University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Colorado School of Mines
|
ER-4609 296
11.3.5 Revegetation (after BLM, 1992)
If the growth medium is to be the tailings, it should
be analyzed and evaluated. It is likely the physical and
chemical characteristics will require some modification to
ensure that the ultimate reclamation goals will be met.
Common amendments include fertilizer, organic material,
limestone (for control of acidity), acidifying agents (for
control of alkalinity), and, in some cases, bactericides to
help control the oxidation of sulfides during the initial
stages of revegetation.
Plant species selected for revegetation should be
adapted to the site-specific conditions in order to fulfill
the ultimate reclamation objectives. Factors to evaluate
include : drought tolerance, rooting depth, hardiness,
metals accumulation, palatability, seed availability,
stabilization ability, ease of propagation, and longevity.
Field trials on test plots during the mine life are often
required to evaluate which species will work best.
Seedbed preparation is the next important phase of
reclamation. Typically, this is performed by standard
agricultural equipment and follows normal practices.
Roughening of the surface to be planted should result in a
firm but friable surface. In many cases, mulching and,
occasionally, irrigation may be used to aid in establishment
of vegetation. It is important to realize that dust must be
controlled during the early stages of revegetation or it
will scour and kill emerging vegetation. Following
planting, the success of revegetation should be monitored to
assure successful reclamation.
|
Colorado School of Mines
|
ER-4609 297
11.4 Revegetation Plan (after BLM, 1992)
A primary goal of a revegetation program is to
stabilize the surface against the long term effects of
erosion. Another major objective is the return of the site
to a productive post-operational use. The revegetation
process begins after the disturbed area has been shaped,
graded, and treated, and the topsoil or other suitable
growth medium is spread and smoothed. The revegetation of
the affected lands shall be accomplished in a timely manner
and consistent with the reclamation plan. Lands which did
not support vegetation prior to mining because of soil
conditions may require no revegetation.
11.4.1 Soils Management (after BLM, 1992)
The use of topsoil or other selected replacement
material as a growth medium to be spread over lands
disturbed by mineral activities during reclamation must be
considered during reclamation planning. The amount and
quality of replacement soils used will have a effect on the
future productivity of the reclaimed lands. Proper soils
management is critical to reclamation success. Some factors
to consider early in the planning process include:
Amount of the topsoil to be saved.
Alternatives to spreading topsoil.
Storage location of salvaged soils.
Protection of stored and salvaged soils.
Direct replacement of the soils.
|
Colorado School of Mines
|
ER-4609 298
Required thickness of replacement soil.
The volume of available soil.
Availability of additional growth media.
A site-specific soil survey should be conducted as a
part of the baseline studies. The soil survey should be
conducted in accordance with the standards of the National
Cooperative Soil Survey (See SCS US DA Handbooks 430 and
436). The purpose of the soil survey is to identify
suitable soils for use as growth media or for other special
purposes. Soil resources within the zones of proposed
disturbance should be inventoried for volume and suitability
prior to the disturbance (see Table 32) . Soils high in
clay content may not be suitable to support plant growth but
may be suitable for use as an impermeable barrier in waste
management. The acceptability of soils is also dependent
upon moisture, organic matter, soluble salts, selenium and
boron content, bulk density and other factors.
ILE 32 Soil suitability for reclamation purposes (after
BLM, 1992)
SOIL
PROPERTY SOIL QUALITY
GOOD FAIR POOR UNSUITABLE
Texture sandy loam sandy clay loam sandy clay
loam silty clay loam loamy sand clay >60%
silt loam clay loam silty clay
Rock & Gravel 0-10 10-20 20-40 >40
(% by volume)
PH 6-8 5-6 4.5-5 >4.5
8-8.5* 8.5-9* >9
Na absor(ption 4 4-8 8-16 >16
ratio, SAR)
Electrical 3 3-7 7-15 >15
Conductivity
(milliohms/cm)
|
Colorado School of Mines
|
ER-4609 299
All replacement soils and material suitable for
reclamation should be salvaged wherever feasible and stored
for later use in reclamation or if conditions permit,
applied directly to recontoured areas ready for reclamation.
Salvaged materials should be properly stored and
revegetated if necessary to protect the stockpiled soils
from erosion. Concurrent topsoil replacement is preferable
to long-term topsoil storage. Studies have indicated that
long-term storage of topsoil may result in the loss of vital
organisms in the soil. Stockpile sites should be located in
areas which will not be affected by future operations and
are easily accessible for removal at the time the soils are
needed.
The appropriate replacement thickness of growth media
is usually based on the amount of available topsoil or
growth media and past experience with application depths.
In general, the poorer the chemical and physical properties
of the spoil or waste materials, the greater the required
depth of the replacement soils. When the availability of
good soil materials is limited, consider the qualities of
the soils available. Generally, a thin layer of topsoil
over unproductive subsoil will result in greater plant
productivity than a thin layer of topsoil alone.
In those cases where the waste materials are finely
textured and exhibit no phytotoxic properties (i.e. highly
acid or saline), about 6-12 inches of replacement topsoil or
other suitable growth medium, if available, should be
sufficient. Coarse textured (rocky) waste or waste
exhibiting phytotoxic properties may require greater
thicknesses and additional treatment. Disturbed areas
containing highly phytotoxic materials may require some form
of mechanical treatment, such as sealing the dump with
|
Colorado School of Mines
|
ER-4609 300
clays, prior to application of the topsoil. Where the
volume of replacement soils is limited variances to the
above recommendations or uneven placement may be justified.
Certain fine grained substrata which exhibit favorable
reclamation properties, may be used to advantage as a soil
medium or as a mantle to cover rocky waste.
Reapplied topsoil or selected subsoils should be tested
for nutrients, Ph, and toxicity factors prior to planting.
11.4.2 Seed Bed Preparation (after BLM, 1992)
The first revegetation step is to prepare the newly
spread soil material for seeding and planting. The soil
material must be permeable enough to absorb precipitation
and to allow for root penetration. In arid zones, seedling
establishment is difficult and highly variable; therefore,
proper seedbed preparation is extremely critical.
Seedbed conditioning provides important benefits for
plant germination, establishment, and long term vitality by
loosening the compacted soil material, providing catchments
to increase water available to plants, and creating
microsites that shelter seeds and seedlings. The seedbed
should be conditioned to collect, hold, and absorb as much
moisture as possible.
Equipment for seedbed conditioning ranges from rippers
and discs or chisel plows to spring tooth harrows and rakes.
After the topsoil is applied and graded, consider
scarifying, shallow ripping, or disking the site to
eliminate compaction and provide for increased infiltration
rates. Ripping or disking will retain water in the seed bed
which is essential to the success of the revegetation. Rip,
|
Colorado School of Mines
|
ER-4609 301
disk or harrow on the contour of the slope to reduce the
effects of surface erosion.
Some important considerations in seedbed preparation
are :
Final shape or landform should be compatible with
the surrounding landforms where practicable. If
possible, natural drainage should not be altered,
except where necessary to protect unstable soils
or tailings or toxic materials areas. The seedbed
should be tested for growth potential prior to
seeding. Capillary breaks may be necessary to
isolate toxic subsoil materials.
Soils and subsoils that have been highly compacted
should be ripped. Subsoils should be ripped prior
to the placement of the topsoil or other growth
media.Rip the mantle when it is relatively dry to
permit shattering beneath the surface. Moisture
content should not exceed field capacity.Ripping
should generally be 2 to 3 feet deep on 2-to 3-
foot centers. A "rule of thumb" is the distance
between rippers should be equal to the depth
ripped. Ripping depth is limited by the
characteristics of subsoil materials, which may
inhibit germination.
11.4.3 Fertilization (after BLM, 1992)
Many disturbed areas and waste embankments may be
nutrient deficient at the time the reclamation is performed
|
Colorado School of Mines
|
ER-4609 302
and may require fertilization to ensure the seedlings
establish themselves. Fertilization is the addition of a
natural or man-made substance to the soil to supply plant
nutrients. Its use can be justified when operations disrupt
the soil balances that affect nutrient availability.
Fertilization can be done before, during, or after seeding,
however it is usually more advantageous to fertilize prior
to seeding. After the initial fertilization and subsequent
establishment of plants, the natural process of nutrient
cycling is expected to maintain the plant community.
Consider the following when fertilizing reclamation
proj ects:
Soil materials should be tested for nutrient
levels prior to fertilization. Only available
nutrients are important. Macronutrient (e.g.
nitrogen, phosphorus, potassium, calcium,
magnesium, iron and sulfur) and micronutrient
(e.g. zinc, boron, selenium) deficiencies will be
determined by the soil sampling.
The nutrient content of bagged and bulk
fertilizers is expressed as a percent of the
content by weight. Example : A 100-pound bag
marked 10-10-10 means 10% nitrogen, 10% phosphorus
, and 10% potash
(P 2O 5) (K 2O 5) .
Equipment to apply chemical fertilizers (common
agricultural fertilizers) range from broadcast
spreaders and drill seeders for dry or granular
fertilizer, subsoil injectors for liquid
fertilizer, and hydro-seeders for applying a
|
Colorado School of Mines
|
ER-4609 303
slurry of fertilizer and water. The application
of biologic fertilizers (manure, compost, etc.)
will require special equipment.
For best results, fertilize before planting, and
harrow or drill the fertilizer into the soil
material to increase the effectiveness of the
fertilizer. If it can be demonstrated that the
seedlings can be established without fertilizer,
consider the application of the fertilizer after
the seedlings are established.
Usually fertilizer applied with a hydro-seeder
will be done in conjunction with seeding. Not
only are fertilizer slurries sometimes
incompatible with organic mulches, but can be
toxic to the seed, and should be applied in
separate operations.
Nitrogen fertilizers should be those that will
release at the time of germination. Losses of
available nitrogen over the winter season may
reach 30%, therefore, adjust application rates to
account for these potential losses.
Adult plants which exhibit a yellowish-green color
and drying of the lower parts of the plant usually
are deficient in nitrogen or iron. Phosphorus
deficiencies in plants often cause a purplish
color in the leaves and the plants display poor
root development, stooling, and spreading.
|
Colorado School of Mines
|
ER-4609 304
Application of low nitrogen levels reduces weed
growth. Higher levels of phosphorus improves root
development with no appreciable increase in the
growth of weedy species.
11.4.4 Soil Amendments (after BLM, 1992)
The use of soil amendments is often an important part
of preparing disturbed areas for revegetation. Most
disturbed sites exhibit soils that have received impacts to
their chemical and/or physical characteristics (e.g.
compaction). These impacts affect the ability of the soil
to function effectively as a growth medium for vegetation
and generally increase the likelihood of soil surface
instability. Soil amendments are natural or man-made
materials incorporated into the soil to improve the soil-
water or soil-air relationships in the soil profile by
altering the chemical and/or physical properties of the
disturbed soils. Soil amendments help provide a suitable
environment for vegetation establishment. Soil amendments
include, but are not limited to: wood chips, calcium
chloride, various organic mulches, gypsum, and lime. When
incorporated into the soil, these materials help mitigate
compaction problems, improve water infiltration, neutralize
acidic or alkaline conditions, modify soil structure, and
enhance water holding capacity while improving drainage.
|
Colorado School of Mines
|
ER-4609 305
11.4.5 Seed Selection and Handling (after BLM, 1992)
The following are some general guidelines for seeding:
Select species from a similar climatic zone and
soil type. Minimum moisture requirements should
determine selection.
Seed a mixture of species. Consider species for
both warm and cool season growth. Use a good
balance of types which provide for the planned
post-reclamation use, such as grazing or wildlife
habitat.
Do not over-seed. Too many seedlings will compete
for available moisture and nutrients.
Protect seeded areas from use until the vegetative
cover is established and self-sustaining.
11.4.5.1 Species Selection (after BLM, 1992)
Selection of adaptable plant species is essential for
successful reclamation. In severe environments such as
deserts, alpine zones, or windy ridgetop exposures, the
number of adaptable species will be less than for sites in
moderate climates. The proper selection of adaptable plant
species will depend on the prevailing climatic and soil
conditions in the project area (see appropriate seed
selection handbook for your area). The seed selection
should be consistent with the Resource Management Plan post
|
Colorado School of Mines
|
ER-4609 307
legumes), litter production, nitrogen fixing capabilities,
etc. Observe native plant species growing in the project
area on both the undisturbed lands and disturbed lands.
The purpose for seeding is to establish ground cover
and protect the soil from erosion and to prevent invasion of
undesirable species. Grass species are best suited to this
because they have a fibrous root system. Sod forming
species are best at reducing erosion.
Consult with BLM specialists, county agents, or other
experts, and appropriate research reports regarding
reclamation research and proper seed selection.
Some general considerations for species selection
follow:
Recommendations for species selection can be
obtained from the BLM, the Soil Conservation
Service (SCS), or the Forest Service.
Elevations and slope aspect are also important
factors that should be considered when selecting
plant species.
The Soil Conservation Service (SCS) Plant Material
Centers has information about seed and seed
dealers. The Centers are an excellent source of
information. Also consider botanical gardens and
native plant organizations as possible sources.
All seed purchased should have species name,
percent germination, percent pure live seed,
percent weed seed and other contaminates,
collection location (especially important for
|
Colorado School of Mines
|
ER-4609 308
native species) and testing date specified on bag.
States require that seed planted within the State
contain no injurious or noxious weeds.
11.4.5.2 Seed Acquisition (after BLM, 1992)
Seed of adaptable plant species may be purchased or
collected from native plants in the vicinity of the project
site. Some guidelines to seed acquisition include :
Purchase seed from dealers with experience in the
geographic area.
If collection of native seeds is viable, locate
appropriate stands of adaptable seed species
before the seed matures and collect the seed only
after it matures. Collect seed in cloth or paper
containers but never seal in plastic bags as this
practice may retain moisture and cause molding of
the seed.
Store the cleaned seed in a cool dry location in
cloth bags. Be sure the germination percent,
collection location, pure live seed, and percent
weed contaminates are specified on the bag label.
BLM may inspect the labels prior to the
application of the seed.
Legume seed and certain other types of seed (e.g.
bitterbrush seed) should be inoculated for best
results.
|
Colorado School of Mines
|
ER-4609 309
11.4.6 Seeding and Planting (after BLM, 1992)
Seeding and planting should be done as soon as the
seedbed preparation is completed, and if possible schedule
the seeding just prior to the longest precipitation period
or when available moisture is most favorable for seedling
establishment. In many locations, seeding prior to snowfall
enhances germination success in the spring. By beginning
the revegetation immediately after the seedbed preparation
is completed, competitive, less desirable species will not
be given an advantage and the seedbed will not degrade
physically or biologically. Quick establishment of
vegetative cover protects the soil from erosion. Seeding
and planting patterns should be designed to best provide the
desired post-mining use.
Seeding rates must be based on pure live seed (PLS)
percentages and seeds per square foot or pounds of pure live
seed per acre. Seeding rates which are too low may result
in sparse stands which may fail to stabilize the site, while
excessive rates waste seed and may result in stagnant,
overly dense stands with reduced plant vigor. The seed
mixtures and application rates should be described in the
plan.
Two basic seeding techniques are drill seeding and
broadcast seeding. The type of seeding to be used is
dependent upon the terrain and species to be used, and both
methods may be employed at the same site. Broadcast seeding
can be divided into ground seeding, aerial seeding, and
hydroseeding. Drill seeding is considered an effective
method of seeding for most grass species, while other
species must be broadcast. If the seedbed is smooth and
free of large rocks, consider seeding the site with a
|
Colorado School of Mines
|
ER-4609 310
cultipacker-type seeder to assure the seeds are evenly
distributed and to control seeding depth. However, if the
site is rough and rocky, a rangeland type drill may be more
effective. Where broadcast seeding is the only alternative,
do the seeding immediately after the site has been prepared
and cover the seed by raking to provide for a good seed-soil
contact.
