g 2
o o
SI
o«
*• F
£z
United States United States EPA/600/9-91/030
Environmental Protection Department of the July 1981
Agency Interior
v>EPA Mine Waste Disposal
Technology
Proceedings: Bureau of
Mines Technology Transfer
Workshop, Denver, Colo.,
July 16, 1981
£ 3 £
I— (0 (0
-------
D
EPV600/9-91/030
July 1981
Information Circular 8857
Mine Waste Disposal Technology
Proceedings: Bureau of Mines
Technology Transfer Workshop,
Denver, Colo., July 16, 1981
Compiled by Staff—Minerals Research
U S Environmental Protection Agency
Region 5, Library (PL-12J)
11 West Jackson Boulevard, 12tn rioor
Chicago, IL 60604-3590
UNITED STATES DEPARTMENT OF THE INTERIOR
James <3. Watt, Secretary
BUREAU OF MINES <§> Printed on Recycled Paper
-------
This publication has been cataloged as follows:
Bureau of Mines Technology Transfer Workshop (1981 :
Denver, Colo.)
Mine waste disposal technology proceedings.
(Information circular ; 8857)
Includes bibliographies.,
Supt. of Docs, no.: I 28.27:8857,
1. Mines and mineral resources—Waste disposal—Congresses. 2. Coal
mine waste—Congresses. I. United States. Bureau of Mines. II. Title,
III. Series: Information circular (United States. Bureau of Mines) ; 8857.
TN295.U4 [TD899.M5] 622s [622'.2j 81-607857 AACR2
-------
PREFACE
This Information Circular summarizes recent Bureau of Mines research results con-
cerning mine waste disposal. These papers represent only a sample of the Bureau's efforts
related to the embankment stability and environmental impacts associated with mine waste
disposal, but they delineate several of the principal areas of this program. The seven
technical presentations reproduced herein were given by Bureau and contractor personnel
at the July 16,1981, Technology Transfer Workshop on Mine Waste Disposal Technology
in Denver, Colo.
Those desiring more information concerning the Bureau's Mine Waste Management
program, Minerals Environmental Technology in general, or information on other specific
research should feel free to contact the Bureau of Mines, Minerals Research Directorate,
2401 E Street, N.W., Washington, D.C. 20241.
-------
Ill
CONTENTS
Page
Preface i
Abstract 1
Bureau of Mines Research in Mine Waste Disposal Technology, by Roger A. Bloom-
field and Richard J. Seibel 2
Disposal of Coal Mine Waste in Active Underground Coal Mines, by Leslie S. Rubin,
Mackenzie Burnett, Al Amundson, Gary J. Colaizzi, and Ralph H. Whaite 8
Factor-of-Safety Charts for Estimating the Stability of Saturated and Unsaturated
Tailings Pond Embankments, by D. R. Tesarik and P. C. McWilliams 21
Probabilistic Approach to the Factor of Safety for Embankment Slope Stability, by P.
C. McWilliams and D. R. Tesarik..... 35
Application of Remote Sensing for Coal Waste Embankment Monitoring, by C. M. K.
Boldt and B. J. Scheibner. 40
Summary of Research on Case Histories of Flow Failures of Mine Tailings Impound-
ments, by P. C. Lucia, J. M. Duncan, and H. B. Seed 46
Summary of Research on Analyses of Flow Failures of Mine Tailings Impoundments,
by J. K. Jeyapalan, J. M. Duncan, and H. B. Seed 54
Controlled Burnout of Fires in Abandoned Coal Mines and Waste Banks by In Situ
Combustion, by Robert F. Chaiken 62
-------
MINE WASTE DISPOSAL TECHNOLOGY
Proceedings: Bureau of Mines Technology Transfer Workshop,
Denver, Colo. July 16,1981
COMPILED BY STAFF—MINERALS RESEARCH
ABSTRACT
This Bureau of Mines publication consists of an overview of the mine waste management
research currently being conducted by the Bureau. The following papers, given at a Tech-
nology Transfer Workshop, emphasize the increasing importance of research related to
the safety and environmental considerations of mine waste disposal in recent years. Current
work related to legislation passed in the last 10 years and their subsequent standards
includes the development of adequate monitoring systems, development of stability and
seepage prediction and control techniques, control of runoff water, and control of leaching
solutions. Selected topics are included here that cover coal mine waste disposal, em-
bankment slope stability monitoring, flow failures of mine tailings impoundments, and
controlled burnout of abandoned coal mine fires. The projects described provide a current
documentation of problems being addressed.
-------
BUREAU OF MINES RESEARCH IN MINE WASTE
DISPOSAL TECHNOLOGY
by
Roger A. Bloomfield1 and Richard J. Seibet2
INTRODUCTION
Waste generated by the mining and milling industries now
total about 2.3 billion tons annually. Most of this waste is
characterized as overburden and tailings or refuse. Although
solid wastes are common to all mining and milling operations,
the technology for handling and disposing in each case may
vary somewhat depending on the specific characteristics of
the waste. For example, if the waste contains a high per-
centage of clay minerals, such as phosphate slimes, the
equilibrium water content may approach 90 pet. If the waste
contains sufficient pyrite, common to coal waste, the runoff
and leachate can be expected to be acidic. Combustion is a
problem sometimes encountered when coal waste is dis-
posed of in a dry condition. Lead-zinc tailings generally av-
erage about 0.3 pet lead and 0.3 pet zinc which may be
soluble in a low pH environment. Where a relatively coarse
waste is produced in conjunction with an underground mine,
backfilling is often considered advantageous. Research de-
scribed herein is aimed at alleviating the structural, environ-
mental, and economic problems associated with the disposal
of all mining and milling wastes.
The need for research on solid mining wastes related to
the structural stability of impoundments can be readily dem-
onstrated by the following examples. In 1966, a coal waste
pile in Aberfan, Wales, failed and devastated a school building
leaving 144 children dead. In 1970, 16 million cubic feet of
saturated tailings and slimes failed and flowed into the lower
levels of the Mufalire mine in Africa, causing extensive dam-
age and the loss of 89 lives. A coal refuse embankment at
Buffalo Creek in West Virginia failed in 1972 sending 21
million cubic feet of water and sludge downstream, leaving
over 100 fatalities, 1,100 injuries, 1,500 houses demolished,
and 4,000 people homeless. Several tailings embankments
have failed in the western United States over the past decade
but have received little publicity since no fatalities occurred?
however, the environmental consequences have been severe
in some cases. Current research related to safe mine waste
disposal is being conducted in the areas of engineering prop-
1 Program manager
2 Chief, Branch of Mine Waste Management Both authors are with the Division
of Minerals Environmental Technology, Bureau of Mines. Washington, D C
erty characterization, waste handling and placement sys-
tems, and stability control and analysis techniques
Research related to the safety and environmental impacts
of mine waste management has become increasingly im-
portant since over 10 Federal Acts have been passed in
recent years, most notably the Federal Water Pollution Con-
trol Act Amendments of 1972, the Resource Conservation
and Recovery Act of 1976. the Surface Mining Control and
Reclamation Act of 1977, and the Federal Mine Safety and
Health Amendments Act of 1977 Current work related to the
Acts and their subsequent standards include the develop-
ment of adequate monitoring systems, development of seep-
age prediction and control techniques, control of runoff water,
and control of leaching solutions
Bureau of Mines research in the mine waste disposal area
is conducted in two Divisions, the Division of Minerals Health
and Safety Technology and the Division of Minerals Envi-
ronmental Technology. The objectives of these programs along
with project descriptions are discussed in the following sec-
tions.
Background
The Bureau of Mines has been active in mine waste dis-
posal research for the past 20 years. Experience gained on
committees such as the ASCE-NSF workshop on Research
Needs for Mining and Industrial Solid Waste Disposal. DOI
Task Force on Coal Waste Hazards, West Virginia Ad Hoc
Commission on Inquiry Into the Buffalo Creek Flood. Inter-
agency Task Force on Reserve Mining Taconite Onland Dis-
posal, NRC-lndustry Uranium Waste Disposal Committee.
Surface Mining Legislative Task Force, DOI Dam Safety
Committee, Industry-Government Working Group for Re-
search on Manganese Nodule Process Rejects, and the EPA
Mining Solid Waste Coordinating Committee provides the
background for a dynamic research program responsive to
both Federal and industrial needs. Over the past decade.
Bureau personnel have been highly active in both national
and international symposia, workshops, and short courses.
-------
Waste disposal practices in both the coal and metal mining
industries have undergone considerable change during the
past 20 years as a result of legislative pressures and changes
in refuse characteristics. For a long time, lump coal was at
a premium for domestic use, and power plants did not use
the fines. The coarse washery waste was dumped onto large
piles and the wash water was discharged into streams. In
response to various state laws in the 1950's and 1960's,
operators began pumping wash water up the hollows behind
the waste piles where, after seeping through the coarse waste,
the water was sufficiently clarified for discharge into the streams.
This solved one problem but created another—saturated,
unstable waste impoundments. Such a situation was re-
sponsible for the Buffalo Creek disaster. This prompted leg-
islative action requiring proper design of waste impoundments
to ensure stability. Meanwhile, the coal industry was undergo-
ing another subtle change. Because steam plants could now
handle fines, fine grinding with flotation was introduced to
promote recovery, which produced a larger percentage of
fine waste and resulted in more impounding structures.
The history of metal mines tailings disposal is similar. Early-
day milling required only coarse crushing followed by gravity
separation on tables or in jigs. Now, with flotation and low-
grade feed material, many mills grind to 60 pet minus 200
mesh and finer. Like the coal industry, many mill operators,
until recently, dumped the tailings into the nearest stream.
This is no longer acceptable, and active tailing ponds now
cover from as little as 1 acre to as many as 5,000 acres,
attaining heights of 300 to 400 feet with projected heights
approaching 500 feet. In general, metal mine tailings ponds
have been better engineered than coal embankments, but
some of these have also failed. As dams become larger, the
need increases for better engineering and operating proce-
dures.
The Bureau of Mines began investigating the properties of
mill tailings hydraulically emplaced in underground mine
openings as early as 1955. The primary reason at that time
for returning the waste material underground Was to provide
backfill for ground support—elimination of surface disposal
was a secondary benefit. In the 1960's, the Bureau began
research on stabilizing mine-waste piles and on eliminating
burning coal-refuse dumps. Since then there has been con-
tinuing emphasis on the solution of waste-disposal problems.
The Bureau has over 100 reports and other publications on
this subject from in-house and contract research.
Objectives
The general overall goals of the Bureau's mine waste dis-
posal research are to (1) define and assess the major struc-
tural stability and environmental problem areas associated
with the disposal of mine and mill wastes for various com-
modities, (2) design and develop control techniques address-
ing these problems and promote their incorporation in industry
practice, and (3) develop alternative disposal practices pro-
moting effective land use and waste utilization.
The following is a description of the specific objectives of
the Bureau's mine waste disposal research.
• Compile a complete inventory of waste disposal sites
and develop an updating system.
• Collect environmental data around several large tailing
ponds and continue long-term monitoring.
• Determine engineering properties, compaction char-
acteristics, and design criteria for waste disposal for
various metal, nonmetal, and coal commodities.
• Develop and demonstrate underground disposal sys-
tems for coal and metal mines.
• Develop alternative techniques for innovative surface
disposal systems and waste utilization methods.
• Develop, test, and monitor control methods to mini-
mize seepage through waste ponds.
• Develop rapid in situ and remote monitoring tech-
niques for determining embankment stability.
• Develop coal waste dewatering systems to improve
the stability of fine coal refuse.
• Develop safety and risk analysis methods for evalu-
ating waste embankment stability.
Implementation
As mentioned previously, the Bureau's research in this area
is carried out in two Divisions. Research conducted in the
Division of Minerals Health and Safety Technology empha-
sizes matters related to embankment stability. The program
is closely coordinated with the Mine Safety and Health Admin-
istration (Department of Labor), primarily through formal joint
ranking sessions where a project prioritization scheme is es-
tablished. Waste management research conducted in the
Division of Minerals Environmental Technology is aimed at
minimizing the environmental effects of mine waste disposal.
Since this is somewhat broad in nature, some stability work
is also performed in this program due to the close relationship
of structural and environmental problem areas. Work in this
program is closely coordinated with other Federal agencies—
the Office of Surface Mining (OSM), Environmental Protection
Agency, and Geological Survey. Joint project ranking ses-
sions are conducted with OSM.
Project Descriptions
The following is a description of mine waste research pro-
jects in the Divisions of Health and Safety and Environmental
Technology listed by in-house and contract categories.
Minerals Environmental Technology, Solid
Waste Management
IN-HOUSE RESEARCH
1. Control of Hazardous Alkalies From Cement Kiln Dust.
Objective: Eliminate an environmental problem in ce-
ment production caused by the alkali and the alkali-free
materials. Presently, 6 to 10 million tons per year of kiln
dust are discarded to dumps or land fills because alkali
renders the dust unsuitable for cement. The discarded
-------
dust is a hazardous waste because the alkali can leach
out and contaminate ground water
2. Characterization of Potential Manganese Nodule Proc-
essing Rejects.
Objective: Verify, improve, and update the information
available on manganese nodules; clarify some of the
terminology used in identification and characterization of
nodules and their processing reject materials; and pre-
pare representative manganese nodule reject materials
for future study. This research is conducted under an
interagency agreement with the National Oceanic and
Atmospheric Administration.
3. Tailings Pond Research and Testing Facility.
Objective: Continue pond liner testing program on sam-
ples installed in industrial ponds. Remove and test liner
samples for properties such as permeability and tear
strength. Install several new liners in operating ponds.
Test dewatenng techniques such as sand bed drainage
and capillary assisted evaporation.
4. Controlled Burnout of Coal Waste Bank Fires at the Moss
No. 1 Plant.
Objective: Assist in technology transfer of the concept
of controlled burnout of a fire in a coal waste bank utilizing
Bureau of Mines developed in situ combustion meth-
odology.
5. Controlled Burnout of Fire at Abandoned Coal Mines-
Calamity Hollow.
Objective: Demonstrate the concept of controlled burn-
out of fire in an abandoned coal mine by utilizing Bureau
of Mines methodology for in situ combustion to deplete
the fuel responsible for the fire in an environmentally
sound manner and utilize the heat produced.
6. R&D Support for Burnout Control of Fires on Abandoned
Coal Mine Lands.
Objective: Develop analytic models for controlled burn-
out, improve methods of high temperature borehole con-
structions and instrumentation for monitoring burnout
control, and evaluate burnout control for coal wastes.
The latter two items will involve activities in the Bureau's
surface trench burn facility
7. Abating Pollution From Lead Tailings Piles in Southeast
Missouri.
Objective: Examine lead mine tailings ponds in the
southeast Missouri lead belt to characterize the nature
and magnitude of compositional and physical makeup of
contaminating constituents and assess potential meth-
ods of rendering these tailings environmentally safe to
regional surface and ground water systems.
8. Pollutant Identification in Auto Shredder Offal Material.
Objective: Select samples of fluff and other residues that
are commonly disposed of in landfills from a number of
shredders throughout the country. Subject samples to
chemical analysis and handpicking to determine their
chemical and physical makeup. Leach samples accord-
ing to proposed EPA standards and analyze the leachate
to determine its composition.
9. Physical and Chemical Characteristics of Manganese
Nodule Processing Wastes
Objective: Obtain samples of waste from the Avondale
Research Center Perform laboratory tests to determine
the following materials properties: (1) Material compo-
sition-mineral and chemical, (2) permeability, (3) settling
rates, (4) size distribution, (5) specific gravity, (6) dens-
ities, (7) electrical conductivity, and (8) strengths
10. Improved Hydraulic Method for Coal Waste Disposal
Objective: Provide an improved method of waste dis-
posal by providing economical slurry transport and a sta-
ble blended waste. Present methods of handling coal
refuse include sludge disposal ponds for fine coal refuse,
which are hazardous because of the hydraulic properties
of the material, and coarse waste dumps that tend to
ignite because of porous nature of the embankments
Recently, some coal cleaning plants have installed de-
watering devices for the fine refuse and combine the
coarse and dewatered sludge prior to disposal. The pro-
cedure is extremely expensive, produces a "bulked"
material for disposal that has poor long-term stability
characteristics, and presents materials handling prob-
lems.
11. Development and Evaluation of Waste Disposal Plans
and Techniques.
Objective: Identify major waste problems through meet-
ings with industry representatives. Maintain industry con-
tacts and perform background and preliminary research
in the development of new research projects
12. Wastewater Monitoring Techniques for Uranium Tail-
ings.
Objective: Simplify and improve data gathering tech-
niques to improve the state of the art of waste water
monitoring and provide a basis on which realistic regu-
lations may be promulgated. In this case the benefit would
apply to industry as well as to applicable regulatory
agencies. Seepage from uranium tailings has been a
quite controversial subject among regulating agencies
and industry. Certain proposed regulations could have a
drastic effect on the economics of uranium mining. Many
uncertainties are due to the many unknowns in water
quality, quantity, and movement in the disposal vicinity.
13. Design of Slime-Sealed Impoundments to Prevent
Ground-Water Contamination.
Objective: Use mill tailings or natural soil rather than
fabric liners to reduce seepage in tailings impoundments.
Determine potential methods for reducing the seepage
beneath tailings ponds through laboratory permeability
testing of mill tailings and natural soil within a tailings
area in combination with additives and compaction tech-
niques. Transfer results to field tests where the actual
seepage for a given area will be measured.
14. Survey Foundry Wastes in Southeastern United States
to Identify Environmental Hazards.
Objective: Initiate a systematic program to identify the
various types of waste materials generated by foundries
in the southeastern United States, determine whether
-------
the wastes currently or potentially pose environmental
hazards, and recommend research to alleviate any prob-
lems.
15. Assessment of Environmental Impacts Associated With
Byproduct Gypsum Stacks From Florida Phosphates.
Objective: Determine the extent and nature of radiation-
related and other environmental problems associated with
the disposal and subsequent leaching of byproduct gyp-
sum, and recommend and initiate research to alleviate
the problems.
16, Borehole Mining Cavities for Disposal of Radioactive
Mine Wastes.
Objective: Investigate aspects of rock mechanics, en-
vironmental engineering, safety engineering, economics,
and nuclear energy regulation relative to adaptation of
the Bureau borehole mining technology to the construc-
tion of storage cavities for solid wastes generated in
uranium in situ leaching operations.
Contract Research
1. Containment Pond Liner Materials Testing.
Objective: Determine performance of typical pond line'r
materials under actual industrial service. Install liners in
operating ponds and test them for material properties
such as permeability and tear strength at various time
intervals.
2. Demonstration and Evaluation of Underground Disposal
of Coal Mining Wastes.
Objective: Conduct a 1 -year demonstration of a system
of underground disposal of coal mining wastes in an
active underground coal mine. Provide an assessment
of technical problems and economics of implementation
of a total system of underground disposal of coal mining
waste in an active underground coal mine.
3. Design and Engineering of a Burnout Control System
for a Coal Waste Bank.
Objective: Utilize the controlled acceleration of a fire by
air injection so that complete fuel burnout is accom-
plished in a time period that is short compared to the
many decades that such fires could continue if left un-
attended. Inject air into and remove combustion products
from the waste bank by a surface exhaust ventilation
system operating through boreholes extending into the
bank.
4. Developing a Slurry Fill for Modified In Situ Oil Shale
Mining.
Objective: Develop an underground slurry filling system
that will minimize surface and underground environmen-
tal disturbance. Specifically, render in situ shale rubble
virtually impermeable and minimize contamination of the
ground water and its aquifers, bind the spent shale to-
gether in a coherent mass sufficiently strong to resist
surface subsidence and eliminate the potential for dif-
ferential settlements of the surface, and dispose of the
maximum amount of surface retorted shale material as
the major constituent of the fill, thereby minimizing or
eliminating disposal of spent shale on the surface.
5. Development of Systematic Waste Disposal Plans.
Objective: Develop a set of guidelines for the mining
industry that provides rationale for the development of
systematic waste disposal plans. The resulting manual
will be suitable for use by Federal and State personnel,
mine operators and planners, and design professionals.
Emphasis will be placed on the impact of a disposed
waste on all aspects of the environment, including, but
not limited to: future land use and economic factors, long-
term stability, hydrology effects, potential contamination
of water supplies and emanation of hazardous sub-
stances. Impact of existing and proposed State and Fed-
eral regulations on the disposal of waste will be
emphasized as well.
6. Mine Waste Location by Satellite Imagery.
Objective: Investigate the potential for using satellite
(Landsat) data for detecting coal and metal-nonmetal
waste and tailings disposal sites. Determine if the band-
ratio method of preprocessing Landsat multispectral data
before classification can be used to detect and identify
waste embankments.
7. Engineering Property Changes and Environmental Ef-
fects on Coal Mine Waste Due to Slaking.
Objective: Determine the changes in physical properties
and chemical environmental impact potential of coal mine
waste dumps and embankments as a function of age,
and evaluate the effects of slaking on the stability of
waste piles and recommend techniques to control the
detrimental changes.
8. Conceptual Designs for Retaining Structures for Open
Pit Backfilling.
Objective: Develop conceptual designs of retaining
structures for constructing backfill areas to dispose of
wastes within open pits. Milling, smelter, and overburden
wastes will be considered. Develop detailed conceptual
designs for two operating open pit mines.
9. State-of-the-Art Environmental Assessment of Onshore
Disposal of Manganese Nodule Rejects.
Objective: Identify all reasonable state-of-the-art and
emerging techniques for onshore disposal of various types
of manganese nodule reiects and develop site selection
design criteria and associated significant environmental
and land use effects for those techniques, sites, and
types of rejects identified.
10. Evaluation of LJxiviation of Mine Waste.
Objective: Determine which types of, and to what extent,
coal wastes contaminate ground water through leaching
of acid- or toxic-forming materials. Collect solid waste
samples and evaluate, through laboratory studies, the
extent to which components of these wastes are leached
by rainwater.
11. Field Test for Environmental Disposal of Mill Tailings in
Surface Backfill.
Objective: Perform a small-scale field test to dispose of
mill tailings along the Coeur d'Alene River valley while
preventing seepage through the waste and into the water
system; conduct geotechnical, hydrologic, and vegeta-
-------
tive experiments at the site; and evaluate feasibility of
full-scale implementation.
12. Investigations of the Flow Characteristics of Mine Tail-
ings.
Objective: Determine the flow characteristics of liquefied
mine tailings and, when a mass of mine tailings flows
and liquefies, the probable extent of travel. Analyze case
studies of mass flow failures of mine embankments.
13. Designs for an Underground Disposal System for Active
Uranium Mines.
Objective: Define environmental stability, health and
safety, and regulatory constraints on backfilling. From
these definitions, develop a number of initial conceptual
designs based on related research and other pertinent
information. As the final phase, develop a detailed design
at a specific site.
14. Disposal of Gold Dredge Tailings.
Objective: Develop and demonstrate methods to partially
level the ridges of coarse gravel left by the dredge waste
conveyor and pump the finer gravel, sand, and soil onto
the smoothened area. This would eliminate ridges of
sterile, washed, coarse gravel that exist from present
methods of operation and place the topsoil on the surface
in a usable condition.
15. Inventory of Waste Embankments, Surface and Under-
ground Openings.
Objective: Compile an inventory of mine waste em-
bankments to form a data base for improving waste man-
agement and waste disposal techniques in the metal-
nonmetal mining industry.
16. Measurement of Seepage Beneath Tailings Ponds.
Objective: Apply the results of the in-house project "De-
sign of Slime Sealed Impoundments to Prevent Ground
Water Contamination" to field conditions where seepage
can be accurately measured to determine the best pos-
sible method for sealing the bottom of the tailings pond
to prevent seepage. Five separate areas are being pre-
pared to test different natural soil-tailing combinations
and compaction to achieve the desired results.
17. Removing Heavy Metal From Runoff Water Draining Lean
Copper-Nickel Ore Stockpiles.
Objective: Develop techniques for removing heavy metal
pollutants from runoff draining copper-nickel lean ore and
waste stock piles. The development of copper-nickel re-
sources in northeastern Minnesota depends upon the
selection of an adequate abatement program to control
heavy metal leaching from exposed waste rock and lean
ore piles. The low grade of the ore deposit makes it
necessary to find an inexpensive control system that will
allow the mining to be economically feasible.
18. Assessment of Ground and Surface Water Effects Around
Coal and Mineral Storage Areas.
Objective: Survey the extent to which storage piles of
coal and minerals release toxic ions or other significant
organic pollutants to ground or surface waters and as-
sess pollution types and levels. Develop mitigative meas-
ures for each mineral commodity for improved control of
runoff.
19. Detection of Lixiviant Excursions With Geophysical Re-
sistant Measurements During In Situ Leaching.
