaluated
 (without experimentation or testing). Figure IV-3 is an example of one of these concepts. At
 the time of the report on that study we were thinking of a foamed-in-place sealant, including
 mechanized means to apply foam to the door while it was off the oven. Implied in our concept
 was (1) compression of the sealant on mounting the door, (2) one-time use of the sealant, and
 (3) mechanized  methods of rapidly placing the material.
    As part of the Concepts Study, members of the EPA organization and the AISI Task Force
 were asked individually to rank  the various concept groupings. This ranking was done by
 working with a criteria listing and a weighting system developed with the cooperation of the
 EPA and AISI representatives. Both the EPA and the AISI members ranked the sealant concept
; as being the best (by a small margin) of all the concepts considered. The higher ranking for the
 sealant concepts was based on their higher rating for "effectiveness to reduce emissions". It
 was indicated that concerns over "operating  cost" and "availability of materials" kept the
 sealant family of concepts from being rated best by a significant margin.
    In Battelle's subsequent estimate of our level of confidence for successful final develop-
 ment and performance of each concept family, sealants were ranked below metal-to-metal
 sealing concepts. Our reservations were primarily in the areas of (1) increased  gap-closure
 capability, and (2) dependability and repeatability.
    It was  Battelle's recommendation (in the Concepts  Study) that "because sealing  systems
 based on  sealants  have the  possibility of completely eliminating door emissions, it  is
 recommended that a laboratory/plant exploratory experimental project be initiated to further
define the problems and  possible  solutions.  It  is suggested that, if there is demonstrable
progress,  the entire approach should  be evaluated in  a feasibility/cost/benefits  analysis
relative to the progress being made  in the development of metal contact seals".
    In the proposal for this present project, it was stated that:
    (1) The "unknowns" that surfaced in the evaluation of sealant systems in the first
       study should be studied in a laboratory research project

-------
    \   Existing Brickwork
                         Gap of Varying Thickness
                           Resulting from Jamb
                         Warpage Fills With Foam
                               Sealing Frame
                                 (Added)
FIGURE IV 3  AN EXAMPLE OF A SEALANT-SEALING CONCEPT PRESENTED IN THE
            FIRST PROJECT (STUDY OF CONCEPTS)

-------
                                         IV-5

    (2) Laboratory testing of sealant systems should be followed by a preliminary evalua-
       tion of candidate sealants at chuck doors
    (3) A feasibility/costs/benefits analysis of sealant procedures will be completed as
       part of Task 4 (if there has been demonstrable progress)
    (4) Because  there is more  uncertainty  in terms  of  the potential for  successful
       development of sealant systems, the expansion  of a possible successful research
       project on sealant systems  into a demonstration project is not included in this
       proposal.
    This proposal was accepted by our Sponsors and this chapter of our report summarizes the
work that was done on sealant research. Details were presented in the working-paper reports
that have been issued to the Sponsors.

                                  Criteria Development
    To guide the research project on sealants, a  list of criteria was developed. This list was ex-
panded during the project and the  final listing stated that an acceptable sealant should be:
    (1) Low in cost per door closure (high-priced sealants can be considered provided
       that the sealants are reuseable)
    (2) Tolerant of short-term heat excursions to 700 K (800 F)
    (3) Capable  of sealing up to a  6-mm (1/4-inch) gap between the door and the jamb
       without any "blowout" by the internal gas pressure.
    (4) Resistant to attack from solvents and tars evolved by the coals
    (5) Completely inert and should not evolve toxic or obnoxious fumes or fumes that
       create an environmental  problem
    (6) Expandable under the action of heat, or it should not shrink or crack (a) upon
       drying, (b) during the duration of the coking cycle, or (c) during a temperature
       excursion
    (7) Capable  of  preventing  leakage  of  emissions (particulates and  gas)  at  all
       times—wet or dry.
    If the sealant is applied  as a slurry or mastic it should:
    (1) Adhere to a vertical steel or cast iron surface operating at 530 K (500 F)
    (2) Dry slowly to allow time to place the sealant and compress the  soft sealant on
       latching the door
    (3) Lose its adherence after drying so that when the door is removed the used sealant
       either falls off or is easily  stripped
    (4) Remain cohesive after drying and  not create  a dusting problem  after drying or
       after use.

-------
                                        IV-6

    The criterion that specified that the sealant should be gas tight was selected as an ideal
condition or goal. Battelle personnel were aware that every charge of coal evolves its own
sealants in the form of condensible tars and vapors. If this specification could not be met in the
laboratory, it was expected that field tests would  determine whether the evolved coke-oven
tars would seal the applied sealant and when this would occur.

                            Summary of Laboratory Results
    In the selection of laboratory test equipment, it was judged that more information could
be obtained if  a thin strip was used  to penetrate or compress a contained sealant rather than
compressing a  sealant between two parallel plates. This approach is analogous to the concept
shown in Figure IV-4 (taken from the Concept Study Report). To simulate this concept, a sim-
ple arrangement was built to evaluate material shrinkage, drying rate, and/or penetration
resistance and  release conditions. This simple equipment is shown in Figure IV-5. To evaluate
gas tightness of both compressed sealants and sealants loosely filling a 6-mm (1/4-inch) gap,
the equipment shown in Figure VI-6 was used.
    A considerable amount of effort was expended in attempts to locate closed-pore foamed
or unfoamed materials for use in the door-sealing application. All candidate sealants failed to
meet  two or  more of the foregoing criteria. With this result, the research was  shifted to
development of a nonfoaming, nonshrinking sealant that would  remain  plastic for a6-minute
dwell  period in the heated equipment shown in Figure IV-5. The 6-minute dwell period was
selected as being about the  maximum amount of time to be  allowed  for both placing the
sealant and mounting the door  on an oven. With the sealant still being plastic upon door
closure, the door edge would penetrate the sealant to effect a seal. The following is a summary
of the results obtained with various materials:
    (1) Sixteen formulations were evaluated with  fly ash as the base. The most promising
       ones had the following formulations:
                                             Volume  Percent
                                      Mix 19     Mix 22     Mix 26
               Fly ash                   40         46        53
               Expanded perlite          40         30        53.6
               Plastic ball clay            20
               Corn starch               -         11         6.4
               Pulverized pitch           —         13         —
               Water                    25         32        31
       A small amount of sodium bicarbonate was added to some of the mixes to create
       early expansion and thereby improve retention in a channel during drying.  Ex-
       pansion was improved, but a layer of subsurface porosity was created at the metal
       interface.