Seeding depth is important for successful germination.
Generally, small seed should be seeded closer to the soil
surface than large seed. Most seeds should be planted from
1/8 inch to 1/2 inch deep, depending upon seed size and
type. Seeding too deeply delays emergence and reduces total
emergence, while seeding too shallowly increases desiccation
and causes faulty root systems. Covering most seed is
important. Some seeds will not germinate when uncovered,
birds and rodents will feed on the exposed seed and seed may
wash away before it germinates.
Steep slopes and rocky soils may prohibit the use of
most mechanical seeding equipment. Where equipment can be
used, seed drills will usually ensure good seeding success.
Special note should be given to the depth of planting. The
appropriate depth of planting for the selected species
should be used.
Broadcast seeding is often required on portions of the
disturbed area. Some types of seed should always be covered
with soil. Some suggested methods are a weighted chain link
fence, light chain, culti-peater, or harrow. Broadcast
seeding works best when done just after completion of the
final earthwork, when the surface is soft.
|
Colorado School of Mines
|
ER-4609 312
Adapted species should be selected for use with
parent material from a similar climate zone.
Care and hardening of the plants should be
considered prior to planting. This can be done by
the supplier.
Adjust time of planting to local conditions.
Site conditions and preparation at the time of
planting.
11.4.8 Mulching (after BLM, 1992)
Mulches can be used in reclamation to stabilize soils
until permanent plant cover becomes established. Mulches
not only reduce or prevent wind and water erosion, a good
mulch cover will protect the seeded area from the severe
effects of heat, cold and drought. Mulching materials can
be organic or inorganic, natural or man-made, soil enriching
or inert. When organic mulches are decomposing they can
create a serious carbon/nitrogen imbalance in the soil and
may require additional nitrogen fertilizer to compensate for
the nitrogen tied up in decomposing the mulch. Annual or
non-competitive perennial cover crops may also be used as
mulch. Commonly used mulches include, straw, hay, jute,
wood chips and other woody material, and synthetic
biodegradable fibers.
Hay and straw mulches should be applied at the rate of
2000 to 3000 pounds per acre. Fiber mulches are best
applied as a hydromulch (in a slurry of water and
|
Colorado School of Mines
|
ER-4609 313
tackifiers) at a rate of at least 2000 pounds per acre. If
seed and fertilizer are added to the hydro-mulch, caution
should be taken to ensure the addition of fertilizer to the
slurry does not make the slurry toxic to the seed. Apply
the hydromulch to a rough surface, such as an exposed road
cut, using suitable tackifiers to keep the mulch in place.
The use of hydro-mulching on cut-banks is effective for
distances up to 150 feet.
Light-colored mulches will reduce summer soil
temperatures while dark-colored mulches raise the soil
temperatures (effective for raising spring soil
temperatures).
The following are suggestions for using mulches :
Commonly used mulches include ; straw, crushed
rock, hay, synthetic mulches, biodegradable fibers
and blankets, wood chips and wood fiber, and jute.
Care should be taken to ensure that hay mulch
does not include noxious weed seeds.
Dark-colored mulch will raise spring soil surface
temperatures.
Light-colored mulches will reduce summer soil
surface temperatures.
Mulching will reduce frost heaving of new
seedlings.
Mulch reduces rain splash, surface wind, particle
movement and other erosional effects.
|
Colorado School of Mines
|
ER-4609 314
Mulch should be applied to a roughened surface.
Do not grade smooth. Apply asphalt or other
suitable tackifiers or crimp mulch into the
surface to keep it in place.
Hay and straw mulches for seeding cover and
erosion control should be applied at the rate of
1,000 to 3,000 pounds per acre. This amount will
provide a 1-to 3-inch deep ground cover.
Mulch can be applied by and on 3:1 or less sloping
sites up to 1 or 2 acres in size. Larger, steeper
sites will require a power blower or mulcher.
These power mulchers have a range of approximately
150 feet from an access road.
Fiber mulches can be applied effectively in a
slurry of water, seed, and fertilizer with a
hydromulcher. In low-precipitation areas, seed
should be applied prior to hydromulching.
Mulching that is crimped into the soil on dry
sites may wick moisture out of the soil in some
conditions.
The use of seeded blankets may be a viable
alternative to separate seeding and mulching,
especially on steep slopes.
|
Colorado School of Mines
|
ER-4609 315
11.4.9 Revegetation of Acidic Mining Wastes
(after BLM, 1992)
Revegetation of acidic mine wastes can pose
particularly difficult long-term reclamation problems.
Acidic mine wastes are toxic to most vegetation. Virtually
all mines which recover ore from sulfide minerals have some
potential for acid mine waste, either as tailings piles,
waste rock dumps, or low-grade ore stockpiles. Acidic
wastes must either be amended chemically or isolated from
the weathering environment in order for ultimate reclamation
to be successful. The exact measures necessary to ensure
reclamation success will depend on a variety of site-
specific factors. Often, acidic mine wastes will require
some form of engineered cover system to isolate wastes from
plant rooting zones. Capillary breaks are effective means
of isolating the waste materials. For a detailed discussion
of this topic, refer to Volumes I and II, Draft Acid Rock
Drainage Technical Guide, prepared for the British Columbia
Acid Mine Drainage Task Force.
11.4.9.1 Lime Amendment (after BLM, 1992)
Inclusion of a lime amendment into the cover system may
help prevent acidification and improve the potential for
revegetation success. Lime amendments may also have other
applications. A lime amendment is particularly effective
when the cover system includes a capillary break from the
acidic materials below. Inclusion of lime into a cover
system is not likely to be effective in reducing acid mine
drainage caused by water infiltration through the cover
system. The waste material can usually release sufficient
|
Colorado School of Mines
|
ER-4609 316
acid to overcome the effect of the amendment.
Where the net neutralization potential (NNP) of waste
rock or one of the cover layers is negative (i.e. the
material is acidic), it may be beneficial to incorporate
lime into the cover system as a neutralizing agent. When
considering lime application it is important to determine if
it will be necessary to add lime above the amount indicated
by the NNP. The NNP is a minimum number and it is often
necessary to substantially increase the amount of lime added
to account for other natural processes, such as
precipitation of iron on the lime, which limit the
availability of the lime for acid neutralization. The rate
of lime incorporation is usually expressed in tons of CaCOg
necessary to effectively neutralize 1000 tons of waste
material.
Lime can be added in several different forms. Slaked
lime (CaO) and hydrated lime (CaOH) are the most effective
neutralization agents. However, the relative abundance and
correspondingly lower cost of limestone (CaCOg) make it more
common for this use. Lime amendments are usually disked or
harrowed into the surface to prevent coating and subsequent
reduction of moisture infiltration. Application of more
than 30 tons of lime per acre may prove to be impractical.
It is also possible to mix lime into waste material in
batches to assure even distribution. It is best to have a
range of sizes present in the lime amendment to ensure
consistent reaction and acid consumption. Normally, an
agricultural grind meets this requirement.
|
Colorado School of Mines
|
ER-4609 317
11.4.9.2 Bactericides (after BLM, 1992)
The oxidation of sulfide minerals is catalyzed by the
bacteria Thiobacillus ferrooxidans. This bacteria can speed
the reaction by several orders of magnitude. The activity
of the bacteria can be limited by the application of
bactericides to pyritic waste minerals and cover systems.
Bactericides should be considered short term remedies.
11.4.10 Test Plots (after BLM, 1992)
It is often appropriate for an operator to install test
plots prior to revegetation of a large disturbed area. This
process will enable the proper seed mixture, fertilization
type and rate, and other soil amendment requirements
identified in the reclamation plan to be evaluated on a
site-specific basis. In addition, it allows for the use of
new and innovative techniques which have not been widely
proven. A major advantage in using test plots is that
failures are much less costly to the operator and the
environment than "real-life" failures. Requirements for
test plots should be developed in conjunction with BLM
renewable resource specialists, such as range
conservationists, wildlife biologists, soil scientists, and
surface protection specialists.
|
Colorado School of Mines
|
ER-4609 319
12.2 Getting Started
Before searching for literature references using WASTE,
it is necessary to be familiar with certain hardware and
software requirements. This program runs on any IBM
compatible computer. The code was written in dBase IV
version 1.1, and compiled using Arago Quicksilver version
2.5. The program is formatted to be used in a single user
environment. The program WASTE, because of the size of the
executable file, is designed to be run from the computer's
hard-drive.
12.2.1 Installation
Three disks contain the files needed to run WASTE.
These disks are located in the thesis pocket. From the C:>
prompt in DOS, create a directory using the MKDIR ^directory
name> command. Enter this new directory by typing CD
<directory name>. Copy the contents of all three disks into
the new directory. From disk 1, copy the files WASTE. EXE,
VALIDITY.MEM, LIBRARYT.DBF, and LIBRARYT.DBT into the new
directory. LIBRARY3.MDX and LIBRARY3.DBF are contained on
disk 2, while LIBRARY3. DBT is contained on disk 3. These
three files should also be copied into the new directory.
The Following files are needed to run the program:
WASTE.EXE The compiled executable file.
LIBRARY3.DBF The database file containing all
references.
|
Colorado School of Mines
|
ER-4609 320
LIBRARY3.DBT The file containing the synopsis of all
references.
LIBRARY3.MDX Multiple index file for library3.dbf and
used to search for references.
LIBRARY3.DBF A temporary file used to store newly
added records.
LIBRARY3.DBT The temporary file used to store the
synopsis of new records.
VALIDITY.MEM This file stores the password necessary
to add, edit, or delete records.
It is important to always keep a floppy disk with
backup copies of the files. When running the program for
the first time, WASTE creates its own multiple index file
(Library3.wdx) plus an index file for each index in
Library3.wdx (Library3.wOO, Library3.W01, etc.). Therefore,
always make a backup of these files as well.
12.3 Running the Program
To start the program, make sure your in the directory
created for the WASTE files, and from the DOS C : > prompt
type the command WASTE and press <enter>. Select the
desired operation by highlighting one of the menu options by
using the arrow keys on the keyboard: ADD; EDIT ; SEARCH;
and QUIT, at the top of the screen. Then press <enter>.
Menu selections are controlled through the use of pop
|
Colorado School of Mines
|
ER-4609 321
up windows. Very few applications require the user to input
any data.
Each pop-up menu has a return option that brings back
the previous menu of the program. A brief explanation of
what each menu item does is provided below.
ADD Allows new records to be added to
the database.
EDIT Allows editing or deleting existing records
SEARCH Allows searching for references in the
database. All records can be searched, or
the search can be limited to articles, books,
proceedings, or Government Documents.
QUIT Exits the Program.
Before the program is made available for searching, a
password should be designated to protect the ADD and EDIT
modes of the program. The ADD and EDIT modes modify the
database files and the index files; and the data therein
maybe corrupted if these operations are conducted by someone
unfamiliar with how references are added and edited. The
search for references with WASTE does not require a
password.
12.4 Setting up a password
Highlight the menu option ADD or EDIT and press
<enter>. The computer will ask for a password. Type
|
Colorado School of Mines
|
ER-4609 322
NWPSSWRD for a new password, and press <enter>. The
computer then asks the user for the old password. Type
WASTE <enter> (the program uses the password WASTE until a
new password is entered) . The computer then asks for the
new password. A minimum of three and a maximum of eight
alphabetical characters will be accepted. A coded version
of the password is saved in the file VALIDITY .MEM. Each
time the EDIT or ADD modes of the program are used, the user
will be required to enter the password.
12.5 The Add Mode
The ADD mode allows the database to be updated with new
references. After the password is entered, the data entry
screen is accessed. To save a new record in the database,
the entry screen will require data in the following fields :
(1) the date field, (2) the title field, and (3) the type
field. If data is not included in any one of these fields,
the entry will not be added to the database. WARNING!
While in the data entry screen, do not press the escape key
or reboot the computer. Doing so may corrupt the database.
To exit the ADD mode, simply enter a blank record. The
computer will ask if you want these records included in the
database. Enter <Y> for yes and <N> for no.
The following rules should be applied for each data
entry field:
ID NUMBER
Do Not Change the ID Number. This number is unique for
each record and is assigned automatically by the program.
|
Colorado School of Mines
|
ER-4609 323
First Author/Editor
This data is not required; however, in order to perform
a name search it should be included. Enter the author's
last name followed by given name and initials. A maximum
length of 60 characters is accepted. Enter only one author
for each field. Use only alphabetical characters, and do
not include dots (.) or commas (,) after or between names.
Second Author/Editor
The same rules apply as for the first author field.
Third Author/Editor
The same rules apply as for the first author field.
TITLE
Data is required. Avoid using hyphens (-). If hyphens
are used, the reference may not be found in a key word
search. A maximum title length of 14 6 characters is
allowed.
DATE (mm/dd/yy)
Data is required. Enter the publication date.
TYPE
Data is required. This entry field determines whether
the reference should be included among articles, books,
proceedings, or government documents. Enter "a" for
article, "b" for book, "p" for proceeding, and "g" for
government document (without the quotes).
INI
For an article, enter the journals name followed by:
|
Colorado School of Mines
|
ER-4609 324
volume number; number; month; and year published; and page
numbers. For example : Mine and Quarry, v 5 no 72 April 1982
p 68-72. When browsing among the references, only the first
60 characters are displayed under JOURNAL. For all other
types (books, proceedings, and government documents) , the
INI field is only displayed while viewing full records. INI
should contain extra information about the publication;
(e.g., additional authors (if more than three), publisher,
ISBN/ISSN Number, Library of Congress Catalog Card Number,
if bibliographical references are included or not, etc. A
maximum length of 254 characters is allowed.
IN2
The IN2 field is only displayed when viewing the whole
record. This field should contain additional publication
information not able to fit into INI.
SUMMARY
The memo field contains the summary/synopsis of the
publication. This can be the author abstract, author
preface, introduction, or a summary written by the person
entering the data.
To open the editing window, place the courser at the
memo marker and press <Ctrl Home>. Type the summary of the
record and save it. Exit by pressing <Ctrl End>.
KEYWORD#
Keywords are necessary for performing a word search.
This program contains six keyword fields, each having a
maximum of 60 characters. Each keyword field is independent
of the others. The program searches for the keywords one
field at a time.
|
Colorado School of Mines
|
ER-4609 325
For example, assume there was a conference about mill
tailings in California that, among other things, covered a
topic on earthquake loading of tailings dams. Further
assume that the first keyword field has the following
keywords : mill tailings conferences California, and that the
second keyword field has the following keywords: earthquake
loading tailings dams. If a search is made using the
keywords : mill tailings California, this record will be
found. However, if a search is made with the keywords:
earthquake loading conference, or tailings dams California,
the record will not be found. It is, therefore, very
important to think through each keyword field. Note that
the word search will also search the title field as an
independent keyword field. Spaces should be used to
separate keywords. Do not include dots (.) or commas (,)
after or between the keywords.
12 .6 The Edit Mode
The EDIT mode of the program allows the database
records to be changed or deleted. After the password has
been entered, enter the 14 digit ID number of the record to
be edited. This number can be seen when displaying the full
record in the search mode of the program. The guidelines
and rules explaining how data should be entered in EDIT mode
are the same as described in section 12.5 for the ADD mode.