Objective: Develop, test, and demonstrate a geophysical
resistance measuring system that can reliably detect the
excursion of a lixiviant having a resistivity half that of the
ground water it replaces when the lixiviant has migrated
half way from an injection welt to a monitor well at a
depth of at least 500 feet.
20. Evaluation of Best Management Practice for Mining Solid
Waste.
Objective: Determine best management practices for mine
waste disposal in the areas of ore (copper, iron, lead,
nickel, molybdenum, and zinc), phosphates, and uranium
by utilizing the results of an extensive monitoring program
for ground and surface water and air quality. (Coopera-
tive agreement with the Environmental Protection Agency)
Minerals Health and Safety Technology,
Mine Waste Stability
IN-HOUSE RESEARCH
1. Evaluation of Filter Cloth for Stabilization of Coal Mine
Wastes.
Objective: Evaluate the criteria for selection of filter cloth
to control seepage in coal mine waste dams. Conduct
laboratory, tests of various filter cloths under simulated
mine waste dam environments. Develop preliminary
guidelines for use of filter cloths in coal mine waste dams
2. Consolidation of Coal-Clay Wastes by an Improved Floc-
culation Technique.
Objective: Demonstrate the technical feasibility of using
an improved flocculation technique to dewater waste coal
sludge generated in coal preparation plants to produce a
consolidated stable waste material containing 50 or more
weight-percent solids that can be safely stored. Through
laboratory investigations, optimize flocculation and con-
solidation parameters, and establish mass flow rates. Based
on laboratory investigations, design and assemble a larger-
scale field test unit. Demonstrate the feasibility of mixing
dewatered coal sludge with coarse coal refuse material
for long-term stabilization of both waste products.
3. Mixing Coarse and Fine Coal Waste.
Objective: Determine optimum mixing ratios of coarse and
fine coal wastes to achieve maximum fill strengths for
surface disposal, and develop a method to mix and trans-
port the mixtures while minimizing segregation. Obtain
samples of coal wastes from preparation plants that mix
fine and coarse coal wastes and that impound fine coal
wastes behind an embankment of coarse wastes. Deter-
mine fill strengths of coarse and fine wastes by laboratory
tests. Prepare a report on the waste disposal practices at
the coal preparation plants visited.
-------
4. Alternative Coal Waste Disposal Methods.
Objective: Complete physical property tests of coarse
anthracite coal waste; collect Shelby tube samples of an-
thracite fine waste and determine their physical properties;
and conduct laboratory model tests stimulating injection
of fine waste particles into voids created by coarse waste.
Injection of fine waste particles into voids is aimed at in-
creasing the support capability of backfilling and other
waste disposal methods.
5. Factor of Safety/Risk Analysis in Tailings Embankments
Design
Objective: Apply techniques of operations research and
statistics to the design of tailing dams; establish a confi-
dence interval or level of uncertainty about the factor of
safety; investigate state-of-the-art sampling procedures
for tailings embankments; construct simplified factor of
safety charts for field applications; investigate industry
sampling procedures used to ascertain the safety of ex-
isting tailings embankments; and develop a statistical pro-
cedure to be applied to onsite embankment maintenance
and control.
Contract Research
1. Safety and Health Problems Associated With Under-
ground Mine Waste Disposal.
Objective: Identify potential safety and health hazards to
underground miners that could result from the under-
ground disposal of mine waste.
2. Disposal of Wastes Over Active Underground Mines.
Objective: Determine if guidelines can be developed for
safe and economical methods for the disposal of wastes
over active underground coal mines. Identify different types
of mine wastes and evaluate their hazards to underground
mining; evaluate underground parameters that may affect
waste pile stability; evaluate peripheral conditions that may
affect the waste disposal area; evaluate probable safety
features that may be incorporated either above or below
ground; and compare available waste disposal techniques
for their applicability.
3. Centrifuge Model Testing of Waste Embankments.
Objective: Determine safety criteria for tailings embank-
ments by simulating field conditions using a centrifuge for
modeling. Conduct tests on a 25-foot-radius centrifuge to
investigate seepage and erosion effects, foundation dif-
ferentials, and other embankment construction problems.
4. Satellite Monitoring of Coal Waste Embankments.
Objective: Demonstrate the use of a satellite communi-
cation link with a remote data collection station located on
a coal waste embankment. Select a suitable satellite and
negotiate an agreement to use it Purchase or rent a data
collection platform for placement at the test site to connect
with the instruments already in place from previous work.
The present data collection station at the test site will
remain in use to check data received from the satellite
and to provide additional long-term information from the
instruments. Compare and evaluate data from the two
data stations to provide information as to the most reliable
and cost-effective system.
5. Critical Parameters for Tailing Embankments.
Objective: Construct probability density functions of soil
parameters for tailings embankments that are represent-
ative of the major mining commodities in the United States.
Collect and categorize engineering parameters of tailings
embankments for future input to slope stability models;
construct probability density functions of the data; and
compute the mean and variance of the data.
6. Compaction Criteria for Metal-Nonmetal Tailings.
Objective: Develop data to determine compaction criteria
for various metal-nonmetal wastes for both surface and
underground disposal. Determine the extent of compac-
tion criteria available through an extensive literature search.
Obtain samples of tailings from various types of metal-
nonmetal mines and conduct laboratory tests to determine
density, grain size, permeability, and static shear. Deter-
mine correlations between laboratory results and actual
field compaction efforts.
Summary
The Bureau of Mines has been active in mine waste dis-
posal research for many years. Current legislative actions
and the concern for structural and environmental integrity of
mine waste management systems have kept this research
area an important part of the Bureau's minerals research
program. The overall objectives and specific projects de-
scribed in this paper provide a current documentation of prob-
lems being addressed. Industry input continues to play a
major role in the formulation of the research programs and
an active technology transfer program is being implemented
to assist in the implementation of results obtained.
-------
DISPOSAL OF COAL MINE WASTE IN ACTIVE
UNDERGROUND COAL MINES
by
Leslie S. Rubin1, Mackenzie Burnett2, Al Amundson3, Gary J. Colaizzi4, and Ralph H.
Whaite5
BACKGROUND
Nature and Extent
Approximately 25 pet of the raw coal extracted from un-
derground mines in the United States is rejected as waste
and deposited on the surface, usually near the preparation
plants6. These accumulations on the surface result from
screening, crushing, sizing, and washing of coal to separate
impurities from the marketable product and are called waste
banks, refuse banks, culm banks, or bony piles. The reject
material contains a certain amount of combustible as well as
pyritic and other objectionable material; it is generally sep-
arated into two classes, coarse and fine refuse. Coarse refuse
accumulations generally resemble conical or ridge-shaped
mounds that may extend up to 700 feet in height and over 1
mile in length. The fine refuse is often impounded in settling
ponds.
The Bureau of Mines estimates that 174,000'acres in the
United States used for disposal of coal processing waste
remain unreclaimed. Burning material is contained in about
290 banks covering approximately 3,000 acres distributed
throughout 13 of the 26 coal producing States6. In the Eastern
Coal Province alone there are at least 3,000 to 5,000 sizable
waste piles and impoundments which cumulatively contain
over 3 billion tons of refuse7. At current coal production rates,
about 60 million tons of mine refuse and processing waste
are generated each year.
Waste banks and impoundments are potential health and
safety hazards. Waste bank slides and impoundment failures
1 Senior engineer, Mechanical Processing and Fluid Transport Division, Foster-
Miller Associates, Inc , Waltham, Mass
2 Division manager, Mechanical Processing and Fluid Transport Division, Fos-
ter-Miller Associates, Inc., Waltham, Mass
3 Chief engineer, Western Slope Carbon, Inc, Salt Lake City. Utah
4 Supervisory mining engineer, Denver Research Center, Denver, Colo
5 Mining engineer, Denver Research Center. Bureau of Mines. Denver, Colo
6 National Academy of Sciences, National Academy of Engineering Under-
ground Disposal of Coal Mine Wastes Government Printing Office, Wash-
ington, D.C., ISB No 0-309-02324-6, 1975, 172 pp.
7 Johnson, W., and G. C. Miller Abandoned Coal-Mined Lands—Nature. Ex-
tent, and Cost of Reclamation BuMmes Spec Pub. 6-79, 1979, 29 pp
have destroyed entire communities. In addition, such banks
are environmentally degrading, create an eyesore, and are
sources of acid drainage pollution. The burning banks con-
tribute significantly to the degradation of surrounding atmos-
pheric conditions and inhibit development and economic
growth. The resulting smoke, dust, and poisonous and nox-
ious gases can be fatal to human, animal, and plant life.
Legislation and Regulations
Recognition of the severity of the solid waste problem in
the United States by the Congress resulted in passage of the
Solid Waste Disposal Act, Public Law 89-272, on October
20, 1965. Under this act the Bureau's responsibility for re-
search and development work on the disposal or utilization
of mineral waste was expanded. It included economic and
resource evaluation studies aimed at delineating factors
causing and contributing to waste disposal problems in min-
eral and fossil fuel industries, and scientific and engineering
research to find ways of utilizing, or otherwise disposing of.
a variety of inorganic waste materials.
The first Federal regulations governing the surface disposal
of coal wastes were promulgated pursuant to the Federal
Coal Mine Health and Safety Act of 1969, Public Law 91-
173. Under the act the Secretary of the Interior was required
to develop health and safety standards for surface disposal
of waste. The first regulation required that waste piles con-
structed after 1971 be located a safe distance from under-
ground mine openings and preparation plants and that new
refuse piles constructed over exposed coal beds be sepa-
rated from the coal by clay or some other inert material. The
second regulation concerned the construction of coal waste
piles, requiring the waste be spread in layers and placed so
as to minimize the passage of air through the piles. They
were not to be constructed in a way that would impede drain-
-------
age or impound water and were to be designed to prevent
sliding and shifting of materials. The third requirement in-
volved inspection and monitoring of all retaining dams if it
was thought that failure of a water or silt-retaining dam would
create a hazard to miners.
In 1966 at Aberfan, Wales, more than 150 residents (mostly
children) in that small community perished, when an 800-
foot-high coal refuse pile slid down the mountainside into the
town. The laws enacted by Pennsylvania, West Virginia, and
Kentucky regarding coal waste disposal were strengthened
in 1968 primarily as a result of the Aberfan disaster. In the
aftermath of the Buffalo Creek disaster in 1972, an inspection
and inventory of all waste impoundments was conducted in
order to determine their stability. Federal inspectors from the
then Department of the Interior Mining and Enforcement and
Safety Administration, who were trained in coal waste pile
stability, made the examinations. In March 1978 this work
was transferred to the Mine Safety and Health Administration
(MSHA) in the Department of Labor. The emergency inves-
tigation of 139 impoundments, initiated after the Buffalo Creek
disaster, resulted in the rating of 8 impoundments as immi-
nent hazards, 66 as obviously deficient, 61 as not obviously
deficient, and 4 as missing data. Under the authority of Public
Law 92-367, The Dam Safety Act, enacted August 8, 1972,
the Army Corps of Engineers conducted a survey of all waste
impoundments in the Appalachian region and reported the
results to the State governors. Because of the potential haz-
ards to miners and the public from improperly constructed
coal waste piles and impoundments, the Department of the
Interior, in 1974, published proposed new and expanded reg-
ulations controlling waste piles in the Federal Register.
The Surface Mining Control and Reclamation Act of 1977,
Public Law 95-87, signed into law on August 3,1977, created
within the Department of the Interior an Office of Surface
Mining Reclamation and Enforcement (OSM), and regula-
tions under the provisions of the Act were promulgated in
December 1977. Under this authority the regulations initiated
by the act of 1969 were upgraded and greatly expanded and
included requirements for underground disposal, wherever
practiced, as well as for surface disposal of coal processing
waste.
Under the act of 1977, the general requirements for the
construction and maintenance of coal processing waste banks
included that the work be done within a permit area and in
accordance with rules established for disposal of under-
ground development waste and excess spoil and to prevent
combustion. The disposal area should not adversely affect
water quality, water flow, or vegetation, should noi create
public health hazards or cause instability in the disposal area,
and should not extend to within 8 feet of any coal outcrop.
Moreover, all coal processing waste banks should be in-
spected, on behalf of the person conducting underground
mining activities, by a qualified registered engineer or other
person approved by the regulatory authority. Furthermore,
coal processing waste should not be used in the construction
of dams and embankments unless it has been demonstrated
to the regulatory authority that the stability of such a structure
is in accord with the regulations.
The requirements for returning coal processing waste to
underground workings are in five parts: (1) Each plan shall
describe the design, operation and maintenance of any pro-
posed coal processing waste disposal facility for approval of
the regulatory authority and the Mine Safety and Health
Administration; (2) Each plan shall describe the source and
quality of waste to be stowed, area to be backfilled, percent
of mine void to be filled, method of constructing underground
retaining walls, influence of the backfilling operation on active
underground mine operations, surface area to be supported
by the backfill, and anticipated occurrence of surface effects
after backfilling; (3) the applicant shall describe the source
of the hydraulic transport mediums, methods of dewatering
the placed backfill, retainment of water underground, and
treatment of water if released to surface streams, including
the effect on the hydrologic regime; (4) the plan shall describe
each permanent monitoring well to be located in the backfilled
area, the stratum underlying the mined coal, and gradient
from the backfilled area; and (5) where applicable the pre-
ceding requirements shall also apply to pneumatic backfilling
operations.
Costs of Waste Bank Reclamation
Costs of coal waste disposal in the United States vary
within a narrow range and depend primarily on the choice of
materials handling systems, topography of the mining oper-
ation, and the distance between the source of waste and the
disposal site. Almost exclusively, waste disposal for under-
ground mining operations is currently accomplished by sur-
face disposal on waste piles near the preparation plant.
Current practices of surface disposal consist of dual waste
disposal systems, one for the coarse waste and one for the
fine waste. Coarse waste is usually handled by truck, belt
conveyor or aerial tram or a combination thereof, and de-
posited directly on the waste pile. Fine waste is normally
transported as a slurry by pipeline to an impoundment. If the
fine waste is completely drained, it can be removed from the
impoundment and deposited on the waste pile.
Typical costs as of 1975 for operating total waste systems
(belt, truck, dozer) range from $0.15 to $0.40 per ton of waste
handled6. Engineering estimates for installing a new system
based on costs of new equipment are considerably greater
Due to inflation and the upgraded requirements promulgated
under Public Law 95-67, costs of surface reclamation by
1980 have risen sharply. While some information on the costs
of disposal of coal processing waste in abandoned under-
ground mines is available; similar data on costs of disposal
in active underground mines are lacking.
Recommendation of National Academy of
Sciences
The inclusion of regulations pertaining to the underground
disposal of coal processing waste that were developed after
passage of the 1977 act may be attributed to a 1975 study
conducted by the National Academy of Sciences (NAS) for
the National Science Foundation6. This study addressed the
-------
10
state-of-the-art technology for mine backfilling concurrent with
coal extraction, and provided an overview of the technical
and economic constraints on underground disposal of coal
mining wastes. The NAS study recommended that under-
ground waste disposal be demonstrated at sites represent-
ative of U.S. coal mining practice in order to establish the
feasibility of this method of disposal of coat mining waste in
the United Slates.
BUREAU OF MINES PROJECT FOR STUDY OF TECHNICAL AND ECONOMIC
EVALUATION OF UNDERGROUND DISPOSAL OF COAL MINE WASTE
In response to the recommendation of NAS and the prom-
ulgation of rules governing the underground disposal of coal
preparation plant waste subsequent to Public Law 95-87, the
Bureau of Mines initiated a detailed study of the technical,
economic, and environmental constraints on underground
waste disposal. A request for proposals to conduct the study
was advertised April 14,1977, and responses were received
from seven engineering firms from various parts of the coun-
try.
The objective of the proposed study was to determine the
specific circumstances under which the disposal of coal mine
waste underground concurrently with extraction is technically
and economically feasible. The work was to be conducted in
two distinct phases.
The contract to perform the work was awarded to HRB-
Singer, Inc., of State College, Pa.8. A final report on the two-
phased investigation that assessed the technical and eco-
nomic feasibility of disposing of coal refuse underground in
active mines was submitted to the Bureau of Mines in January
1980.
During phase I, which included a literature survey, coal-
producing regions of the United States were identified in which
the problems of surface disposal of coal refuse were most
severe. Criteria used in identifying the regions included reg-
ulatory constraints, availability of disposal sites, and other
social, economic, and physiographic factors such as the amount
of coal refuse produced in the region, population density,
precipitation, local relief, steepness of slopes, volatility and
sulfur content of the coal including the refuse, urban area
density, number of coal preparation plants, and number of
burning refuse banks and impoundment failures reported in
the region over the past decade.
From an initial list of 20 regions, three were selected as
having the most severe problems of surface refuse disposal:
southwestern Pennsylvania, southern West Virginia, and
southwestern West Virginia combined with eastern Kentucky.
For each of these three regions, a model mine was postulated
that typified the geologic conditions, mining operations, eco-
nomic factors, and environmental conditions that prevail in
each area. Costs of surface disposal were also estimated for
each region..
The mine model that was developed for southwestern West
Virginia and eastern Kentucky, the "Kanawha Region," was
8 Bucek, M. F , J. K Clauser, J A. Schad, and N K Chukravorti Technical
and Economic Evaluation of Underground Disposal of Coal Mining Wastes
HRB-Singer, Inc (State College, Pa.), Final Rept Contract J0285008, Jan-
uary 1980, 349 pp ; BuMines OFR 80-015, available for consultation at the
Bureau of Mines facilities in Albany, Oreg., Avondale, Md , Denver, Colo ,
Pittsburgh, Pa., Reno, Nev, Rolla, Mo, Salt Lake City, Utah, Spokane,
Wash., Tuscatoosa, Ala., Minneapolis, Minn., and Central Library, U S De-
partment of the Interior, Washington, D.C.; DOE facilities in Carbondale, III,
and Morgantown, W. Va.; and from National Technical Information Service,
Springfield, Va., PB 80-154768.
selected as the basis for a detailed analysis of underground
disposal systems in phase II. The Kanawha Region was se-
lected because of its past and present problems of refuse
disposal (Buffalo Creek is located in this region), and because
the topography, mine size, mine type, and mining operations
of this region are similar to conditions existing elsewhere in
West Virginia, eastern Kentucky, and Virginia.
In phase II of this 2-year study, various methods used to
stow refuse underground were reviewed and evaluated, and
three methods that would be most suitable for implementation
at the model mine were selected. For each of these methods,
a conceptual design of a stowing system was developed.
The stowing systems developed included: (1) A mechanical
backfilling system in which coal refuse is transported from
the preparation plant to the mine site by truck, introduced to
the underground workings by gravity feed via a service shaft
and conveyor belt, and emplaced in a mined out area by
battery-powered scoop cars; (2) a pneumatic backfilling sys-
tem in which the refuse is transported from the preparation
plant to the mine entry by truck, introduced to the stowing
area by rail cars, and emplaced in the mined out voids by
pneumatic stowers; and (3) a direct hydraulic system whereby
refuse in the form of a slurry is transported from the prepa-
ration plant to the underground stowage area hydraulically in
a pipeline. The three methods were selected not only on the
basis of their appropriateness for the model mine, but also
on the basis of the extent to which the methods exemplified
the major methods of backfilling employed currently through-
out the world. Environmental factors associated with each
type of backfilling, the technical feasibility of each design
concept, and the costs of implementing each conceptual de-
sign were assessed.
All three systems of stowing refuse underground in active
mines are more expensive than surface disposal of refuse in
the Kanawha Region. Estimated costs of surface disposal
(1978) ranged from $2.46 to $2.61 per ton of refuse, whereas
estimated costs of mechanically stowing 1 ton of refuse were
$4.14; pneumatically stowing 1 ton of refuse, $7.67; and hy-
draulically stowing 1 ton of refuse, $4.31. The costs of hy-
draulic stowing were somewhat unrealistic because the
preparation plant of the model mine was located more than
3 miles from the mine site. If the preparation plant were lo-
cated closer to the mine, the per-ton cost of refuse disposal
would be considerably lower.
The members of the evaluation team determined that the
elements of each conceptual design were acceptable and
their costs reasonable. While the model mine does not com-
pletely encompass the complexity of variable conditions from
mine to mine in relation to the full range of backfilling options,
it does provide some useful guidelines to the mine operator
in efforts to comply with regulations at an acceptable cost.
-------
11
DETAILED DESIGN AND DEMONSTRATION OF UNDERGROUND DISPOSAL OF COAL
MINING WASTES—BUREAU OF MINES REQUEST FOR PROPOSALS
The Bureau of Mines, as part of its minerals environmental
technology program and continuing search for more definitive
answers to the underground disposal of coal wastes, decided
to initiate a detailed mine site specific design of a system of
underground disposal of coal wastes; and implement a co-
operative demonstration of the system in an active under-
ground coal mine.
The scope of work was divided into two distinct phases:
Phase I—detailed design of underground coal waste disposal
system; and phase II—demonstration and evaluation of the
underground disposal system. It was required that the dis-
posal system design include the handling of the solid waste
(coarse and fine) generated as a result of the underground
coal mining and associated coal preparation processes.
It was required also that the proposal include a description
of the proposed demonstration mine site for development of
the detailed design; and a discussion of the proposed con-
ceptual disposal system including method or methods (hy-
draulic, mechanical, pneumatic, etc.) for the proposed mine.
The mine description included its location and geologic set-
ting, mining methods, current waste disposal methods, pro-
duction, coal preparation, characteristics of coal wastes, etc.
Phase I—Design
During phase I, which was to be completed in 6 months,
it was required that the contractor prepare a detailed mine
site specific design including calculations, drawings, and costs
for the subsequent demonstration of the complete under-
ground disposal system. In developing the design the con-
tractor was required to assess or incorporate the following
considerations: (1) Physical and chemical characteristics of
coal wastes produced by the mine to determine material flow
characteristics, wear rates, etc.; (2) availability of required
equipment and technology to introduce and operate the sys-
tem; (3) extent to which current operating methods in the
mine may be changed or disrupted in order to accomodate
the disposal system; (4) any anticipated decrease in mine
productivity due to the introduction of the disposal system,
(5) effects on health and safety of production personnel as
well as those engaged in implementing the proposed disposal
system; and (6) associated benefits, if any, such as strata
control or increase in recoverable coal due to the under-
ground disposal of the coal wastes.
The overall compatibility between the current mining sys-
tem and the underground disposal scheme was to be as-
sessed. This assessment was to include an analysis of safety
and conformance of all equipment and design criteria re-
quired by all applicable State and Federal laws and regula-
tions, including the Coal Mine Health and Safety Act, Surface
Mining Control and Reclamation Act, etc.
A comprehensive monitoring system was to include ma-
terial flow quantities, placed material density, compressibility,
mine roof support, effects on ventilation, dust, noise, and
where applicable effects upon mine water handling, influx,
discharge, etc.
During phase I, the contractor was required to obtain or be
actively engaged in acquiring all the necessary State and
Federal permits for implementing the underground disposal
system demonstration.
Phase II—Onsite Demonstration
During this phase, the contractor will conduct a demon-
stration of the complete underground waste disposal system.
The demonstration shall be full scale in order to perform an
evaluation of the technical and economic success of the un-
derground disposal system.
The work will include the construction of all facilities nec-
essary to integrate the approved waste disposal system into
the active underground mine. All equipment and installations
shall comply with all Federal and State laws and regulations
All monitoring and control systems shall be thoroughly checked
to accurately reflect the performance of the entire system.
An accurate daily log shall be maintained and include all
pertinent cost data, from which the cost per ton of waste
placed underground shall be calculated. From this data, any
loss or gain in productivity shall also be measured and re-
ported. In addition to environmental effects, the final report
shall summarize all findings, results, conclusions, and rec-
ommendations including sketches and design drawings as
needed for clarification.
Two Contracts Awarded for Phase I
Only two responses were received and both proposals were
evaluated by a committee of technically qualified Bureau.
OSM, and MSHA employees assisted by Bureau contracting
personnel. It was decided to fund both proposals because
they represented alternative methods, which could be quite
important to the mining industry. Both contracts were awarded
on September 27, 1980.
One contract was awarded to GEX Colorado. Inc . Pali-
sade, Colo,, operators of underground coal mines in Palisade.
working as a team with Michael Baker, Jr.. Inc.. a diverse
civil engineering firm in Beaver, Pa. The proposal submitted
by this team covering phase I was concluded with a com-
prehensive detailed design of an underground disposal sys-
tem for the Roadside mine in Mesa County, Colo., about 12
miles northeast of Grand Junction, Colo, The plan involved
hydraulic underground disposal of the 1,750 tons of refuse
produced daily by the Roadside preparation plant Complete
details of the proposed demonstration work were given in a
final report by the GEX Colorado—Michael Baker, Jr., team.