-------
                                               IV-7
                        Existing Brickwork
1. Foamed-in-place or pliable material would be applied to the
  channel frame by mechanical means mounted on the door
  machines. Likewise, mechanical means would be required to
  remove the material when it is no longer functional.
2. Preformed foam shaper would be foamed at an off-battery
  side, brought to the door machines, stored in suitable
  facilities, and applied to the channel by mechanical means.
                                              May be Shimmed and
                                              Grouted to Make Flat
                                              if jamb Is Badly Warped
                                               Pliable Material of
                                               Foamed in Place or
                                               Preformed Foam
                                               Deforms Around
                                               "T" Member
           FIGURE IV-4  SEALANT APPLIED BETWEEN THE DOOR AND JAMB SURFACES
                        (CONCEPT TAKEN FROM THE PHASE I REPORT)
                                Door Edge
   Contact Thermocouple
         Sealant
       Channel

Sealant End Dams
  (Removed)

       Hot Plate
               FIGURE IV-5.  SIMULATED JAMB CHANNEL (DOOR EDGE EMBEDDED
                             IN THE SEALANT)

-------
                                 IV-8
FIGURE IV-6. PHOTOGRAPH OF A LABORATORY ARRANGEMENT USED TO
            SIMULATE THE JAMB/SEAL RELATIONSHIP AT COKE OVENS

            This equipment was used to evaluate the leak tightness (gas flow) of
            sealants when compressed between the seal and the jamb.

            (See page IV-11 for comments on this equipment)

-------
                                     IV-9
(2) Seven formulations were evaluated with  minus 0.8 mm  (minus 20-mesh) steel
   plant dust as a base. The base material consisted of waste in the form of basic oxy-
   gen furnace dust, blast furnace dust, dried blast furnace sludge, and mill scale in
   the proportions produced at the Kaiser Steel Corporation. All of the mixes were
   bonded with corn starch. Some had additions of expanded  perlite or sodium
   bicarbonate. None of the mixes based on steel plant dust was judged to be
   acceptable.
(3) As a reference for sealant development, a quantity of luting material was provid-
   ed by the Koppers  Company  which has a luted-door coke plant in St.  Paul,
   Minnesota. This Koppers mix  consisted  of a combination of C&L clay (Cedar
   Heights Clay Company), coal, coke breeze, and water. As used in the test-channel
   arrangement, this mixture became  hard and brittle in a 6-minute dwell period.
   Additions of expanded perlite or sodium bicarbonate resulted in good plasticity
   for 6 minutes but markedly increased the  porosity. All variations of the Koppers
                                                       r
   mix gave a dirty release but none of them shrank on drying. The lack of shrinkage
   led to an evaluation of clays as a sealant base.
(4) C&L clay, without additions, was plastic after a 6-minute dwell period in a 420 K
   (300 F) channel, but shrank 17  percent. Upon  reuse with some fresh clay, the
   shrinkage was lowered to 3 percent.
   Western bentonite and Southern bentonite tested at 420 K (300 F) shrank 18 and
   15 percent, respectively. Upon  reuse of the Southern bentonite with some fresh
   material, the shrinkage was still high (13 percent). Plastic  ball clay with an addi-
   tion of some expanded perlite shrank 16 percent  in a 420  K (300 F) channel.
   Reused C&L clay was superior to the bentonites in resisting shrinkage during dry-
   ing. With 2 parts of used C&L clay to 1 part of fresh clay, an  addition  of 50 volume
   percent of coke breeze resulted in a small amount of shrinkage only across the
   top. of a 420 K (300 F) channel. At 60 volume percent coke breeze, shrinkage was
   absent but the sealant was hard after a 6-minute  dwell period.
   From  ancient times, potters have  added carbonaceous  materials to clays to
   reduce shrinkage on drying.  It is apparent that the early developers of coke-plant
   luting material had known or rediscovered this effect.
(5) With a basic mix consisting of equal volumes of C&L clay and coke breeze in the
   proportions of 1 part used to 1 part fresh, additions of corn starch  or expanded
   perlite with only 0.4 volume percent corn starch resulted in good plasticity, no

-------
                                          IV-10

        shrinkage, some debris on release, and scatter pores. In a 420 K (300 F) channel,
        the best composition with these ingredients was Mix 70:
            2 parts of used Mix 69 (equal volumes of fresh C&L clay and coke breeze)
            1 part fresh C&L clay
            1 part coke breeze
            6 v/o corn starch
            25 v/o water.
        In tests conducted at 530 K (500 F) it was found necessary to increase the amount
        of corn starch to retain plasticity for the desired 6-minute period. The mixture
        recommended for gas-tightness tests and field tests consisted of:
            2 parts used material
            1 part fresh C&L clay
            1 part coke breeze
            8 v/o corn starch
            28 v/o water.
    (6) Figure IV-6 shows a simulated door/jamb system that could be pressurized with
        nitrogen to evaluate the  leak tightness of  sealants after the  door  edge was
        embedded at a load of 13 kN per linear meter (75 pounds per linear inch). All of
        the sealant mixes tested became  porous after drying and leaked gas. A review of
        recorded observations showed that all of the mixes showed evidence of veining
        at the hot-metal interface. The veins formed passages for leakage of the nitrogen.
    (7) The  cost of the materials  in the sealant recommended for field trials was es-
        timated as follows:
   Mix     v/o    v/o of Dry Mix                    Dry Mix Materials Cost
1 p clay     50   100(0.5)/1. 08=46.3  (1286 kg/m')($0.011/kg)(0.463 mVm3 mix)   = $ 6.55/ml
                                [(180.3 lb/ft')($10/ton)(0.463 ftVff' mix)/2000 =           $0.19/ft']
1p breeze   50   100(0.5)/1. 08=46.3  (844 kg/m')($0.033/kg) (0.463 mVm3 mix)    = $12.90/m'
                                [(52.6 lb/ft')($30/ton)(0.463 ft'/ft' mix)/2000 =           $0.37/ft']
8 v/o starch    8"" 100(0.08)71.08=7.4  (631 kg/m')($.209/kg)(0.074 m'/m' mix)     = $ 9.76/m!
                                [(39.4 lb/ft')($190/ton)(0.074 ftVft3 mix)/2000 = _  $0.28/ft']
                                                                     $29.21/m!  [$0.84/ft']
      (a) Of the above.