If the DELETE option is selected, the program will ask
for the ID number and that record will be displayed. The
program will ask "Delete this Record? (Y/N)". Enter <Y>, if
the record is to be deleted, if not, enter <N>. A maximum
of 20 records can be deleted each time. To discontinue the
|
Colorado School of Mines
|
ER-4609 326
DELETE process, press the enter key when asked for the ID
number. The program will then ask "Permanently remove the
records marked for deletion? (Y/N)”. If the answer is <Y>,
all records marked for deletion will be deleted from the
database. If the answer is <N>, there will be an
opportunity to recall the records one by one. The program
will ask "Recall which record number:". Enter there record
number of references to be recalled to the database.
12.7 The Search Mode
The user can search for references within all the
records in the database, or limit the search to articles,
books, proceedings, or government documents. Limiting the
search will increase the speed of a name of word search.
After selecting the location to make the search (the
whole database, articles, books, proceedings, or government
documents) , it is necessary to select the type of search
desired. Three types of searches are available : (1) a name
search, (2) a word search, and (3) browse. Each of the
search alternatives are explained below:
12.7.1 Name Search
This option is used to search for an author/editor
name. It is only possible to search for one author/editor
at a time. Enter the name (last name and given names or
initials) of one author or editor.
If the program finds, for example, five references by
this author/editor, the following message will be displayed:
"5 references with author :" <name> (where <name> is the
|
Colorado School of Mines
|
ER-4609 327
name used in the search) . A menu pops up that enables the
user to display the references found or return to the
previous menu.
If the display option is selected, all the references
found, up to a maximum of 1000, will be displayed, page-by-
page. At the bottom of each page the user will be given the
following options: to quit; to continue the display; to
display the previous page; or to display a specific full
record. Enter: <Q> to quit; <P> to display the previous
page; <ENTER> to continue the display; and <record number>
to display the full record. In most cases, the full record
also contains a summary of the of the reference.
12.7.2 Word Search
This option enables the user to search keywords. After
selecting this option, the user is asked to enter keywords
for which to search. Enter the keywords in priority order,
with the most important first. Spaces should be used to
separate keywords. Do not include dots (.) or commas (,)
after or between keywords.
If the program finds any references, the following will
be displayed: "# of references with keywords:" <keywords>.
A menu pops up that allows the user to display the
references found.
If the display option is selected, all the references
found will be displayed page-by-page.
If the program does not find any references with the
keywords entered, it will eliminate the last keyword and
look for the others. For example, assume that the following
keywords were entered: tailings dam design drains. If the
|
Colorado School of Mines
|
ER-4609 328
program does not find any references with all four keywords,
it will eliminate the last one and look for tailings dam
design. If no references are found with this set of
keywords, it will look for tailings dam. Assume that five
references with this set of keywords is found. The
following will be displayed: "No references with keywords :
tailings dam design drains. 5 references with keywords :
tailings dam. The user will once again have the option of
displaying the references found, or to return for a new
search.
At the bottom of each page, the user is given the
following options : to quit; to continue the display; to
display the previous page; or to display a specific full
record. Enter: <Q> to quit; <P> to display the previous
page; <ENTER> to continue the display; and <record number>
to display the full record.
12.7.3 Browse
This option allows the user to browse in the entire
database between selected years. When this option is
selected, the user is asked to enter the years between which
to browse. Enter years between 1970 and 1999.
All the references found, up to a maximum of 1000, will
be displayed, page-by-page. At the bottom of each page, the
user is given the following options : to quit; to continue
the display; to display the previous page; or to display a
specific full record. Enter : <Q> to quit; <P> to display
the previous page; <ENTER> to continue the display; and
<record number> to display the full record.
|
Colorado School of Mines
|
ABSTRACT
The object of this research program is to simulate gravity flow of ore in
ore passes by development and implementation of a Discrete Element Methodology. Two
major concerns in this regard are (1) the designed load capacity of gate assemblies at the
bottom of ore passes and (2) hang-ups. Either gate failure or hang-up removal can result
in the sudden, uncontrolled spillage of a large amount of rock and subsequent fatalities
and injuries to miners.
Two-dimensional Discrete Element Method (DEM) simulations were performed
in order to study the effects of ore characteristics and ore passes configuration on gravity
flow of ore in ore passes. The study shows that ore shape can significantly affect both
static and dynamic loads on the ore pass gate assembly.
In addition, the ore pass configuration and the existence of a Dogleg are shown,
having a major influence on the flow behavior of ore. When the simulation results are
compared with experiments and the classical approach (such as the Jansen's formula)
DEM calculation have shown promsing results. Visualization techniques, which benefit
from today's advancements in computer technology, are applied to accommodate better
understanding of the results.
The results of these analyses have proven that the DEM is a powerful tool for
modeling the loading on ore pass gate systems and gravity flow of ore in ore passes.
|
Colorado School of Mines
|
2.9 Wear and lining of ore pass walls 36
3. THE DISCRETE ELEMENT METHOD 37
3.1 Introduction 37
3.2 Governing equation 39
3.3 Cluster theory 3 9
3.4 Contact detection determination 41
3.4.1 Contact detection of two ore-disk particles 42
3.4.2 Contact detection of an ore-disk particle with a boundary 44
3.5 Contact forces 47
3.5.1 Automatic detection of possible contact 47
3.5.2 Grid method 49
3.5.3 Determination of contact coefficients 5 6
3.5.4 Contact force calculations 60
3.5.5 Time integration and critical time step selection 63
3.5.6 System stability, energy checking 65
4. MODELING GRAVITY FLOW OF ORE WITH DEM 69
4.1 Ore shape characteristics and size distribution after blasting 70
4.2 ORE-CLUSTER SHAPE AND SIZE DETERMINATION FOE DEM ANALYSIS 76
4.3 Ore pass configurations 81
4.4 Ore pass simulations, load on bottom gate 82
4.4.1 Effect of ore pass inclination 83
4.4.2 Effect of ore-cluster shape 8 6
4.4.3 Friction effect 90
4.4.4 SlZE-DEVIATION EFFECT 93
4.4.5 Dogleg effect 95
4.4.6 Effect of initial dumping of ore material 99
4.4.7 Effect of contact stiffness 102
4.5 Ore pass simulations, flow of ore in ore passes 105
V
|
Colorado School of Mines
|
1
CHAPTER 1
INTRODUCTION
Many industries store and handle solid materials in bulk form. When the volume
of the solids is large, industry invariably relies upon to gravity induce the solids to flow
out of storage, through channels and reactors. These materials (ore, cement, flour,
polypropylene) are generally described as bulk solids or materials. The gravity flow of
billions of tons of materials occurs annually in thousands of installations. For example,
the mining industry relies on gravity flow in block and sublevel caving, as well as in
storage and loading bins.
Underground mining can be described as an exercise in bulk materials handling
where men and equipment produce ore, Blight et al. (1994). In this context an ore pass is
an element in the handling system, which is also comprised of stopes, a crusher, and ore
bins, loading pocket and the mine shaft. The configuration, size and purpose of ore passes
vary widely. They may be located at the surface or underground, vertical or sloped,
straight or doglegged, circular or rectangular in cross-section, designed for intermediate
storage, or as a chute. Figure 1.1.
Chronologically, an ore pass is the last development constructed prior to
producing the ore. Given time constraints from the pressure to produce, industry often
pays inadequate attention to details concerning ore pass location and design. By contrast,
other elements in the handling system are rigorously designed as permanent features
throughout the economic life of the mine. Consequently, it is not surprising that ore
passes frequently present problems or even fail.
|
Colorado School of Mines
|
3
1.1 Problem statement
An important class of industrial problems involves the control and prediction of
flow of solids that behave in a highly discontinuous manner. The material flow problems
associated with mining disciplines often require sophisticated computer simulation tools
in order to develop a solution.
One example of this is the flow of ore in an ore pass. Although ore passes, chutes
and gate systems for mining of metal and non-metal mines must meet the requirements
specified in the U.S. Code of Federal Regulations (CFR), part 57, 75, recent structural
failure of ore pass linings and gates have underlined the lack of adequate ore pass design
standards available to both U.S. Mine Safety and Health Administration (MSHA) and
mining engineers.
Beus et al. (1997) stated that a review of MSHA statistics for period of 1975-1995
show that nearly 75% of injuries in U.S. underground metal mines are directly or
indirectly related to pulling or freeing of ore pass chutes, the use of hand tools in ore
passes, falls of broken rock in an ore pass and structural failures of chutes or gates and
ore pass walls.
The MSHA database (1987-1996) on accidents and fatalities associated to the
above mentioned problems related to ore passes, reports a total 743 nonfatal accidents
(e.g., permanent disability, injury, occupational illness) and eight fatalities, MSHA
(1998).
1.2 Proposed research
The main objective of this research is the development, application and validation
of a numerical methodology to study the operational performance of ore passes which
include better understanding of ore flow behavior in ore passes and conditions for
|
Colorado School of Mines
|
4
occurrence of hang-ups. Experiments by analogy with flow of other materials, full-scale
field studies and numerical modeling are the potential methods. Analogy of ore flow with
other materials such as sand could not be done precisely because of ore shape, size,
discharge rate and boundary conditions in ore pass operations are not analogous to
experiment conditions. Full-scale field studies are generally very expensive, they create
interruptions in regular mining operation and production, and could provide only
localized information. Among different methods of numerical modeling, the DEM due to
its ability to simulate accurately the mechanical behavior of granular ore materials has
been selected. The current research has been focused in two major areas, which are
summarized below:
1.2.1 Development and application of the 2-D discrete element method
♦ Development and customization of an existing two-dimensional DEM computer code
Mustoe (1998) for ore passes analysis
♦ To model general shaped ore material, development and application of ore-cluster
shaped particle (overlapping of rigid disk-shape particles).
♦ Improvement in modeling techniques for boundaries of problems (translational and
angular movement of ore pass gates).
♦ Development and implementation of a model for static and dynamic frictional
behavior.
♦ Improvement in simulation efficiency and accuracy by checking the stability and
equilibrium of system during the simulation period.
|
Colorado School of Mines
|
5
1.2.2 Application of dem to ore passes
♦ Understanding the ore pass design procedures and the key parameters required in
computer simulation.
♦ Determine general shapes and size distributions of ore after blasting.
♦ Generate the ore pass geometry (main ore pass body, feeder, dogleg, gate), different
shape and size ore (dry with negligible cohesion), with defined characteristics.
♦ Draw conclusions about the modeling of ore passes with DEM.
1.3 The discrete element method
Discrete Element Methods are a family of numerical procedures specifically
designed for simulating the mechanical behavior of systems of discrete, interacting
bodies. It should be noted that many finite element, boundary element and Lagrangian
finite difference programs have interface elements or “slide lines” that enable them to
model a discontinuous material to some extent. However, Cundall (1989), proposed that
the term of “discrete element method” should apply to a computer program only if the
algorithm:
♦ Allows finite displacement and rotations of discrete bodies, including complete
detachment, and
♦ Recognizes new contacts automatically as the calculation progresses.
The DEM explicitly models the dynamic motion and mechanical interaction of
each particle (body) throughout a simulation and provides a detailed description of the
positions, velocities and forces acting on each particle at discrete points in time during
the analysis. These discrete points are called time steps.
|
Colorado School of Mines
|
6
Up to the present, two conferences (1989, 1993) have been organized on DEM.
Their proceedings provide good sources of information on discrete element methods and
their applications in applied science and engineering.
The foundation of the DEM was first developed in Peter Cundall’s Ph.D. thesis at
the Imperial College in London, (1971). Then researchers such as Burman (1971),
Rodriguez-Ortiz (1974) and Hocking (1977) introduced more aspects of DEM in
different, related engineering problems.
Around the mid-1970s, use of DEM became more generalized at University of
Minnesota through a research effort supported by Corps of Engineers (Cundall and
Cundall et al. 1974; 1975). Cundall has continued his contribution to DEM modeling in
two fields (Itasca consulting group):
♦ Polyhedral blocks in two and three dimensions, UDEC (1980) and 3DEC (1985)
♦ Particles modeled as disks and spheres, PFC2D and PFC3D (1995)
Hocking, Mustoe and Williams in the early and mid-1980s made a major
contribution to DEM modeling. They initiated introduction of three-dimensional contact
of polyhedral blocks, fracture of brittle plates, and the generation of ice force-
displacement laws for ice-structure interaction problems, Hocking et al. (1987), Mustoe
et al. (1987), Williams et al. (1985).
Mustoe continued his research on DEM, working on a plate and beam element to
simulate deep-ocean pipes in two dimensions: Mustoe (1989), Mustoe et al. (1992),
Mustoe (1992).
Hustrulid (1993) and Zhang (1996), used beam formulation respectively for their
Master's thesis (modeling conveyor belt system) and Ph.D. thesis (modeling of flexible
boundaries for two dimension material compaction). Zhang (1993) and Mathews (1994)
completed their master’s degree by implementation of a two-dimensional DEM disk
element respectively for hydraulic and particle flow problems.
|
Colorado School of Mines
|
7
Hustrulid (1997) developed a computational methodology for modeling large-
scale sublevel caving with a three-dimensional Discrete Element Method (sphere
element). Mustoe and Griffiths (1998) investigated the development of a Discrete
Element Method within a Finite Element framework to model the mechanical behavior of
an isotropic elastic continuum. Mustoe has done research on a superquadrics shape
particle as well, Mustoe et al. (1993; 2000).
Williams and Pentland (1989) at the Massachusetts Institute of Technology (MIT)
pioneered the use of superquadric and hyperquadrics elements. They used DEM in an
environment of parametric shape, Barr (1981), to analyze the dynamic impact of a ball on
a two-by-four piece of wood and to design a chair. Williams continued his contribution to
DEM by developing a new contact detection scheme based on a heap- sorting algorithm
and an object representation method for better contact resolution between arbitrary
geometries, O Connor et al. (1993), O'Connor (1996), Williams et al. (1995a). Williams
also has investigated the formation of coherent structures in deforming granular materials
by use of their code "MIMES", Williams and Rege (1995b; 1996).
Bruno (1996), by using DEM, modeled the influences of saturation and flow rate
on the episodic progression and stabilization of sanding cavities in oil wells. At the
Geomechanics Department of Sandia National Laboratories O'Connor et al. (1997)
investigated the combination of a Finite Element flow model (based on Darcy flow
formulation) with sets of DEM formulation on modeling of sand production in oil
recovery process.
Sawamoto et al. (1998) proposed a new analytical approach for assessing local
damage to reinforced concrete structures subjected to an impact load of a rigid and
deformable missiles by applying DEM. Furthermore, the various impact response
characteristics and failure mechanisms, such as impact forces, penetration behavior,
reduction in missile velocity and energy transfer process (which are difficult to obtain
experimentally) are analytically evaluated.
|
Colorado School of Mines
|
8
Ting et al. (1989) presented a comprehensive study of soil mechanics by using a
two-dimensional disk. With further research indicating the importance of rolling
mechanism of deformation in granular systems consisting of perfectly round particles,
researchers focused attention on the use of ellipse-shaped particles in DEM numerical
modeling, Ng (1992), Rothenburg and Bathurst (1991; 1993), Ting (1991; 1992), Wei
(1991). The main challenge facing simulation of elliptical particle interactions would be
the development of a reliable contact detection algorithm. Ting et al. (1993a; 1993b)
presented the methodology for contact detection between two ellipses, an ellipse and a
boundary in two-dimensions. Researchers from the Russian Academy of Science used
assemblies of elastic elliptic particles to study the mechanics of disordered (amorphous)
and ordered (crystal) bodies: the glass-liquid transition, irreversible deformation and
intermediate (liquid-crystal) state, Berlin et al. (1996). Ouadfel and Rothenburg (1999)
modified the DEM program TRUBAL originally written to simulate the behavior of
assemblies of spheres for an inter-ellipsoid contact detection algorithm. The modified
program was used to perform deviatoric and axi symmetric compression tests on 1000
prolate (increasing elongation ratio, while maintaining the particle volume) spheroids in
periodic space. They emphasized that the numerically obtained stress-strain curves
conformed with experimental evidence both qualitatively and quantitatively.