Work on the second contract is more advanced and is
therefore the one about which this paper is primarily con-
cerned. It involves a method of underground disposal of prep-
aration plant waste in an active mine that utilizes both hydraulic
and pneumatic systems. The contract to prepare this design
was awarded to Foster Miller Associates, Inc. (FMA)/Western
Slope Carbon, Inc. (WSC).
-------
12
The WSC Hawk's Nest mine near Paonia in western Col-
orado was selected as the site for the demonstration The
Federal leases that make up the Hawk's Nest mine property
lie along the north fork of the Gunnison River about 1.5 miles
east of Somerset, Colo. Elevation varies from 6,140 feet along
the river to 8,200 feet on the northern edge of the leases.
Both phases of the proposed work are to be managed by
FMA with technical and labor support from WSC.
Design Submitted by Foster-Miller
Associates/Western Slope Carbon, Inc., for
Demonstration at the Hawk's Nest Mine,
Paonia, Colo.
The underground waste disposal system being considered
for the WSC Hawk's Nest mine utilizes hydraulic transport,
in-mine dewatenng, and pneumatic haulage. This disposal
system is summarized in the generalized flowsheet of figure
1. Fine refuse (minus 35 mesh) and coarse refuse (plus 35
mesh) exit from the wash plant as two separate slurries. Both
slurries are transported 4,400 feet through separate pipelines
to the underground dewatering-pneumatic feeding station.
Here the refuse slurries are dewatered on separate com-
ponents and the solids fed to a pneumatic feeder for transport
and placement in the appropriate mine location. Screen drain-
age and centrifuge effluent are returned to the wash plant,
closing the water circuit.
The remainder of this paper describes: (1) The mine refuse,
(2) the waste disposal system and equipment selection, (3)
the influence of the disposal system on mine productivity, and
miner health and safety, and (4) the phase II program for
system installation and evaluation.
WASH PLANT
FINE REFUSE
SLURRY
FINE REFUSE
OEUATERING
CLEAN
• COAL
COARSE REFUSE
SLURRY
COARSE REFUSE
[WATERING
PNEUMATIC
SYSTEM
TO MINE
STOWING
Figure 1.—Simplified flowsheet for underground
disposal system.
Geology at the Hawk's Nest Mine
The WSC Hawk's Nest Mine is located at the eastern end
of the Somerset District, a part of the Paonia Coal Field.
Locally, the coal bearing strata is about 2,500 feet thick and
is divided into four members, in ascending order. (1) Rollins
sandstone member, (2) lower coal member (Bowie), (3) upper
coal member (Paonia), and (4) barren member. The main
coal beds within WSC's lease are found in the upper coal,
or Paonia member. Of the four minable seams in this mem-
ber, WSC is presently mining the E seam (Johnson's Hawk's
Nest or F bed) and plans future development of rock tunnels
to the underlying Wild seam and D seam.
The E seam is an attrital coal with a seam thickness ranging
from 4.5 to 12 feet. Floor sediments are generally carbona-
ceous shale. Roof sediments vary from carbonaceous shales
with mterbedded siltstones to good roof-supporting sand-
stone. The Wild seam (Johnson's E seam) ranges from 5.5
to 10 feet in thickness above and below a minable coal. The
coal invariably exhibits numerous splits and generally has
poor roof and floor sediments, usually carbonaceous shales
and interbedded siltstones. The D seam is an attrital coal
with thicknesses ranging from 6 to 14 feet. This coal seam
directly overlies the Bowie member and generally has very
good floor and roof sandstones.
Physical and Chemical Characteristics of
Coal Mine Wastes at Hawk's Nest Mine
The average run-of-mine feed rate for which the wash plant
was originally designed is 400 tph. Twenty to 25 pet of the
feed is discharged from the wash plant as waste WSC pre-
sently utilizes two independent systems for disposing its min-
ing waste. One system handles that waste material which is
35 mesh by 0 (fine refuse), approximately 15 pet of the total.
The second system handles the coarse or 6-inch by 35-mesh
refuse.
Fine refuse as indicated in the flowsheet of figure 2 is a
combination of cyclone overflow, sieve bin drainage, and
drainage from the refuse dewatering screen. This 5- to 6-pct-
solids waste is concentrated in a static thickener to a 25- to
30-pct-solids consistency, transferred to a thickener under-
flow sump, and pumped underground to clarifying lagoons at
rates of 12 to 14 tph (solids). The fine t efuse typically contains
30 pet ash forming materials and averages 10,000 Btu/lb.
Coarse refuse is the reject material from the jig which has
been sized and dewatered on a 5- by 12-foot double deck
screen with 1-inch and 35-mesh cloth openings. The surface
wet plus 35-mesh waste discharges into a refuse storage
hopper for truck loading and transport to the Delta County
landfill site. The landfill site is located in an arid desert-like
region, 26 miles from the mine. Present coarse refuse dis-
posal costs are approximately $3.67 per ton for loading, trans-
porting, spreading, and covering.
The average hourly production of coarse refuse measured
during June, July, and August 1980, equaled 87 tph. The
abrasiveness of the coarse refuse as measured by the silica
and alumina content (42 and 10 pet respectively) was high.
-------
13
Water, after being exposed to coarse refuse for 1 week, reached
a pH of 8.25. Corrosion of coarse refuse slurry equipment
should not be a problem.
Conceptual Design
The proposed waste disposal system will utilize hydraulic
transport and pneumatic conveying for transportation and
placement of coal wastes underground. The proposed com-
bination of hydraulics and pneumatics utilized the best fea-
tures, cost and safety, of each transportation system.
The coal waste disposal system has been divided into
two subsystems.
1. The above ground system is responsible for ma-
terial sizing, slurry formation, and slurry pumping.
It must fit into the existing wash plant with minimal
changes, and permit continuous material han-
dling and storage during emergency situations.
2. The below ground system transports, dewaters,
and pneumatically stows waste material. This
system fits into the existing mine and mine plan,
is flexible enough to permit movement of the de-
watering-pneumatic feeding station, and does not
interfere with the present mining methods.
The waste disposal system's conceptual design is illus-
trated in figure 3. Refuse, separated from clean coal in the
Baum jig, is removed by draining bucket elevators and is
discharged onto a double-deck dewatermg screen with 2-inch
and 0.5-mm cloths. The 2-inch by 0.5-mm waste discharges
into the coarse refuse storage hopper. Plus 2-inch waste is
fed to the coarse refuse cone crusher for reduction to a 2-
inch top size.
During underground disposal system operation, coarse re-
fuse stowed in the storage hopper is conveyed to the coarse
refuse slurry tank where it mixes with water and is pumped
underground by a pair of centrifugal pumps.
In a parallel fashion, the fine refuse (thickener underflow)
is pumped underground through a separate pipeline. Ration-
ale for two separate pipelines is discussed later.
Below ground, the fine refuse is clarified and dewatered
by a solid bowl scroll type centrifuge. Coarse refuse is de-
watered on a single deck screen (35-mesh cloth). Effluent
from the centrifuge and drainage from the screen are col-
lected in a common sump and returned to the coarse refuse
slurry tank at the wash plant. Surplus water,drains back to
the static thickener closing the water circuit and permitting
the eventual capture of the screen drainage solids (minus
35-mesh particles) as thickener underflow. The underground
dewatered solids are fed to the pneumatic feeder by conveyor
and pneumatically transferred approximately 1,000 feet for
placement in the mine.
When required, flow from both lines can be diverted to an
underground emergency storage sump to facilitate pipeline
drainage.
When the underground mine disposal system is not op-
erational, the coarse refuse belt conveyor's direction of travel
is reversed and coarse refuse is discharged into 20-ton trucks
for surface disposal. The fine refuse during this period of time
will be pumped to existing in-mine fine refuse disposal ponds
Above Ground System
As illustrated in the conceptual design flowsheet (fig 3) by
the letters E (existing equipment), M (equipment requiring
modifications), and N (new equipment to be purchased) most
of the major above ground system components presently
exist. Major new components include (1) Flop gate actuator.
(2) coarse refuse crusher, and (3) coarse refuse slurry pump
Coarse Refuse Crusher
Present state-of-the-art hydrotransport technology limits
particle top size to one-third the pipeline inside diameter
Since the mine's refuse has a top size of 5 to 6 inches and
the coarse refuse slurry pipeline has a 6-in-ID, crushing of
the coarse refuse to a 2-inch top size is required
The comminution device necessary to reduce the run-of-
mine refuse at the mine in preparation for hydrotransport
meets the majority of the following criteria: (1) Processes
highly abrasive material, (2) generates a minimum of fines,
(3) has capability to reduce 20,000-psi compressive strength
material, (4) reduces minus 8-inch material to 100 pet passing
2 inches, and (5) capacity to be a minimum of 25 tph.
A cone crusher is the only reduction device that addresses
the majority of the criteria. A 2-ft-head-diameter cone crusher
of the secondary type has been specified for the circuit and
is used in conjunction with a scalping mechanism to remove
the near size, minus 2 inch, material in its feed The crusher
is set at 1 inch and draws approximately 30 hp at 25 tph
Coarse Refuse Slurry Tank
Thickener underflow at the mine was previously transferred
to a thickener underflow sump (2,500 gal) before being pumped
underground for disposal. Bypassing the thickener underflow
sump and pumping the thickener underflow directly under-
ground frees the thickener sump for use as a coarse refuse
tank. The modifications required to complete this tank trans-
formation included: (1) Pumping thickener underflow directly
underground, bypassing the thickener sump, (2) provide sump
with a jet type inlet for water return, inlet is opposite suction
side of the coarse refuse solids pump, and (3) provide ov-
erflow baffles and drainage pipe for water return to the thick-
ener. The coarse refuse mixing tank is illustrated in figure 4.
Coarse Refuse Slurry Pump
The pressure required to pump coarse slurry through a given
pipeline is a function of slurry flow rate, slurry concentration.
slurry specific gravity, particle size distribution, pipe size, length,
and friction. Presently there is no general theory which com-
bines these parameters in a correctable fashion to actual
data.
At the (FMA) coarse slurry transport facility, studies were
conducted on a 300-ft-long, 6-in-ID pipeline to determine the
pressure requirements for coarse slurry pumping. Results of
these tests indicated a 4,400-ft-long, 6-in-ID pipeline with a
-------
75G
(30, S)
-------
Figure 2.—Preparation plant flowsheet.
KEY
T TONS PER HOUR
G GALLONS PER MINUTE
S SOLIDS
SM SURFACE MOISTURE
M U.S. STANDARD SIEVE MESH
1 RAW COAL TRUCK BIN (EXISTING)
2 42- by 84-INCH INCLINED FEEDER
4 30-INCH PLANT FEED CONVEYOR (6 INCHES BY 0)
7 BAUM JIG
8 CLEAN COAL FIXED SIEVE
9 ONE 8- BY 20-FOOT DOUBLE DECK CLEAN COAL SCREEN
10 ONE WEMCO 1300 CENTRIFUGE
11 CLEAN COAL CRUSHER
12 FINE COAL SUMP AND PUMP
4 ONE 5- BY 12-FOOT DOUBLE-DECK REFUSE SCREEN
16 100-TON REFUSE BIN
17 REFUSE BIN GATE
24 THREE 20-INCH CLASSIFYING CYCLONES
25 19-FOOT, HIGH-CAPACITY STATIC THICKENER
26 THICKENER UNDERFLOW PUMP
27 CLARIFIED WATER PUMP
28 FINE COAL SIEVE BEND
29 ONE EBW-36 CENTRIFUGE
31 30-INCH CLEAN COAL STORAGE CONVEYOR
32 CLEAN COAL STORAGE BIN
33 CLEAN COAL BIN GATE
34 PLANT FLOOR SUMP PUMP
60 6- BY 9-FOOT HINGED STATIONARY GRIZZLY
61 DIVERTER GATE
62 24-INCH STOKER CONVEYOR
64-1 DIVERTER GATE
64-2 DIVERTER GATE
65 REFUSE CRUSHER
66 THICKENER UNDERFLOW SUMP AND MINE PUMP
71 42-INCH CRUSHER FEED CONVEYOR
72 CRUSHER (EXISTING)
-------
REMOVAL BY TRUCK
t - EXISTING EQUIPMENT
M - MODIFIED EQUIPMENT
N - NEW EQUIPMENT
T - TONS PER HOUR
G - GALLONS PER MINUTE
2-3T
645-8536
Figure 3.—Underground disposal system conceptual design.
-------
17
SYSTEM CURVE FOR SLURRY
MAXIMUM CONCENTRATION
COARSE REFUSE
SLilRRY PUMP
Figure 4.—Coarse refuse slurry tank.
180-foot drop in elevation will require a pump output pressure
of 260 psi to hydraulically transport 90 tph of coarse solids.
The pumping system consists of three centrifugal pumps
in series driven by one 350- to 400-hp variable speed motor.
The pumps are all Ni-hard with wearing plates and suction
head liners.
The need for variable speed drive is the result of pumping
downhill. Water alone with a 180-foot head will drain from
the coarse refuse slurry tank at a rate of 845 gpm. Turning
on the slurry pumps at full speed with only water will drast-
ically increase the water flow rate to a level which cannot be
handled by the underground water return pump.
Figure 5 illustrates the pump curve superimposed on the
system curves for water and maximum slurry concentration.
Point A represents the steady-state operating condition for
maximum slurry concentration transport. Point B represents
the operating condition for water only at the same pump
speed (N1). At this condition the water flow rate QB is sig-
nificantly greater than the water return pump capabilities that
are bracketed by flows QR and QR'. To drop the water only
flow rate to a condition that can be met by the water return
pump, the operating condition must be moved to point C. This
can only be attained by reducing the pump speed to N2.
During nonsteady state conditions when the pipeline is
being filled or emptied of solids, intersection of the pump and
system curve must be maintained between flow rates QR and
QR'. This constantly changing condition requires a control
loop that monitors the slurry flow rate and adjusts the pump
speed accordingly.
Figure 5.—Pump curve, system curve relationship.
Pipelines
The underground waste disposal system utilizes three sep-
arate pipelines; one each for the fine slurry, coarse solids'
hydraulic transport, and return water. The pipe inside diam-
eters will be 4, 6 and 8 inches respectively. Initially, the pipe-
lines are mild steel construction. Aluminum could be used to
reduce pipe weight and simplify the labor requirements for
installation. Pipes are joined with Victauhc-type couplings.
Slurry pipeline bends have a radius equal to or greater than
eight times the pipe diameter. During the first year of oper-
ation (phase II), sections of various specialty pipes resistant
to wear will be inserted in both slurry lines. Visual and ana-
lytical examination of these pipe segments will provide pipe-
line life-cost ratios
The fine slurry and coarse solids will be handled separately
to minimize the fine refuse dewatering equipment and total
system cost. Presently, fine refuse is concentrated in the
static thickener to a 25 to 30 pet (by weight) solids consistency
and discharged at rates of 225 to 250 gpm (12 to 14 tph of
solids). If thickener underflow were mixed with the coarse
slurry to form a 20 pet (by volume) solids concentration, the
anticipated 100 tph of total refuse would require 790 gpm of
water. Once below ground at the dewatering station, the com-
bined slurries would report to a horizontal screen for coarse
particle removal and dewatering. Assuming little or no particle
attrition, the screen drainage would contain 12 to 14 tph of
solids and 700 to 750 gpm of water. One centrifuge can
separate and dewater thickener underflow solids at slurry
rates of 225 to 250 gpm with 12 to 14 tph solids. At solids
-------
18
rates of 12 to 14 tph, with a hydraulic loading of 750 gpm,
three to four centrifuges are required for slurry processing
The cost of a separate 4-inch pipeline for fine refuse transport,
including slurry pump and installation is approximately $70,000.
The alternative capital cost of two to three additional centri-
fuges for the single pipeline system is $500,000 to $700,000.
The dual pipeline system is less costly and the space re-
quirements for the underground dewatering station are con-
siderably smaller.
The initial pipelines are 4,400 feet long with a vertical drop
of 180 feet from beginning to end The first 200 to 250 feet
of pipeline running from the wash plant to the return entry is
uphill at a 5-pct grade. Once inside the return entry, the
pipeline slopes downhill. In the winter months the outside
temperature can reach - 30° F. Pipeline freezing is prevented
by venting the apex of the pipelines and draining the exposed
pipelines back into the wash plant. Pipelines inside the mine
need not be buried as the return air is warm enough to prevent
freezing.
Underground System
Figure 6 conceptually illustrates the position of the under-
ground system equipment within a crosscut of a room-and-
pillar mine. Fine and coarse refuse materials are respectively
dewatered by a solid bowl centrifuge and a single deck screen.
Centrifuge effluent and screen drainage are transferred to a
control sump by low suction head pumps (pumps insensitive
to air ingestion). The control sump maintains a sufficient leve
of water to prevent pump cavitation. Control of the water level
is achieved with a level sensor, controller, and anticavitating
control valve on the discharge side of the water control pump
The dewatered solids are discharged onto a common con-
veyor and fed to the pneumatic feeder.
Coarse Refuse
Coarse refuse dewatering is accomplished on a single-
deck, horizontal screen. The necessary screening area is a
function of particle size distribution, water flow rate, solids
flow rate, and cloth opening. For a water and solids flow rate
of 1,000 gpm and 100 tph respectively, and a 35-mesh cloth,
a 5- by 16-foot screen is sufficient.
The screen has an overall height of 4 feet, 8 inches, the
lowest profile screen of all major screen manufacturers. The
cloth is fabricated from stainless steel profile wire panels of
1/8-inch grizzly rod construction.
Fine Refuse
Dewatering of fine refuse presents a more formidable prob-
lem. Many types of equipment (belt filters, vacuum filters,
centrifuges, and filter presses) have been evaluated and utilizeo
both on a laboratory and commercial basis for fine refuse
dewatering. Since fine refuse dewatering is conducted un-
derground, equipment height and required floor space are
important design criteria. Of the present state-of-the-art de-
watering components, the solid bowl scroll type centrifuge is
considered most beneficial from a height, floor space, and
performance basis.
The solid bowl centrifuge being considered for fine refuse
processing has a 36-m-diameter bowl measuring 96 inches
in length. Its success in achieving effluent clarity (1,000 ppm
or less solids) and its production of a 35- to 50-pct moisture
cake has been continually demonstrated
Pneumatic System
The pneumatic system has been designed to transport 100
tph of mining waste with 10 to 15 pet moisture for a distance
of 1,000 feet. Components that comprise this system include-
(1) Infeed conveyor, (2) airlock feeder, (3) hydraulic power
pack, (4) blower, (5) pipeline, and (6) deflector
The mfeed conveyor is a chain flight conveyor driven by a
variable speed hydraulic motor The conveyor speed is varied
automatically in response to the pneumatic system back pres-
sure. The feeder is an eight-pocket, rotary airlock feeder drive
by a slow-speed, high-torque hydraulic motor Hydraulic power
for both these components is provided by an electrically driven
hydraulic power pack. This power pack is equipped with relief
valves, controls, and pressure sensing switches for meas-
uring and correcting equipment overload.
The blower is electrically driven by a 750-hp motor and is
provided with acoustical protection such as filters, air cham-
ber baffles, and an acoustically clad housing. The pipeline is
10-inch Scd. 40 mild steel pipe capable of conveying 200,000
to 300,000 tons of semiabrasive shale Use of lightweight
pipes at the end of the pneumatic pipeline will facilitate pipe
movement
Evaluation of Disposal System
Monitoring Material Flow
During the first year of operation the waste disposal system
will be equipped with instrumentation for monitoring material
flows and system performance. All online continuous meas-
urements will be documented on a multichannel chart re-
corder. Those variables to be monitored and the type of
instrumentation which will be utilized are summarized in table
1.
Periodic sampling around individual components will pro-
vide further equipment and system evaluations Those com-
ponents and associated streams which will be periodically
analyzed include: (1) Solid bowl centrifuge-feed, cake, and
effluent, (2) dewatering screen—solids and drainage, and (3)
crusher-feed and product.
Sampled streams will be primarily analyzed for moisture
content and size distribution. As content. British thermal unit
value and sulfur content will be included in the analysis on
a less regular basis.
Effects on Mine Operation
In general, it appears that installation of the proposed dis-
posal system will not materially affect the overall operation
of the mine. System components located in the wash plant
and on the surface will be operated by the wash plant crew
as part of the normal operating shift. Underground equipment
will be operated by two additional miners. Productivity on an
overall mine basis will, however, be virtually unaffected as
-------
19
POLYMER
TANKS WITH
AGITATORS
POLYMER
PUMP
DECK DEWATERING
SCREEN
BLOWER
PNEUMATIC
PIPELINE
SOLID BOWL
CENTRIFUGE
FINE —^- -
AND COARSE >SJ'
REFUSE SLURRY ',
LINES i|
RETURN
WATER SYSTEM
PILLAR
FEEDER HYDRAULIC POWER RETURN CONTROL
PACK WATER SUMP
PUMP
Figure 6.—In-mine dewatering-pneumatic feeding system.
the two men operating the dewatering-pneumatic system will
be replacing mine-waste truck drivers.
The pipelines are small and will not interfere with coal
haulage. The dewatering-pneumatic feeding station will be in
a vacant crosscut; thus, it will not interfere with ventilation or
material transport. Within the disposal section the pneumatic
pipeline will be handled by the additional pneumatic equip-
ment operator and a part-time swing man from the mining
crew.
In-mine waste disposal at the HNM may improve coal re-
covery as present coal recovery in room and pillar operations
averages 50 to 75 pet. WSC anticipates future efforts with
longwalls will increase coal recovery to 85 or 90 pet.
Effects on Mine Personnel
Compared with shuttle cars, battery scoops, and convey-
ors, hydraulic haulage and pneumatic conveying are believed
to offer the safest method for transporting refuse. Waste is
TABLE 1.—System instrumentation
Variable
Coarse refuse slurry flow rate . . . _--_.-_
Fine refuse slurry flow rate . . . . - . . _ . .
Return water flow rate ... . . .
Coarse refuse slurry pump discharge pressure
Fine refuse slurry pump discharge pressure
Coarse refuse slurry concentration
Fine refuse slurry concentration '
Coarse refuse slurry pump power consumption . .
Pneumatic svstem power consumption . .
Instrument
Magnetic flow meter and ultrasonic flow meter
Watts transducers
-------
20
TABLE 2.—System operation measurements
Measurement
Physical:
Density of material as placed - . . - .
Roof support
Subsidence
Compressibility
Environmental
Dust
Noise .
Ventilation .. .....
Hydrology Water quality and flow
Method
In-place sampling
Surface measurements
In-place sampling
GCA-RAM 1 dust monitor
Sound-level meters
Anemometer
oH meter and laboratorv analysis onlv reouired if water discharae is sianificant
contained within a pipeline and no moving parts can interfere
with mining personnel.
Backfilling with pneumatics can generate dust in two ways.
The rock being conveyed is broken by impact on pipe tran-
sitions and in-line collisions, and high velocity air ejected from
the nozzle at the end of the pipeline can raise dust from the
roof, floor, and ribs. By combining pneumatics with hydraulic
haulage and controlling the dewatering before pneumatic
conveying (combined refuse moisture of 10 to 12 pet), refuse
generated dust can be substantially reduced. Dust created
by high-velocity air and solids impacting dry surfaces will have
to be controlled Dual split ventilation will most likely be used
to control the dust if necessary.
The pneumatic blower is a source of noise It is acoustically
clad and located around a corner from the dewatering-pneu-
matic feed control station. At the discharge nozzle of the
pipeline, the airstream is directed away from the operator and
when the system is fully loaded, the discharge is relatively
quiet. In any event, it is not necessary for the operator to be
in constant attendance at the discharge nozzle. Noise gen-
erated by the remaining underground equipment will be mon-
itored during Phase II. If noise levels are excessive, steps
will be taken to correct this situation
Effects on Mine Conditions
The mine may be affected by the backfill operation in three
separate categories: (1) Physical stability, (2) environmen-
tally, and (3) hydrologically.
During Phase II, specific measurements will be taken pe-
riodically before and during actual system operation These
measurements and methods of obtaining them are listed in
table 2.
Outlook for the Foster-Miller Associates/Western
Slope Carbon, Inc., Design
The information to be gained from this joint Bureau of Mines
and WSC program will benefit all underground coal produc-
ers. Operation of the combined system of hydraulics, pneu-
matics, and underground dewatering will provide operating
experience in most transport and stowing mechanisms which
can be utilized for underground disposal of mining waste. The
benefits of such a system are numerous. Environmental im-
provements will result in the elimination of:
1. Unsightly surface disposal sites
2. Smoke and fumes from burning waste heaps.
3. Dust brought into the air by wind erosion.
4. Muddy acid runoff water
In addition to these improvements, there are numerous
potential mining benefits'
1. Reduction in disposal costs.
2. Improved coal recovery and mine productivity
3. Improved roof control.
4. Reduction in subsidence.
The economics and benefits of underground disposal of
mining waste are site specific and will vary from mine to mine
The potential gams are encouraging and all underground coal
producers owe it not only to the general public but to them-
selves to evaluate the potential for disposing mining wastes
underground.