        Wet  mix is about 80 v/o of dry mix; cost of wet mix  is:
                             = $36.5Vm',or      = $
        With clay at $0.022 per kilogram ($20 per ton) the cost is increased to $45.20/m'
        ($1.28/ft3) of wet mix. Reuse of 1 or 2 parts of used sealant to 1 part of fresh sealant

-------
                                         IV-11

       looks promising to reduce the cost. However, repeated reuse needs to be in-
       vestigated to learn whether or not the starch content of the fresh portion needs
       to be increased to maintain the delayed hardening after the mix is applied to the
       hot door.
           If the assumption is made that the industry will keep  the machinery  that
       slides the door downward a fraction of an inch prior to latching, then the channel
       at the top and bottom of the door needs to be wider than the channel on the
       sides of the door. For a door having a channel 2.5 cm (1 inch) deep, 51 cm (20 in-
       ches) across at the top and bottom of the door, 4.3 m (14 feet) on the sides of the
       door, and horizontal channels are 2.5 cm (1 inch) wide, and the vertical channels
       are 1.3 cm (1/2 inch) wide, the volume of  sealant required is 0.0034 m' (0.12 ft').
       With clay at $0.022 a kilogram ($20 a ton), the indicated mater/a/ cost for sealing a
       door is 15 cents. This is for a mix that consists of all virgin materials.

                  Conclusions of the First Phase of the Laboratory Work
(1) A near-standard luting  mixture (clay  and coke breeze), to which corn starch has  been
   added to slow the drying rate, meets all of the selected criteria except that it is not gas tight
   after drying.
(2) It is recommended that a fresh mix of clay/coke breeze/corn starch/water be tested at a
   coke  plant to determine whether the oils and tars evolved from the  coals will  plug the
   porosity that develops in the dried sealant.
(3) Of all the mixtures examined, the modified luting mixture (given above) will be the lowest
   cost per unit volume and per door closure.
(4) Luting mixtures containing corn starch give off  a faint odor of burnt starch when heated to
   530 K (500 F) and higher. In the field testing it should be determined whether this odor is
   detectable.

                              Summary of Field Evaluations
    The major objectives of field testing a sealant were:
    (a)  Determine whether sealant (porous after  drying) would prevent emissions, i.e.,
        would  the liquids evolved from the coal plug the "pores" of the sealant and pre-
        vent emissions in a usable length of time
    (b)  Determine whether gas leakage is actually blocked when emissions are not visi-
        ble at the sealant location.

-------
                                        IV-12

    The sealant mixture taken into the field consisted of:
        1 part minus 3.4 mm (6-mesh) coke breeze
        1 part pulverized refractory clay (Cedar Heights Clay Company)
        9 volume percent corn starch (of the above)
        33 volume percent water (of the dry mix)
    In weight percent this mixture consists of:
        37.6 percent coke breeze
        57.3 percent clay
        5.1 percent starch
        32.0 percent water (of the dry mix)
    This above percentage of water gives a rather stiff sealant suitable for hand application
with a putty knife. For the test, the jamb and door seal of a chuck door were scraped clean of
carbon. This cleaning operation opened up gaps as much as 0.6 mm (0.024 inch) in width along
the top and side of the closed chuck door. Just after the leveling operation, a thick layer of
sealant was troweled only on the right top corner of the frame of the chuck-door housing, in
line with the metal seal strip on  the door. The door was then closed and the chuck door was
observed during the first 30 minutes of the coking cycle. Emissions were observed from every
location except where the sealant had been applied. The sealant dried slowly and the internal
pressure of 59 Pa (0.24 in. of water) or higher did not force any visible emissions through the
dried sealant. This was judged to be an encouraging result, but not necessarily a significant
result. It was judged that duplication of the result at the bottom of a coke door (because of the
higher  pressure drop across the sealant) would, however, be a significant result.
    To complete a test on part of an operating door,  it was necessary to arrange for a known
gap size (to be filled with sealant prior to closing the door) between the seal and the jamb, i.e.,
a manufactured gap. Consideration was given to placing a metal "spacer" in one of the lower
corners of either a Koppers or Wolff-type door. Coke-plant superintendents, however, could
not work up any enthusiasm for this approach. About this time it was learned that personnel of
the Toledo coke plant of Interlake, Incorporated, were experimenting with an innovative seal
consisting  of a 2.5-cm-wide (1-inch-wide) asbestos braid. This plant was visited and permis-
sion was obtained to test sealants. Interlake,  Incorporated, furnished a drawing of this seal
(Figure IV-7).
    By removing some of the adhered  and hardened tar on the inboard face of this seal, we
were able to develop a 3-mm (1/8-inch) depression or "gap" over a length of 15 cm (6 inches)
at the left-hand bottom corner of the door.

-------
Slot (or
Pins Guide Pins
\
/
' \ / '
CJ
SIDE VIEW
                                                                        Oven Door
                                                                           Jamb
                           Seal Packing
                          Retainer
                            Bars

                              Steel Bearing Strip

                                          Nut
                                        (Welded)
                                        Adjusting
                                        Set Screw
                                       Approx. 9"
                                         Centers
                                                                         Lining
                                                                        Retainer
                                                                                           J. M. Asbestos
                                                                                        #103 Sheet Millboard

                                                                                           1/16" Gaskets
           FIGURE IV-7. INTERLAKE INCORPORATED DRAWING OF A PROPRIETARY DOOR
                       SEAL, I.E., A MODIFIED FIXED-EDGE SEAL

                       Adjusting set screws are in addition to the standard adjusting bolts.