Although by use of non-circular-shape particles (e.g., elliptical, oval) the shape
and fabrics of granular materials can be modeled closer to reality, complexity and
computational time are increased significantly. As an example, the contact detection of an
ellipse-ellipse contact location, Ouadefl (1999) requires the solution of a quadric-
equation vs. the first-order equation for circular particle/particle contact detection.
Previous work by a number of researchers has demonstrated that the excessive
rotation of circular particles causes unrealistically low global strength. One possible
solution, constraining all/some rotations, or artificially increasing polar moment of inertia
may dramatically increase the computed global friction, at the same time enforcing the
fact that these types of constrain underlie the premise of capturing realistic particles
|
Colorado School of Mines
|
9
interaction by DEM. This fact encourages researchers looking for an alternative and
better way to model particle shape, Thomas and Bray (1999).
As a solution to the above-mentioned problem, Potyondy et al. (1996) presented a
computational methodology for simulating inelastic deformation and fracture of rock (in
which a densely packed assembly of arbitrarily sized circular particles bonded together at
their contact points represents the rock). These bonds preclude sliding and limits the
allowable magnitudes of normal tension and shear force acting at the contact. A potential
application of this model is to estimate the state of damage surrounding an excavation, if
there are enough data on properties of rock formation, Potyondy and Cundall (1998).
Thomas and Bray (1999) presented a disk cluster which is a particle assembly
consisting of a group of individual disks rigidly and permanently connected into an
environment of DDAD (Discontinuous Deformation Analysis of Disks), an implicit
method which employs minimization of potential energy and the penalty method to solve
for the displacement of disks).
The research described in this thesis is development and implementation of a two-
dimensional DEM code, which is, used a cluster of cemented rigid disks and it is applied
to study the gravity flow of ore in ore pass system. These disk clusters more accurately
model the non-spherical shape of granular materials and exhibit fewer tendencies to
topple or rotate excessively. Because the DEM cluster algorithm is based on circular disk
geometry retains the advantages such as, simplicity of modeling the contact detection and
force calculation without significant reduction in computational speed.
1.4 Gravity flow of materials
The movement pattern of granular material under gravity is of great importance to
practical mine design. Mining relies on gravity flow in block caving and in ore passes, as
|
Colorado School of Mines
|
10
well as in storage and loading bins. In some cases, such as caving method, gravity flow
significantly affects the dilution of the drawn ore, the recovery of the ore body and
eventually, the overall efficiency of the caving method.
It should be emphasized that the term " gravity flow", means an uninterrupted
movement of coarse material and is a completely different phenomenon than flow of
liquids.
The first significant studies related to the storage and flow of bulk solids were
reported at the end of nineteenth century. That work originated from the need to store
large quantities of grain and was concerned mainly with wall pressure affecting the
structural design of silos and bins. The most famous researcher is Janssen (1895), who
developed his formula for prediction of the wall pressure.
His method is based on the concept of differential slices of infinitesimal thickness
and finite cross-section and perimeter. Janssen's analysis is based on three major
assumptions:
♦ That the stresses are uniform across any horizontal section of the material,
♦ That the vertical and horizontal stresses are the principal stresses, and
♦ That granular material is cohesionless, Nedderman (1992).
Jenike (1954; 1964) was another pioneer who investigated the flow of granular
materials. His work led to the postulation of a flow factor for flowability of channels, as a
ratio of consolidation pressure in a channel to obstruction pressure. The smaller the flow
factor, the better channel is categorized regarding the flow of solids.
Kvapil (1965a; 1965b) and Janeiled (1966) conducted experiments using a very
simple vertical glass model with a horizontally layered white and black filling. His
observations led him to develop successive ellipsoid extraction theory. After opening the
|
Colorado School of Mines
|
11
bottom outlet of a bin or bunker already filled with a granular material, the material
begins to flow out under the influence of gravity. After a certain time, all discharged
material will have originated from within an ellipsoid-like zone (ellipsoid of motion).
Material between this ellipsoid and another will have loosened and moved, but will not
travel to the bottom outlet (Lorig et al., 1995).
Peters (1982) has done some experiments with a 15-foot high-test facility. He
concluded that the draw envelope has a cylindrical midsection and an ellipsoidal top and
bottom, rather than the traditionally accepted constant eccentricity ellipsoidal form.
Chen (1984) tried to model the gravity flow of material by using stochastic
theory. In his model, the downward particle flow is represented in terms of an equivalent
upward-biased random flight of voids originating at the draw points.
Block caving methods are important mining techniques for the extraction of
relatively low-grade, wide expanded ore bodies, where ore panels, or ore blocks are
undercut to induce caving of ore, which is drawn off from below (Hartman, 1987). The
daily production from block-caving operations throughout the world is approximately
370,000 tons per day. See Laubscher (1994).
Unlike many underground mining methods, block caving requires extensive
development of infrastructure such as extraction of drifts; draw raises and crushing
facilities before the start of production. This makes the prediction of the cavability of
rock formations and gravity flow of caved rocks the primary factors in design process.
This subject has been studied in the past by a variety of methods, including:
♦ Experiments by analogy with the flow of other materials (e.g., sand) in bins and
bunker.
♦ Full-scale field studies
♦ Large -scale physical model
♦ Numerical modeling
|
Colorado School of Mines
|
12
Researchers argued that the flow of caved rock could not be simulated precisely
by the theories developed by the flow of other materials such as sand. This is because the
ore shape, size, discharge rates and boundary conditions in the real case were not
analogous to experiments condition, Yenge (1980).
The full-scale field studies of block caving are generally very expensive and can
provide only localized information for a specific mine. Marker recovery is currently the
only affordable testing procedure of caved ore in situ. It involves positioning of markers
about 1m length of electrical cable (with density slightly lighter than ore) in the ore body
prior to blasting and locating them as they are extracted. Markers may provide the
researchers with initial location of the markers and when and where it is extracted (speed
of flowed rock) without any information on the path traveled throughout the fragmented
rock, Lorig (1995); McNearny (1991).
McNearny (1991) performed large-scale physical models to study the drawing
behavior of a rock mass mined by the block caving, with half-sized brick as ore with an
undercut composed of one-inch nominal diameter rock.
Four models were constructed, using a steel and Plexiglas frame twenty feet long
by fifteen feet high and up to three feet the thickness. An approximate weight of fifty tons
of material was contained in each model. Additionally, by use of UDEC (Universal
Distinct Element Code), numerical modeling of each test was conducted.
DEM simulations by McNearny were concerned primarily with the initiation of
caving and did not consider the details process of gravity flow. He concluded that the
establishment of an arch could form the caving mechanism. Once the sides of the arch
were undercut by draw, material would fall into the void, forming a new arch, see Figure
1.3.
The introduction of fast computers with large memory provides an opportunity for
researchers to use DEM to develop quantitative relationship of the shapes and sizes of
motion and loosening in block caving. In 1985, a UDEC model was set up to model the
|
Colorado School of Mines
|
13
effects of an assisted caving operation within Schist and Gneissic country rock (Guest
and Cundall, 1992). Modeling indicates that the undercut could be initiated by the VCR
(Vertical Crater Retreat blasting) stoping and that over the area of undercut, caving would
propagate a distance of some fifteen meters before a stable arch would develop.
Lorig et al. (1995a; 1995b) presented results from an examination of numerical
modeling by DEM code at South African and Chilean caving mines with the objective of
testing the capability of this approach for prediction of cavability and reasonable
estimation of fragmentation. They concluded that PFC (Particle Flow Code, 2D and 3D)
is unlikely to be used directly in design and operation in the near future due to both
mechanical limitation (lack of knowledge about the material properties) and
computational limitation (speed and memory). They suggested that based on numerical
modeling results and calibration a simpler model should be developed, and then the
results would be used directly in design and operation.
Hustrulid (1997) presented the development and application of a three
dimensional (spherical particle) DEM code to study the material flow in sublevel caving
operations used by LKAB in their iron mine (Kiruna). Because of the size and
complexity of the large scale sublevel caving, he suggested that further research should
be involved in a more focused investigation of some sub-problems in the sublevel caving
that requires shorter CPU simulation time. One example of sub-problem can be the
simulation of LHD taking scoops of ore from muck pile.
Behavior of granular materials in a rotating drum has been of great technological
interest (e.g., to the pharmaceutical industry) for more than 200 years, Jaeger et al.
(1996). Nakagawa et al. and Yamane et al. (1998) have investigated the dynamic angle of
repose and steady particulate flows in rotating drum experimentally and numerically by
DEM. Using MRI (Magnetic Resonance Imaging) they found a good correlation between
the results from DEM simulation and experiments.
Ristow (1998) stated that: "Since a complete theoretical description for the
dynamics of granular materials is still in its infancy, numerical simulations are a very
|
Colorado School of Mines
|
14
valuable and sometimes even necessary tool to determine the static and dynamical
properties in granular systems."
Ristow studied the flow of granular materials in conical hopper and rotating drum
by use of DEM. He emphasized that in the case of a conical hopper, there is a good
agreement between results from two-dimensional simulations with their three-
dimensional counterpart, thus justifying the use of a two-dimensional computer program
in many similar cases.
Rong (1997) in his investigation performed a series of DEM simulation to
elucidate the mechanical behavior of particulate materials and their effect on wall
pressure in a silo during the filling and discharging. After using a moving average
method, the pressure distribution vs. height of silo was represented by a least squares fit
to the Janssen theory.
Two main numerical methods, Finite Element Method (FEM) and Discrete
Element Method (DEM) have been used in recent years to model the behavior of solids in
silos.
Holst et al. (1999a; 1999b) presented the results of an international collaborative
project to evaluate the capabilities of these two methods in assessing the flow of materials
in silo in 2D dimensions. For this purpose, a well-defined problem description of filling a
silo with a dry, cohesionless, granular solid was devised and sent to a large number of
researchers all over the world. Based on results from thirty-eight FEM and sixteen DEM
calculations, they concluded that FEM give smoother curves for wall pressure vs. silo's
height than DEM calculation.
It can be argued that the high scatter in DEM is a real outcome of the force
transmission systems in granular solids. If this is true, it indicates that a huge number of
particles are needed that providing meaningful predictions of silo phenomena without
artificial smoothing. One shortcoming of the continuum analysis (FEM) is that the silo
filling process cannot really be modeled. Finally because of the discrepancy in the
|
Colorado School of Mines
|
15
submitted results they argued that neither method of analysis is yet sufficiently developed
to capture all the essential features of granular solids behavior in silos.
Cleary (2000) presented results of investigation on modeling the flow of materials
in three case studies of dragline excavators, mixing in tumblers and centrifugal mills by
DEM. He emphasized the importance of particle shape and blockiness of the materials in
three mentioned case studies.
Hambley and Sigh (1983) and Hambley (1987) gave comprehensive design
guidelines of ore passes in open pit mines. Ferguson (1991), in response to request from
the Mining Research Directorate of Canada documented a design rational for new ore
passes and presented a review of current practices and procedures.
Blight and Haak (1994) from their experiments on the ore pass model (Figure
1.2), concluded that pressure on the gate of ore pass could be predicted by the Janssen
equation for inclined silo. Also decrease in ore pass inclination from 90 to 50 degree can
reduce the maximum impact factor (maximum dynamic load / weight of material) on
bottom gate of ore pass from 4.09 to 1.09.
Goodwill, et al. (1999) proposed that in a good design practice with use of Jenike
Flow Function (measured in the lab by running uniaxial shear test cell, "Jenike Shear
Tester") and flow factor (determined from Jenike chart), could eliminate or at least
substantially reduce, the severity of hang-up and wear problems often encountered in ore
pass operation.
Beus and Ruff (1997) reported an investigation by NIOSH (National Institute for
Occupational Safety and Health) on development of a mine hoist and ore pass research
facility in Spokane, Washington, to monitor hoisting and ore pass operation. One of the
main research tasks for this center was mentioned as: development of computer models to
analyze design and muck flow in ore pass to identify potential safety problems.
Beus et al. (1997) reported the investigation of hazards in and around ore passes
in hard rock mining sections. They employed risk assessment methods such as fault-tree
analysis to pinpoint the most probable reasons for ore pass failures and related injuries or
|
Colorado School of Mines
|
16
fatalities. The report also documents the construction of a full-scale muck-up typical of a
mine ore pass and support system to develop an installation for the measurement of static
and dynamic muck loads that can be implemented in the field.
Larson et al. (1998) at the Spokane Research Laboratory (NIOSH) investigated
the use of PFC3D to simulate the flow and interactions of rock particles to determine static
and dynamic load on an ore pass bottom gate. They modeled three different ore passes
(vertical, inclined eighty degree, doglegged at eighty degree) hit by a single rock particle
of 30 cm diameter, showing the reduction in impact load as the inclination ore pass is
increased. They concluded that in obtaining a good estimate of peak dynamic load in
computer simulation, it is necessary to model the fall of the largest size particle in ore
pass.
Beus et al. (1999) reported the results from computer modeling of ore pass system
and field measurements. They used a gate assembly offset 2.4 meter from the
longitudinal axis of ore pass. As a result, they measured impact factors at range of 1.06-
1.33 compared to average value of 4.09 reported by Blight and Haak (1994). They further
discussed the overestimate measurement of dynamic load by computer modeling and
difficulties in determination of shape functions, damping coefficient and relative stiffness
of rocks and surrounding walls of ore pass.
Stewart et al. (1999) reported the construction of a one-third-scale ore pass model
in the Spokane laboratory (NIOSH). This model will be used to investigate the static and
dynamic loads on alternative chute designs and a providing a means of testing of hang-up
removal methods. He proposed that despite the existence of excellent guidelines, hang
ups, failures and other ore pass problems still occur frequently in underground mines.
At the recommendation of the National Institute for Occupational Safety and
Health (NIOSH) current research has been conducted through Western Mining Resource
Center (WMRC) at Colorado School of Mines. Mustoe (1999-2000) presented the results
on numerical modeling of ore passes with two-dimensional superquadric and three-
dimensional ellipsoidal particles. He reported that the two-and three-dimensional DEM
|
Colorado School of Mines
|
20
CHAPTER 2
ORE PASSES
2.1 Introduction
Ore passes are widely used in the mining industry because of their advantages of
high productivity, high efficiency, low cost, need for less equipment and breadth of
geologic application. Not only do productions in most deep underground mines depend
on ore passes, some open pit mines use them as the primary choice of ore transportation.
An ore pass is much the same as a tunnel, with a very large height-to-diameter ratio.
Ratios up to 100:1 are not unusual for ore passes.
The configuration, size and purpose of ore passes vary widely. They may be
located at the surface or underground, vertical or sloped, straight or dog-legged, circular
or rectangular in cross-section, designed for intermediate storage, or as a chute, or be part
of a mining method (Glory hole method). Figure 2.1 shows the different configurations of
ore passes, Goodwill et al. (1999).