-------
21
FACTOR-OF-SAFETY CHARTS FOR ESTIMATING THE
STABILITY OF SATURATED AND UNSATURATED TAILINGS
POND EMBANKMENTS
by
D. R. Tesarik1 and P. C. McWilliams2
ABSTRACT
The factor of safety, the traditional measure of stability for
earth embankments, is presented graphically in factor-of-
safety charts. Factor-of-safety contours are drawn for ho-
mogeneous earth embankments for two contrasting situa-
tions: no phreatic surface and a phreatic surface that as-
sumes 10 percent freeboard. The factor of safety can be read
directly from the charts if the physical properties of the soil
and the geometry of the embankment are known. The curves
provide a quick first approximation of the "most stable" and
"least stable" condition of the embankment. The factor-of-
safety calculations were done by the Simplified Bishop
Method of Slices and a series of least-squares curve fitting
steps were used to present the factor of safety in final chart
form.
Introduction
The exact determination of the factor of safety for a soil or
tailings embankment generally requires an extensive physical
property sampling program and the use of a computer model.
Precalculated slope stability charts assuming homogeneous
materials and uniform conditions provide a quick and easy
method of approximating the factor of safety.
The Bureau of Mines has developed factor-of-safety charts
for estimating the stability of tailings pond embankments that
are easy to read, contain a wide range of slope angles, con-
sider pore water pressure, and cover a wide range of internal
friction angles.
This paper discusses slope stability charts in general, the
computer model that was used to calculate the factor of safety
from the Simplified Bishop equation, and the procedure used
to construct the charts. Several examples of how to use the
charts and the assumptions accompanying their use are in-
cluded. The charts (examples appear in the appendixes3)
provide estimates of stability for personnel responsible for
judging the stability of soil and tailings embankments in the
mineral industry.
Slope Stability Charts
Slope stability charts reduce a multidimensional problem
that includes the following listed parameters into a two-di-
mensional graphic display for quick and easy reference:
F = factor of safety.
•y = unit weight of the soil in Ib/ft3.
H = height of the embankment in feet.
<|>' or = internal friction angle in degrees".
c' or c = cohesion in Ib/in2 or Ib/ft2.
_ _u_ _ pore pressure ratio at a point, where u is
r" ~ ^h ~ the pore pressure at that point and h is
the depth of the point below the soil sur-
face.
3 = slope of the embankment in degrees,
sometimes expressed as increments in
the x and y directions respectively; for
example, 2:1 = 26.57°.
_ I. depth factor where L is the distance from
~ H the top of the embankment to the stiff
base (fig. 1).5
' Mathematician
2 Mathematical statistician
Both authors are with the Spokane Research Center, Bureau of Mines, Spo-
kane, Wash.
3 The complete set ot charts are in BuMines Rl 8564, "Factor of Safety Charts
for Estimating the Stability of Saturated and Unsaturated Tailings Pond Em-
bankments," (in press)
4 The prime (') symbol indicates that the parameter is in terms of effective
stress
5 Values of F for D = 1 25 were calculated but have been eliminated from the
charts Investgation of 2,409 values of F shows that values of F for D =
1 50 are more critical (than for D = 1.25) 81 percent of the time When
values of F for D = 1 25 are more critical, the difference is usually in the
third decimal place
-------
22
Trial circle
center
Control grid
stiff base-
Figure 1.—A typical embankment cross section and
failure surface.
Fixed parameter:
Slope
0 , degrees
Figure 2.—Factor of safety, F, contours, Singh for-
mat.
Previous authors have combined several parameters into
different forms such as c'F-yH (V[),6 c/^Htan* (2), or cfyH
((3-10). From the available literature reviewed, the chart for-
mat presented by Singh (W) for dry embankments is probably
the easiest format to use if the factor of safety is desired,
given that all other parameters are known (fig. 2). The charts
in this publication employ a "Singh-like" format for both dry
and saturated embankments.
grid, the program computes a minimum radius; that is, a
radius that will just touch the slope and a maximum radius
based on the geometry and boundary conditions of the slope
profile. A series of factors of safety is computed starting with
a radius slightly smaller than the maximum radius. Each suc-
cessive radius is reduced by approximately one-eleventh of
the difference between the minimum and maximum radii (1_).
Data Processing
The Simplified Bishop Equation
The computer program (1J used to determine the minimum
factor of safety for given slope conditions is based on Bishop's
Simplified Method of Slices. The program divides the slope
cross section into slices. The number of slices is approxi-
mately equal to the number input by the user. This number
may differ slightly from the final value used because of the
boundary conditions used by the program. Eighty slices were
input for the analysis used in this paper. The failure surface
is assumed to be an arc of a circle with a minimum penetration
depth equal to 20 percent of the embankment height (for
example, 20 feet for a 100-foot-high embankment).
The slice boundaries are determined before any trial circles
are analyzed, so the number of slices in each sliding mass
varies with the radius of the trial circle. A sample failure sur-
face and slices are shown in figure 1.
For each set of parameters input, slope geometry, 4>', c',
and -y, 1,331 trial circles were examined to determine the
critical circle. This was accomplished by using the control
grid option of the program (fig. 1). At each point on the control
6 Underlined numbers in parentheses refer to items in the list of references
preceding the appendixes
The equation which is solved to determine F is the sim-
plified Bishop equation (equation 1). Slice geometry and the
forces on a typical slice are shown in figure 3
F =
1
X W, sin a,
+ W,(l-rjtan
(1)
sec a,
1 +
tan ' tan a,
where
c' = cohesion of soil,
b, = breadth, ith slice,
-------
23
Trial-
circle \
center \
Fixed parameters
-Phreotic
surface
bi = Breadth
Ej,Ej,| = Horizontal forces
Xj,Xj,.| = Resultant vertical shear forces
W| = weight
Si = Shear force
Pi =Total normal force
' are held constant, then the factor of safety i's a function
I —
o 10 20 30
FACTOR OF SAFETY, F
Figure 4.—c'/-yH expressed as a function of factor
of safety, F, for (j>, slope and depth factor, D, fixed.
of the geometry of the embankment and sliding mass (4).
This is seen by multiplying equation 1 by (H-H) (H-H) and
expressing W, as yb,-h, The resulting equation is
F —
1
b, h,
seca,
(4)
tan <)>' tan a,'
Note that ru, = 0 for embankments containing no water By
inspecting equation 4, it is obvious that the critical terms F,
cV-yH, and ' are rather entwined, and some manipulation
is required to "free" these quantities. In order to express c'
-yH as a function of H = P0 + P,F + P2F2, where P0, P,, and P2 are
the equation's coefficients as shown in figure 4. Values of c'
-/H were chosen as 0.0, 0 025, 0.05, 0.075, 0.1, 0.2, and 0.3.
The values in the range of 0.0 « cV-yH =s 0.1 were chosen
to be consistent with previously published literature (4, 9). In
order to attain a factor of safety of at least 3.0 (a predeter-
mined minimum F to display on the charts), the ratio was
extended to include 0.2 and 0 3 In some cases, values of 0 4
and 0.6 were used.
The information contained in the family of plots of figure
4 was then rearranged in the more convenient format of
"Singh-charts." Factor-of-safety values were contoured with
the dependent variable still c'/-yH, but *' became the new
independent variable. The rearrangement of information was
achieved by:
1. Fixing F at a desired contour value (say F = 1!1).
2. Selecting 4>' to correspond to a member of the "'-fixed"
-------
24
05
r cV? H = P0 + P,F+P2F2
10
20
30
' , degrees
40
50
Figure 5.—Factor of safety contours with slope and
depth factor as overriding parameters.
graphs of figure 4 and substituting the F-contour value into
the quadratic equation, giving the appropriate c'/yH value.
The F-contour value is this ', c'/-yH pairing.
3. Varying <(>' over its domain of values, giving a complete
curve for this F-contour.
4. Setting F = "next desired contour value" and repeating
steps 1 through 3. Thus, the final product (fig. 5) is a Singh
chart with fixed overriding parameters of slope and depth
factor, D.
Construction of the Factor-of-Safety
Contours with a Phreatic Surface
To construct charts containing pore pressure, the value
ru in equation 4 must be considerd since it no longer has a
value of 0. For F to depend on geometry alone, ru must be
constant along with the ratio c'/-/H. The following alternatives
were considered in dealing with pore pressure.
1. Let the user calculate an average ru for the entire em-
bankment. Bishop and Morgenstern have devised an aver-
aging technique that has been shown to be accurate in many
problems (4).
2. Assume a fixed phreatic surface in which the headwater
is a distance of H/10 feet from the top of the embankment—
called 10 percent freeboard (fig. 6). Recall that ru = u/(-y-h,)
= 62.4-d,-cos2 6/(rh,). By holding -y constant and substituting
62.4-d,-cosa 6/(-y-h,) into equation 4, it is seen that F is a
function of the geometry of the embankment and sliding mass
alone, provided that the ratio c'/H is also held fixed.
Regarding option 1, it was found that for the case of ho-
mogeneous embankments with a phreatic surface due to
steady state seepage, the values of F obtained when using
an average rb value differed, in some cases, significantly from
the value of F obtained when ru look on its corresponding
values for each slice. Further, the value of F was sometimes
quite sensitive to small changes in ru, necessitating the cal-
culation of ru to the nearest hundredth place for accuracy.
This would require either a large number of charts with ru in
increments of 0.01 or interpolation by the user.
Because of these reasons, and in order to eliminate the
calculation of an average rut the second alternative—to make
y a fixed parameter and limit the steady-state seepage charts
to the condition of 10 percent freeboard—was chosen.
The method used for generating the curves containing pore
pressure is the same as described in the preceding section,
except that c'/H is now plotted on the y axis instead of cV-yH
since -y is now a fixed parameter. The location of the phreatic
surface for input into the Bishop computer code was deter-
mined by running a finite-element program. Figure 6 illus-
trates the phreatic surface for an homogeneous slope of 2.5
to 1.
A natural question is whether linear interpolation is valid
for the charts—for example, if y = 95, what does one do?
Because of the discreteness of the Bishop process, one does
not find a smoothness between points. Thus, interpolation
should be done with caution, particularly if F varies signifi-
cantly from chart-to-chart.
Using the Slope Stability Charts
FOR EMBANKMENTS WITH NO PHREATIC
SURFACE
If the factor of safety is desired for an embankment with
no phreatic surface, the following steps are necessary:
1. Calculate c'/-yH.
2. Determine the slope for the embankment and if the em-
bankment is constructed on a stiff base.7 If the embankment
is on a stiff base, use the charts such as shown in appendix
A. If the stiff base is approximately H/2 feet below the em-
bankment, use the charts such as shown in appendix B.
3. Locate the appropriate chart by using the information in
the upper right-hand comer of the chart.
4. Find where the ordered pair d/, c'/-/H intersects the
factor-of-safety contours. This is the critical factor of safety
desired.
For Embankments With 10 Percent
Freeboard
1. Calculate c'/H.
2. Determine the slope for the embankment, the density
of the soil, and if the embankment is constructed on a stiff
base. If the embankment is on a stiff base, use the charts
such as shown in appendix C. If the stiff base is approximately
7 A stiff base is assumed when the shear strength of the material on which the
embankment is constructed is higher than the shear strength of the material
in the embankment
-------
25
/././// y
Figure 6.—Phreatic surface computed by the finite-element method.
H/2 feet below the embankment, use the charts such as
shown in appendix D.
3. Locate the appropriate chart by using the information in
the upper right-hand corner of the chart.
4. Find where the ordered pair <$>', c'/H intersects the factor-
of-safety contours to determine the critical safety factor.
Example 1
The following physical parameters are part of the data col-
lected by the Bureau of Mines from West Virginia coal refuse
embankments (5).
Y««, = 90.74 Ib/ft3
y,,^ = I0l.i2lb/ft3
c' = 3.2 Ib/in2,
V = 33.48°.
If an embankment 100 feet high with slope 1.5 to 1 and no
phreatic surface is to be constructed on a stiff base, the factor
of safety is obtained as follows:
100ft = 0.051.
2. The slope is 1.5 to 1 and the embankment is to be
constructed on a stiff base.
3. The appropriate chart has 1.5 to 1 in the upper right-
hand corner (fig. A-2). The ordered pair ($', c'/yH) = (33.48°,
0.051) lies close to the midpoint between the contour lines
of F =-1.6 and F = 1.8, yielding a factor of safety of ap-
proximately 1.7
To verify the result, program Bishop was run for this case.
The result (F = 1.729) verifies the factor of safety obtained
via the charts.
Example 2
Suppose an embankment with the same physical proper-
ties and of the same geometry as in example 1 is to be
constructed; however, a phreatic surface at 10 percent free-
board is anticipated. The following steps are taken:
-J- I 144^1 , 100ft = 4.61£
2. The slope is 1.5 to 1 and the density is rounded to 100
IbJ
ft3'
3. The appropriate chart has y = 100, and 1.5 to 1 in the
upper right-hand corner (fig. C-2). The factor of safety cor-
responding to the ordered pair (33.48°, 4.61 7-) is approxi-
mately 1.18.
The results obtained from running the computer code for
this specific case yield 1.183 as a factor of safety. The dif-
ference between the chart value for F and the value obtained
from the individual computer run can be attributed to a com-
bination of the following:
1. Variation introduced by the procedure used to construct
the charts.
Ib
2. Variation introduced by using 100 — instead of 101.12
Ib
ft3'
Note the impact of using the phreatic surface—the factor of
safety is substantially reduced from 1.73 to 1.18.
Summary
The factor of safety representing critical circles for homo-
geneous earth slopes is displayed in an easy-to-use plot form
for dry embankments and saturated embankments having 10
percent freeboard. The factor of safety can be read directly
from the charts if the physical properties of the soil and ge-
ometry of the embankment are known. This form was
achieved by organizing the factor-of-safety values using a
8 A comparative computer run was made using 90 74 Ib/ft3 for density above
the phreatic line and 101.12 Ib/ft3 for density below the phreatic line The
factor of safety was 1.205.
-------
26
series of least-squares curve fit routines. The soil parameters
and slope boundaries include:
Slope 1:1, 1.5:1, 2:1, 2.5:1
5:1
Effective angle of internal 0° to 50°
friction, '
cV-yH (dry embankments) 0.0 to 0.6 (dimensionless)
c'/H (wet embankments 0.0 to 80.0 (Ib/ft3)
with 10 percent
freeboard)
Depth factor, D 1.00,1.50
The charts should be a useful tool in initial embankment
design or field work where detailed stability analysis is not
imperative. However, the charts are not meant to replace the
use of slope stability models when time and resources are
available.
References
1. Bailey, W. A. Stability Analysis by Limiting Equilibrium.
C. E. Thesis, Massachusetts Institute of Technology, Boston,
Ma., 1966, Appendix C, pp. 64-68.
2. Bell, J. M. Dimensionless Parameters for Homogeneous
Earth Slopes. Soil Mech. and Foundations Div., ASCE, v. 92,
No. SMS, September 1966, pp. 51-65.
3. Bishop, A. W. The Use of the Slip Circle in the Stability
Analysis of Slopes. Geotechnique, v. 5, No. 1, 1955, pp.
7-17.
4. Bishop, A. W., and N. R. Morgenstern. Stability Coeffi-
cients for Earth Slopes. Geotechnique, Institution of Civil
Engineers, v. 10, No. 4, 1960, pp. 129-150
5. Busch, R. A., R. R. Backer, and L. A. Atkins. Physical
Property Data on Coal Waste Embankment Materials. BuMmes
Rl 7964, 1975, 142 pp.
6. Cousins, B. F. Stability Charts for Simple Earth Slopes
Geotech. Eng. Div., ASCE, v. 104, No. GT2, February 1978,
pp. 267-279.
7. Morgenstern, N. R. The Department of Civil Engineering,
The University of Alberta, Edmonton, Alberta, Canada Pri-
vate communications, February 1979. Available upon request
from D. R. Tesarik, Spokane Research Center, Bureau of
Mines, Spokane, Wash.
8. O'Connor, M. J. EBA Engineering Consultants, Ltd ,
Calgary, Alberta, Canada. Private communications, April
1979. Available upon request from D. R. Tesarik, Spokane
Research Center, Bureau of Mines, Spokane, Wash.
9. O'Connor, M. J., and R. J. Mitchell. An Extension of the
Bishop and Morgenstern Slope Stability Charts. Can. Geo-
tech. J., v. 14, No. 1, 1977, pp. 144-151.
10. Singh, A. Shear Strength and Stability of Man-Made
Slopes. J. Soil Mech. and Foundations Div., ASCE, v. 96,
No. SM6, November 1970, pp. 1879-1892.
11. Spencer, E. A Method of Analysis of the Stability of
Embankments Assuming Parallel Inter-Slice Forces. Geo-
technique, v. 17, No. 1, 1967, pp. 11-26.
-------
Height of
embankment, ft
Density,
Ib/ft3
Effective
cohesion of
soil, Ib/ft2
KEY DIAGRAM
27
Slope of embankment
40
35
30 -»
25
20
15
10
Height of embankment
0
6 24
, degrees
32
"Effective angle of internal friction
Figure A-1.—Example stability charts for embankments with no phreatic surface and depth factor = 1.0.
-------
28
0.50
.HO
0', degrees
Figure A-2.—Factor of safety—1.5:1 slope, no phreatic surface, D = 1.00.
-------
Height of
embankment, ft.
Density,
Ib/ft3
Effective
cohesion of
soil, Ib/ft2
40
35
30
25
20
15
10
KEY DIAGRAM
29
Slope of embankment
Height of embankment
16 24
', degrees
32
40
•Effective angle of internal friction
Figure B-1.—Example stability charts for embankments with no phreatic surface and depth factor = 1.50.
-------
30
0.60
u
10
20 30
', degrees
50
Figure B-2.—Factor of safety—1.0:1 slope, no phreatic surface, D = 1.50.
-------
31
KEY DIAGRAM
Slope of embankment
Free water surface
40
Height of
embankment, ft
Effective
cohesion of
soil, Ib/ft2
I/IO Height of
embankment
16 24 32 40
', degrees
•Effective angle of internal friction
Figure C-1.—Example stability charts for embankments with a phreatic surface associated with 10 percent
freeboard and depth factor = 1.00.
-------
32
50
40
30
20
10
0
20 30
, degrees
40
50
Figure C-2.—Factor of safety—1.5:1 slope, phreatic surface with 10 percent freeboard, density = 100 Ib/
ft3, D = 1.00.
-------
33
KEY DIAGRAM
-Height of
\embonkment.
ft.
Effective
cohesion of
soil, lb/ftz
40
35
30
25
= 20
15
10
Slope of embankment
Free water surface
40
I/IO Height of
embankment
-Effective angle of internal friction
Figure D-1.—Stability charts for embankments with a phreatic surface associated with 10 percent free-
board and depth factor = 1.50.
-------
34
35
30
25
20
15
10
20 30
', degrees
Figure D-2.—Factor of safety—1.0:1 slope, phreatic surface with 10 percent freeboard, density = 70 Ib/
ft3, D = 1.50.
-------
35
PROBABILISTIC APPROACH TO THE FACTOR OF SAFETY
FOR EMBANKMENT SLOPE STABILITY
by
P. C. McWilliams' and D. R. Tesarik2
INTRODUCTION
The Bureau of Mines is investigating the risk analysis, prob-
abilistic approach to computing the factor of safety for the
following reasons:
1. The Federal Mine Safety and Health Administration
(MSHA) has imposed slope stability analysis as a require-
ment for the construction of coal tailings embankments (9).3
2. The President's Federal Coordinating Council for Sci-
ence Engineering and Technology reported on Improving
Federal Dam Safety in 1977. A prime research goal was to
pursue a risk analysis approach to dam design
3. To resolve inconsistencies between the calculated factor
of safety and the corresponding success or failure of the
embankment.
In fact, theoreticians in soil mechanics have been pursuing
a probabilistic approach to the factor of safety of an em-
bankment or dam for the past 10 years. The motivation for
this work is in contrast to the current practice of determin-
istically computing a factor of safety and treating it as an
absolute with no regard of its inherent statistical variability
Stated simply, the applied soils engineer works with one pa-
rameter—factor of safety—while the theoretician would im-
pose an additional parameter; the dispersion of the factor of
safety. Given these two parameters, one may then compute
the probability of failure as a measure of slope stability Fur-
ther, the two-dimensional modeling of Fellenius, Bishop, et
al. (5) is challenged by a three-dimensional approach which
postulates a cylindrical slip surface (10). Not surprisingly,
three-dimensional modeling increases costs, for extensive
soil sampling and laboratory testing are requisite to the re-
quired analytic techniques. A standard deviation of the factor
of safety can be computed for both the two-dimensional
model (via propagation of error) and the three-dimensional
model. Requisite to these computations is some knowledge
of the variability of the soil parameters—internal angle of
friction, cohesion, and soil density
1 Mathematical statistician
2 Mathematician
Both authors are with the Spokane Research Center Bureau of Mines.
Spokane, Wash
3 Underlined numbers in parentheses refer to items in the list of references at
the end of this report
The question remains as to whether implementing these
techniques is justifiable from both a scientific and economic
view. A final point, the practitioners currently using determin-
istic two-dimensional analysis indirectly make allowances for
the variability of the factor of safety by using conservative,
rather than average, estimates of the soil parameters.
Background
For at least the past 10 years, geotechnical journals and
books have been advocating an increased usage of proba-
bility and statistical techniques (1, 5, 10-1_1). In particular,
the factor of safety, the current measure of the stability of an
earth slope, has been challenged as to its adequacy. Slopes
with a seemingly "safe" factor of safety have failed, while
others whose factor of safety was at best minimal have not
failed; thus, there is some validity in remvestigatmg the prob-
lem. The real question is—does a better technique now exist
that should replace the current factor-of-safety criteria for
embankment design acceptance?
The factor of safety, F, is defined as-
F =
Resisting Moment
Overturning Moment
— of a circular failure mass
(fig. D.
Note that the two most popular techniques for computing
the factor of safety both utilize the two-dimensional slip circle.
Fellenius (or the Swedish method) first introduced this con-
cept, most practitioners today use the modified Bishop-Mor-
ganstern algorithm (5-6).
After considerable deliberation, this work was restricted to
the probability aspect only; leaving the area of risk analysis
for others to pursue. It was felt that introducing the economic
arguments is premature at this time. Further, the probability
work provides the necessary basis for continuation into risk
analysis, if so desired.
-------
36
Radius
Distribution of F
W|= weight of i slice
N = normal force, i slice
i
S ^resisting shear force
Figure 1.—Circular failure mass.
Population Distributions and the Probability
of Failure
The Bureau has done considerable research relative to the
factor of safety. The problems are by no means resolved at
this time, but much interesting work has been done and is
worthy of consideration. The fundamental point is that prac-
titioners have relied on a single quantity—factor of safety—
as the measure of the stability of an embankment. The factor
of safety obtained from field sampling is, however, an esti-
mate of the true factor of safety, and one should consider the
concept of a population of factor-of-safety values, with a
mean, F, and a standard deviation, F.4 If one is skeptical
regarding a population of safety factors, reference is invited
to papers by Lee and Singh (7) who sent the same soil sam-
ples to 28 different soils laboratories in the Los Angeles area.
The coefficient of variation5 for cohesion, c was 20 to 40 pet;
for the internal angle of friction, <{>, from 10 to 20 pet, and for
soil density, y, from 0 to 5 pet. Besides laboratory variability,
there is the natural variability inherent in the soil itself. Thus,
many authors are proponent of a probability of failure ap-
proach.
Since the embankment should fail if the factor of safety
S 1.0, the probability of failure equals the area of the pop-
ulation distribution to the left of F = 1.0 (fig. 2):
It is worth noting that current practitioners indirectly account
for the variability in F by using conservative estimates for
input, giving a conservative (F) value fo the factor of safety.
Thus, in meeting a requirement such as "factor of safety must
exceed 1.5," the industry uses F rather than F; thus effectively
sliding the population of the right, and thus decreasing the
probability of failure as illustrated in figure 3.
All of the preceding is fine, but the theory is only of practical
' To avoid notation confusion, statisticians use "s" for standard deviation and
soil engineers use "s" for strength—deference is made to soils engineers
Thus, s = soil strength and F = standard deviation of the factor of safety
Probability of failure
F = 1 0
Figure 2.—Concept of probability of failure.
value if the population assumptions are valid. Next, consider
the practicality of obtaining F and determining the population
of F.