-------
                                       IV-14

    Just prior to latching the door with the asbestos seal, sealant was troweled onto the in-
board seal surface in the area of the 3-mm (1/8-inch) depression. When the door was latched,
excess sealant was forced out sideways, indicating that the sealant had filled the gap area. On
the next coking cycle, no emission leakage was seen at the sealant location. At the end of the
coking cycle, the "used" sealant was easily removed from the door for examination. The
sealant had hardened in use and had changed color from gray to black. Microscopic examina-
tion indicated that the pores or gas passages in the sealant had been plugged with tars. The
portion of the sealant removed was about 3-mm (1/8-inch) thick indicating that the gap had
truly been filled with sealant on closing the door.
    In a second experiment, the localized sealant test was repeated and, in addition, sealant
was plastered on the seal at other locations where emission leakage was observed in the first
coking cycle. In this  experiment, there were no visible emissions from any location where
sealant had been applied. To test for possible gas leakage, a torch  flame was directed at the
seal area at 15-minute intervals for the first 2 hours of the coking cycle. No combustion could
be initiated in any area where sealant had been applied,  but other areas along the door did ig-
nite. No odor from the heated starch was detectable.
    It was concluded that these were significant experiments and that:
    (1) Sealants prevent emissions even after becoming  dry and porous.
    (2) The tar  from  the coal seals the pores in the luting material to the extent that it
       prevents gas leakage.
    (3) Additional  research  should  be  completed  in the  laboratory equipment  to
       evaluate the performance of sealants with a higher water content, i.e., sealants
       that are either more pumpable or can  be sprayed or flung into position on  the
       door.
    (4) Experiments should be completed to  determine to  what degree wet sealants
       thermally shock a heated cast-iron jamb; i.e., investigate whether the use of wet
       sealants could cause cracking of cast iron jambs.
    (5) The approach of pressing a door or jamb-mounted knife edge into a channel
       filled with sealant should be abandoned.
    (6) Emphasis should be placed on mechanized placement of sealant inboard of ex-
       isting seals on coke-oven doors; i.e., apply the sealant in the gas passage depth in
       the seal area. This approach is shown in Figure IV-8.
    Of the many original reservations that Battelle  researchers had about the use of sealants
on coke-oven doors, a major concern was the realistic position that some small percentage of
doors would (a) not be given the proper amount of sealant, or (b) the  application of sealant

-------
                                         IV-15
                                               With the door off the oven,
                                               apply the sealant in the
                                               indicated locations all around
                                               the door. Then latch the door.
                                             SEALANT
        Fixed-
        Edge Seal

    FIGURE IV-8. ILLUSTRATION OF THE CONCEPT OF "INBOARD LUTING", I.E., THE
                APPLICATION OF LUTING MATERIAL IN THE GAS PASSAGE
                OF EXISTING EQUIPMENT
 would be inadvertently "missed" on some doors. Without a backup-sealing system, any errors
 in sealant application would become painfully apparent  only after the coal had entered the
 oven, i.e., the emissions from a poorly luted door could be huge and would probably repre-
 sent a serious fire hazard in terms of probable fire damage to nearby end-closure elements.
 The concept  of using sealants inside the gas passage areas of existing seals gives the sealant
 sealing system a backup metal-to-metal sealing system and would eliminate one of the reser-
 vations we had about sealant systems.

                      Summary of the Follow-On Laboratory Tests
Variations in Water Content in Sealants
    A few experiments were conducted to learn the effect of water content on the unsup-
ported adherence of the most promising sealant mix when the mix was applied to a hot, ver-
tical steel surface. The mix consisted of 1 part of 1.68-mm (12-mesh) fireclay (Cedar Heights

-------
                                        IV-16

Clay Company), 1 part coke breeze, and 8 volume percent corn starch. Batches of about 1 liter
(about 1 quart) were dry mulled for 15 minutes and wet mulled for 10 minutes. With 37 volume
percent water, the consistency was equivalent to earlier mixes that had been prepared with a
similar clay (used by Koppers Company, St. Paul) and 28 volume percent water. As with the
earlier mixes, shrinkage of the new  mix was zero when it was dried in a 530 K (500 F) steel
channel.
    Some characteristics of the new mixes were as follows:
Water Content, v/o
37
43
49
54
58
KSPL*
Mix Consistency
Mod. soft putty
Soft putty
Stiff cream
Soft cream
Slimy
Slimy
Stickiness
None
Somewhat
Sticky
Very sticky
Very sticky
Very sticky
Slump
Tendency
None
None
None
Somewhat
Slumps
Slight
Surface
Water Sheen
None
Slight
Moderate
High
High
High
*Koppers Company, St. Paul loam consisting of clay, coke breeze, and coal.

    Samples of about 16 cm3 (1 in.3) were flung against a vertical steel surface heated to 420 K
(300 F). Above a water content of 49 percent, rebound of the mix was very high. The adhered
splats ranged in thickness from about 6 mm (1/4 inch) for the mix with 37 percent water to 2
mm (1/16 inch) for those above 40 percent water. All of the splats adhered for 6 minutes at
which time they were dry and  the test was terminated.
    When flung against the steel surface at 530 K (500 F), the results were as follows:
                                    Splat Thickness
Water Content, v/o
37
43
49
54
58
KSPL
mm
6.35
3.18
3.18
1.59
1.59
3.18
inch
1/4
1/8
1/8
1/16
1/16
1/8
Adherence, minute
1/4
1
1-1/2
7 Splat dry,
7 test terminated
7
    Application of a 13-mm-thick (1/2-inch-thick) layer to the steel plate at 530 K (500 F) was
attempted. The adherence of mixes with a water content of up to 54 percent ranged from nil to
about 1 minute. The mix with 58 percent water and Koppers'  loam separated from the steel
surface when the second thin layer was applied to build up the thickness. All of the mixes were
adherent when applied as a single 3-mm-thick  (1/8-inch-thick) layer. Good adherence per-
sisted up to 15 minutes at which time the test was terminated.
    The results of the above experiments suggest that the thinner mixes (high water content)
had better adherence to a vertical,  heated steel surface. Additional study is recommended.

-------
                                        IV-17
    Application of the mixes by flinging resulted in excessive rebound from the steel surface.
Rebound was greater and adherence was poorer when the mixes were flung against the steel
surface at the  higher temperature, 530 K (500 F). Building up a thick layer by flinging may
result in premature release of the initial layer. In the event that "thick" layers prove to be a
problem under production conditions, the gas passages can be "filled" to lower the thickness
(and material requirements) as shown in Figure IV-9. There are other options or alternatives
that can be explored.
    Adherence was good for all of the mixes when they were applied as a thick layer to the
steel surface heated to 420 K (300 F). Adherence of thick layers was poor when the mixes were
applied to the steel surface heated to 530 K (500 F). However, 3-mm (1/8-inch) layers were very
adherent. This suggests that the mixes may be applied as adherent thick layers to a steel surface
at 530 K (500 F)  if the mixes are pressed so that they retain intimate contact with the steel until
the mix interface loses its bulk water. Steam generated at the interface is suspected as the
cause for poor adherence. Additional research would be required prior to any demonstration
effort.
        Luting Material

      Added "Filler"^
      Piece      .  N
                         \
   Fir IIRF IV 9  EXISTING SEALS CAN BE EQUIPPED WITH A "FILLER" SECTION TO
   FIGURE IV-9. E*™l™™™             MATERIAL USED IN INBOARD LUTING

-------
                                         IV-18

    For a given dwell period on the hot steel surface, plasticity of our mix was higher than that
of Kopper's loam. Comparison with Koppers' loam showed the need and effectiveness of corn
starch in our mix in promoting prolonged plasticity when the mix was in contact with a hot
surface.