2.2 Ore passes in open pit mines
In open pit mining, the ore is generally mined at the bottom of the pit, and then
hauled to the surface for processing purpose. Uphill haulage under load represents a
costly and time-consuming item for the operation. Sometimes increases in fuel cost and
|
Colorado School of Mines
|
24
2.3 Ore passes in underground mines
Ore passes have traditionally been the provinces of underground mining. If an
underground mining is considered to be a bulk materials handling system where men and
equipment produce ore, then the ore pass is one element in the system, which also
consists of stopes, a crusher, ore bins, loading pockets and the mine-shaft.
Chronologically, the ore pass is the last section in a development plan; it is
constructed just prior to producing the ore. It is a customized feature that, subjected to
time constraints and pressure to start production, could result in inadequate attention to
detail concerning ore pass location and design.
It should be noted that, in underground mines, if the primary ore pass becomes
inoperable because of structural failure, hang-up or any other reasons, production of the
entire mine (or at least of that specified zone) will come to a halt. These types of
blockages or failures require extensive rehabilitation and extraordinary expenditures
while causing production interruptions and lost revenue as the fixed operating costs
continue to occur. Ideally an ore pass should be considered a permanent opening, which,
along with the shaft, comprises the main artery to sustain the economic life of a mine.
Figures 2.3, shows a schematic of an ore pass with branches in an underground mine.
2.4 Ore passes configurations
Ore passes are simply vertical, or steeply inclined, tunnels. They are generally
long compared to their cross-sectional dimensions. The usual shapes are rectangular,
square, and circular. Circular cross-sections are becoming more popular with the advent
of raise boring machines and their ability to dig harder rock formations. Cross-sectional
areas have been reported up to 50 ft2 in underground mines.
|
Colorado School of Mines
|
26
Faced with the higher production rate and the larger ore size of open pit mines,
circular ore passes over 30 ft in diameter and 600 ft in length have been mentioned,
Pfleider (1961). The length of the ore pass depends upon the distance between sublevels
and the excavation methods available.
An inclined ore pass, 12 X 8 ft rectangular cross-section and length over 1100 ft
has been reported in China, Li (1980). It is strongly recommended that the extent of each
ore pass should restricted to the minimum possible that will meet the requirements of
transferring and storing ore. The main reason for this restriction is that hang-ups in long
ore passes are difficult to locate and, subsequently, dislodge.
The inclination of ore passes should normally lie between the angles of 60° and
75°. The lower limit may be reduced to 55° and the upper bound may be extended to 83°.
The upper limit is based on the fact that inclination close to 90° result in the free fall of
rock blocks with consequent damage to chutes and control devices.
Ore passes, which are likely to transfer material with moisture content greater
than 10%, should be inclined to the upper limit of 70°. Finger raises should be inclined at
least 60° in such a way to guarantee free flow without causing high velocity impacts on
the main pass wall. Thus, the finger raise inclination and angle of intersection with the
main ore pass should be designed in a manner so that ore move easily from the finger
raise rolls onto the main ore pass, Ferguson (1991).
2.5 ORE PASSES DESIGN
In design, the location and spatial properties (size, shape and length) of the ore
pass are the first parameters for which values must be considered. Since the ore pass
being only one component of a mine haulage system, the location and spatial
characteristics cannot be selected independently. It should be noted that design of an ore
|
Colorado School of Mines
|
27
pass is dependent on the physical characteristics of the mine and also the mine operator’s
philosophy.
For example, there might be several satisfactory locations for an ore pass at a
mine, but the one finally selected would be dependent on the mine operator’s objectives
and performance. One of the key issues in this decision is the shape and method of
drawing ore from the ore pass. Generally speaking, the most important factors in ore pass
design are maximum ore size, percentage of fines in the ore, variation in characteristics of
ore being taken from different parts of ore body and rock mechanics consideration.
The common types of ore pass design may be categorized as follows, Goodwill et
al. (1999):
♦ Rock slide (Figure 2.4.a). In this type of design, ore slide down on the footwall and
the ore pass does not operate full of ore. Therefore, less steep angles are required to
promote flow. A chute test could be conducted to determine the minimum slope for a
bed of ore to slide on various lining materials.
♦ Ore pass with mobile reclaim equipment, (Figure 2.4.b). Usually ore is dumped into
the ore pass through a grizzly to eliminate large boulders. Then from the base of the
pile, a scoop tram removes ore. In this type of design, the ore pass would be partially
full all the time. For reliable flow, a 25° forward slope and ore with top size limited to
1/10 of ore pass diameter has been suggested.
♦ Ore pass with Dogleg and continuous drawing (Figure 2.4.c). Existence of the Dogleg
will prevent the direct impact of ore on the gate assembly and its mechanical feeder
system. The level in the ore passes rises and falls providing surges of storage
capacity, but should be maintained at least two diameters above the Dogleg.
|
Colorado School of Mines
|
30
2.6 Flow patterns of ore in ore passes
The main objective of ore pass design is to provide a reliable flow of material that
prevents obstructions. A.W. Jenike was one of the pioneers in the study of gravity flow
of materials. Based on his finding and others, a classification of flow of ore can be
presented, Figure 2.5, Goodwill et al. (1999):
♦ In mass flow, we assume an ore pass consists of a tall vertical cylinder and a
sufficiently steep and smooth hopper. Then ore flows without any inactive or dead
regions. Mass flow has some advantages, such as uniformity of flow, absence of
channeling, hang-up or flooding. Non-segregating storage ability is obtained in a
first-in first-out regime. Mass flow is recommended for handling ores, which contain
large amounts of cohesive fines under full or partly full operation most of the time.
This is because liner abrasion due to impact of the falling stream of rock can be very
severe.
♦ In funnel flow, the hopper is not sufficiently steep and smooth to force ore to slide
along the converging hopper wall. Funnel flow is useful for the handling of very hard;
abrasive, lumpy solids because in funnel flow there is little wear of hopper walls. Ore
and muck with high amounts of cohesive fines may arch and/or rat hole severely in
funnel flow design.
♦ Expanded flow is actually a combination of the above-mentioned cases. It consists of
a mass flow hopper below and a funnel flow channel on top. This type of flow is
recommended for handling coarse ore containing 10% or less -5 mesh fines.
|
Colorado School of Mines
|
33
2.7 B lockage of ore in an ore pass fhang-up) and its prevention
For an ore pass system to fulfill its function, the transfer of ore should occur in a
regular fashion and in such a way that the full size of the opening is employed
effectively. The bulk ore consists of particles with different sizes and shapes; they are
usually wet or damp and often contain clays.
As a consequence, blockage of materials, or hang-up, can frequently occur. There
are two major types of hang-ups as follows, Hambley (1987):
♦ Hang-up where large-sized boulders become wedged together to form interlocking
arches. This occurs when the relatively few larger fragments form stable hang-up
arrangements in the ore pass. The occurrence of this type of hang-up could be
enhanced by abrupt changes in ore pass geometry. The possibility of forming such
arches depends on the percentage of large size particles in the material handled, on
the size of the particles relative to the size of the ore pass and outlet, on the shape of
the rock fragments, and on the velocity profiles across the flowing ore.
♦ The cementation effect of fine and sticky particles (d<0.01 in.) may create a cohesive
arch type of hang-up. The presence of moisture in the bulk ore mass will also increase
cohesive resistance of arch. The mining industry relies on a few functional design
factors to avoid blockage of materials in an ore pass. They are mostly based on
empirical equations and elementary mechanics of materials. These factors are
summarized in Table 1.1.
|
Colorado School of Mines
|
35
Tanner (1995) describes an accident in which two miners were trapped in a
sudden flood of ore during hang-up removal. One of them was fatally injured, and
production was halted for a while. If the hang-up occurs at the draw point, it is relatively
easy to remove. Otherwise, the location of the hang-up is determined by use of helium
balloons. The methods for dislodging hang-ups include:
♦ Secondary blasting. There are many blasting techniques utilized to dislodge hang-ups.
Nevertheless, mining engineers hesitate to use these methods initially because of their
hit-and-miss approach and because of the potential danger to personnel and
equipment. The former U.S. Bureau of Mines even developed a “Hang-up Clearance
Module” that fires an explosive charge toward a hang-up, Hambley (1987). It should
be noted that where cohesive arches are involved, blasting can actually compact the
arches if the explosives are not laid out properly.
♦ Mechanical push. This is the most elementary method- releasing hang-up by use of a
crowbar. The modern version of the crowbar is the hydraulic arm. Scraper winches
can be used to remove packed material from ore pass walls by pulling back the blade
from dump point to draw point.
♦ Air Canon and Sonic Gun. The air Canon hits the blockage of ore with a short burst
of pressurized air. This method can work well on densely compacted material, both
fines and blocky masses. The Sonic Gun uses ultrasonic waves to remove the hang
ups. In both cases, it is necessary to have an unobstructed view of the blockage in
order to use the tool properly.
♦ Water Jet. A high-pressure water jet can be used to remove cohesive arches.
However, use of this method is limited because of the possibility of ore flood / mud
rush, Ferguson (1991).
|
Colorado School of Mines
|
36
2.9 W ear and lining of ore pass w alls
Most ore passes are unlined. This means that the ore pass is located in a
competent rock, whose abrasiveness and hardness are higher than the ore being handled.
Unlined ore passes can last for many years. Some researchers have proposed that the
existence of sticky material in ore could reduce the wear effect, but there is still a
possibility of hang-up, Hambley (1987).
Abrasive wear of an ore pass can be caused either by direct impact or by sliding
friction. Clearly, the likelihood of direct impacts of ore on walls is reduced as the ore pass
diameter is increased. Lining is necessary if:
♦ The ore is hard and abrasive, and the ore pass is located in a soft or jointed rock.
♦ There is ground water present, with ore being mined from an open pit mine in a very
cold climate. Here, freezing of ore and, as a consequence, hang-up may result.
The type of lining system is based on the type of ore and the characteristics of the
rock mass around the ore pass. Typical lining materials are concrete (plain or reinforced)
rails and the abrasion-resistant cylindrical steel shell, Goodwill (1999).
|
Colorado School of Mines
|
37
CHAPTER 3
THE DISCRETE ELEMENT METHOD
3.1 Introduction
Discrete element methods are the numerical procedures for simulating the
complete dynamic behavior of discrete, interacting bodies. Each individual particle
(body) with general shape (deformable or rigid) based on its unique characteristics could
be subjected to gross motion. Engineering problems such as the flow of granular
materials, which exhibits very large-scale discontinuous dynamic or static behavior,
cannot be solved with a conventional continuum-based approach such as the Finite
Element Method. DEM has provided a numerical means for analyzing the progressive
movements and interactions of bodies in granular assemblies. Its algorithm applies
Newton’s second law to each particle within the system. The continual movement of each
body results from the non-equilibrium of different forces exerted on it. DEM explicitly
models the dynamic motion and mechanical interaction of each body at discrete points in
time, with each point being termed a step. For this purpose, integration of equations of
motion and contact laws are necessary. This is the heart of each DEM code and is the
most time intensive part (computational timing). Generally, three steps for each discrete
element program may be described, Hustrulid (1997).
♦ Initialization. Define the boundaries, discrete element locations and velocities, and
material properties for the system.
|
Colorado School of Mines
|
39
3.2 Governing equation
In order to simulate the gravity flow of materials, Newton’s second law is applied
to each particle in the system. Theoretically, the dynamic equilibrium can be written for a
mass as:
mx = '£ F x , "V> = I X (31)
I 0 = Y Mo (3.2)
Where m is the mass, x and ÿ are respectively the x and y components of
translational accelerations, Fx is the x component of resultant forces, Fy is the y
component of resultant forces, / is the mass moment of inertia, 6 is the angular
acceleration and MG total moment with respect to center of mass.
3.3 Cluster theory
DEM modeling of granular media is often performed using a system of 2D
circular (disk) elements. Disk elements offer several advantages: relatively short
computation time, and simplicity of contact detections and contact force calculation. But
the disks can roll excessively, and they demonstrate lower peak friction angles than real
materials, see Thomas and Bray (1999). Basically, it is difficult to simulate real material
behavior when the ore is even slightly angular. In the current research, a cluster version
of DEM code has been developed. The Cluster-2D code combines the simplicity of
dealing with circular disks and the accuracy of modeling ore shape. Each cluster consists
of the desired number of disks, connected rigidly together in a specified manner. Figure
3.2 shows a rectangular-like cluster of six disks.
|
Colorado School of Mines
|
40
M = ï>,. (3 3)
z=i
r
r
/ = Z Mid; 4" 171: —-- (3.4)
1=1 y
IG ~ ’ (3 5)
Where m; is the mass of the ith disk, M is the mass of cluster, I is the mass moment
of inertia of cluster about its center of mass (G), Ig is the polar moment of inertia of
cluster, 0 is the angular acceleration of the cluster, Mg is the total moment of cluster with
respect to the center of mass, d; is the distance from the G (center mass of cluster) to the
centeroid of i* disk, and r; is the radius of i* disk.
Figure 3.2 Cluster of Six Circular Disks
In the Cluster-2D code, all the overlapping areas of different combinations of
disks for each individual cluster have been considered.
By using an explicit time step algorithm, Mustoe and Williams (1989), the
acceleration, position and velocity are updated for each individual cluster within a small
time increment A/. This can be presented as follows:
|
Colorado School of Mines
|
47
At the two edges of each boundary, it is assumed that there are fixed ghost disks
with radius of r =(0.0001)d. After contact detection at the end points, the algorithm
measures contact forces based on penetration of the disk particles at these ghost disks.
This method eliminates the double calculation of contact forces at the edges.
3.5 Contact forces
The contact forces between interacting bodies or between bodies/wall are
modeled with a contact law, which has components in the normal and shear directions.
The most important part of contact force calculation, is fast and accurate contact
detection followed by evaluation of contact forces. In the following sections, the basic
concepts surrounding these steps are explained.
3.5.1 Automatic detection of possible contact
The heart of each DEM algorithm is the contact search procedure. Contact
searching within a system of N bodies is a very time consuming computation. It is well
known that with a simple direct searching procedure, the total number of computational
N(N -1)
efforts would be —^ — -. Note that after one disk-particle has been checked against
another disk-particle, that disk would be deleted from the checking list.
For a system of several hundred or thousands of bodies, this is a prohibitive
computational burden, Mustoe and DePoorter (1993). It is therefore critical to reduce this
time-consuming effort by using subdivision techniques within a problem domain.
Theoretically, the grid size depends on total particle number, particle shape, and
distribution of particles within the system. Subdivision techniques, or grid cell method,
|
Colorado School of Mines
|
48
were first employed by Cundall (1971, 1974). He reported use of a grid with 208 cells
providing coverage of the problem domain. This number was chosen because of
computer memory and screen resolution limitations. However, in 1978, in an effort to use
a more optimal grid size, he introduced the BALL code with a grid size at least four times
that of the largest particle in the system, to limit maximum number of particles in one cell
to four, Cundall (1978).
This method limits the maximum number of grids per particle to four. Relying on
the same logic, Taylor and Preece (1989) and Greening (1996) used a grid cell size equal
to that of the largest particle.
Zhang (1993, 1996) emphasizes that optimum cell size can be calculated by
running the code several times with different grid cell sizes and finding the minimum
CPU time. Based on this logical analysis, he arrives at a grid size as large as the particle.
In his case, a given particle would be in, at most, four cells.
Mathews (1994) uses a grid system with uniform spacing, the grid size being 2.5
times the size of the radius of the smallest particle. His size constraint limits, the
maximum number of particles occupying any given grid cell at any given time to nine.
This type of spacing reduces the number of contact searching operations to
^C.n(n - 1), where n is the number of particles in one grid cell at each time steps which
is less than or equal to nine and C is the total number of grid cells containing particles.
Hustrulid (1997) uses a grid size equal to the maximum particle diameter. He
postulates that this method makes the algorithm both simpler and faster.