Propagation of Error for Fellenius and
Bishop Equations
Now F can be computed from either Bishop's equation or
from Fellenius' equation. Since Fellenius' equation has F in
a closed rather than iterative form, it is chosen for illustration:
cL + tan
("/ A cos e, - u,
y X-A, sin 9,
and computation of the variance of F gives:
+ F*
Newly defined terms are:
covariance terms.6
variance = square of standard deviation,
-y = soil density,
A, = area of ilh slice,
u, = pore pressure for the i'h slice,
cs, $?, 72, u? = variance of soil parameters, and
F, = partial derivatives with respect to soil pa-
rameters.
Due to independence of most of the parameters or to the
relative small magnitudes of the resulting terms, the covari-
ance terms can be ignored, leaving the preceding first-order
approximation for F2. It is important to note that to find F, one
must input the variances of the soil parameters. Good esti-
mates for c, ~4>, and "y are usually available. Harder to estimate
are the u, values. The Bureau has incorporated the preceding
formulation in the modified Bishop computer program.
5 Coefficient of variation =
standard deviation
mean
.100.
6 F2 was computed by statistical propagation of error techniques (3).
-------
37
Indirect shifting
of distribution
by using r rather
than ? as criteria
Probability of failure
f >15
Figure 3.—Current practice of establishing factor
of safety.
Possible Reduction of Factor-of-Safety
Requirements
Knowing F may relax the requirements for the minimum
acceptable factor of safety. Consider two alternative designs
with the following statistics:
F-value
f
Pr (failure)7
Design one
1.60
.30
2.2 pet
Design two
1.40
.15
.4 pet
Without knowledge of F, by using today's deterministic ap-
proach and accepting F greater than 1.5, design one would
have to be chosen. But using probability of failure as the
criteria, design two is deemed best. For, from the perspective
of probability of failure, design two is five-times less likely to
fail as is design one. Thus, acknowledging that variability in
the factor of safety exists and using that quantity accordingly
could result in a change of criteria that many practitioners
would favor.
Sensitivity Analysis
An ongoing effort is to measure the dependency of the
standard deviation of the factor of safety, SF, on the respective
standard deviations of <)>, c, y, and u. Although the sensitivity
analysis is not complete, changes in soil density, i.e., through
compaction, seem to be prominent in reducing the factor of
safety. Final documentation of this project will, of course,
include a complete sensitivity analysis and corresponding
recommendations.
Population Estimates from Field Data
A direct approach to finding the population distribution of
F was to sample a tailings embankment and actually generate
a family of F-values. Due to funding consideration, this ex-
periment was conducted in a simulated sense; that is, tailings
7 Probability of failure assumes a normal distribution for F.
material contained by a very stable earthen embankment was
sampled. The conjecture was what if the contained tailings
material was used as an embankment wall; what would the
stability picture be? The stability of the pond tested was not
addressed, the important point was to determine the shape
and dispersion of the population F-values; for example, an-
swer the question as to whether one can assume a normal
distribution, a log-normal distribution, or what, if any conven-
tional distribution does fit the situation?
In order to generate the necessary population distributions,
some 50 Shelby tubes were used to gather the tailings ma-
terial. Direct shear tests were run to obtain a family of , c,
and 7 values. For each data input, a factor of safety was
computed via the modified Bishop program. In all, 74 factor-
of-safety values were derived for the simulated embankment.
No prior work of this kind, carrying the work to the concluding
distribution of the factor of safety, was found.
Figure 4 and summary statistics represent the data.
Summary statistics are:
Sample mean = 1.53
Sample standard
deviation = 0.40
Sample size = 74
Probability of failure =
8/74 ~ 11 pet
Maximum F = 2.35
Maximum F = 0.56
Coefficient of variation =
26.1 pet
The following conclusions would seem in order:
1. The distribution does show central tendency; thus, nor-
mal-distribution theory may be appropriate.
2. The variability is large, as seen by the large probability
of failure (again, this is a simulated case).
Note that it is not suggested that the industry will ever be
required to draw large populations of F-values; rather this is
background work requisite to using F properly. As mentioned,
the preceding is an artificial case; the Bureau intends to re-
peat the experiment in a more realistic environment
20
15
I 10
o
LLt
oe.
0
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2 2 2.4
FACTOR OF SAFETY
Figure 4.—Histogram of F-values for a simulated
embankment.
-------
38
Three-Dimensional Modeling
Another major thrust has been to change the factor of
safety to a three-dimensional cylindrical surface model. Ang,
Cornell, and Vanmarcke (2, 10) have written a series of ar-
ticles over the past 10 years on this subject. A very brief
sketch of this rather complex technique follows:
1 . The basic premise that failures occur in three dimensions
not two, thus a cylindrical surface is a better model to pos-
tulate.
2. The factor of safety is still defined as the ratio of the
resisting to the overturning moment (fig. 5).
3. The ensuing formula for the factor of safety is:
_ =
b M0 ~
suLrb
Wab
where:
Mr
M0
R9
= resisting moment,
= overturning moment,
= average design value of undrained shear
strength,8 and
= contribution of end sections of failure mass
to the resisting moment.
4. By propagation of error the preceding formula for Fb
produces:
subLr
Wa
where:
8 The analysis can also be performed in terms of effective stress
FB = standard deviation of the three-dimensional factor
of safety, and
sub = standard deviation of the shear strength in the
embankment.
5. In order to find su and sub, one must measure strengths
in three dimensions along the embankment Further, Van-
marcke QO) specifies that the usual variance (and thus stand-
ard deviation) computations for strengths must be reduced
to account for the high intercorrelation between adjacent
strength readings.
6. The variance reduction procedure requires an iterative
search for "best" spacing of borings on the embankment
surface. Average values of strength and end resistance, plus
autocorrelation values for variance reduction work, are input
for computation of Fb and its standard deviation.
7. Finally, the modeling finds a critical length, bc, which is
the most likely failure length for the hypothesized cylindrical
model.
Failure mass
A= Area
Geometry of failure mass
Cross section, failure mass
B= length of embankment L=arc length of failure surface
b = length of failure mass r = radius of failure mass
Re = resistance on end sections W= weight of failure mass per
of failure mass unit length
a = horizontal distance from trial
center "O" to center gravity
of the failure mass
Figure 5.—Three-dimensional failure mass showing geometry of failure mass and cross section of failure
mass.
-------
39
The preceding outlines the procedure being field tested by
contractors for the Bureau of Mines. Besides usual borehole
work, a Dutch cone penetrometer was used to find the nec-
essary supplementary borings. Preliminary results are en-
couraging; a final report will be forthcoming at the project's
termination in the fall of 1981. It is not fair to make rigid cost
comparisons at this time, for the initial field work is experi-
mental and not production efficient. However, it will cost more
to obtain field data of this detail—the question is, of course,
is the product worth the expenditure?—a question for the
future.
In deference to the preceding work, the cylindrical model
may be better than the slip circle model, but the cylindrical
model has shortcomings too; which one would be amiss to
ignore. The "real-world" failure mechanism need not be a
rotating cylinder either; it has been hypothesized that a model
of a "sliding trough" with probability assigned on a particle
basis would be better modeling than either the slip circle or
the cylindrical model (a three-dimensional random-walk con-
cept (4)). So the future will undoubtedly bring more evolution
of ideas regarding "best" modeling for embankment slope
stability.
Summary
The following represent current state of the art of the factor-
of-safety work as perceived by the authors:
1. A strong conviction that some measure of the variability
of the factor of safety should be used—ranging from some-
thing relatively simple, such as the coefficient of variation, to
the probability of failure approach.
2. It is not yet established whether three-dimensional mod-
eling should replace the current two-dimensional slip-circle
modeling.
a. The theoretical work will and should continue—it's
neither complete at this time nor has it been honed as
a practical field tool.
b. At this time there is no justification for recommending
a change in MSHA's current criteria. Only when it can
be proven that failed embankment would have been
properly predicted by three-dimensional modeling will
there be universal acceptance for a model change.
c. However, a designer with real doubts about an em-
bankment's safety would be advised to use the three-
dimensional modeling now.
3. The Bureau will continue its pursuit of both major thrusts;
inclusion of the variability of the factor of safety and inves-
tigation of three-dimensional modeling.
4. The Bureau of Mines has, available for distribution, a
computer program which outputs the usual factor of safety
(Fellenius, Bishop) plus the standard deviation of the factor
of safety using first-order approximations. To use this added
feature, the user must input estimates of the standard devia-
tions of the soil parameters. The program can be obtained
by contacting the Spokane Research Center.
References
1. Alonso, E. E. Risk Analysis of Slopes and Its Application
to Slopes in Canadian Sensitive Clays. Geotechnique, v. 26
No. 3, 1976, pp. 453-472.
2. Ang, A. H-S., and C. A. Cornell. Reliability Bases of
Structural Safety and Design. J. Structural Div., ASCE, ST9,
Sept. 1974, pp. 1755-1769.
3. Deming, W. E. Statistical Analysis of Data. Dover Pub-
lishing Co., New York, 1964, pp. 37-48.
4. Feller, W. Probabilistic Theory and Its Application. J.
Wiley & Sons, Inc., London, 1950, pp. 279-306.
5. Harr, M. E. Mechanics of Paniculate Media. McGraw-
Hill Book Co., Inc., New York, 1977, pp. 427-442.
6. Jubenville, David, Chen, and Associates. Limit Equilib-
rium Slope Analysis and Computer Software. Denver, Colo.,
1978, pp. 1-16.
7. Lee, K. L., and A. Singh. Report of the Direct Shear
Comparative Study. Soil Mechanics Group, Los Angeles Sec-
tion, ASCE, November 1968, pp. 1-38.
8. Riggs, J. L. Engineering Economics. McGraw-Hill Book
Co., Inc., New York, 1977, pp. 469-537.
9. U.S. Code of Federal Regulations. Title 30—Mineral
Resources; Chapter I—Mine Safety and Health Administra-
tion, Department of Labor; Subchapter 0—Coal Mine Health
and Safety; Part 77—Mandatory Safety Standards, Surface
Coal Mines and Surface Work Areas of Underground Coal
Mines; Subpart C—Surface Installations; Sections 77.215(h)
and 77.216-2(a)(13). Federal Register, v. 40, Sept. 9, 1975,
p. 41776-41777.
10. Vanmarcke, E. Reliability of Earth Slopes. J. Geotech.
Eng. Div., ASCE, v. GTII, November 1977, pp. 1247-1263.
11, Yong, R. N., E. Alonso, and M. M. Tabba. Application
of Risk Analysis to the Prediction of Slope Stability. Can.
Geotech. J., v. 14, No. 540 1977, pp. 1-16.
-------
40
APPLICATION OF REMOTE SENSING FOR COAL WASTE
EMBANKMENT MONITORING
by
C. M. K. Boldt1 and B. J. Scheibner2
INTRODUCTION
Remote sensing, as related to mine waste embankment
monitoring, is defined as the gathering of information without
direct human contact. In this report two forms of remote sen-
sing were studied; instrument data gathering by satellite and
aerial photogrammetry.
The speed of technological advance is incredible. Over 140
years ago the use of photography was first documented; 19
years later in 1858, a ballon floating several hundred meters
over Paris took the first aerial photograph. It was not until
early World War II that aerial photography was considered
more than flights of fancy, when the Germans were believed
to have relied heavily on information gathered from inter-
preting air photos taken of the entire Western Front every 2
weeks (1).3
The first U.S. satellite, Explorer I, was launched on January
31, 1958. As early as 1960, satellites were being used to
transmit weather data, and in 1962, the first 'telecommuni-
cations satellite was put into orbit. Since these early days,
satellites have achieved an ever-increasing role in our lives,
and have become more and more diversified in their uses.
The launch of the first Earth Resources Technology Satellite
(ERTS-1), now the Landsat series of satellites, ushered in a
new type of technology—that of managing and inventorying
earth resources by merging space and remote-sensing tech-
nologies. Not only is the satellite capable of mapping large
areas through photography, but it can also relay information
from point to point on the earth's surface. In May 1974, the
Synchronous Meteorological Satellite/Geostationary Opera-
tional Environmental Satellite (SMS/GOES) was launched,
the first of a series. This is a no-charge operational system
allowing any user to request access from the National Oceanic
and Atmospheric Administration (3).
Since the launch of these satellites, they have been used
for various studies including monitoring weather conditions
in remote, high mountain areas (2), automatic data collection
1 Civil engineer.
2 Geologist. Both authors are with the Spokane Research Center, Bureau ot
Mines, Spokane, Wash.
3 Underlined numbers in parentheses refer to items in the list of references at
the end of this report.
from hydrologic stations (4), and monitoring global volcanic
activity (6).
General
The mining of earth resources is a geographically inflexible
industry in that ore is where you find it, not where it is con-
venient, environmentally favorable or profitable to mine. A
hundred miles of winding mountain roads can separate one
mine from another or they can be so close together their
waste piles touch. There are over 6,000 coal mines through-
out the United States and the Department of Labor, Mine
Safety and Health Administration (MSHA), is one agency
required by law to inspect them. Improving the inspection and
procedures and decreasing the workload by using remote
sensing techniques is the objective of this research.
The Bureau of Mines has two projects in the remote sen-
sing field as it pertains to mining. One of these is a recently
completed contract that studied the use of aircraft mounted
cameras and the results of photogrammetry, or obtaining
measurements by use of photography. The second project
is an ongoing contract to study the effectiveness of using an
instrumentation system wired to remote data collection sta-
tion to record embankment data.
Aerial Monitoring
The first project, an aerial monitoring contract (J0188027),
was completed by Chicago Aerial Survey (CAS) of Des Plaines,
III., titled "Improving Surface Coal Disposal Site Inspections."
The objective was to determine and expand the data collec-
tion capabilities of aerial photogrammetry on actual coal waste
embankments. A previous Bureau of Mines project titled "Rapid
Monitoring of Coal Refuse Embankments" by CH2M Hill, which
studied an active landslide and two coal waste embankments
had indicated aerial reconnaissance could be used to monitor
embankment movements and changes (5). Table 1 lists major
differences between the CAS and CH2M Hill projects.
-------
41
TABLE 1.—Differences between CAS and CH2M
Hill projects
CAS
Used vertical photography.
One flyby necessary per site photo-
graphed.
Accuracy of ± 0.3 foot with modifi-
cation of elevation reading net-
work.
Targets were set adjacent to the
embankment as bench marks
with elevations read anywhere on
the embankment.
Monitored 15 coal waste sites in
West Virginia and Kentucky be-
tween March and December
1979.
CH2M
Used convergent and vertical pho-
tography combination.
Three flybys necessary per site
photographed
Accuracy of ±0.15 foot.
Targets were set on the embank-
ments and read directly for eleva-
tions only at that point
Monitored two coal waste sites in
West Virginia monthly between
July and December 1979
Chicago Aerial Survey and its subcontractor, Dames &
Moore, monitored 15 coal waste sites in West Virginia and
Kentucky each month for 10 months (fig. 1).
Black-and-white and color-infrared (CIR) photographs were
taken and analyzed for each site by CAS while Dames &
Moore verified the conditions on the ground and solicited
industry and MSHA inspector comments. Distortion in the
photographs was corrected to produce a scalable photograph
or orthophoto. From these, computer assisted stereoplotters
were able to pick off elevations at predetermined points form-
ing a 100-foot-grid pattern over the embankment face, crest,
and impoundment. CIR photographs were taken seasonally
to compare resolution and clarity with those of black-and-
white photographs.
Major conclusions at the end of the project showed aerial
photogrammetry (fig. 2) is useful as a supplement to existing
MSHA inspection procedures. The interpretation of the or-
thophotos can offer cross sectional profiles, volume esti-
mations, vertical movement, and objective documentation over
the life of a mine. Seepage areas, erosion gullies, diversion
systems, and overall views of the site can be examined first
on the photographs then in an organized pattern over the
waste site. It is believed this type of monitoring can be useful
on high-hazard embankments when the inspector and mine
personnel need more quantitative and qualitative information
readily available.
Accuracy of the system is dependent on altitude, terrain,
sun angle, atmospheric conditions, and human error. With
the interpreter picking only optimum points from which ele-
vations can be read instead of reading elevations every 100
feet whether the ground is visible or not, the accuracy ap-
proaches the ± 0.3-foot range. This compares with a ± 0.6-
foot accuracy achieved with the 100-foot-grid pattern. It was
estimated that present MSHA inspection procedures cost $
470 per site. Using aerial photogrammetry would increase
the cost to a little over $500 to fly, develop, and interpret the
photographs per site with a one time survey fee of $2,000 to
establish ground targets. Many more embankments could be
inspected per day by using aerial photogrammetry allowing
inspection personnel more time to interpret photographs and
to concentrate on problem areas rather than make an indi-
vidual inspection of each site.
Satellite Monitoring
The purpose of this project is to develop a system that can
monitor the stability of one or more coal waste embankments
using various sensors and data acquisition equipment. It con-
sists of two phases, phase I being the installation and initial
monitoring of selected instruments at a test site. The raw data
are then sent to the user's receiving station via commercial
telephone lines. Phase II, using the same test site and em-
bankment instrumentation, will test the principle of relaying
data to the user via a satellite and a central receiving station.
During phase I, a test site was selected at a West Virginia
coal mine and 17 instruments were installed on a 200-foot-
high embankment. Three multiposition borehole extensom-
eters (fig. 3) measured vertical movement; three biaxial tilt-
meters, one inplace inclinometer, and one traversing probe
inclinometer measured horizontal deformation; and seven
vibrating wire and two resistance piezometers measured water
levels within the embankment. A pond level sensor, a V-
notched weir with a recorder for monitoring seepage on the
face of the embankment, and meteorological instruments were
also installed. Fourteen instruments are located along the
crest of the embankment, and three vibrating wire piezom-
eters at various levels down the face of the embankment.
Each instrument is located in a drill hole and protected
above ground by a covered standpipe. Two covered cable
junction boxes with lightning protection circuits connect the
sensors via cable to the onsite data station housing a signal
conditioner, power supply unit, data logger, printer, modem,
and telephone. Data were scheduled to be received at the
user's receiving station at 2-day intervals. Incoming raw data
are routed directly through a minicomputer for reduction to
usable information. A paper tape of the raw data is also made
at the test site to maintain a check on the accuracy of the
transmitted data.
Several problems were encountered which included main
power shutdowns, telephone lines knocked down, and in-
strument failures in the trailer. This resulted in only three sets
of data being received for a total of 61 days of data collection.
However, even though a relatively stable embankment was
selected for the test site, and with the limited amount of in-
formation available, trends could be noted. These trends in-
cluded correlations between piezometer readings and time
of rainfall, while extensometer data suggested a slight surface
settlement of the crest.
Phase II, which will be completed by 1984, will demonstrate
the use of a satellite data collection system to monitor the
stability of a coal waste embankment. This self-contained
system will consist of a data collection platform wired to the
existing instruments with an antenna and a battery-solar power
source (fig. 4). This method has been used since 1974 by
Landsat users to monitor global volcanic activity, and since
1976 by the U.S. Geological Survey to monitor stream flow
and snow pack data in remote regions (3). Data would be
relayed directly from the embankment to the GOES satellite,
eliminating onsite power and telephone lines that are costly
to install in remote or difficult access areas. Another purpose
for using the satellite is the speed with which an inspector
can obtain data during the rainy season. From this viewpoint,
-------
42
Figure 1.-Typical camera mount aircraft over a coal waste embankment.
-------
43
SEDIMENT CATCHMENTS
Figure 2.—Aerial photograph of a monitored coal waste embankment. (Approximate scale: 1 inch = 250
feet.)
several dams in potentially hazardous or populated areas
could be monitored from a central district office If one dam
out of several shows indications of problems, that dam can
be checked first.
The data collection platform is simple to use The em-
bankment sensors are connected to the data collection plat-
form which is powered by a battery-solar array combination
Data from the sensors are transmitted via an antenna to the
satellite, and then to the central data acquisition center. From
here they can be sent to the user via mail or telephone link,
depending on the urgency of the data.
Use of the GOES and the associated data collection station
facilities is provided free of charge by the National Oceanic
and Atmospheric Administration (3). This was one of the rea-
sons for selecting this satellite, the other being that one of
its main functions is data collection.
Cost of the data collection platform is about $5,000. This
includes the platform, cables, battery-solar array and an-
tenna, though one antenna could serve several platforms
depending on their locations Unmstalled instrument costs for
this project range from $1,000 for a resistance piezometer.
$4,000 for an extensometer, to $12,000 for an mplace incli-
nometer.
The basis of this test was to define the most suitable in-
struments to use in an embankment that could adequately
monitor horizontal and vertical movements, and water levels
If these water levels and movements could be ascertained.
then in an actual monitoring situation only two to five instru-
ments would be needed, depending on the size of the em-
bankment, its condition, end its location in relation to populated
areas, transportation facilities, or utility centers. Though the
initial cost of the instruments may appear high, their use as
a tool for an early warning of movement within an embank-
ment, the capability of providing more frequent monitoring,
-------
44
Figure 3.—Multiposition borehole extensometer in standpipe with cover.
and the opportunity for maintaining permanent records tor
review and study would seem to override that disadvantage.
Conclusions
In using the remote sensing techniques, the following sit.
uation is possible.
A high-hazard rated embankment in a remote area of the
Appalachian region has been outfitted with instruments re-
laying data through the GOES satellite. The instruments are
a piezometer, inclinometer, and a pond level sensor. This
same embankment is also being monitored by use of aerial
photogrammetry. Flights over the site have been scheduled
by the inspector every 3 months, making sure a flight occurs
right before and after the wet season, probably December
and March. Using the information from these two techniques
the inspector feels confident the embankment, though show-
ing creeping movement on the order of 6 inches every 12
months is in no imminent danger of failure. The pond level
is slowly rising since mining activity has increased due to a
rise in the spot market price.
An unseasonal rainfall dumps an inch of rain in the area
over a 3-hour interval; the inspector, knowing the pond level
is already rising, requests updated data through the satellite
system. The pond level sensor relays information that indi-
cates the pond is now within 18 inches of the crest and the
inclinometer has detected an increase in deformation indi-
cating a possible mass movement along the dam face. The
inspector requests data on the embankment every hour.
Twenty-four hours after the initial rainstorm, the piezometer
notes a rise, of water within the embankment but the pond
level sensor records dropping water levels, while the incli-
nometer indicates a slight settling of the embankment.
At this point the inspector ascertains that the slope stability
is adequate and there 19 no immediate danger of dam failure.
Information is now requested every 6 hours until the inspector
feels the embankment has stabilized. However, due to some
changes having taken place and the dropping water levels,
the inspector schedules an aerial flight the following day. The
-------
45
GOES satellite
Transmit
Transmit
Data
acquisition
center
Convertible
collection
platform
In place Extensometer
Power inclinometer
Test set >*•««•«•••• source
Figure 4.—Typical GOES convertible data collection platform installation.
Tiltmeter
data gleaned from the photographs show a well-formed ero-
sion gully at one of the abutments, the diversion ditch clogged
in spots with rain washed debris, and the embankment ele-
vation unchanged. With this information in hand the inspector
decides to visually inspect the embankment, going over repair
requirements with the operator.
From this scenario it can be seen that remote sensing
would be a powerful tool for mine inspection and mine com-
pany personnel. By monitoring internal changes through in-
strumentation and transmitting the data via satellite, and
observing external surface movement by use of aerial pho-
togrammetry the user can more readily and accurately eval-
uate waste embankments even in the most remote regions
of the country. It is not recommended that these two remote
sensing techniques be used on every embankment. Rather,
they should be seriously considered for high hazard em-
bankments which the inspector and mine personnel feel with
added information could provide an early warning system.
This will increase chances of correcting a hazardous situation
before it advances to failure.
Though this paper has discussed remote sensing systems
only in connection with coal waste embankments, the sys-
tems could be used for other types of embankments such as
metal-nonmetal mine waste embankments and tailings ponds.
References
1. Janza, F. J. (ed.). Manual of Remote Sensing. American
Society of Photogrammetry, Falls Church, Va., v. 1-2, 1975,
2144pp.
2. Kahan, A. M. Monitoring Cloud-seeding Conditions in
the San Juan Mountains of Colorado. ERTS-1, A New Win-
dow on Our Planet. U.S. Geol. Surv. Prof. Paper 929, 1976,
pp. 214-216.
3. National Oceanic and Atmospheric Administration.
Geostationary Operational Environmental Satellite/Data Col-
lection System. NOAA Tech. Rept. NESS 78, July 1979, 35
PP-
4. Paulson, R. W. Use of Earth Satellites for Automation
of Hydrologic Data Collection. U.S. Geol. Surv. Circ. 756,
1978, 7 pp.