Effect of Water Content on Sealant Shrinkage
    The 1-liter (1-quart) batches which were prepared for evaluating the effect of water con-
tent on adherence to a heated, vertical steel surface, were also used to evaluate the effect on
shrinkage during drying.
    A steel channel at 530 K (500 F) was filled with the loam and pressed flat with a spatula. The
channel was 2.86 cm (11-1/8 inches) deep, 2.82 cm (1-7/64 inches) across the top, and 2.74 cm
(1-5/64 inches) across the bottom. Temporary steel end dams (coated with silicone spray;
removed after filling the channel) made  a cavity 3.18 cm (1-1/4 inches) long. The dry sealant
plugs were measured after a dwell of 3/4  hour in  the heated channel  with  the following
results:
Domi vvrain 01 Lmeu nug
(top swell)
Water Content, v/o
37
43
49
54
58
KSPL(al
BL(h)
cm
0
0.16
0.16
0.16
0.16
0.08
0
in.
0
1/16
1/16
1/16
1/16
1/32
0
Bottom
cm
2.70
2.70
2.70
2.70
2.70
2.66
2.70
in.
1-1/16
1-1/16
1-1/16
1-1/16
1-1/16
1-3/64
1-1/16
Top
cm
2.74
2.78
2.74
2.78
2.78
2.66
2.74
in.
1-5/64
1-3/32
1-5/64
1-3/32
1-3/32
1-3/64
1-5/64
Shrinkage, %
Bottom
1.5
1.5
1.5
1.5
1.5
2.9
1.5
Top
1.4
0.02
1.4
0.02
0.02
4.3
1.4
(a) Koppers Company, St. Paul loam consisting of clay, coke breeze, and coal.
(b) Bethlehem Steel Corporation loam.

Measurements were made to an accuracy of 0.04 cm (1/64 inch) which is about 1.5 percent of
the width of the plugs. Although there was a small amount of shrinkage, especially at the bot-
tom of the  plugs, the degree of shrinkage was not influenced by the water content of the mix-
es. Bethlehem's loam (BL) had shrinkage similar to our mixes. Koppers' loam (KSPL) definitely
had higher shrinkage than our mixes.

-------
                                         IV-19

    Subjective judgment of the characteristics of the mixes was as follows:
Water
Content, v/o
37
43
Af\
49
54
58
KSPL(i"
Dl 
bl_

Mix Consistency
Very stiff putty
Soft putty
f* * t f
Stiff cream
Soft cream
Slimy
Slimy
f** • f f
Stiff cream

Sag Tendency
None
None
None
Slight
Very
Moderate
None
Surf arc
fc<»^ii • Q^%,
Water Sheen
None
Slight
Moderate
High
Very high
Very high
Moderate

Stickiness
None
Moderate
Sticky
Very sticky
Very sticky
Very sticky
Moderate
                                           -----          • v I
(a) Koppers Company, St. Paul loam consisting of clay, coke breeze, and coal
(b) Bethlehem Steel Corporation loam.

Overall, the characteristics of KSPL most closely matched our mix containing 58 volume per-
cent water; BL most closely matched our mix containing 49 volume percent.

    Comments.  The  results  were  considered  encouraging  enough  to  complete  a
feasibility/costs/benefits analysis for Task IV. In any development/demonstration of inboard
luting, it will be  necessary to develop means of applying thick deposits of luting material.
Development and demonstration of inboard luting was not included in our original proposal
for a follow-on project.

Preliminary Evaluation of the Possibility of Cracking
Cast-Iron Jambs as a Result of Using Wet Sealants
    Gray cast iron  jambs are usually replaced only because of crack formation, in some in-
stances, after years (up to 20 years) of service. The number of jambs replaced  per year because
of cracking is a small percent of the total in service. Battelle researchers have examined jambs
that were partly cracked and noted that, in all cases examined, the crack was extending inward
(horizontally) starting from the outboard edge of the jamb web. The failure (cracking) of cast
iron is almost always in tension. Due to the temperature gradients at jambs, it  is to be expected
that the outer edge of the jamb web is normally under some tension.
    Because of the existing jamb-cracking problem, coke-plant superintendents are wary of
procedures that chill or thermally shock their cast iron jambs. In addition,  they are wary of
procedures that would appear to insulate the exposed portion of jambs. An example of major
chilling  of jambs  is  the rapid quenching that occurs during heavy rainstorms. A minor jamb
chilling  effect results from the practice of using water hoses to put out fires at  poorly fitting
coke-oven doors. An example that  includes both chilling and insulating might occur  upon