In this thesis, a grid cell slightly larger than the maximum ore-disk particle is
used. The grid cell size is chosen, first of all to ensure that a disk-particle will only map
into at most four grid cells, secondly speeding up the algorithm and minimizing the
maximum number of particles in any individual grid cell.
|
Colorado School of Mines
|
49
3.5.2 Grid
method
As the name grid method implies, the problem domain is divided into equal size
squares. The domain can be adjusted as the simulation proceeds or it can be fixed at the
start of the simulation. Given the nature of ore passes, and in an effort to avoid dynamic
allocation/de-allocation of memory, in this thesis the problem domain is fixed at the
beginning of the simulation.
The choice of grid cell size can directly affect the performance of the grid
algorithm. If the cells are too large, there will be too many particles in each cell; in this
case, the search could approach the maximum number (N2). If the cells are too small, a
single particle may be mapped to several cells, increasing the required memory since
more cells must be searched for possible contacts. In this case, computational time will be
increased.
The extents of simulation are defined with pre-introduced minimum (xmin, ymin)
and maximum (xmax, ymax) coordinates of a box. The area of the box is divided into
geometric square cells (grid cells), which have a width of dgrid, Figure 3.6.
dgrid = 2.02 * radi _ max (3.29)
The parameter rdi max is the pre-defined maximum radius of the ore-disk
particle. The number of grid cells in x and y directions are calculated so that the grid
system can extend through the entire simulation. The number of grid cells in the x
direction, denoted by max i, is calculated as
xmax - xmin
max / = (3.30)
dgrid
Where, |_xj represents the smallest integer greater than or equal to x.
|
Colorado School of Mines
|
52
Table 3.1 Ore-Disk Particle Locations within a Grid System
Grid cells
Particle
i (18,23), (19,23)
j (19,22), (19,23), (20,23), (20,22)
k (20,21), (21,22), (21,21)
1 (18,21)
As demonstrated by this table, not only does the contact search require more
computational time, it is also possible that two particles overlap in more than one grid
cell. Contact forces can then be added unrealistically several times to the certain particles.
Researchers have tried different solutions to overcome this obstacle.
Mathews (1994) introduces the idea of marking the contact point between two
particles and comparing the grid cell with the original grid system for the particles. But
this solution is very path-dependent and does not match well to ore pass simulation.
In this thesis, we depart from the traditional grid method and instead of marking
all grid cells; a disk particle overlaps to a marking cell, only and if only to the cell where
the center of the disk particle is located, Figure 3.8. The coordinates of the particle’s
center determine its grid cell number. The grid indices respectively are called icel and
jcel in the x and y directions. The icel index is calculated as:
body(p xpos) - xmin
icel = (3 31)
dgrid
|
Colorado School of Mines
|
54
The real advantage to mapping each particle is not only in decreasing the required
memory, but also speeding up the contact checking process. Since one of the major
inefficiencies of the traditional grid method is the requirement of checking the entire grid
system for finding the possible contact, this is very expensive with respect to
computational processing time.
But in the modified grid cell method only four grid cells (right hand side and
top) of the marked grid cells need to be checked for possible contact. Figure 3.9
illustrates a marked grid cell with a particle being checked to find the possible contacts.
As shown, the first code checks the possible contact between the particles in the same
grid cell (marked cell). Then only half of the adjacent cells need to be checked for
possible contacts. The other half of the cells will be checked if and when they become
marked cells. Using this contact search method eliminates the possibility of multiple
contact checks of particles.
As mentioned earlier in this thesis, cemented clusters of ore-disk particles are
used; therefore within this algorithm we must ensure that two ore-disk particles in contact
are from two different clusters. After updating the position of the individual particle in
each time step, their location in the grid cell system must be checked and re-mapped.
Only when a particle has moved between grid cells (its center having moved from one
grid cell to another), must the linked list of particles be maintained. Note, similar
approaches have used by Hustrulid (1997).
|
Colorado School of Mines
|
56
3.5.3 DETERMINATION OF CONTACT COEFFICIENTS
The elastic part of contact forces can be presented either by a linear-spring or a
non-linear spring (Hertz type of contact), Brogliato (2000). In Hertz theory, the
assumption of elasticity is used to derive the normal stiffness of the contact between two
deformable bodies. Classically, the term of normal coefficient of non-linear spring is
calculated as:
kh = ^ E 2r---------
3(E(l-u22) + £2(l-yi2))
Where (Ei, E 2, Vi, V2), represent the elastic properties of the contacting bodies and r is a
function of the curvature radii of the contacting bodies.
But in the current study, due to lack of available quantitative data, the linear
shear spring and a dumping model (Kelvin-Voigt elements) have been chosen. Because
the contact stiffness of a material is dependent on size, some researchers have tried with
experiments to express it as a function of size and shape. For a spherical rock, Larson et
al. (1998) calculates kn as:
kn = ArE (3.33)
Where, r (in inches) is the radius of the sphere, E (in psi) is Young's modulus and A is a
dimensionless constant which a least-square method whose value was computed as
0.19052.
However, this method is very specific and needs data from experiments. In this
work, the normal contact stiffness, kn can be estimated based on the maximum allowable
overlap between two particles defined by the user and an approximation of
(6max,)
maximum possible speed of ore particle in an ore pass simulation, Figure 3.10. The
maximum overlap between particles is determined by the stiffness kn of the spring in the
|
Colorado School of Mines
|
60
of data required to describe the variations in e (Larson et al. 1998) and results from
previous simulation, the coefficient restitution of e=0.2 is used.
3.5.4 Contact force calculations
In the Discrete Element Method, contact forces play a very important role. They
result from particle/particle or particle/boundary collisions. Contact mechanics between
interacting bodies are modeled with a spring (linear elastic component) and a dashpot
(viscous damping component).
Some researchers prefer to use Hertz theory (non-linear normal stiffness). But in
the current study, due to lack of available quantitative data of the stiffness and dumping
coefficient between rock particles, the linear shear spring and dumping model (Kelvin-
Voigt elements) have been chosen.
The shear contact force component is governed by a Coulomb friction model,
which is defined with a friction coefficient and a shear spring. In this thesis, the shear
stiffness is set equal to the normal stiffness although the static and dynamic friction
coefficients are set equally (i.e. kn=ks =k;fis = fid = ju ).
The normal, instantaneous elastic contact force either between two particles or
between one particle and a boundary (wall) is calculated as:
(3 43)
Where, 6 and kn are derived in sections 3.4.1, 3.4.2 and 3.5.3 and Mnormal
direction at the point of contact. The viscous damping force, which contributes to energy
loss in the system, can be expressed as:
(3.44)
|
Colorado School of Mines
|
62
program but it underestimates friction, especially in the case of a very low relative shear
velocity.
In this thesis, the relative tangential velocity of particles at point of contact in the
shear direction over the collisions period, acts as an incremental shear spring that stores
energy from the relative motions and represents the elastic shear deformation of the
contacting bodies.
For this purpose the Cluster-2D code tracks the contacts every time step and is
tagged differently if the current contact is a new or old contact. The current shear
displacement is calculated as follows:
V„=((Vpl-V pi).s)s (3.46)
If the contact is new, then the relative shear displacement at time step Af is calculated as:
(3.47)
If the contact is considered old, then the relative shear displacement at time step A/ is
calculated as:
(3.48)
Then the shear force as a result of this relative shear displacement is calculated as:
(3.49)
This slip friction force is calculated as
F¥ = ?.M (3 50)
The magnitude of shear force from equation [3.49] is compared to the slip friction
force from equation [3.50]. If it is larger than the slip friction force, then it should be
|
Colorado School of Mines
|
63
scaled down to the maximum slip friction. In this situation, the particles slip relative to
each other. Generally, the friction force is added to the forces and moments acting on the
particles in contact. Then particles velocity and position are updated, and this process
continues at each time step until the pre-defined number of time steps has reached.
3.5.5 Time integration and critical time step selection
As mentioned in Sections (3.2 and 3.3} for a system of discrete particles Newton's
second law is applied to each particle:
(3 51)
E M „ = ^ (3.52)
at
Where H \s angular momentum with respect to the center of mass, applied to each
particle, and subscripts n denotes quantities defined at time t = tn.
The next piece of any algorithm is the numerical time integration scheme. This
scheme performs dynamic updating of the particle's velocity and position throughout the
duration of the DEM simulation.
The usual time integration scheme implemented in most DEM code is the explicit
Euler central difference procedure, Cundall (1974); Zhang (1993); Mathews (1994);
Hustrulid (1997). However, some researchers use non-explicit time-step schemes,
Greening (1996) and O'Connor (1996).
The choice of using an explicit time step over an implicit one is based on several
factors. First of all, use of the implicit method requires that the system behavior is not
path dependent, Cundall (1974). This assumption is rarely overcame with DEM
modeling. The discontinuous behavior as a result of contact between particles will not be
|
Colorado School of Mines
|
64
modeled accurately if the time step as an implicit scheme is greater than the duration of
the contact. In current research, the explicit central difference scheme is used.
It should be noted that the choice of time step is very important. Too large a time
step can produce instability of system. Conversely too small a time step presents accurate
results but its impact on computational time could make it infeasible.
Therefore it must be less than some critical value to ensure the stability and
feasibility of the modeling. This critical time for a simple elastic model (spring-mass
system) is defined by:
Ter=— (3.53)
COq
(3.54)
Where (o0 is the natural circular frequency of a simple spring-mass system, k is
the stiffness of and m is the minimum particle mass. For a stable condition, the time step
is defined as:
A/<7„ (3.55)
Cundall (1978) suggests that 10% of critical time is probably safe for most of the
DEM problems, but 20%-50% may be used with caution for loosely packed particles.
Zhang (1993) in the hydraulic jump simulation uses 5% of critical time increment in
order to ensure stability. Mathews (1994) and Hustrulid (1997) suggest using less 20%
and 10%, respectively. In this thesis, a factor less than 10% of critical time has provided
stable results.
|
Colorado School of Mines
|
65
3.5.6 System stability, energy checking
The numerical stability of a DEM ore pass system simulation is verified by
energy checking. The total energy of a particle at a given time step is presented as:
ET = —mv2 + —I02 +mgh + — kS 2 (3.56)
2 2 2
? * "
Where m is the mass of the particle, v is transnational velocity, I is moment of inertia of
mass with respect to its mass center, 9 is angular velocity, g is gravitational acceleration,
h relative height with respect to calculation datum, kn is stiffness of the particle and Ô is
penetration.
If there are no mechanical losses of energy (such as damper, friction, etc.) and no
energy is added, the total energy of the system should conserve. For verifying the
correctness of the code, the total energy for the special case of 14 clusters of single-size
disks inside a box without any mechanical energy losses is investigated, Figure 3.12.
All of the clusters (single disk) have an initial velocity of Vx=0.0, Vy=4.0 m/s.
Each disk has a radius of 0.2 m and a mass density of 1273 kg/m3; this makes the total
energy of the system at the beginning 40312 N-m, gravity and friction assumed to be
negligible. The simulation result for the total energy (sum of kinetic and potential energy)
is shown in Figure 3.13.
|
Colorado School of Mines
|
69
CHAPTER 4
MODELING GRAVITY FLOW OF ORE WITH DEM
This chapter describes the research tasks performed during the period of
investigations. Due to a lack of complete understanding of theory behind the dynamic and
the static properties of granular materials, numerical simulations are very powerful tool
and sometimes the only choice to study the mechanical behavior of granular systems.
Ristow (1998) emphasized that in case of a conical hopper, there is a good
correlation between results from two-dimensional simulations with their three-
dimensional counterpart. Due to the similarity with the configuration of ore passes (very
long in one dimension compared to their cross sectional) loading on ore passes gates
should be comparable in 2D and 3D simulations. In qualitative terms there is agreement
in predicting similar hang-up conditions in 2D and 3D simulations, Mustoe (2000). The
above rational justifies the application of a two-dimensional computer program to study
the gravity flow of ore in ore passes. Note, to quantify the differences between two-and
three-dimensional ore flow needs further investigations.
A 2D numerical model based on Discrete Element Method (DEM) has been
developed to study different problems concerning gravity flow of ore in ore passes. By
using this new tool it is hoped that a better understanding of what helps and what disrupts
the performance of the gravity flow of ore can be obtained. This model with using a rigid
and cemented cluster of disks can simulate different ore shape and its angularity as well.
The modeling parameters for the ore and ore pass would be as follows:
|
Colorado School of Mines
|
70
♦ Ore pass inclination, coefficient of friction (ore/wall), different dumping points
design, stiffness of wall and existence of Dogleg, and,
♦ Ore shape and size distribution, coefficient of friction (ore/ore and ore/wall) and
stiffness of ore material.
The simulations focused on the effects of ore pass configuration and ore
characteristics on the most important aspects of ore pass design as follows:
♦ The dynamic and static loads on ore pass gate assemblies, and,
♦ The flow regime of ore in ore passes with specific interest to creation of hang-ups.
The terms dynamic and static loads refer to the maximum normal impact load on
bottom gate of the ore pass and the steady vertical load on the bottom gate (proportional
to the weight of ore). Given the importance of ore shape and size distribution after
blasting, these topics will be discussed first.
4.1 Ore shape characteristics and size distribution after blasting
Measuring the post-blast fragmentation of rock mass formation is very
unpredictable and reveals difficulties even within a mining site. As a result, several
researchers have proposed techniques to facilitate such measurements. Singh and Appa
(1979); Singh et al. (1986); Chavez (1996), have studied the following issues:
♦ Different rock mass properties in mining blocks.
|
Colorado School of Mines
|
71
♦ Changes in strength (compressive, tensile, shear) of ore and waste even, within the
same block, as a result of either inclusion of different sets of minerals or ground
water.
♦ The presence of different sets of joints and fractures in mined blocks.
♦ Different degrees of weathering.
The prediction of shape and size distributions of ore after blasting, are very
important and critical steps in most mining operations, which strongly influences the
subsequent steps of digging, hauling, transporting, and crushing. Furthermore, proper
consideration of ore/waste size and shape distributions can bring tremendous savings in
energy and cost to a mining operation. See Scott et al. (1996) and Napier et al. (1996).
The primary method for determining the size distribution of ore after blasting is
sieve analysis Dick et al. (1973), Bhandari and Vutukuri (1974), Singh et al. (1980),
Maerz et al. (1987). However this method is slow, expensive and sometimes due to the
very large size of the fragmented rock, technically impossible.
Researchers attempt to predict the size distribution of ore after blasting either
from the blasting parameters and the rock mass properties (using empirical formulas), see
Lovely (1973) and Stagg et al. (1992) or from computer simulation, see Gama (1984) and
Kuszmaul (1987).
Yu et al. (1996) presented experimental results to describe the relationship
between impact energy and specific surface area of crushed rock as a measure of size
distribution of the blasted rock. Prasad et al. (1996), proposed that both grindability and
blastability may be seen as the energy spent in the process of getting a desired product
size from a specific sized rock mass. They proposed that comminution properties
established through laboratory studies might be useful in predicting blasting-induced rock
fragmentation. The main obstacle with these methods is, they do not measuring the actual
fragmentation size. Therefore they should be calibrated by applying the appropriate
proportionality constant for different mining operations.
|
Colorado School of Mines
|
72
Carter (1973), Aimone and Dowding (1983) tried manually to estimate the size
distribution of blasted ore from different pictures of blasted piles of ore. Their efforts was
continued by Gozon (1986) using image analyzing computer software. But these methods
consider only the visible part of fragments, not the sections overlapped by other
fragments. This represented a serious sampling bias, Maerz et al. (1996).