5. Roth, L. H., J. A. Cesare, and G. S. Allison. Rapid Mon-
itoring of Coal Refuse Embankments. CH2M Hill (Redding,
Ca.), Final Rept. Contract H0262009, June 1977, 113 pp.;
BuMines OFR 11-78, available for consultation at Bureau of
Mines facilities in Denver, Colo., Minneapolis, Minn., Bru-
ceton and Pittsburgh, Pa., and Spokane, Wash.; Department
of Energy facilities in Carbondale, III., and Morgantown, W.
Va.; National Mine Health and Safety Academy, Berkley, W.
Va., and National Library of Natural Resources, U.S. De-
partment of the Interior, Washington, D.C.; and from National
Technical Information Service, Springfield, Va., PB 277 975/
AS.
6. Ward, P. L., and J. P. Eaton. New Method for Monitoring
Global Volcanic Activity. ERTS-1, A New Window on Our
Planet. U.S. Geol. Surv. Prof. Paper 929,1976, pp. 106-108.
-------
46
SUMMARY OF RESEARCH ON CASE HISTORIES OF FLOW
FAILURES OF MINE TAILINGS IMPOUNDMENTS
by
P. C. Lucia,1 J. M. Duncan,2 H. B. Seed2
INTRODUCTION
In the past decade there has been a dramatic increase in
the size of tailings dams and mine waste impoundments.
Tailings dams are now included among the largest dams in
the world, and a 700-foot-high tailings dam is now in the
planning stages (7).3 Studies conducted after the failure of
the coal waste impoundment at Buffalo Creek, W. Va., in
1972, indicated that a great many of the mine waste im-
poundments constructed at that time received no engineering
design consideration Q3). Recent history indicates that the
consequences of failure of a large mine waste deposit can
be disastrous from the standpoints of both loss of life and
environmental damage.
Liquefaction and flow of liquefied tailings or other types of
waste deposits have resulted in many deaths. Seismically
induced failures in Chile at the Barahona and El Cobre tailings
dams resulted in over 200 deaths. In Africa, the failure of the
Mulfulira and the Bafokeng tailings dams resulted in a total
of about 100 deaths. The failure of the coal waste dam at
Aberfan, Wales, resulted in over 140 deaths, many of them
school-age children between the ages of 7 and 10. The failure
of the coal waste dam at Buffalo Creek and the subsequent
deaths of over 100 people brought the problem of mine waste
disposal to the public's attention in the United States.
Severe environmental pollution can develop depending on
the nature of the tailings and its proximity to rivers and streams.
The failure of a phosphate tailings pond in Florida polluted
the Peace River for a distance of about 120 km. In Japan,
the Mochikoshi tailings dam failed during a 1978 earthquake.
The flow of cyanide-laden tailings polluted a river for about
30 km and destroyed the marine life in a bay.
It is clear that the consequences of failure can be disastrous
both from physical and environmental loss in some cases.
There are many other cases where a failure results in only
a temporary loss of storage to a mining company. A careful
review of many failures has resulted in empirical and theo-
retical methods by which the consequences of a failure can
1 Presently employed by Converse-Ward-Davis-Dixon, San Francisco, Calif
2 University of California at Berkeley.
3 Underlined numbers in parentheses refer to items in the list of references at
the end of this report.
be evaluated. This paper will summarize the case histories
studied and an empirical approach to the assessment of con-
sequences based on those case histories.
Soil Behavior at Liquefaction
Any mine waste material, regardless of origin, can be clas-
sified according to the principles of soil mechanics. Tailings
are frequently angular, bulky grained sand, and silt-size par-
ticles. It has been known for some time that sand and silt
particles are susceptible to rapid and large reduction in strength
due to very minor disturbances if they are deposited in a
loose condition and they are saturated.
Tailings are commonly deposited using hydraulic methods,
where the particles separate by size due to gravity in a pe-
ripheral discharge system or by cycloning. As they settle from
the water in which they were transported, tailings often ac-
cumulate in loose deposits, and silt sizes may form a met-
astable honeycombed structure, as shown in figure 1.
Rapid loading of this type of soil structure, either by seismic
or by static means, results in a rapid buildup in pore pres-
sures. The induced shear stresses are resisted at the points
of contact between soil particles because the water in the
voids has no shear strength. The soil particles move under
the shear stresses and tend to density, which results in
compression in the pore fluid. This transferring of compres-
sive stress to the pore fluid results in reduced compressive
forces at interparticle contacts, and a consequent weakening
of the soil structure. After a small amount of strain (less than
1 pet) particle-to-particle contact may be lost, and complete
breakdown of the structure may occur. This phenomenon,
which results in nearly complete loss of strength, is called
liquefaction. The stress-strain curve in figure 2 illustrates the
loss of strength resulting from liquefaction.
The breakdown in structure and loss of point-to-point con-
tact results in the induced shear stresses being transferred
to the fluid, which has no shear strength, only viscous shear-
ing resistance. The soil mass is then driven by forces for
which it does not have sufficient resistance, resulting in the
-------
47
,x—Typically less than 1 percent strain
Peak strength
-Residual strength
Figure 1 .—Honeycomb structure of loosely depos-
ited angular bulky grained soils.
onset of flow of the mass. While the soil particles are in
motion, it may be imagined that they form a "minimum re-
sistance" structure. During flow, at any given time, there is
always some soil-to-soil contact, and this results in some
small shear resistance for the flowing soil mass. This shear
strength is termed the residual strength; the magnitude of the
residual strength is a function of soil type, initial density, and
rate of flow. The flowing soil mass will come to rest when the
shearing resistance in the soil due to its residual strength is
equal to the shear stress.
Case Histories
In order to develop a rational approach to predicting how
far tailings will flow if failure occurs, case histories of flow
Strain
Figure 2.—Typical stress-strain curve for loose bulky
grained soil.
failures were collected. Due to the reluctance of many mining
companies to release information on previous failures, the
case history data were supplemented with information on flow
failures involving similar types of soils. Table 1 summarizes
information on the failure of the dams or waste impoundments
used in this study.
Typical of the extremes of behavior noted in the study are
the cases shown in figures 3 and 4. Figure 3 shows the failure
of a gypsum tailings pond (9) in Texas where the liquefied
tailings flowed over an essentially flat slope and came to rest
with a slope of about 1 pet on the surface of the tailings. The
other extreme is the liquefaction and flow of the coal waste
dump at Aberfan, Wales (8). In this case, the l^uefied ma-
terial flowed down a 12° slope, partially covering a school at
the bottom of the slope. The 'flowing material covered the
hillside at a relatively uniform thickness over the distance of
TABLE 1.—Data on case histories
Dam
Barahona
El Cobre (Old)
El Cobre (New)
Hieno Viejo
Los Maquis
La Patagua
Cerro Negro
Bellavista
Ramayana
Tailings Dam
Bafokeng
Gypsum
Mochikoshi
Phosphate
Tip No 7
Tip No 4
Abercynon
Blackpool
Cholwich
Louisville
Jupille
Fort Peck
East Chicago
Koda Numa
Uetsu
Location
Chile
Chile
Chile
Chile
Chile
Chile
Chile
Chile
Chile
Southwest United States
South Africa
Texas
Japan
Florida
Aberfan, Wales
Aberfan, Wales
Abercynon, Wales
England
England
Kentucky
Belgium
United States
United States
Japan
Japan
Probable cause
of failure
Seismc
Seism™
Seismic
Seismic
Seismic
Seismic
Seismic
Seismic
Seismic
Seepage
Seepage
Seepage
Seismic
Seepage
Static
Static
Static
Static
Static
Seepage
Static
Static
Static
Seismic
Seismic
Height,
m
65
35
15
5
15
15
20
20
5
44
20
11
32
4
37
46
37
40
46
31
46
69
2
3
10
Total quantity
of material
stored, tons
NA
76 x 10s
50 x 105
NA
60 x 104
NA
79 x 105
70 x 105
NA
NA
22 x 107
70 x 106
82 x 103
NA
43 x 10s
1 7 x 107
NA
NA
NA
1 0 x 106
60 x 105
NA
NA
NA
NA
Material
involved in
failure, tons
NA
1 9 x 105
50 x 105
1 2 x 103
30 x 104
50 x 10*
1 2 x 105
1 0 x 105
20 x 102
2.0 x 102
5.2 x 106
20 x 105
1 4 x 103
80 x 106
1 9 x 105
NA
1 8 x 106
1.5 x 10"
25 x 104
1 0 x 106
1.5 x 10s
5.0 x 108
NA
NA
NA
Travel
distance,
km
NA
12
12
1
5
5
5
25
NA
24
45
.3
30
120
6
.7
6
1
2
.1
6
4
02
02
11
NA Not available
-------
48
flow. In one case, essentially no damage resulted from the
failure; in the other case over 100 lives were lost.
The collection and understanding of the behavior of liq-
uefaction failure can serve as a guide to the engineer in
evaluating the potential consequences of failure of tailings
impoundments. The case histories collected during this study
are discussed briefly in the following sections, considering
first the tailings dams, and then waste impoundments and
other earth structures
Chilean Tailings Dam
The Barahona tailings dam in Chile failed in 1928 (1) during
an earthquake. The material flowed down a 9° slope, even-
tually getting into a river. A total of 54 lives were lost.
Dobry and Alvarez (G)discussed the failure of 10 tailings
dams in Chile due to an earthquake in 1965. The tailings
flowed distances ranging from 1 to 12 km. The downstream
geometries and final slopes on which the tailings came to
rest are unknown.
Tailings Dam (Southwest United States)
A copper tailings dam in the southwestern part of the United
States failed due to excessive seepage and subsequent ero-
sion. The loss of confinement resulted in the liquefaction and
flow of the tailings. The tailings flowed in rivers and streams
about 24 km from the disposal site. The tailings that remained
in the pond stopped flowing and became stable when the
slope angle of the tailings was about 1.5°.
Bafokeng, South Africa
The Bafokeng platinum tailings pond failed in 1974 (3) due
to excessive seepage. The liquefied tailings flowed down a
mine shaft downstream from the dam, resulting in the deaths
of 11 men. The tailings flowed over a 1° slope to the Kwa-
Leragane River about 600 m from the dam. Most of the tail-
ings that reached the river were carried to the Vaalkop res-
ervoir about 45 km away. When the tailings stopped, the slope
of their surface was about 1.3°, measured from the river back
into the pond.
Figure 3.—Failure of a gypsum tailings pond.
-------
49
Figure 4.—Failure of a coal waste dump at Aberfan, Wales.
-------
50
Gypsum Tailings, Texas
The failure of a gypsum tailings pond in 1966 (9) is shown
in figure 3. The dike around the pond failed due to excessive
seepage. The liquefied tailings flowed a distance of about
300 m from the embankment and the failure extended about
100 m back into the pond (fig. 3). The tailings came to rest
at aslope of about 1°.
Mochikoshi Tailings Dam, Japan
The Mochikoshi gold tailings pond failed in 1978 (10) during
a major earthquake. The tailings pond was formed by three
small dams. Two of the dams failed during the earthquake,
resulting in the flow of tailings down the hillside which was
somewhat steeper than 20°. The tailings polluted a river and
a bay about 30 km from the site. The flow of tailings from the
pond stopped when the tailings in the pond became stable
at a slope of about 4° to 5°.
Phosphate Tailings Pond, Florida
A phosphate tailings dike failed in 1971 (4), polluting the
Peace River in Florida over a distance of about 120 km.
Phosphate tailings differ from most other tailings in that they
are clay-size rather than bulky silt-size particles. They fre-
quently have water contents of several hundred percent and
little if any residual shear strength. Therefore, they flow much
the same as would water.
slope to a depth of 3.0 to 3.6 m over a distance of about 120
m. In October 1968, a similar waste deposit at Cholwich
failed. The flowslide traveled about 180 m down a 6° slope
and came to rest with a T slope in the waste.
Carbide Lime Tailings Pond, Louisville
In 1963, a carbide lime tailings pond in Louisville, Ky.,
failed, resulting in a flowslide. Carbide lime differs from tail-
ings in that it is not a waste product of a mining operation
but rather a byproduct of the production of acetylene gas.
The failure in the pond was due to excessive seepage and
the subsequent erosion of the embankment. The flow oc-
curred over flat land and involved the entire pond. The liq-
uefied tailings became stable when the slope of the tailings
was about 1.5°.
Jupille, Belgium
A fly ash deposit failed in Jupille, Belgium, in 1961 resulting
in 11 deaths (2). The waste deposit was formed by dumping
the ash and allowing it to fall at its angle of repose. The
liquefied material flowed down a slope as steep at 18° and
traveled a distance of about 610 m. Almost all the material
in the waste deposit flowed to the bottom of the slope, and
was believed to have traveled at speeds of 110 to 160 km/
hr.
Fort Peck Dam
Aberfan, Wales
In 1966, a coal waste dump, tip no. 7, failed (8). The sub-
sequent liquefaction and flow of the coal waste resulted in
144 deaths in the village of Aberfan. This failure was one of
three that had occurred in the area over a period of years.
In December 1939, a similar tip had failed at Abercynon,
about 8 km from Aberfan, under very similar conditions. The
material flowed down a 12° slope, traveling about 610 m and
depositing material about 6 m thick over the slope. At Aberfan,
in 1944, tip no. 4 failed just upslope from the one that failed
in 1966. In 1944, the liquefied material flowed down a 12°
slope about 610 m and blanketed the slope to a depth of
about 4.5 m. The failure in 1966 was similar in all respects
to the two previous failures, only in this case the village was
within 6110 m of the tip that failed, and the failure thus resulted
in a tremendous loss of life.
Blackpool and Cholwich, England
Fort Peck Dam is a hydraulic fill dam constructed in a
manner similar to a tailings pond. The method of deposition
results in coarser particles near the edge of the embankment
and finer particles in the core. More importantly, the resulting
structure was composed of bulky grained particles in a very
loose state, similar to tailings. The dam failed in 1938 (5).
The resulting flowslide traveled about 480 m from the toe of
the dam over nearly level ground. The liquefied material came
to rest at an average slope of about 2.5°.
East Chicago
Peck and Kaun Q2) described a flowslide that occurred in
uniform fine sand during construction of a dock wall in 1946.
The sand was initially placed under water behind a dock wall.
Removal of the wall and the subsequent loss of support in-
itiated the flowslide. The flowslide stabilized when the slope
in the liquefied material was about 4°, and the area over which
it flowed was essentially level.
In the china clay industry in England, the clay-size particles
are removed for use in making china, leaving sand and silt-
size particles as a waste product. In Blackpool, a waste de-
posit failed in October 1967 (2), resulting in a flowslide down
a 7° slope. When it came to rest, the material blanketed the
Uetsu Railway Embankment
A sand fill placed to serve as a railway embankment failed
during the 1964 Niigata earthquake (14). The embankment
-------
51
was constructed through a rice field and the bottom portions
of the embankment were saturated. The liquefied material
flowed about 120 m over ground which sloped at 2°, and
came to rest at a slope angle of about 4°.
Koda Numa, Japan
A small railway embankment at Koda Numa, Japan failed
during the 1968 Tokachioki earthquake (1J_). The soil was a
fine to medium sand. The embankment liquefied and flowed
in both directions from the centerline, over level ground. The
liquefied material flowed about 18 m, coming to rest at slope
angle of about 4°.
Summary of Case Histories
A summary of the final stable slopes and downstream ground
slopes for the case histories discussed is presented in table
2, for those cases where this information could be deter-
mined. With the exception of the coal and china clay waste
deposits, none of the liquefied soils were able to sustain
slopes greater than about 5°. At Barahona, the downstream
slope was about 9°, and the liquefied tailings did not have
sufficient strength to stop on this slope. At Mochikoshi, the
downstream slope was in excess of 20°; the material that
remained in the pond was only able to sustain a slope of 4°
to 5°, indicating that the 20° slope was far in excess of a
slope that the liquefied tailings could sustain.
While the data are limited, they indicate a consistent pattern
of behavior. Liquefied soils have a low shear strength that
enables them to come to rest on some small slope. In cases
where the downstream ground slopes were steeper than 4°
only the unsaturated coal and china wastes came to rest. All
of the liquefied saturated tailings materials continued to flow
until they reached inclinations of 1 ° to 4°
Simplified Procedure for Predicting the
Distance of Flow
A review of the case histories indicates that liquefied soils
can sustain themselves on mild slopes in a stable condition.
An idealized cross section is shown in figure 5. The actual
shape at the toe of the slope is shown in a dashed line with
the idealized shape shown by solid lines. Available data on
the postfailure conditions usually included a, p, L, and ma-
terial index properties such as water content. To quantify the
behavior observed in the case histories, the shear strength
required to give a factor of safety of 1.0 was calculated. The
assumption was made that at the moment the flow stopped,
the shear strength of the liquefied mass equaled the shear
stresses induced by the low slope angle. This assumption
neglects inertia forces; however, they may be negligible at
low velocities just before flow stops.
In calculating the shear strength of the soil, the assumption
was made that the behavior was undramed, and that the
strength could be represented as an equivalent shear strength,
Su. This back-calculated strength represents the residual
strength of the previously discussed "minimum resistance"
structure. The residual strength of the following mass may
vary from place to place through the mass due to different
degrees of drainage, and variations in particle size or void
ratio, among other factors. The shearing resistance also var-
ies with rate of flow due to viscosity. Calculating the shear
Figure 5.—Idealized cross section.
TABLE 2.—Postfailure conditions for case histories
Dam
Barahona
Tailings Dam
Bafokeng
Gypsum
Mochikoshi
Phosphate
Tip No. 7
Tip No. 4
Abercynon
Blackpool
Cholwich
Louisville
Jupille
Fort Peck
East Chicago
Koda Numa
Uetsu
Location
Chile
Southwest United States
South Africa
Texas
Japan
Florida
Aberfan
Aberfan
Abercynon
England
England
Kentucky
Belgium
United States
United States
Japan
Japan
Material type
Copper tailings
Copper tailings
Platinum tailings
Gypsum tailings
Gold tailings
Phosphate tailings
Coal waste
Coal waste
Coal waste
China clay
China clay
Carbide fine tailings
Fly ash
Clay to fine sand
Fine sand
Fine sand
Fine sand
Downstream
ground slope,
degrees
9
on
1
0
0(2)
NA
12
12
12
7
6
0
>18
0
0
0
0
Final slope
of liquefied mate-
rials, degrees
(')
15(2)
1 3
1
4 to 5 (2)
(1)
12
12
12
7
7
1 5
(')
25
4
4
4
Calculated
residual shear
strength, psf
NA
50
15
20
210
NA
375
330
450
140
340
53
NA
250
20
25
35
NA Not available.
1 Material traveled in rivers or other waterways
2 Measured slopes on materials remaining in tailings pond.
-------
52
strength based on the geometry when the mass comes to
rest makes it unnecessary to consider viscous and inertial
effects.
Various modes of failure were considered to find the most
critical. It was found that considering shear along the base
and active pressure at the back of the liquefied wedge re-
quired greater shear strength than the most critical circular
or noncircular slip surfaces, and this mechanism was there-
fore used to develop dimensionless stability charts. The con-
dition analyzed is shown in figure 5. For this condition the
equation of stability can be written in dimensionless form if
it is assumed that the factor of safety is unity:
W
:sinp -
S,,-L
"yHT2cosp
-cosp
^-cosp = 0 (1)
and this equation can be solved in terms of the dimensionless
parameter, N0, where
where -y = total unit weight of tailings,
HT = as shown on figure 5,
Su = as residual shear strength,
in which N0 varies with a and p as shown in figure 6. Given
values of a and p, the stability number, N0, can be determined
directly from figure 6. The shear strength, Su, can be calcu-
lated using equation 2 if -y and HT are known. The shear
strength values determined for the case histories studied are
shown in table 2.
1,000.0
100.0
10.0
1.0
I I
1 I
-1.0 1.0 2.0 4.0
Slope angle a, degree*
Figure 6.—Slope stability chart.
10.0 20.0
The distance through which liquefied tailings may flow be-
fore coming to rest (or "freezing") can be estimated based
on these charts. This requires knowledge of Su, the residual
strength after liquefaction. At the present time, there are no
laboratory procedures for measuring this important property
of tailings materials. Studies are now underway to develop
suitable procedures. At the present time, the most appropriate
procedure appears to be to estimate the residual strength of
the tailings during flow using back-calculated values for sim-
ilar materials.
The value of p the downstream slope angle, also must be
known in order to estimate the flow distance. Where down-
stream conditions are not uniform, it is necessary to use
judgment to determine an average value of p for the area
downstream where the liquefied tailings would flow. When
the value of L has been calculated, using the steps described
in the following, the value of p should be reviewed to confirm
that it is representative of the ground slope within the flow
area.
When values of Su and p have been established, the flow
distance, L, can be estimated using the steps outlined. Be-
cause the flow distance is affected by the volume of material
which liquefies, it is necessary to use a number of trials to
calculate L, and it is convenient to use a chart of the type
shown in figure 7 to determine the distance for L for which
the requirements of shear strength, bed slope, and flow vol-
ume are all satisfied simultaneously.
The steps for plotting the curves shown in figure 7 and
determining L are as follows:
1. Using figure 6, determine the value of N0 for a number
of assumed values for a. For each of these values of N0,
calculate HT using this formula:
HT =
N0SU
(3)
Strength curve
•Volume curve
1.
•i Stable condition (or given f and C
"•table
Figure 7.—Prediction of distance of flow.
-------
53
2. Plot the values of HT versus the corresponding values
of a on a diagram like that shown in figure 7. These values
of HT and a will define a "strength curve" which slopes down
to the right.
3. Estimate the volume of tailings that would be involved
in the flow, V,. As shown in table 1, in many cases, the volume
of tailings which flow is considerably less than the total vol-
ume of the pond. However, if the objective is to determine
the greatest possible flow distance, the maximum possible
volume of tailings should be considered. The data in table
1 indicate that in some cases, especially those where the
tailings were extremely fluid, the entire volume of tailings in
the pond did flow. Therefore, in the absence of evidence to
the contrary, it appears that the most appropriate assumption
will often be that 100 pet of the tailings will flow.
4. For a number of assumed values of a, calculate HT using
this formula:
HT =
where
A,2HC2 + A2V, - A3HC
/ tana
A — I
1 \tana - tanp
2tan2a
tana - tanp
tanp
tana - tanp
and V, = volume of material which flows.
(4)
(5)
(6)
(7)
(8)
The values of HT and the corresponding values of a will
define a "volume curve" which slopes up to the right, as
shown in figure 7.
5. Where the strength curve and the volume curve intersect,
all conditions with regard to strength and geometry are sat-
isfied simultaneously, with a factor of safety equal to one.
The values of HT and a at the point of intersection are those
corresponding to the limiting stability conditions. The value
of the flow distance L (fig. 5) can be calculated using the
expression
L =
HT - Hc
tana
(9)
This procedure is based on case histories of failures where
the liquefied tailings became stable after flowing over slopes
of less than 3° or 4°. Steeper slopes may require different
analyses where viscosity, dynamic effects, drainage, and
other factors are considered. It appears unlikely that tailings
materials will come to rest on slopes steeper than about 9°.
In such cases, it should be expected that flow will continue
until a flatter area or a body of water is reached.
Conclusions
A review of case histories of failure shows that liquefied
mine tailings composed of sand and silt sizes have some
small residual strength after liquefaction, and they will come
to rest at slopes of 1° to 4°. Mine tailings such as phosphates,
which consist of clay-size particles and have water contents
of several hundred percent, flow in much the same way as
water when loss of impoundment occurs.
A simplified procedure has been developed for predicting
how far tailings will flow in case of failure when the slope of
the surface on which the material flows is less than about 4°.
The procedure requires that the residual shear strength of
the liquefied tailings be known. At present, the residual
strength after liquefaction may be estimated based on values
back-calculated from field experience. Further research is
required to develop laboratory procedures for measuring this
important property of mine tailing materials.
References
1. Aguero, G. Formation de depositos de relaves en el
mineral del Teniento. Anales del Institute de Ingenieros de
Chile, No. 5, 1929, pp. 164-187.
2. Bishop, A. W. The Stability of Tips and Spoil Heaps.
Quarterly J. Eng. Geol. v. 6, 1973, pp. 335-376.
3. Blight, G. Personal communication, 1979. Available
upon request from J. M. Duncan, Berkeley, Calif.
4. Bromwell, L. Personal communication, 1978. Available
upon request from J. M. Duncan, Berkeley, Calif.
5. Casagrande, A. Role of the Calculated Risk in Earthwork
and Foundation Engineering. J. Soil Mech. and Foundations
Div., ASCE, v. 91, No. SM4, July 1965, pp. 1^0.
6. Dobry, R., and L. Alvarez. Seismic Failures of Chilean
Tailings Dams. J. Soil Mech. and Foundations Div., ASCE,
v. 93, No. SM6, Proc. Paper 5582, November 1967, pp.