-------
                                        IV-20

compressing a wet sealant against warm jambs. The water contained in the sealant will cool the
surface of the jamb. After the sealant has dried, it would tend to insulate a part of the jamb
surface.
    As part of the sealant-testing task, it was decided to find out to what degree the applica-
tion of water-tempered sealant would change the stress level at the surface of the seal-mating
portion of jambs. To approach these conditions, sealant was rapidly applied to the top web of
the cross piece of the heated jamb in the physical model. This location was chosen because it
had the desired accessibility and the  required surface thermocouples and strain gages. The
data obtained indicated that in about 1 minute following application of the sealant, the jamb
surface temperature had  dropped to a low of 430 K (310 F) from the  original steady state
temperature of 520 K (470 F).  Following the period of flash cooling, the sealant dried slowly
and the jamb temperature returned to 520 K (470 F) in 30 minutes. The maximum temperature
change then was 90 K (160 F). Temperature-compensated strain gage readings taken over the
maximum temperature-differential period indicated that there had been a short-period in-
crease in tensile stress of 11.7 MPa (1700 psi). As the temperature of the jamb surface returned
to the original temperature,  the surface stress level also returned to the original level.
    A tensile stress spike  of about 11.7 MPa (1700 psi) for every application of wet sealant  is
considered to be a very minor increase in tension at the seal-mating surface. It was concluded
that this  periodic minor straining  should not result in the development of cracked jambs.
Stated another way, it is not expected that the application of wet sealant to the seal-mating sur-
face will  initiate cracks starting at this  location. This is a judgment conclusion  based upon the
reasoning that the seal-mating portion of the jamb is near the neutral plane of the jamb;  i.e.,
the seal-mating surface is normally  at a low  service-stress level. The  work of Kattus and
McPherson*  in  thermal shocking  mechanically loaded samples of metal indicated that the
number of cycles to cracking was a function of the level of the bending stress on the samples.
The  higher  the  bending load on steel and cast iron samples, the fewer  the number of
quenching cycles [700 to 300 K in 15  seconds (800 F to 75 F in 15 seconds)], and vice versa. At
low levels of bending stress, notched samples of cast iron would either not crack or would re-
quire a very large number of cycles to initiate cracking. All coke-oven jambs are hotter on the
back side. Under conditions  where these jambs are constrained (actual amount of inward jamb
bowing is less than the bowing that would occur naturally as a result of the thermal gradient),
the backside jamb surface is under maximum  compression and the exposed outboard web
surface is under maximum tension. The seal-mating surface is located between these two ex-
tremes and is near the cross-over point between tension and compression (the neutral plane)
*Kattus, j. R., and McPherson, B., "Properties of Cast Iron at Elevated Temperatures", ASTM STP No. 248.

-------
                                        IV-21

and it is, therefore, only under some low service stress (mild compression or mild tension). It is
expected that mild chilling of a seal-mating surface that is not at a high stress level will not
result in a crack formation.
    If the coke-producing industry accepts the concept of inboard luting and proceeds with a
program of development and implementation; it may be advisable to measure the service
stress level on operating jambs to  verify our judgmental conclusion.
    It is probable that the implementation of inboard luting will raise the average temperature
of jambs, but by  only a small amount. The portion of the jamb (underneath the seal) that
would be insulated by the dried luting material is not now in a good position to lose heat. The
degree of temperature rise is, therefore, expected to be small, but should be determined and
evaluated in any demonstration project.

                      Overall Summary of the Laboratory Research
                                   and Field Tests
    The results of  the laboratory research and field tests lead to the following overview:
     (1) The results of all of the laboratory and field work are regarded as being favorable
        and in line with our objectives. Our interpretation is  that inboard luting has the
        potential  of being developed into an alternative sealing method. However, the
        work  that was completed can only be considered  as "scoping research" in-
        asmuch as various  details that could  improve/optimize the results were  not
        evaluated. Further work was held in abeyance because (a) we had reached the
        end of the funds allotted to this part of the project and (b) we need to learn the
        judgments of our Sponsors (EPA/AISI). Additional optimizing research/develop-
        ment in sealant characteristics would be required if any organization elects to
        carry on with this approach. As one general example, clay from  only one deposit
        was evaluated, and no attempt  was  made to control  the mesh size of the
        materials  used  in preparing  the mixtures. It is probable that the introduction of
        some portion of finely ground coke breeze would lower the required amount of
        water.
     (2) In the approach being recommended, the sealant would be applied (mounded)
        into the gas passage area between the existing metal  seal and the door plug. This
        would, in most instances result in a wider cross section of sealant than the 25-mm
        (1-inch) width that was tested in the field. It is judged that the emission tightness
        exhibited by a  25-mm-wide  (1-inch-wide) layer of sealant containing 33 volume
        percent water  will  be duplicated by a wider cross  section  of sealant having  a

-------
                                       IV-22

      higher water content. The application of sealant into the gas passage of the door
      rather than application on the jamb is expected to have advantages because this
      area is normally cooler than the jamb, especially after the door has been off the
      oven for a normal period.
   (3) The  application of sealant  onto the heated jamb of the physical  model of the
      door/jamb indicated that the tension in the surface  layer of the jamb in the
      sealant area was momentarily increased by about 11.7 MPa (1700 psi). Overall,
      the contact of the wet and slowly drying sealant to a heated jamb is a mild cool-
      ing effect as compared with a rainstorm. On the other hand, the application of
      the sealant could be presented as a mild and short sprinkle  of rain occurring
      every 16 to 24 hours. It is judged that the use of sealant sealing will cause crack-
      ing only if the jamb surface is already at a high level of tension. On some jambs, it
      is estimated that there can be high tension only on the outboard edge of the web
      of the. jamb. The sealant-mating portion of the jamb, however, is closer to what
      would  normally be the compression side of the jamb. From this general evalua-
      tion, we expect that the use of a sealant will not cause jambs to crack. This is not,
      however, a warranty; additional research would be advisable.
   (4) Overall, it was judged that sufficient progress has  been made to complete a
      feasibility/costs/benefits analysis  on sealants. In this regard it should be noted
      that  the inboard luting approach has the potential of eliminating the need for
      manual cleaning of both the jamb and the gas passage on doors.

                                  Feasibility Analysis
    In the  Work Plan dealing with this project, it was stated  that a feasibility/costs/benefits
analysis of sealant procedures would be completed if there had been demonstrable progress
in this  research  project. We feel  that  this progress has  been  demonstrated short  of
development/demonstration efforts at coke plants. This portion of the report deals with  an
analysis of the overall potential of use of sealants.
    As it stands now,  it is judged that the approach is technically feasible. It has a basic attrac-
tion in that it appears that it could be implemented rapidly. Emission control on existing older
ovens (with the use of sealants within the existing metal seals) would be significantly better
than continued operation with the existing metal seals.  Short of a demonstration project, the
pertinent questions at this time are:
    (1) Will the sealant  method give better emission control (at end closures) than the
       upgraded metal seals that are being developed?

-------
                                         IV-23


    (2) What are the relative costs of sealant sealing and upgraded metal seals?

    (3) For what period of time will upgraded metal seals give  acceptable emission-
       control performance?

    (4) Are there coke plants where the sealant approach might be readily accepted (and
       developed and evaluated) because of special circumstances?