Another method of measuring fragmentation (size and shape distribution) is to
acquire digital images of rock fragments and then process these images using digital
image processing techniques. The main steps in this process are as follows.
♦ Photographing the pile of rock in different angles.
♦ Conversion of pictures to a gray-scale image.
♦ Enhancing and segmenting the image.
♦ Special computer digitizing process.
♦ Material sizes, elongation (aspect ratio), shape measurement; see Maerz et al. (1987),
Wang et al. (1996).
There are a number of systems and software that can be used to quantify
fragmentation from digital pictures. The main obstacles of these systems are the existence
of fines and the different standards for the shape and size measurements around the
world.
Fines (d<l mm) are usually underestimated in image processing systems, either
because they cannot be seen on surface exposure or because the collection of fines can be
misrepresented as a large fragment, Kemeny (1999).
After performing the second step of image processing (conversion of gray-level
images into binary ones), either by a semi-automatic or a fully automatic method, the
material size, elongation (aspect ratio) and shape factor are measured.
|
Colorado School of Mines
|
73
Unfortunately there is no single measurement standard; as a consequence,
different measurement methods produce extremely different results, causing confusion
about material size and shape, Wang et al. (1996). Figure 4.1 illustrates a typical example
of this problem by applying different measurement methods to an orthogonal triangular
shaped aggregate. Here a, b, ç, d, and e refer respectively to chord sizing in certain
scanning direction; two Feret diameters; equivalent circle (e.g. by area); maximum
diameter; and equivalent ellipse.
As it has been mentioned earlier, optical systems tend typically to overestimate
the central tendency of size distribution, and underestimate the distribution of variability,
Maerz and Zhou (1998). Different image analysis systems attempt to address this
problem with performing curve fitting with some known distribution (e.g. Rosin-
Rammler and sieve analysis). The Rosin-Rammler equation is:
y = 1- exp (4.1)
Where y is the cumulative percent passing, x is the particle size, xc is the
characteristic size that 63.2% of materials are passing and n (uniformity index) is a
parameter describing the spread of the distribution.
The characteristic size needs to be measured by sieve analysis; n is 3.0 for a very
uniform size distribution and 0.75 for a well-graded size distribution, WipFrag
Calibration (2000). Based on the blasting and mining methods characteristics, the
uniformity index can be calculated with empirical formulas, Ryan (1998).
|
Colorado School of Mines
|
74
Figure 4.1 Various Size Definitions for an Object (Shaded)
(After Wang et al., 1996)
With respect to size and shape distributions of blasted ore/waste, Kvapil (1992)
presents a simplified version of different flow pattern of materials in mining industries
(Figure 4.2). Type I shows blasted ore (coarse materials) with large spherical pieces of
more or less the same size and shape. Type II is representative of an almost uniform size,
but different shape of coarse materials. Type III indicates a composition of large
fragments, chippings, and sand. Type IV presents a blasted ore characterized by a mixture
of large blocks, medium-sized fragments, chippings, sand and/or rock fines, or clays. The
mechanical behavior of types III and FV change considerably based on the percentage of
fines and moisture content.
Up to an inclination of 40°, the blasted materials require mechanical
transportation. With a higher inclination of ore pass, gravity flow can be used as a driving
force for flow of materials. In Figure 4.2, GF illustrates the range of gravity flow
applicability, where A is the inclination range of ore pass for types I and II only. It
suggests steeper ore passes for types III and IV.
|
Colorado School of Mines
|
76
Yet all researchers agree that the concepts of shape and size are interrelated. The
length and width are referred to as sizes (2D dimensions), but they may also be used as a
factor to describe the shape of particles (elongation/aspect ratio).
Along with this logic, Wang et al. (1996) introduced shape reflection, noting that
the measured sizes should crudely describe the shape, which in practice means that
overall properties of ore shape, such as elongation and angularity, are possible to infer
from that crude description.
It needs to be noted that terms of angularity in a broad definition means how
much a shape deviates from a spherical shape by having sharp ends. This concept is
absolutely different from roughness around the boundary of ore particles due to their
physical properties.
4.2 Ore- cluster shape and size determination for dem analysis
According to these considerations, a method for shape description of an ore-
cluster should meet the following basic conditions:
♦ The method should have a shape reflection definition where the measured ore-cluster
is unique, independent of its rotation (rotational-invariance).
♦ The method should be repeatable with very low boundary roughness sensitivity.
In this research using the previous discussion and results from literature reviews
(e.g. Wang et al., 1996 and Maerz, 1998) three ore-cluster shapes are defined and applied
in different simulations. These are crude shape description of the ore-cluster and even can
be used in combination to model a specific type of ore materials. Figures 4.3-4.5 are
|
Colorado School of Mines
|
80
For calculation of the shape reflection, a rectangle is assumed to be circumscribed
to the longest dimension of ore-cluster. The aspect ratio (elongation) of the ore-cluster
would be e = —.
b
The index for the angularity of an ore-cluster is defined as the ratio of the area of
cluster to the area of circumscribed rectangle. This value varies from 1.0 (for a shape
without abrupt angularity or rectangular-like clusters) to 0.5 (a shape with extreme
angularity such as bisectors triangular-like cluster).
The width of the ore pass is related to the largest dimension of its flowing ore.
Hambley (1987) and Ferguson (1991) suggested that application of an empirical ratio
— > 5 to prevent interlocking hang-ups. Here, D is the diameter, or side length of the ore
d
pass and d is the greatest dimension of the largest rock block. However despite the
existence of such empirical guidelines, we encounter hang-ups, failures, and other ore
pass problems frequently in underground mines. See Stewart et al. (1999) and MSHA
database (1987-1996).
Interlocking hang-ups are chance occurrences of stable arrangements of the
relatively large size fragments of rock in the ore pass. Their probability of occurrence
depends on, size distribution of the ore materials, ratio of ore pass width (diameter) to the
maximum size of ore, and the shape of rock fragment, Hambley and Singh (1983). A
larger ore pass will results in an increase in construction and probably maintenance costs,
whereas the production of smaller ore sized after blasting requires higher operating costs
(finer fragmentation at the face, or use of a primary crusher).
Reports from mines by Hambley and Singh (1983) and Ferguson (1991) reported
cases of hang-ups at mines with ratio of higher than eight (— >8). Based on the above
d
mentioned and previous experience from simulations, the maximum dimension of the
largest ore-cluster particles (cohesionless and dry) in the DEM was determined to be
|
Colorado School of Mines
|
81
about 0.1 m. For an ore pass with its width equals to 1 m, this gives a ratio of — = 10 in
d
simulation of ore passes in this research.
With the assumption of a uniform size distribution in a good blasting practice
(WipFrag calibration, 2000) I have used a maximum size-deviation of 15% in the current
research. As a result, the smallest ore-cluster is 30% smaller than the size of the largest
ore-cluster in the system.
In this study whenever the ore-cluster shape is not specified, the Trapezoidal-
like cluster is the primary choice. This is because the Trapezoidal-like cluster geometry
lies between the other two specified cluster geometries, namely Rectangle and the
Triangle ore-clusters.
4.3 ORE PASS CONFIGURATIONS
A review of the literatures shows that there is no limit to the ratio of
(height/diameter) of ore pass. Based on the methods of mining, ore body configuration,
structural geology of mine block, this ratio can be determined. Ferguson (1991) reported
that extremely high ratios, such as 100:1, are not uncommon, especially in ore passes in
open-pit mines.
With respect to the available information from different ore pass operations and
the maximum size of an ore-cluster about 0.1 m, a ratio of 25:1 (height/width of ore pass)
for the simulation purpose is selected.
Figure 4.7 shows the selected ore pass configurations. Primary simulations
indicated that a height to diameter ratio of 25:1 allows enough time to flow of ore-cluster
for building up the incremental friction force. Note the DEM simulations have a run time
of 1.5-3.0 days (CPU time), in order to more closely model the actual loading/dumping of
ore materials, and also to avoid non-physically high dynamic load due to direct impact of
|
Colorado School of Mines
|
83
Typically, in any DEM simulations, the loads are calculated in a sample time
interval of approximately 10'3-10'5 second. Conversely most experiment data recording
systems can only record data at a range of 100-300 measurements per second. (This
average may drop to 30-60 measurements per second for a chart recorder). To prevent
spiky behavior of load-time history data obtained from a DEM simulation and making it
more comparable to data from measurement, the time-averaged load concept is
introduced. In this method, the DEM load calculations (impact factor graphs) are
smoothed over intervals (0.01-0.02) seconds, see Mustoe, (2000).
The following sections describes the DEM simulation results performed to
study: a) effect of ore pass inclinations, b) different ore-cluster shapes, c) various
coefficients of friction, d) size-deviations, e) change in inclination of Doglegs, f)
different coefficients of ore particle/wall contact stiffness, and g) initial launch velocities
of ore materials.
4.4.1 Effect of ore pass inclination
The inclination of the ore pass must be sufficiently great enough that material can
flow easily. This would favor higher inclination (750-90°). Conversely, in order to reduce
the dynamic load on the ore pass gate, the inclination should be as low as possible
(Ferguson, 1991). Note, more interactions of ore with ore pass walls; will cause the ore to
lose more kinetic energy, resulting in lower impact velocity and load.
Figure 4.8 illustrates the dimensions of the ore passes and the three different
inclinations (60, 75, and 90 degree).
|
Colorado School of Mines
|
89
Experiments and data from active ore passes demonstrate that when an ore pass is
used as a storage facility (even temporarily) the dynamic load on the bottom gate of the
ore passes will be decreased after a few dumps of ore. Based on this phenomenon, which
is called a cushioning effect, the rule of thumb is that when an ore pass acts as an
intermediate storage facility, it should never be left empty (Ferguson 1991).
Figure 4.11 displays the effect of ore-cluster shape on dynamic and static load on
ore pass gate and cushioning effect. Because of the high angularity of the Triangular ore-
clusters a bed of cushioning ore that acts such as shock absorber is developed very
quickly. As discussed in previous chapters, when the ore contains sticky materials (clays)
due to presence of moisture, the ore pass cannot be used as a storage facility and needs to
empty as fast as possible. Otherwise, as a result of compaction and cementation of ore
materials inside the ore pass, a hang-up may block the ore pass.
|
Colorado School of Mines
|
93
Table 4.3 Summaries of the Effect of Coefficient of Friction on Measured Load on
Bottom Gate of Ore Pass
Coefficient of Maximum Dynamic Maximum Averaged Static
Friction (Load/Weight) Dynamic (Load/Weight)
(Load/Weight)
0.25 5.01 2.53 1.29
0.50 4.17 1.63 0.91
0.75 1.88 1.22 0.49
4.4.4 SlZE-DEVIATION EFFECT
Changing the scenario from a uniform ore size to a distributed ore size has various
effects on the dynamic and static load acting on bottom gate of ore passes. With the
assumption of a very good blasting operation (Figure 4.5) a maximum 15% size deviation
is applied to the simulations. That means the smallest size ore-cluster would be 1/3 of the
largest ore-cluster in a system.
To simulate the effect of a size deviation, a 75° inclined ore pass has been filled
within 10 equal dumps of Trapezoidal-like ore-clusters. These simulations were
performed with mono size, 10%, and 15% ore size deviations respectively.
Figure 4.14 presents the effect of different size distributions on static and dynamic
loads. If there is a large range of size distribution of ore after blasting, then the smaller
size could sit among the larger in a more closed-packed arrangement. This increment in
compaction would cause more interaction of ore-clusters with each other and wall too.
This would also help to mobilize the friction forces and lower the static forces. With
higher size deviation the reduction in static load more is pronounced.
|
Colorado School of Mines
|
95
Table 4.4 Summaries of the Effect of Size Deviation on Measured Load on Bottom Gate
of Ore Passes
Size Deviation Maximum Dynamic Maximum Averaged Static
(Load/Weight) Dynamic (Load/Weight)
(Load/Weight)
Mono Size 3 33 2.80 0.82
10% 2.80 1.68 0.80
15% 2.51 1.29 0.75
4.4.5 Dogleg effect
As discussed earlier, higher inclined ore passes generate free and fast flowing ore.
However, the faster ore flow may cause severe damage to the ore passes walls and gates.
To accomplish a balance between these two opposing effects, an ore pass can be designed
with a Dogleg section, which is an abrupt change in the inclination of ore pass.
Adding a Dogleg section to an ore pass system, will allow the mine to benefit
from the free and fast flow of ore and a reduction of dynamic load. This can also help to
simplify and reduce the cost of the design for the gate facilities. On the other hand in
some cases a sudden change in the inclination of an ore pass may create other problems,
such as hang-ups.
For the purposes of simulation, a Dogleg section is inserted at 75% of the ore pass
height with an inclination towards the left side of the ore pass. The simulation was
repeated three times with inclinations of 20, 30, and 40 degree for the Dogleg section
respectively (Figure 4.15).
|
Colorado School of Mines
|
96
One Dump of Ore Material
1.6 m
t 2-5m
30 Degree 40 Degree
Figure 4.15 Different Configurations of Dogleg (Not to Scale)
Figure 4.16 shows the effect of a Dogleg part on dynamic and static load.
Compared to the case of without Dogleg there is a significant reduction in maximum
impact factor with an increase in inclination of Dogleg (e.g., from 4.35 to 2.47), because
ore losses most of its kinetic energy during collision with the wall at the Dogleg part.
Figure 4.17 is an enlargement of Figure 4.16 and shows this loss of kinetic energy
very clearly. The ore somehow trapped in Dogleg part, helping to decrease the push back
spring effect (the sudden rise and fall in dynamic load) for the 40-degree inclined Dogleg.
The slow-down effect of Doglegs sometimes could create hang-up in ore passes.
The angle of inclination of Dogleg plays a very important role in this matter. In the
section related to study of the ore flow this subject will discuss in details.
|
Colorado School of Mines
|
99
Table 4.5 Summaries of Measured Dynamic and Static Loads for Different inclination of
Doglegs
Dogleg Inclination Maximum Dynamic Maximum Averaged Static
(Degree) (Load/Weight) Dynamic (Load/Weight)
(Load/Weight)
Without Dogleg 7.36 5.41 1.09
20 6.43 5.23 1.06
30 5.51 3.95 1.14
40 4.21 2.77 1.09
4.4.6 Effect of initial dumping of ore materials
The dump point design is a very crucial factor in the design sequence of ore
passes; it dictates size and initial velocity of ore. Generally there is a grizzly at the dump
point. This provides a measure of control over the largest fragment of rock that enters the
ore pass. When rocks are too large to pass through the grizzly bars, they need to be
broken at the top of grizzly. Hence, if a grizzly were used, the spacing between bars
would be as great as the largest estimated dimension of fragmented rock (Hambley,
1987). Primary studies show significant reduction in dynamic load on ore pass is gained
with a small reduction in the initial launch velocity.
To study launch velocity effects, three ore passes have been filled with one dump
of ore with three different dump point designs. Figure 4.18 respectively shows these three
dump designs as (a) without feeder, (b) with feeder, (c) with automatic feeder, which let
it to be completely opened within 0.34 seconds.
|
Colorado School of Mines
|
102
4.4.7 Effect of contact stiffness
As discussed in Chapter 3 (Sections, 3.5.3; 3.5.4; 3.5.5), the contact stiffness
between the interacting particles in the DEM simulation affects the amplitude of the
contact forces and time-stepping scheme (system stability). For the case of a collision of
one ore-cluster with the wall of an ore pass if we ignore the loss of energy (due to friction
and damping), the maximum normal impact force is defined as:
4.2
Where m is the mass of an ore-cluster, kn is the normal contact stiffness, vimp is
velocity of ore-cluster exactly before impact. If mass and impact velocity are held
constant then the variation of the maximum normal dynamic forces is proportional to
Spiky behavior, push back spring effect of the computed load-time history of data
obtained from DEM simulation is another concern for performing the following
simulations.