237-260.
7. Gifford, F. Personal communication, 1980. Available
upon request from J. M. Duncan, Berkeley, Calif.
8. Her Majesties Stationary Office. Report of the Tribunal
Appointed to Inquire Into the Disaster at Aberfan on October
21, 1966. London, 1967, 215 pp.
9. Kleiner, D. E. Design and Construction of an Embank-
ment Dam to Impound Gypsum Wastes. Proc., 12th Internal.
Cong, on Large Dams, International Commission on Large
Dams, Mexico City, 1976, pp. 235-249.
10. Marcuson, W. F., R. F. Ballard, and R. H. Ledbetter.
Liquefaction Failure of Tailings Dams Resulting From the
Near Izu Oshima Earthquake, 14 and 15 January 1978. Proc.
6th Panamerican Conf. in Soil Mech. and Foundation Eng.,
Lima, Peru, International Society of Soil Mechanics and Foun-
dation Engineers, 1979, v. 2.
11. Mushina, S., and H. Kimura. Characteristics of Land-
slides and Embankment Failure During the Tokachioki Earth-
quake. Soils and Foundations, v. 10, No. 2, February 1970,
pp. 39-51.
12. Peck, R. B., and W. V. Kaun. Description of a Flow
Slide in Loose Sand. Proc. 2d Internat. Conf. on Soil Mech.
and Foundation Eng., International Society of Soil Mechanics
and Foundation Engineers, 1948, pp. 31-33.
13. Wahler, W. A., and D. P. Schlick. Mine Refuse Im-
poundments in the United States. Proc. 12th Internat. Cong.
on Large Dams, International Commission on Large Dams,
Mexico City, 1976, pp. 279-319.
14. Yamada, G. Damage to Earth Structures and Foun-
dations by the Niigata Earthquake, June 16, 1964, in JNR.
Soils and Foundations, v. 6, No. 1, January 1966, pp. 1-13.
-------
54
SUMMARY OF RESEARCH ON ANALYSES OF FLOW
FAILURES OF MINE TAILINGS IMPOUNDMENTS
by
J. K. Jeyapalan,1 J. M. Duncan,2 and H. B. Seed2
INTRODUCTION
While earth dam engineering has evolved with the devel-
opment of a large body of published theory and engineering
practice, tailings embankment design and construction has
received relatively little geotechnical engineering input until
recently. Thousands of large and small mine waste piles and
impoundments that have received no engineering attention
are scattered throughout the United States and other parts
of the world. Because of the poor quality of construction and
maintenance, many of these dams have failed, and the ex-
istence of these potentially hazardous impoundments is of
considerable concern to the public and to the mining indus-
tries. A characteristic common to most tailings dam failures
is that the mine tailings tend to liquefy and flow over sub-
stantial distances, with the potential for extensive damage to
property and life. Failures of El Cobre Dam, Chile (1965);
Aberfan, Wales (1966); Buffalo Creek, W. Va. (1972); and
Mochikoshi Dam, Japan (1978) are examples of such cata-
strophic dam incidents. In those four incidents, 450 lives were
lost, and the loss of property was approximately $200 million.
Table 1 summarizes some of the master failures that have
been described in the literature.
In order to be able to assess the potential for damage in
case of such a failure, it is necessary to be able to predict
the characteristics of the flow and the possible extent of flood
movement. Developing procedures for such analyses was
the purpose of the research summarized in this paper. The
study involved the following series of investigations:
1. Review of the available literature on the behavior of
liquefied tailings and similar earth materials.
2. Review of the available methods of analyzing the flow
of fluids from behind a breached dam.
3. Development of new analysis procedures to study the
characteristics of the flow of mine tailings after loss of im-
poundment.
4. Experimental study of the flow of highly viscous fluids
1 Presently at Department of Civil Engineering, Texas ASM University, College
Station, Tex.
2 University of California at Berkeley
released suddenly from impoundment to check the theory
previously developed.
5. Application of the analysis procedures to field cases for
which sufficient details were available to permit comparison
between calculated and observed behavior.
This paper deals briefly with the development of an ana-
lytical procedure for predicting the extent of flow of liquefied
tailings if a failure should occur and comparison of the the-
oretical results obtained with field observations in three cases
where such flows developed.
Rheological Properties of Liquefied Tailings
In order to analyze the flow phenomena associated with
events of this type, it is necessary to use a suitable rheological
model to represent the behavior of tailings during flow. Sev-
eral rheological models were reviewed and an appropriate
model for the liquefied tailings was chosen.
The subject of behavior of loose sands under undrained
loading has been studied thoroughly and is well understood.
Typical behavior of a loose sand under undrained loading
conditions is shown in figure 1. Because the volume of such
a saturated sample of loose sand can not change during the
undrained shear test, the contractive volume change tend-
ency causes a large positive change in pore pressure, as
shown in the lower part of figure 1. The peak strength of the
sample is reached at a small value of axial strain (typically
less than 1 percent), and the shearing resistance then de-
creases rapidly to a small residual value. At strains larger
than about 3 percent, the shear strength remains constant
at the residual value.
Mine tailings consist predominantly of sand and cohesion-
less silt sizes, and it would be expected that their behavior
under undrained loading conditions would be similar to that
of loose sand.
The data described indicate that neither loose sands nor
tailings lose all strength when they fail. Instead, their shearing
resistance decreases to a finite low value and at a given
-------
55
a: o 2
O "X
13 min
0.20 sec
Stress-strain curve
1
1
5 10 15
AXIAL STRAIN, percent
4 r
Pore pressure curve
0 5 10 15
AXIAL STRAIN, percent
Figure 1.—Behavior of loose sand under undrained
loading (after Castro, (3)).
strain rate remains constant thereafter. However, the residual
shear resistance increases with strain rate. Therefore, a
Bingham3 plastic model as shown in figure 2 was chosen to
represent the behavior of tailings materials during flow. The
3 Reference to specific trade names is made for identificatidn only and does
not imply endorsement by the Bureau of Mines
mathematical representation of this type of behavior is ex-
pressed by the equation.
T = Ty + rip
for values of
> r
(1)
where T, and TIP are referred to as the Bingham yield strength
and plastic viscosity. This model can also be written as an
apparent viscosity model as shown in figure 2. The ranges
of values expected for the Bingham parameters for tailings
can be estimated by comparing the characteristics of these
deposits with the properties of similar materials. Based on
such comparisons, the probable ranges of yield strength and
plastic viscosity of types of tailings other than phosphate
tailings are shown in the spectra in figure 3 (U).4
Dimensionless Numbers for Typical Tailings
and Associated Flow Regimes
Whether the flow of liquefied tailings will be laminar or
turbulent can be determined using the procedure suggested
by Hanks and Pratt (8) which is based on determination of
the Reynolds and Hedstrom numbers for the liquefied ma-
terial. Typical ranges of various parameters, flow velocities,
depths of flow, total unit weights, yield strengths, and plastic
viscosities for liquefied mine tailings are given in table 2. The
probable values for minimum and maximum dimensionless
parameters, Reynolds number, and Hedstrom number are
also listed in this table. The probable ranges of Reynolds
number for flows of phosphate tailings and other tailings are
plotted with the probable ranges of Hedstrom number in figure
4. It is apparent from this plot that, based on the Hanks and
Pratt criterion, the flow of phosphate tailings would be ex-
pected to be turbulent, whereas flows of other types of tailings
would be expected to be laminar. Therefore, different types
4 Underlined numbers in parentheses refer to items in the list of references at
the end of this report
TABLE 1 .—List of failures of tailing dams
Name and location of dam
Barahona, Chile
Old El Cobre, Chile
New El Cobre, Chile
Hierro Viejo, Chile
Los Maquis, Chile
La Patagua, Chile
Cerro Negro, Chile
Bellavista, Chile
Ramayana, Chile
Bafokeng, South Africa
Buffalo Creek, West
Virginia
Aberfan Tip 7, Wales
Gypsum, Texas
Mochhoshi, Japan
Jupille, Belgium
Phosphate, Florida
Year of
failure
1928
1965
1965
1965
1965
1965
1965
1965
1965
1974
1972
1966
1966
1978
1961
1971
Consequences
of failure
54 killed
210 killed
Pollution
Pollution
Pollution
Pollution
Pollution
Pollution
12 killed
118 killed
144 killed
Pollution
Pollution
11 killed
Pollution
Type of
tailings
Copper
Copper
Copper
Copper
Copper
Copper
Copper
Copper
Copper
Platinum
Coal
Coal
Gypsum
Gold
Ry ash
Phosphate
Method of
construction
Upstream
Upstream
Upstream
Upstream
Upstream
Upstream
Upstream
Upstream
Upstream
Upstream
Tip
Tip
Upstream
Upstream
Tip
Centerlme
Cause of
failure
Seismic
Seismic
Seismic
Seismic
Seismic
Seismic
Seismic
Seismic
Seismic
Seepage
Erosion
Seepage
Static
Seismic
Static
Seepage
Height
of Dam.
m
61
35
15
5
15
15
20
20
5
20
18
37
11
32
45
4
Volume
flowed,
million
tons
4.0
20
5
001
03
.05
12
10
.0002
52
55
.2
.2
14
.3
.8
Distance
inundated,
km
NA
12
12
1
5
5
5
25
NA
45
64
6
3
30
6
120
Source
Dobry and Alvarez (5)
do.
do.
do.
do.
do
do.
do.
do.
Midgley (14), Blight (2)
Seals, Marr, Lambs
(16)
Aberfan Tribunal (10)
Kleiner (12)
Marcuson (13)
Bishop (1)
Wahler and Schlick
(18)
-------
56
r)p=plastic viscosity
<
UJ
r;a= apparent viscosity
TABLE 2.—Summary of flow parameters of
typical liquefied tailings
Parameter
Probable minimum
value
Probable maximum
value
PHOSPHATE TAILINGS
Total unit weight pounds per cu-
bic foot
Yield strength . pounds per square
foot. . . .
Plastic viscosity . pound-second
per square foot
Flow depth . feet . . .
Flow velocity . feet per second
Reynolds number
Hedstrom number
80
40 x 10 4
20 x 10-"
2
5
40 x 104
8.0 x 103
100
40 x 10^
2.0 x 10-"
5
50
20 x 10s
1 0 x 105
OTHER TAILINGS
7-SHEAR STRAIN RATE, rod/sec
Figure 2.—Bingham plastic model for liquefied tail-
ings.
VISCOSITY, e/sec
/ 107 B Dow corning grease
Total unit weight . pounds per cu-
bic foot
Yield strength . pounds per square
foot
Plastic viscosity . pound-second
per square foot
Flow depth . feet
Flow velocity feet per second .
Reynolds number . . .
Hedstrom number . .
90
20
2
5
5
10
100
110
150
100
50
20
300
350
YIELD STRENGTH, Ib/ft2
8000
Probable range
for
liquefied tailings
105
10'
10'
101
Asphalt at 47'C
Natural mudflow
Wet cement mortar
Machine oil
4000
2000
1000
500
250
150
Probable range
for
liquefied tailings
U Water at 20*C
1 c/sec 2xlO"5 Ib sec/ft2
1 c/sec 0.001 Pascal sec
Figure 3.—Viscosity spectrum and yield shear strength spectrum.
20
Hard clay
Very stiff clay
Stiff clay
Medium clay
Soft clay
Very soft clay
-------
57
of analyses would be applicable to phosphate tailings on the
one hand, and all remaining types of tailings on the other
hand.
Analysis Procedures Developed
Hydraulic wave theory was extended to analyze the laminar
flow of tailings from behind a breached tailings impoundment.
Various analysis procedures were developed and the appli-
cability of these for different conditions is summarized in table
3. The analytical results can be presented in the form of
prediction charts based on dimensionless resistance param-
eters R and S in the case of flow on planes:
where
R =
and
(2)
(3)
in which g = acceleration of gravity, H0 = initial height of
impoundment, and y - total unit weight of material. Typical
predictive charts are shown in figures 5 and 6. For the more
complex case of flow in prismatic valleys, a computer program
(TFLOW) is required.
,06
10
to
S.
2io3
o
z
10
10
10
Eatlmated range for
phosphate tailing*
Turbulent
Laminar
-Critical boundary
determined by
Hanka and Pratt
(fi)
-Eatimated range for other
liquefied talllnga
io2 io3 io4 10*
HEOSTROM NUMBER
io6
0 010
008
006 •
004
002 -
0 002 0 004 0 008 0 008
R-viscous parameter
0 010
Figure 5.—Variation of inundation distance with
resistance.
0.010
008 -
006 -
o
c
0
004
.002
0 002 0.004 0.006 0.008 0.010
R-viscous parameter
Figure 4.— Variation of critical Reynolds number Figure 6.— Variation of freezing time with resist-
with Hedstrom number. ance parameters.
-------
58
1 iom
11m
i
300m *•]
Volume displaced = 80,000 to 130,000 m3
-profile after failure
I
Figure 7.—Case 1—gypsum tailings dam failure.
Flume Studies
In order to assess the validity of the procedures developed
for the analysis of flow failures of mine tailings impoundments,
a series of flume studies was performed. The details of the
experiments and comparisons of the experimental results
with the results of calculations are given by Jeyapalan (1J_).
There was close agreement between the analytical results
and the laboratory test data.
Case Studies
In this section, the applicability of the charts and the com-
puter program TFLOW for analyses of flow failures of mine
tailings impoundments is illustrated through examples. Cal-
culated results are compared with field observations for two
cases where the flow was laminar, and the suitability of avail-
able hydraulic flood routing computer programs for analysis
of turbulent flows is illustrated by comparison of calculations
with observations for a case where the flow was turbulent.
Laminar Flow
The flow of most liquefied tailings would be expected to be
laminar. Two cases of this type are discussed in the following
paragraphs.
TABLE 3.—Summary of various solutions and
their applicability
Solution
Ritter (15)
Dressier (6)
Tip theory (this study)
Perturbed solutions (this study)
Do.
Do.
Do
Applicable analysis
Flow of mviscid fluids in wide horizontal
channels
Turbulent flow of water and phosphate
tailings in wide horizontal channels
Laminar flow of viscous fluids in wide hor-
izontal channels
Laminar flow of viscous fluids in wide hor-
izontal channels
Laminar flow of most tailings in wide hori-
zontal channels
Laminar flow of most tailings in wide slop-
ing channels
Laminar flow of most tailings in prismatic
sloping channels
Case 1
Kleiner (1j?) described the failure of a gypsum tailings im-
poundment in east Texas in 1966 The geometry of the site
is shown in figure 7. The tailings were nonplastic silt with an
average field water content of about 30 percent. Using a total
unit weight of 90 Ib/ft3, a simple slope stability calculation was
done to determine the probable yield shear strength of the
material. The value calculated by this means was 20 Ib/ft2.
Using the viscosity-water content correlation developed as
part of the research study, the probable viscosity of the ma-
terial was estimated as 50,000 centipoise (1 Ib sec/ft2). Using
these values it was found that for the east Texas tailing im-
poundment, R = 0.0005 and S = 0.0006. For these values
for R and S, the dimensionless inundation distance, x, = 54
and the freezing time, t, = 130, were obtained from the charts
shown in figures 5 and 6. These dimensionless values were
converted to dimensional values in Jeyapalan (V1_) and the
results obtained from this analysis are compared with field
observations in table 4.
This case was also analyzed using the computer program
TFLOW for the case of a finite impoundment volume to better
simulate the field conditions. This analysis gave an inundation
distance of 470 m and a freezing time of 85 sec. It may be
seen in table 4 that both of these values are smaller than
those calculated using the dimensionless charts, because the
charts do not consider the finite volume of tailings.
Since there was considerable lateral spreading in the di-
rection perpendicular to the main direction of the flow, the
field observations indicate a lower value for the inundation
distance in comparison with the analysis. However, the over-
all agreement between the calculations and the field obser-
TABLE 4.—Comparison of theoretical results and
actual observations for gypsum pond failure, east
Texas
Flow
characteristics
Inundation distance, xt,
meters
Freezing time, t;, seconds
Mean velocity, um*, meters per
second
Observed
values
300
60 to 120
2 5 to 50
Theoretical
results from
charts
550
132
42
Results
using TFLOW
for finite
impoundment
470
85
55
NOTE—The notations x', K, and um* are used here for consistency with other
publications on this subject, where x,, t, and u, refer to dimensionless
parameters.
-------
59
vations is reasonable. The degree of agreement obtained
clearly depends to a great extent on the estimated values of
np and ry.
Case 2
The Aberfan Tribunal (U)) reported a detailed investigation
of the Aberfan disaster that occurred in Wales. Similar cal-
culations were performed using the charts and the computer
program TFLOW. A triangular channel with slide slopes of
one vertical on two horizontal was used to represent the gully
in which flow occurred.
The results from the chart analysis are compared in table
5 with the field observations and the results obtained by the
use of the computer program TFLOW. The inundation dis-
tance and freezing time from the chart analysis are consid-
erably longer than the actual values, but the results calculated
using the program TFLOW are in good agreement with ob-
servations. The larger inundation distance and freezing time
from the chart analysis may be attributed to the implicit as-
sumption of no channel side frictional resistance to flow. This
case illustrates the importance of considering the cross sec-
tional shape of the channel in the analyses.
1,600
rom analyses using
program GVFP (17)
0 4
STATION, miles
Figure 8.—Buffalo Creek flood—maximum water elevations.
-------
60
TABLE 5.—Comparison of theoretical results and
actual observations for Aberfan slide, Wales
Flow characteristics
Inundation distance, xf, meters . .
Freezing time, tj, seconds ....
Mean velocity, tC, meters per
second
Observed
value
600
120
50
Theoretical
results
using
charts
1,700
260
73
Results
using
TFLOW for
a triangular
channel
670
116
6.0
Turbulent Flow
If the flow will be turbulent, as is likely with very fluid tailings
materials such as phosphate tailings, existing flood routing
computer programs (e.g., GVFP (V7) and Fread (7)) can be
used for the analyses. These programs for turbulent flow
analysis incorporate resistance to flow through the use of the
empirical Manning's n relationship rather than friction factors
or fluid viscosity. Values of Manning's n (9) have been de-
termined by laboratory tests Using water, and it is not clear
whether these same values are applicable when the fluid
involved in the turbulent flow is not water. At the present time,
it appears that the best approach may be to use slightly higher
values for Manning's n than those applicable for water, for
purposes of analyzing flows of fluids such as phosphate tail-
ings.
The use of hydraulic flood routing computer programs
(GVFP (117) and Fread (7)) for analyses of turbulent flows is
illustrated in this section by application to a flood consisting
of water and coal waste (Buffalo Creek).
Case 3
The Buffalo Creek coal waste embankment failure released
a mixture of coal waste in a large amount of water which
apparently flowed much like water. The flood produced by
this failure was turbulent, and the program GVFP (V7) was
used to analyze the characteristics of this flood as a part of
this research study. The maximum calculated water eleva-
tions during the flood was compared with field observations
in figure 8. Also, computed travel times for the flood peak are
compared with the observations in figure 9. It may be seen
that the agreement between analyses and observations
shown in these figures is good. '
This case illustrates the applicability of available flood rout-
ing programs for analyses of turbulent flood flows to flow of
highly fluid tailings deposits. These programs can be con-
6
£4
UJ
O.
O 2
O
O
From analyses using
program GVFP (17)
0 4 8 12
STATION, miles
Figure 9.—Comparison of flood travel time from analyses and observation for Buffalo Creek flood.
-------
61
veniently used with no modifications for the analysis of tur-
bulent flows that are likely in the case of very fluid tailings
such as phosphate tailings or the fluid mixture of water and
coal waste which flowed down Buffalo Creek.
Conclusions
The main conclusions drawn from the research study are
as follows:
1. Instances of flow failures of mine tailings impoundments
indicate that the failure of these structures has considerable
potential for damage to life and property in many cases.
2. The behavior of tailings materials during flow can be
represented with reasonable accuracy by the Bingham plastic
Theological model.
3. The currently available computer programs (GVFP (1_7)
and Fread (7)) can be used without modification for analyses
of potential inundation zones likely to result from turbulent
flows of very fluid tailings such as phosphate tailings and the
mixture of water and coal waste which flowed at Buffalo
Creek.
4. The analysis procedures developed during this research
study can be used for analyses of flow failures in more highly
viscous tailings which undergo laminar flow. These proce-
dures are applicable for flow of tailings on horizontal and
sloping planes and in prismatic valleys. The analyses can be
performed using charts in the case of flow on planes, and by
means of a computer program (TFLOW) in the case of flow
in prismatic valleys.
5. The flume experiments conducted as part of this re-
search study indicate that the analysis procedures developed
are reliable. Application of these analysis procedures to field
cases also showed good agreement between calculations
and field observations. Thus, the analysis procedures de-
veloped appear to provide a useful means for assessing the
potential inundation regions downstream of mine tailings im-
poundments.
References
1. Bishop, A. W. The Stability of Tips and Spoil Heaps.
Quarterly J. Eng. Geol., v. 6, 1973, pp. 335-376.
2. Blight, G. B. Personal communication, 1980. Available
upon request from J. M. Duncan Berkeley, Calif.
3. Castro, G. Liquefaction of Sands. Ph.D. Thesis, Harvard
Univ., Cambridge, Mass, January 1969, 231 pp.
4. Davies, W. E., J. F. Bailey, and D. B. Kelly. West Vir-
ginia's Buffalo Creek Flood: A Study of the Hydrology and
Engineering Geology. U.S. Geol, Surv. Circ. 667, 1972, 32
PP-
5. Dobry, Ft., and L. Alvarez. Seismic Failures of Chilean
Tailings Dams. J. Soil Mech. and Foundation Eng., ASCE,
U. 93, No. SM6, November 1967, pp. 237-260.
6. Dressier, R. G. Hydraulic Resistance Effect Upon the
Dam-Break Functions. J. Res., National Bureau of Standards,
v. 49, No. 3, September 1952, pp. 217-225.
7. Fread, D. L. The NWS Dam-Break Flood Forecasting
Model. Report from Office of Hydrology, National Weather
Service, Silver Spring, Md., September 1978, 33 pp.
8. Hanks, R. W., and D. R. Pratt. On the Flow of Bingham
Plastic Slurries in Pipes and Between Parallel Plates J. of
Soc. Petrol. Eng., 240 December 1967, pp. 342-346,
9. Henderson, F. M. Open Channel Flow. MacMillan Pub-
lication, New York 1966, 522 pp.
10. Her Majesties Stationary Office. Report of the Tribunal
Appointed to Inquire into the Disaster at Aberfan on October
21, 1966. London, 1967, 151 pp.
11. Jeyapalan, J. K. Analyses of Flow Failures of Mine
Tailings Impoundments. Ph.D. Dissertation submitted to
University of California, Berkeley, August 1980, 298 pp.
12. Kleiner, D. E. Design and Construction of an Embank-
ment Dam to Impound Gypsum Wastes. Proc. Mexico City,
12th Internat. Cong, on Large Dams, International Commis-
sion on Large Dams, 1976, pp. 235-249.
13. Marcuson, W. F., R. F. Ballard, and R. H. Ledbetter.
Liquefaction Failure of Tailings Dams Resulting From the
Near Izu Oshima Earthquake, 14 and 15 January 1978. Proc.
6th Panamerican Conf. in Soil Mech. and Foundation Eng.,
v. 2 Lima, Peru, December 1979.
14. Midgley, D. C. Hydrological Aspects and a Barrier to
Further Escape of Slimes. The Civil Engineer in South Africa,
June 1979.
15. Ritter, A. Die Fortpflanzung der Wasserwellen (Prop-
agation of Waterwaves). Zeitschrift des Vereines Deutscher
Ingenieure, v. 36, No. 33, 1892, pp. 947-954.
16. Seals, R. K., W. A. Marr, and T. W. Lambe. Failure of
Dam 3 on the Buffalo Creek Near Saunders, West Virginia.
Report to Committee on Natural Disasters, National Academy
of Engineering, Washington, D.C., 1972.
17. U.S. Army Corps of Engineers. Gradually Varied Flow
Profile Program. Hydrologic Engineering Center, Corps of
Engineers, Davis, Calif., 1978, 32 pp.
18. Wahler, W. A., and D. P. Schlick. Mine Refuse Im-
poundments in the United States. Proc. 12th Internat. Cong.
on Large Dams, Mexico City, International Commission on
Large Dams 1976, pp. 279-319.
-------
62
CONTROLLED BURNOUT OF FIRES IN ABANDONED COAL
MINES AND WASTE BANKS BY IN SITU COMBUSTION1
by
ROBERT F. CHAIKEN2
ABSTRACT
A novel approach to eliminating environmental and public
safety hazards that are associated with fires in abandoned
coal mines and waste banks involves the use of in situ com-
bustion technology developed by the Federal Bureau of
Mines to accelerate the burning of the wasted coals in place.