    Definitive answers  to  Questions 1,  2, and 3  require the testing of both approaches
(sealants and improved metal seals) in comparable situations and the development of a con-

sensus decision based on overall merit. On the other hand, one evaluation approach that can

be taken is to assume that a sealant will match upgraded metal seals in emission-control per-
formance. This approach shifts the emphasis to developing an answer to Question 2. An

attempt was made to answer this question in a definitive way, but we encountered an  un-
known or a yet-to-be-established cost element. This cost element is whether or not there is a
need to employ additional workers just to apply sealants on doors.


    As background to the unknown cost element, it may be helpful to know that every coke
battery has two working sides and has at least one "benchman" worker per side of the battery,
per shift. There are at least 2 benchmen per battery per shift, or a total of 8 for the 4-turn  (24-
hour per day) coke battery operation. The duties of these individual workers are listed below

(with comments):
           Duties                                    Comments

(1) Shovel coke spillage and clean  (1)  Some batteries have a large amount of spillage and
   the bench.                       others do not.
(2) Manually clean the jamb      (2)  Where batteries do not have jamb-cleaning machinery
   while the door is off the oven.      (and this is all batteries less than 5 meters tall), this is a
                                    difficult task. Performance of this task varies from "not
                                    done" to a "reasonable amount of effort". Where the
                                    work load (defined later) is heavy at a particular
                                    battery, this task is often omitted or neglected.

(3) Manually remove at least      (3)  This task has many of the difficulties (and worker
   some of the tar and carbon         resistance) associated with jamb cleaning. In many
   that collected in the gas-          instances, this work is done only periodically (once a
   passage area on the door.          week). This is acceptable from the sealing-
                                    performance viewpoint. So long as carbon is not
                                    allowed to build up to the point where it interferes
                                    with the door fitting into the oven or interfering with
                                    the seal mating with the jamb, emission control is not
                                    affected. However, allowing the buildup of tar (which
                                    converts to adhered carbon) only makes the cleaning
                                    job more difficult when the carbon must be removed.
                                    Some companies are developing periodic high-
                                    pressure, water-jet cleaning of the gas-passage area.

-------
                                        IV-24
             Duties                             Comments
(4) Chip out the hard carbon      (4) The rate of carbon buildup differs among plants.
   that builds up on the oven sill      Need for chipping can be "seldom".
   below the door.
(5) Assist in bottom latching of     (5) The need for this assisting duty depends on the condi-
   doors when problems occur.       tion and age of the equipment.
    At the 6-meter batteries, machinery has been installed to eliminate Duties 2 and 3 above.
At some locations, the benchman is still needed to "touch up" the cleaning of the jamb when
the machinery is awaiting adjustments.
    In the recommended approach of applying sealant inboard of the existing  seals, it is ex-
pected (but not proven) that one application per cycle will eliminate the need for the arduous
duties of manually cleaning the jamb and door (Duties 2 and 3 in the above listing). Because
sealant cannot be applied manually (as is done in outboard luting), it is assumed that this
operation will be at least partly mechanized. In one approach, the worker would coat the door
areas  by using a version of a pneumatic spray gun. It would be necessary to install an elevator
or stairs arrangement that would allow the worker to approach his work areas near the top of
the doors.
    At least in theory, it would seem appropriate to substitute/exchange an equipment-
supported, inboard-luting duty for the two manual  (and difficult) cleaning duties. Theory may
be appropriate at some coke plants where there are relatively few ovens per battery and where
the coking rate is low. On a 50-oven battery making 24-hour coke, the number of door
openings per side is an average of 17 per shift, or about a half hour apart. On the other hand, at
a 90-oven  battery making 16-hour coke, the  average number of door openings per side is
about 45 per shift or about one every 10 minutes. In most instances, the door-removal rate is
lower than the average time (i.e., faster operation) because of  rest breaks and the need to
"pick  up" the time  lost in operating delays. There  is,  therefore, a rather wide range of
working-rate or work load  for bench workers at coke batteries. As understood by  Battelle
researchers,  the work load on the bench is the subject of negotiation between the company
and the local union  representatives. Coke-plant supervision and personnel did not feel that
they were in a  position to predict the outcome of  negotiations dealing with the possible in-
troduction of sealants.
    As might be expected, the cost of  using sealants will  be the highest where additional
workers are  required. Conversely,  the lowest cost will exist where an agreement is reached
that the existing benchmen would apply the sealant in exchange for eliminating the duties of
manual cleaning of jambs and doors. Because all upgraded sealing approaches represent an
increase in cost, the most acceptable emission-control solution  will be that  with the highest

-------
                                         IV-25

relationship of performance to cost. The cost aspects of the use of sealants are considered in
the following section.

Cost of Sealant Approach Including
Additional Benchmen

    In some instances, it may be necessary to consider the addition of a benchman to apply
sealants. A cost approximation for this situation is as follows:

    Assumptions

    (1) Adding a helper benchman represents adding 8 additional workers per battery.
       Battery operation is a 4-shift-per-week system and there are two sides to a battery.
       It is judged that an additional day-turn man would be required to prepare the
       sealant. The total increase in crew size would then be 9 workers per battery.
    (2) Cost per worker added will average about $20,000 per year, including all employ-
       ment costs.  Lutermen sealing  doors with  outboard luting (slow coking)  are
       presently paid about $16,000  per year (gross), including incentive pay. Employ-
       ment costs approach 30 percent or more. The additional labor cost for inboard
       luting workers would be 9 X $20,000, or $180,000 per year.
    (3) In late 1976, the selling price of a complete mixing, pumping, and pneumatic
       spraying package was $15,000. This unit will deliver 0.85 cubic meters per hour (30
       cubic feet per hour). A heavy  application of sealant at any one door could take as
       little as 20 seconds of time depending upon positioning of the worker. Two spray-
       ing systems are required per  battery. Including installation and winterizing, the
       installed  cost for both units would total about $60,000 or more. The cost of equip-
       ment or equipment modifications to allow the worker to reach the upper parts of
       the door (while it is off the oven) may cost $40,000 installed. Total installed cost is
       taken as $100,000.
    (4) The  coke  battery  being  considered  has  50  ovens  and  produces  290,000
       tonnes(320,000 net tons) of coke in a high-demand year.
    (5) The cost of the sealant material is about 10 cents per door closing, or about $5,000
       per year on the above battery.
    (6) Door-seal maintenance costs  are lowered by $20,000 per year. Coke-producing
       companies report  a savings in  maintenance costs  and possibly an increase in
       door-plug life in comparing batteries having luted doors with those having metal

-------
                                     IV-26

    seals. These are, however, only generalizations inasmuch as no attempt has been
    made by these companies to pinpoint the actual amount of savings. The above
    figure is, therefore, a guess—it might be on the low side or it could be on the high
    side.
 (7) The after-tax cost of capital is  taken as 10 percent. This is a variable depending
    somewhat on the profitability of the company and the  industry and may, for this
    industry, be a low figure.
 (8) The depreciation life of the new equipment (installation costs included) is taken
    as 10 years with no salvage expected. Straight-line depreciation is used.
 (9) For purposes of  illustration only a simple approach is used. No consideration,
    therefore, was given to (a) rate of inflation, (b) time value of money*, or (c) the
    effect of decreases in production  rate during low-demand  periods.

 Case A: Equipment Installed, 9 Additional Workers Hired

 Yearly Expenses:
 (a) Increased labor costs (including cost of employment)   '               $180,000
 (b) Sealant cost                                                          5,500
 (c) Cost of capital (10 percent of $100,000)                                  10,000
 (d) Depreciation expense (10 percent of $100,000)                          10,000
                                                                      $205,500
    Savings on metal-seal maintenance                                   $30,000
    Estimated net cost per year                                          $175,500
    Estimated cost per ton of coke                             61
-------
                                         IV-27

    Case B: Equipment Installed, One Additional Worker
            Hired to Prepare Sealant
    Yearly Expenses:
    (a) Increased labor cost                                                $20,000
    (b) Sealant cost                                                          5,500
    (c) Cost of capital                                                       10,000
    (d) Depreciation expense                                                10,000
                                                                          $45,500
       Savings on metal-seal maintenance                                   $30,000
       Estimated net cost per year                                          $15,500
       Estimated cost per ton of coke                            5.5c/tonne of coke
                                                                  (5
-------
                                       IV-28

    It appears that the use of a sealant could keep warped end closures operating under good
emission control. In addition, it is conjectured that:
    (a) Pneumatic luting of ovens that are  presently being manually luted could improve
       their emission-control performance, and
    (b) The pumping equipment recommended for this overall approach probably can be
       modified or augmented to become a superior (more labor-efficient) equipment for
       use in patching behind jambs. In this new application, refractory mortars or insulating
       mortars  could be substituted for the low-cost mixtures recommended for inboard
       and outboard luting.

Conclusions
    The following conclusions have been drawn:
    (1) Acceptance of the use  of sealants (if this method can be developed and implemented)
       will depend upon whether it is possible to introduce  the use of sealant applying
       equipment without materially increasing the number of workers on the bench.
    (2) There are thought to be coke plants operating under special circumstances such that
       the plants would have a special need for the use of sealants.

                                 Recommendations
    The following recommendations are made:
    (1) The Sponsors should  give  consideration to the  development, testing, and
       demonstration of inboard luting. The coke-producing industry  (members of the
       AISI and  companies belonging to other associations) should be polled to deter-
       mine which companies are interested in the inboard luting approach.
    (2) Individual companies should not attempt a trial-and-error approach to inboard
       luting without the support of additional laboratory work. The interested com-
       panies should pool  their  resources to give inboard luting a strong, technically
       oriented  development effort. This effort should include  (a) testing to determine
       whether  compressing wet sealants against heated jambs  will or will not result in
       cracking  of gray iron jambs, and (b) development of sealants and application
       methods  that will result in the adhered thickness required.

-------
                         	IV-29	

                                TECHNICAL REPORT DATA
                         (Please read Instructions on the reverse before completing)
1. REPORT NO.
 EPA-600/2-78-189
2.
                           3. RECIPIENT'S ACCESSION-NO.
  ITLEANDSUBTITLE Development and Demonstration of
Concepts for Improving Coke-oven Door Seals:
Interim Report
                            . REPORT DATE
                            August 1978
                            , PERFORMING ORGANIZATION CODE
|7.AUTHOR(S)

A.O. Hoffman, A. T. Hopper, and R. L. Paul
                            8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Battelle Columbus Laboratories
 505 King Avenue
 Columbus, Ohio 43201
                            10. PROGRAM ELEMENT NO.
                            1AB604C
                            11. CONTRACT/GRANT NO.

                            68-02-2173'
 12.SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC 27711
                            13. TYPE OF REP.ORT AND PERIOD COVERED
                            Interim; 8/76-7/78	
                            14. SPONSORING AGENCY CODE
                             EPA/600/13
 is.SUPPLEMENTARY NOTES IERL-RTP project officer is Robert C.  McCrillis, Mail Drop 62,
 919/541-2733.
 16-ABSTRACTThe report gives pre-engineering analyses, evaluations, and recommenda-
 tions in an ongoing research project dealing with the development of a retrofittable
 concept for minimizing emissions from door seals on coke ovens.  It includes evalua-
 tions drawn from tasks dealing with mathematical and physical modeling, and from a
 task dealing with field-data collection and field experiments. Based on these results,
 the recommended metal-to-metal sealing system includes: a simplified-shape seal,
 a new improved procedure for mounting and adjusting seals, and high-strength heat-
 resistant materials.  The recommendations were approved by the sponsors, and
 engineering tasks are in progress.  It is recommended that as new door jambs are
 installed, they be allowed to assume their natural curvature resulting from operating
 temperature gradients, while simultaneously restricted to prevent hourglassing,
 and that they be ferritic ductile iron castings.  Limited experiments with luting com-
 pounds were encouraging. If developed further, this might be an attractive alternative
 particularly for older batteries.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                          b.lDENTIFIERS/OPEN ENDED TERMS
                                         c.  COSATI Field/Group
 Pollution
 Iron and Steel Industry
 Coking
 (Doors
  eals
 Packings
                Pollution Control
                Stationary Sources
                Coke Ovens
                Door Seals
                Luting Compounds
13B
11F
13H
13M
11A
  DISTRIBUTION STATEMEN1


 Unlimited


   Form 2220-1 <9'73)
                Unclassified
                                                        - (This Report)
21. NO. OF PAGES

   120
                20. SECURITY CLASS (Thispage)
                Unclassified
                                         22. PRICE

-------