This technology would be used under exhaust ventilation
control conditions that would allow for total management of
the hot gases produced. Combustion stoichiometries would
be optimized to minimize unburnt combustibles and to max-
imize the heat content of the gas products, which will be
exhausted at one or more fan locations. When necessary,
scrubber systems would be employed to remove air pollu-
tants, such as sulfur dioxide; heat utilization systems (process
heat, steam, and electricity) would also be employed to offset
operational costs. Ultimately, complete burnout would solve
the fire and acid water formation problems of the abandoned
coal mine or waste bank.
Pertinent technical data from the burning of tonnage quan-
tities of coal and coal refuse under simulated in situ conditions
are discussed in terms of burnout control. Based on these
data a field trial of the concept is being prepared at an existing
abandoned coal mine fire site (approximately '\Vz acres in
extent) located 6 miles from the Bureau's research facilities
at Bruceton (Calamity Hollow mine fire project). The field
burnout ventilation system has been designed to handle ex-
haust gas temperatures of approximately 1,400° C (approx-
imately 2,600° F) and thermal power output levels of ap-
proximately 5 MW. In the present configuration, the heat will
be wasted through the fan exhaust stack.
Introduction
A significant aspect of all past and current coal mining is
wasted coal. Wasted coal occurs in abandoned mines, which,
1 This paper is a summary and update of BuMines Rl 8476, "Controlled Burnout
of Wasted Coal on Abandoned Coal Mine Lands," by Robert F Chaiken,
1960.
2 Supervisory research chemist, Pittsburgh Research Center, Pittsburgh, Pa
because of the exigencies of underground mining operations,
often contain as much coal as was extracted; and coal refuse
piles, which are accumulations on the surface of reject ma-
terial from coal preparation plants and from underground
mining operations. Individual refuse piles (or waste banks)
may contain millions of cubic yards of solid waste, with a
combustible content ranging from 15 to 50 pet by weight.
Fires continue to occur in this wasted coal, introducing
significant hazards to public health and safety, such as
emissions of toxic and obnoxious fumes to the atmosphere
and destruction of residential and commercial buildings. Once
established, these fires can smolder for decades, and extin-
guishing them by conventional methods of sealing, dig-out,
and quench is costly and hazardous (&-1_i).3
The magnitude of the problem can be appreciated by ex-
amining the cost estimates for extinguishing the wasted coal
fires on abandoned lands (7). In a 1968 survey by the Bureau,
292 waste banks were found to be on fire and involved about
270 million tons of refuse and 3,200 acres of land. In a 1977
survey, 261 coal deposits were classified as burning. It has
been estimated that the cost of extinguishing the 292 burning
waste banks would be $468 million, and the cost of controlling
the 261 fires in active coal deposits would be $75.6 million.
These cost estimates probably should be reexamined and
brought up to date as they undoubtedly do not represent
current pricing structures. These estimates are for fix-up
only—they do not remove the fire potential of the wasted
coal, nor do they include the effective value of the wasted
coal if it could be utilized during the extinguishment and rec-
lamation process.
Although the intrinsic public health and safety benefits of
controlling fires on abandoned coal mined lands cannot be
denied, neither can the high costs of conventional control
techniques be ignored. Such costs present a serious con-
straint on how much and how fast corrective action can be
accomplished.
3 Underlined numbers in parentheses refer to items in the list of references at
the end of this report.
-------
63
Burnout control of fires in wasted coal is a novel concept
that offers a possibly more cost effective alternative approach
to conventional fire control methodology. The concept in-
volves acceleration of the wasted coal fire to burn the fuel
completely, but to do so in an environmentally and econom-
ically sound manner. The in situ combustion control tech-
niques currently being developed by the Bureau are expected
to accomplish this goal, and at the same time allow for pos-
sible utilization of the thermal energy produced during burn-
out. This paper describes the burnout control concept and
a. Plan view of multiple borehole system
Exhaust hole
® Inlet air hole
b. Side view of "blind" borehole system
Air flow
Exhaust hole _/_ \ i _i
Figure 1.—Two possible borehole arrangements for
burnout of a coal waste bank fire.
the status of some of the Bureau's efforts at applying it to
coal mine and refuse bank fires.
Burnout Control of Wasted Coal
The burnout control concept involves the controlled ac-
celeration of the fire by air injection so that complete fuel
removal is accomplished in a relatively short time compared
to a "normal" waste coal burn time (i.e., many decades). By
developing burn channels or zones in the burning wasted
coal (i.e., waste bank or underground mine), it should be
possible through exhaust ventilation techniques to effect a
controlled complete burnout of the coal and all other nearby
combustibles, such as carbonaceous rock materials and py-
rites. The process can perhaps be best envisioned from figure
1 which refers to a waste bank on fire. Figure 1 b depicts an
exhaust borehole driven into a waste bank. By sucking on
this borehole a negative pressure is maintained within the
waste bank which causes ambient air to permeate into the
bank. The air flow will accelerate the burning while all the
combustion gases will be exhausted through the borehole.
In this case the air flow will be controlled by the amount of
suction and the natural permeability of the waste pile. Alter-
natively, as shown in figure 1a, air inlet pipes can be inserted
into the waste bank to enhance the effective permeability of
the refuse.
Figure 1 a would also be applicable to underground mine
fires, where the boreholes now extend into the burning en-
tries. The nature of mine fires is such that propagation of the
fire always proceeds in the direction of the air flow; hence
through proper positioning of the air inlet holes, it is possible
to direct the fire underground.
A simple burnout exhaust ventilation scheme is shown in
figure 2. Here, a basic exhaust blower/duct system incor-
Infiltration air
/\
Air dilution
inlet
t
<-r.
" ; /
i
(Flue gas
J" i fc
i
i
Afterburner
iii
A & A
/:
t / /~
^
Exhaust
stack
Controlled
damper
Blower
Refactory lined
outlet from spoil
bank Water spray
Controlled
supplemental air inlets
Figure 2.—Schematic of burnout exhaust ventilation system.
-------
64
porates an afterburner to insure complete combustion of the
gases leaving the underground fire, and a water spray/air
dilution system to cool the hot gases before they enter the
blower. With the ventilation scheme shown in figure 2 the
heat would be wasted through the stack. Alternatively, one
or more heat extraction devices, such as a steam boiler or
air/air-heat exchanger, could replace the air and spray cooling
method shown so that the heat generated during burnout
could be utilized.
The burnout control concept has a number of distinct ad-
vantages:
1. The affected coal mine workings or refuse bank will be
at negative pressure, relative to ambient; hence little or no
fumes will be emitted to the atmosphere except at the fan
exhaust points.
2. Accumulation of all the fumes at fan exhaust points will
enable postburn incineration of the exhaust to insure com-
plete combustion of carbon monoxide and unburnt soot and
hydrocarbons to carbon dioxide and water. If required, scrub-
ber treatment can also be applied to remove air pollutants,
such as sulfur dioxide and particulates.
3. Controlled air injection might enable the burn to be car-
ried out under oxidation conditions favorable for sulfur dioxide
to react in situ to form solid sulfates, for example, calcium
sulfate which would remain in the ground.
4. The heat of combustion of the burning fuel will appear
as sensible heat in the exhaust products—perhaps at a tem-
perature as high as 1,000°C (1,832°F). This heat can be
recovered onsite for local use such as production of steam,
hot water, process heat, and electricity.
5. The complete burnout of combustible material (carbo-
naceous material and pyrites) in a mine or waste bank will
finally solve the environmental problems of an active fire. In
contrast, fires extinguished by wetting and sealing leave
wasted coal with its potential for reignition and acid water
formation.
6. The solid residue from the complete burnput of a coal
refuse bank is red dog, a gravel substitute with commercial
value.
The described advantages of burnout control of wasted
coal would suggest that the technique could be a panacea
for local environmental and energy problems. Although as
yet there has been no complete demonstration of burnout
control in an actual abandoned mine or waste bank, several
successful large-scale experiments have been carried out
with coal and coal waste under simulated in situ burning
conditions. Also, a field trial of the burnout of an abandoned
coal mine fire is currently being prepared.
Simulated In Situ Combustion of Coal
The Bureau background for burnout control of wasted coal
stems from its studies of the burning of coal underground
(2-4). A novel sealed surface trench burn facility was de-
veloped that enabled tonnage quantities of coal (or coal
waste) to be burned under simulated in situ combustion con-
ditions utilizing a ventilation scheme similar to that depicted
in figure 2. A photograph of the facility, which is located at
the Pittsburgh Research Center, is shown in figure 3. Figure
4 is a detailed diagram of the trench as set up for a simulated
coal mine entry fire.4
In this case the trench contains approximately 45 tons of
rubblized coal (Pittsburgh seam) with a 1-ft2, 35-foot-long
axial channel formed through the center of the coal to rep-
resent the mine entry. The exhaust fan (15,000 scfm, 18-inch
H2O head capacity) maintains a negative pressure in the
trench which induces air flow to the channel through the air
inlet pipes to sustain burning along the walls of the channel.
At the same time the fan exhausts the gaseous combustion
products from the end of the channel. While the fire studies
that have been carried out in the surface trench facility cannot
be reviewed here in their entirety, some results pertinent to
burnout of a mine fire and a coal refuse pile are discussed.
During one simulated mine entry fire experiment of 33-hour
duration, the thermal output of the in situ combustion process
was readily controlled between 1.0 and 1.7 MW, and the
exiting combustion products were at very high temperature—
about 1,600° C (2,900° F). Except for sulfur dioxide, the ex-
haust gas was exceptionally clean in terms of low concen-
trations of carbon monoxide, nitrogen oxides, and particu-
lates. Figure 5 shows data time plots obtained in this
experiment from measurements of the flue gas at a position
about 20 feet downstream of the channel exit (that is, follow-
ing some dilution with air). The exhaust temperatures at this
point are approximately 1,100° C (2,000° F), It is noteworthy
that while the observed sulfur dioxide emission (1,100 ppm)
is commensurate with the sulfur content of the coal (2 wt-
pct), the observed nitrogen oxides emission (110 ppm) is
considerably lower than what might be expected on the basis
of the fuel-nitrogen content (1.5 wt-pct). The nitrogen oxides
emission is actually about a factor of 5 less than values re-
ported for pulverized fuel combustors. Similarly, the partic-
ulate emissions appeared to be quite small.5
During the entire experiment (ignition, steady burning, and
cool down), the exhaust ventilation control system provided
total management of the combustion gas flow. Despite air
leakage into the trench and total burnout and cave-in of a
portion of the trench, no combustion products escaped to the
atmosphere except through the exhaust system.
Another in situ burning experiment of approximately 75-
hour duration was also carried out with 90 tons of bituminous
refuse obtained from a waste bank that is currently on fire.
Proximate analysis indicated the refuse was comparable to
a 75 pet ash coal with a heating value of 2,800 Btu/lb. In the
simulated refuse fire experiment, the 1-ft2 ventilated axial
channel served only as a zone for ignition. After the refuse
was ignited the inlet air lines were closed, and combustion
air was permeated into the trench through the top exposed
surface of the coal waste under the negative pressure in-
duced by the fan.
Figure 6 shows data time plots for this controlled burnout
experiment. They are somewhat similar to the coal channel
4 All dimensions shown in figure 4 are in metric units.
5 In a second sealed trench burn tests involving a channeled solid block of
coal, results similar to the rubblized coal burn were obtained, and in this
case the measured paniculate emissions at the fan exhaust stack averaged
only 0.008 lb/10° Btu over a 45-hour burning period,
-------
65
Figure 3.—Surface trench burn facility at Bruceton.
burning data previously described. The thermal power level
of the exhaust is in the range of 0.5 to 0.8 MW. The 600° C
(1,112° F) exhaust temperature shown in figures, which was
measured 20 feet downstream of the channel exit, is about
one-half the values recorded directly at the channel exit. This
indicates the occurrence of considerable dilution of the com-
bustion products by air leaking directly into the thermal breech
line The observed elevated oxygen concentrations in the
exhaust (approximately 15 pet at the 20-foot-downstream
station) also reflects this air leakage. Without dilution by this
air leakage the exhaust temperature and oxygen concentra-
tion would be approximately 1,000° C (1,832° F) and ap-
proximately 8 pet, respectively, in keeping with values ex-
pected from complete combustion of the fuel (1j. It should
be noted that air dilution that decreases the exhaust tem-
perature also increases the mass flow rates; hence the ther-
mal power level of the exhaust is relatively invariant to the
air dilution. The low CO levels observed in the exhaust is
further evidence that complete combustion was achieved.
It is also seen from figure 6 that the pollutant emissions,
NOX and SO2, are quite low indicating a very clean burning
system. Both the NO, and SOZ are less than what would be
anticipated based upon the fuel-sulfur and fuel-nitrogen (2.2
pet and 1.6 pet, respectively on an ash-free, moisture-free
basis). In this case of the NO* it is speculated that the in situ
burning geometry promotes the following redox reaction to
occur:
CO
In the case of SO2 it is speculated that the high effective
ash content of the coal waste acts as a solid scrubber system
for the SO? produced by burning; eg.,
Ca O(s) +
Mg 0(s) +
+ SO,
+ SO
»CaSO,(s)
>MgS04(s).
In this case the solid sulfates formed would remain un-
derground as part of the fused ash (or red dog). Chemical
analyses of the residue to obtain a mass balance on the sulfur
have not been carried out as yet
As in the case of the simulated coal mine entry burn, the
paniculate emissions from the in situ burning of coal waste
were quite low. In fact they were essentially negligible based
upon the simple sampling system employed at the fan ex-
haust stack.
One set of conclusions that might be drawn from the sur-
face trench burn experiments is that the controlled in situ
burning of coal and coal waste can be carried out efficiently,
cleanly, and with total management of the heat and gases
produced. However, surface subsidence cannot be truly sim-
ulated in the surface trench experiments. Questions still to
be answered on the long-term control of an actual burnout
system are (1) what are the effects of crevices and channels
-------
Cod-lined igniter section
Ignition gate
Servo vofces
Arrtet section
Inferno) water spray
(for quenching)
Thermal breech line
At each station
la, 2a, and 3a
5 thermocouples
I gas analysis
I static pressure
I bidirectional velocity
turbine meter
thermocouple
bidirectional flow
Refractory cement cover
Grade elev,
319.28 m~\
Smoke analyzer and oxygen probe
for servocontrol of excess air
t Typical section through
coal trench
At each station
lc,2c,3c,4c,<»A 5c
14 thermocouples
I gas analysis
I to 3 static pressures
I to 3 bidirectional flows
I water spray nozzle
Firebrick
Concrete
Stack probes
I smoke analyzer
I gas analysis
I bidirectional velocity
I static pressure
I fly ash
Fon
Grade etev,
PROBE KEY
Thermocouple
Bidirectknol flow
a Amubor flow
Gas analysis
Static pressure
Water nozzle
Exhaust line probes
lamubor flow
4 thermocouples
I static pressure
Figure 4.—Simulated coal mine entry fire: diagrammatic setup of experiment (all dimensions in metric units).
-------
67
ipoo-
500-
IN SITU COAL
TEMPERATURE
ENTHALPY
3 MO M*
CARBON MONOXIDE
The abandoned coal mine fire is in a shallow turn-of-the
century drift mine in the Pittsburgh seam, and has been smol-
dering for about two and one-half decades (fig. 7). In 1964,
the Bureau attempted to extinguish the fire by digging an
isolation trench (outcrop to outcrop) and surface sealing, but
these efforts were not totally satisfactory. Isolation of the fire
by the trench barrier to about 1 Vz acres of coal seam was
successful, but surface sealing failed to extinguish the fire;
primarily due to inadequate upkeep of the surface seal.
In carrying out the field trial, the site was first surveyed
(December 1979) to provide a base map on which lease lines
were drawn. Site preparations involved cutting access roads
for both light- and heavy-duty truck traffic, grubbing and cleaning
the 3-acre work area, and leveling and slagging the ground
for the proposed burnout system. A cyclone fence was in-
stalled around the site to provide for public safety and to
prevent vandalism.
Concurrent with the site preparations, an exploratory drill-
ing program was begun to accomplish several objectives: (1)
Boreholes were required to determine the exact location of
the isolation trench, the original highwall, and the under-
ground high-temperature zones; (2) the boreholes would help
BITUMINOUS REFUSE
TEMPERATURE
60
70
80
Figure 5.—Data-time plots for burnout of a simu
lated coal mine entry fire.
on the waste bank permeability and the effective burn volume;
(2) what are the effects of roof falls on the pressure drop
requirements for burning mined out entries, and would this
lead to deterioration of the combustion efficiency; and (3)
what surface subsidence effects will be experienced from
complete burnout?
These questions are important, but they are site-selective
and can only be answered through field trials in actual aban-
doned coal mines and actual coal waste banks.
Field Trial of Burnout Control
To answer the question of long-term control of a burnout
system, a full-scale field trial of the concept is currently being
developed at the site of an abandoned coal mine fire located
at the Calamity Hollow section of Jefferson Borough, Pa.,
about 6 miles from the Pittsburgh Research Center. The work
efforts to date have involved mostly site preparations, de-
signs, procurements, and constructions. Startup of the burn-
out operations is scheduled for June 1981. This section of
the paper is essentially a brief report on the status of the
Calamity Hollow fire project.
20-
10-
0
OXYGEN
10
20
30
50
60
70
80
CARBON MONOXIDE
/-4616ppm (ovtrog*)
10 20 30 UO 50 60
OXIDES OF NITROGEN
70
80
10 20 30 "JO 50
SULFUR DIOXIDE
60
70
80
20 30 HO 50
TIME, hours
60
70
80
Figure 6.-
lated coal
-Data-time plots for burnout of a simu-
waste bank fire.
-------
68
Parkinson Estate
Outline of pro/ect area.
Isolation _
trench barrier
test holes
SW property
line of former \
Parkinson
private pr°Perly
I EG END
Pittsburgh Coal Bed ^
Preproject surface "
evidence of fire
O IOO 20O 3OO
Scale, feet
Surface Sealed Area
Direction of Propagation
project
Figure 7.—Plan view of Calamity Hollow mine fire project site.
-------
69
in establishing the mining pattern (note mine maps are un-
available for the site) and the air communication between
various underground zones; (3) sensors in the boreholes will
be used to monitor the progress of the fire; and (4) boreholes
will serve as inlets to supply combustion air to the fire, and
to control the path of the fire during burnout.
With exploratory drilling completed (about 100 4-inch holes),
temperature measurements and air communication tests were
carried out to check the condition of the old mine workings.
A small mobile fan setup was built (400 scfm at 60-inch H2O
head capacity) and used to draw air from the mine while flows
were measured at surrounding boreholes. With these data a
location was selected for a large 48-mch-OD, 25-ft-deep ex-
haust hole. This hole was augered into a warm zone (ap-
proximately 100° F), about 75 feet from a communicating high
temperature zone (approximately 800° F). A large diameter
(44-inch-OD, 36-inch-ID) water-jacketed manifold pipe, which
was previously fabricated and lined with insulation brick, was
installed in the hole. This pipe will serve as the mam exhaust
hole for the hot gases produced during burnout.
The ventilation system designed for Calamity Hollow is
shown in figure 8. Heat generated during burnout will be
wasted through the stack. To date, fabrication of the high-
temperature elbow and horizontal ductwork has been com-
pleted and the components installed. The first section of duct-
work leading off the elbow connection is refractory lined and
will serve as an afterburner to complete combustion of the
exhaust gases as required. The ductwork design includes
instrument-activated automatic valves for controlling both the
flow of hot exhaust gases and the cooling of these gases
with ambient air. The instrumentation and control systems
are currently being assembled.
The diesel generator (350-kW capacity) and the exhaust
fan (25,000 scfm, 40-inch H20 head capacity) are mounted
on a 50-ton-capacity flatbed trailer. The entire mobile unit is
now positioned on temporary foundations at the site to receive
the horizontal ductwork. The fan-motor-generator assembly
is located on the cold side of the isolation trench barrier to
protect it from potential subsidence. Not shown in figure 8
are two additional items that are planned for incorporation
into the burnout system: (1) An accordian type expansion
joint which will allow for relative displacement of the ductwork
in all directions; and (2) a drop-out pot interposed between
the fan and the horizontal ductwork. The drop-out pot will
serve to catch large particles that may be sucked from the
mine and, if required, as a scrubber device to control air
pollutant emissions.
It is anticipated that major component fabrication and in-
stallation of the ventilation system will be completed in April
1981; instrumentation and controls will be completed during
May 1981; and start up of the actual burnout will be sometime
in June 1981. It is anticipated that the system will be capable
of burning as much as 1,500 Ib of coal per hour, developing
5 MW of thermal power in the exhaust. At this rate, the es-
timated 4,000 tons of coal remaining on the "fire" side of the
isolation trench at Calamity Hollow would be consumed in
about 235 days. However, it is most likely that the burnout
will be stopped before that time 'in order to uncover the coal
and examine in detail the underground residues. In this way
detailed information will be obtained on the underground zones;
for example: (1) The size and geometry of burning from a
single exhaust hole; (2) the effect of high temperatures on
the remaining coal pillars and surrounding rock strata; and
(3) the physical and chemical nature of the residues that
remain underground.
Summary
A novel approach to controlling the environmental and pub-
lic safety hazards associated with fires in abandoned coal
mines and refuse piles has been described. This approach
62-in diam
35ft-
Refroctory
lined
55 in
Additional
duct
lOlin
4ft
40-in diam
— 36-in diam
30-in diam
\.
Q
-in diam—"iHH
30-in
Bleed-in air
damper
46-in diam
Diesel generator
Flue line to mine
ot 30-f t level
Figure 8.—Calamity Hollow burnout ventilation system.
-------
70
involves complete burnout of the wasted coal in place under
conditions that allow for total management of the hot gases
produced. Ultimately, complete burnout would solve once and
for all the fire and acid water formation problems of these
fires on abandoned mine lands. In addition, utilization of the
heat produced during burnout (e.g., to produce steam, elec-
tricity, or process heat) could more than offset the costs of
constructing and operating a burnout exhaust ventilation sys-
tem.
Data from large-scale simulated in situ combustion exper-
iments strongly support the feasibility of the burnout concept;
however such experiments do not adequately address po-
tential problems from subsidence. A field trial of burnout con-
trol of an actual abandoned coal mine fire is currently being
prepared to answer many if not all the questions raised in
this report. This field trial (Calamity Hollow mine fire project)
is expected to begin burnout operations in June 1981.
References
1. Chaiken, R. F. Controlled Burnout of Wasted Coal on
Abandoned Coal Mine Lands, BuMines Rl 8478, 1980, 23
PP-
2. . Heat Balance in In Situ Combustion. BuMines
RI8221, 1977, 11 pp.
3. . In Situ Combustion of Coal for Energy. BuMines
TPR84, 1974, 12pp.
4. Chaiken, R. F., L. E. Dalverny, M. E. Harris, and J. M.
Singer. Simulated In Situ Coal Combustion Experiment. 4th
Ann. Underground Coal Conversion Symp., Steamboat
Springs, Colo., July 1978. Sandia Laboratories, Albuquerque,
New Mex., pp. 515-526.
5. Chaiken, R. F., J. M. Singer and C. K. Lees. Model Coal
Tunnel Fires in Ventilation Flow. BuMines Rl 8355, 1979, 32
PP.
6. Griffith, F. E., M. O. Magnuson, and G. J. R. Toothman.
Control of Fires in Inactive Coal Formations in the United
States. BuMines Bull. 590, 1960, 105 pp.
7. Johnson, W., and G. C. Miller. Abandoned Coal-Mined
Lands—Nature, Extent, and Cost of Reclamation. BuMines
Spec. Pub. 6-79, 1979, 29 pp.
8. Magnuson, M. O. Control of Fires in Abandoned Mines
in the Eastern Bituminous Region of the United States. A
Supplement to Bulletin 590. BuMines 1C 8620, 1974, 53 pp.
9. Magnuson, M. O., and E. C. Baker. State-of-the-Art in
Extinguishing Refuse Pile Fires. 1st Symp. on Mine and Prep-
aration Plant Refuse Disposal. Nat. Coal Assoc., Coal and
the Environment Tech. Conf., Louisville, Ky., October 1974,
pp. 165-182.
10. McNay, L. M. Coal Refuse Fires, An Environmental
Hazard. BuMines 1C 8515, 1971, 50 pp.
11. Myers, J. W., J. J. Pfeiffer, E. M. Murphy, and F. E.
Griffith. Ignition and Control of Burning of Coal Mine Refuse.
BuMines Rl 6758, 1966, 24 pp.
U s Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, l^tn MOOT
Chicago, !L 60604-3590
. GOVERNMENT PRINTING OFFICE: W9I - 548-187/40640
------- |