To simulate the effect of contact stiffness, a vertical ore pass has been filled
within 10 equal dumps of Trapezoidal-like ore-clusters. The simulations were repeated
with contact stiffness of 1.2 X 108 N/m and 6 X 108 N/m respectively.
Figure 4.20 shows an increase of about 12% in maximum impact factor for the
stiffer case, as it was expected the increase in dynamic load because of loss of kinetic
energy is much less pronounced from theory (equation 4.2). With an increase in stiffness
of system, the push back spring effect has been decreased too. Figure 4.21 (an
enlargement of Figure 4.20) shows this effect clearly. This means overall system after
settlement acts stiffer such as a block. To examine validity of this observation, we must
compare the strain energy of systems in two cases. The strain energy for a system of
linear spring is defined as:
|
Colorado School of Mines
|
106
flow of materials in ore pass. The measured parameter would be the time history of mass-
flow of ore-cluster, which reaches the bottom gate level.
4.5.1 Friction effect
Friction has a major impact on flow of ore in ore passes. Note, friction forces
along with damping forces, are responsible for a significant of loss in kinetic energy of
gravity flow of ore. In DEM simulations of dry and cohesionless materials this effect
becomes much more pronounced.
To simulate the influence of friction on gravity flow of ore an inclined (75°) ore
pass has been filled within 10 equal dumps of Trapezoidal-like ore-clusters. These
simulations are performed with the coefficients of friction equal to 0.25, 0.50, and 0.75
respectively. The bottom gate is opened after about 44 seconds of simulations. Figure
4.24 shows the snapshots of the ore flow after opening the gate at the end of simulation
(time: 48.5). It shows the effect of high coefficient of friction and clogging of ore at the
bottom ore pass where the coefficient of friction equals to 0.75. Also clearly illustrates
effect of coefficient of friction on angle of repose.
|
Colorado School of Mines
|
110
It seems Triangular ore-cluster as result of their high angularity can somehow
lock with each other and continue to flood-flow. Conversely, Trapezoidal ore-clusters,
with their medium angularity and higher aspect ratio, show a lower tendency to flood-
flow behavior and as a result, are more likely to create a hang-up. Disk-clusters have a
tendency to flow freely and create a denser flow. The results somehow emphasize that
with the exemption of extreme case (Single Disk), which the different shapes of ore-
clusters are more likely to create a hang-up rather than to speed up the flow of ore.
To investigate this assumption, another set of simulations with the smaller
distance (1m) of horizontal drift from the bottom gate of ore passes were performed. The
goal was to find out which of the different ore-cluster shapes is more vulnerable to hang
up creation. Figure 4.28 shows snapshots of the simulations at the time (t =53.6 sec)
where all the different shape ore-clusters initiate hang-ups. It seems Trapezoidal-like ore-
clusters with their medium angularity and higher aspect ratio create hang-up faster than
other shapes of ore-clusters.
|
Colorado School of Mines
|
113
4.5.3 Dogleg effect
As discussed earlier (Section, 4.4.5), Dogleg are sometimes required, either due to
mine ore transportation plans, and/or to simplify the ore pass bottom gate design.
Previous experience has proved that Doglegs affect the flow of in ore passes.
To investigate the effect of the Dogleg, the ore passes configurations (Figure
4.15) are filled within 10 equal dumps of Trapezoidal ore-cluster and then suddenly (t=
48.5 sec) the bottom gate is opened to allow the mass flow of ore at the drift under the ore
passes. For comparison purposes this simulation also has been repeated for the same ore
pass without Dogleg.
Figure 4.29 shows the results of mass-flow measurements for different
configurations of Doglegs. An ore pass without the Dogleg part leads the flow of ore
outside, quickly and smoothly. Increasing the inclination of a Dogleg causes reduction in
flow of ore out of ore passes. Even at 40° inclinations, blockage of ore has completely
stopped the flow of ore out of ore pass within two seconds and 35% of ore has blocked in
ore pass. Figure 4.30 illustrates the snapshots of ore passes at the end of simulation and
creation of hang-up for the 40°-inclined ore pass.
Figure 4.31 shows the creation of a highly compressive stressed zone
(interlocking arch) due to weight of ore-cluster at the top of Dogleg part around the hang
up area.
|
Colorado School of Mines
|
117
occurrence of individual measurement values exhibits a recognizable pattern). Bury
(1975). In all the simulations presented the ore-clusters were generated based on random
generation of center of mass of clusters within a boundary box (polygon, dump region).
Figure 4.32. The reason for this decision is that we want to check these models against
the reality, to determine whether they are faithful and accurate enough for the practical
purpose, see Kreyszig (1993) for more details. The position of each individual ore-
clusters within each dump is based on generation of a psuedo random real numbers
between 0.0 and 1.0. Therefore the initial conditions of ore-clusters within an individual
dump are designed randomly.
1.6 m
Figure 4.32 Random Generation of Ore-Clusters within a Polygon "One Dump"
(Not To Scale)
In order to investigate the rank of sensitivity of creation of hang-ups at the Dogleg
parts, four more simulation of flow of ore in ore passes (with 40 degree Dogleg) with
slightly different positions of ore-clusters within each individual dumps have been
|
Colorado School of Mines
|
118
performed. For this purpose, in each of the simulations different sets of random numbers
were applied. Figure 4.33 shows the time history of mass flow of ore materials out of the
bottom gate of ore passes for the above mentioned. It can be concluded that in all the
cases, the existence of Dogleg part slows down the flow of ore out of ore passes by 35%-
40%, and in three out of five simulations a stable hang-up was created. Figure 4.34
illustrates the snapshots of flow of ore at the end of simulations for different initial
conditions of ore at each dump. Note, during these simulations all the other conditions
(e.g. number of ore-clusters in each dump and total weight of ore-clusters are identical).
Sensitivity Analysis For Creation of
Hangups
7000 -i
6000 -
g 5000 -
& 4000 -
| 3000 -
è 2000 -
1000
-
o -
48 50 52 54 56 58
Time (sec)
Run 1 Run 2 Run 3 Run 4 Run 5
Figure 4.33 Effect of Different Initial Conditions of Ore-Clusters at each Dump on Mass
Flow of Ore
|
Colorado School of Mines
|
121
In a closed-looped of mineshaft and ore pass facility; a front-loader unloads ore
through a grizzly in an underground silo. From that point, ore is transported to the top of
ore pass by a skip (0.84 m3). At a pre-designed height, the skip is automatically tipped
over and the ore falls into the ore pass.
Based on available information (Scott and Dresher, 1999), the ore material was
highly angular crushed-limestone with the average size of 38 mm and dry unit weight of
15.5 kN/m3, which has loaded into the ore pass with 21 equal dumps. The angle of
internal friction was assumed to be equal to the angle of repose (34 degree); where the
coefficient of friction for simulation purposes would be equal to 0.675.
In order to compare the results from the Hanson site with simulation results, the
test results needed to be displayed continuously and without delays. For this purpose, the
actual time duration for the Hanson test was scaled down to the simulation time defined
in the full DEM computation. Further scaling was done with the respect the total weight
of crushed-limestone. The total weight of the crushed limestone in test (916) has been
estimated, based on averaging the total normal load acting on the bottom gate of the ore
pass after four dumps.
Figure 4.36 illustrates the comparison of static loads. In the beginning, there is a
good match between results experiments and DEM simulations, with the final results
showing about a 16% deviation. In comparison of static loads, the spiky behavior of
DEM simulation results should not be considered. The reasons for overestimate of static
load by DEM can be summarized as follows:
♦ The actual filling factor of skip, for the test (916) was not available.
♦ Difficulties in measurement of forces from strain gages Scott and Drescher, (1999).
♦ Approximations in estimating crushed-limestone characteristics and corrugated metal
walls of ore pass.
|
Colorado School of Mines
|
123
4.7 Janssen differential slice method
Granular materials, or bulk solids as they are sometimes called, can be defined as
any material composed of many individual solid particles, irrespective of particle size,
shape. Thus the term granular material embraces a wide variety of particle dimensions
and shapes, ranging from the mixtures of ore and waste rock after blasting to the finest
icing sugar.
Janssen introduced his method in 1895 according to a series of approximations.
The original version of analysis is approximate and most design manuals present a set of
empirically derived correction factors for use in conjunction with the predictions.
His method is the basis of the recommended procedures in most, if not all,
national codes of practice for bunker design. Researchers such as Blight et al., 1994
attempted with some success to find some kind of correlation between their experiment
results and Janssen prediction of load on ore pass walls.
4.7.1 Janssen’s analysis
This method is based on the concept of differential slices of infinitesimal
thickness and finite cross-section and perimeter. His analysis is based on three
assumptions:
♦ That the stresses are uniform across any horizontal section of the material,
♦ That the vertical and horizontal stresses are the principal stresses, and
♦ That granular material is cohesionless, Nedderman (1992).
|
Colorado School of Mines
|
125
F3 = -2twgwôz (4.6)
F4 = 7tw2yôz (4.7)
Because we are dealing with a cohesionless material, there is a linear relationship
between constant shear stress and normal stress, Nedderman (1991). With this
assumption, and through integration of equation of equilibrium, the normal stress on the
bottom and sidewall of a bunker/ore pass (with inclination of/? ) respectively are written
as:
f -Iflkz \
yw$m(P)
\ — e wSin(P)
cr = (4.8)
2kju
\ j
^
f -Iflkz
yw sm(/3) l_e wSin(fi)
(4.9)
2ju
v y
Where fi is the coefficient of friction between the ore and the wall, and k is the lateral
(bulk) pressure ratio.
4.7.2 DEM WALL PRESSURE EVALUATION BASED ON JANSSEN'S EQUATION
DEM can be used to model the mechanical behavior of the material at the
individual particle level, and thereby elucidate the relationship between the bulk
properties of the material and the underlying interactions among the constituent
properties. Therefore, we need an alternative to obtain the macroscopic bulk properties of
the granular material from those of the constituent particles.
Based on the above argument a bulk pressure ratio at each time step is evaluated
from DEM calculation of stresses for all the individual particles in the system as follows:
|
Colorado School of Mines
|
Where N is total number of particles in the system, cr^. is the x-component of stress on
particle i, and a 1^ is the y-component of stress on the same particle.
Evaluation of the global coefficient of friction (for assumed continuum to
compare versus Janssen formula), from DEM calculation of shear and normal forces at
the individual points, are based on the moving average of wall pressures and the least
square method, for more details see Rong (1997) and Nazeri et al. (2001).
To evaluate the applicability of results from DEM simulation versus Janssen's
prediction of wall pressures in ore passes a well-known example was selected. The
example has been submitted to 130 research groups around the world, through an
international attempt to assess the current state of art in FEM and DEM modeling of silo
problem, see Holst et al, (1999).
The problem was to fill a silo of a defined geometry (Figure 4.38) with the 10,000
disk-particles with the mean diameter of 10 mm and a size deviation of 5%. The density
of the solid particles was 1190 kg/m3. The individual particle/wall properties are shown
in Table 4.6, Rong (1997).
|
Colorado School of Mines
|
131
CHAPTER 5
CONCLUDING REMARKS AND FUTURE WORK
This chapter describes the conclusions obtained from computer simulation of ore
passes and outlines recommendations for future work.
The starting point for this study was a comprehensive survey of the relevant
literature related to the ore pass design, operational guidelines and hazards to personnel
or mining facilities reported due to ore pass malfunctions. The safety problems related to
ore pass operation in underground mines where identified as a major point of weakness in
mine ore transportation system.
A review of the U.S. Mine Safety and Health Administration (MSHA) database
for period of 1987-1996 has reported 8 fatalities and 16 permanent disabilities out of 743
accidents in metal and nonmetal underground mines. The statistics show that the majority
of accidents relate to hang-up removal, gate and wall failure, and uncontrolled flow
ore/waste from ore passes.
5.1 Conclusions
This research demonstrates that DEM can be used to model gravity flow of ore in
ore passes and the loading on gate assembly with sophisticated two-dimensional
numerical simulations. Results from this research will help mining engineers to better
|
Colorado School of Mines
|
132
understanding of gravity flow of ore in ore passes and to a more economical and safer
design of the ore pass gate assembly.
The overall goal of this research work was to develop a numerical modeling
capability that would be used to improve the current ore pass design guidelines,
operational performance and reliability.
The major contributions of this research are as follows:
♦ Identification of the critical issues regarding ore pass design, performance and safety
through a comprehensive literature survey.
♦ Development and implementation of a DEM methodology for the engineering
analysis of ore passes. The methodology employs a two-dimensional algorithm,
which can predict the flow behavior of ore materials through the ore passes and
determine the static and dynamic loads on chutes and gating system of ore passes.
♦ Development and application of rigid clusters of disk particles to better modeling of
different ore shapes.
♦ Determination of shape and size distribution of ore after blasting and introduction of
appropriate shape factors for the numerical simulations.
♦ Application of the specialized DEM analysis computer code to study the effect of
some of the major ore pass design parameters including:
Ore material shape and size distributions.
- Ore pass configurations.
Ore/Ore pass walls coefficient of friction.
- The presence of a Dogleg and its configuration.
- The different designs of a dump point.
The stiffness of ore/ore pass walls.
The cushioning effect of ore material resting in the bottom of ore pass after a few
dumps of ore material.
|
Colorado School of Mines
|
133
♦ Comparison of the DEM results to large scale ore pass experimental test data and
classical approach such as Janssen's formula.
♦ Compatibility of DEM simulation results to other DEM codes.
♦ Simulation results demonstrate the necessity to revise of some empirical formulas
such as — = 5, and consideration of the effect of different ore shapes and friction.
d
The present Cluster-2D code is a complex multi-step modeling method, which
allow its users to incorporate the mechanical behavior and characteristics of ore/ore pass
in numerical simulations. The features include:
♦ The preparation of ore (dry and cohesionless) and ore pass geometry and their initial
conditions.
♦ Consideration of global parameters for simulations such as, acceleration of gravity (X
and Y components), friction, drag coefficient, coefficient of restitution, coefficient of
normal and shear stiffness, damping coefficient.
♦ Consideration of system stability and forces equilibrium within a period of
simulation.
♦ Accurate modeling of friction forces of ore/walls.
♦ Application of clusters of cemented rigid disks to model different ore shapes.
♦ Determination of normal and shear loads on ore pass walls at any contact locations
(ore-cluster/walls) within different time steps.
♦ Measurement of lateral (bulk) pressure ratio at different time steps.
♦ Determination of mass-flow of ore through ore pass bottom gate as a function of time.
♦ Accurate and efficient visualization of two-dimensional discrete element results to
view flow of ore in ore passes.
|
Colorado School of Mines
|
134
5.2 Future work
Although this research presents a significant development in numerical modeling
of gravity flow of ore in ore passes, several advancements still need to be made to
provide a more complete and realistic simulation tool for ore pass design:
♦ DEM simulations are very computationally intensive (an average of 1.5 days for each
of the performed simulations on a 950 MHZ Pentium III), thus parallel
implementation of the code may assist the simulations of more complex ore shapes,
and a larger number of ore-clusters.
♦ Establishment of a more logical and realistic relationship between coefficients of
contact stiffness and damping with ore/ore pass walls characteristics (e.g. Young's
modulus, Poisson' ratio, rock toughness).
♦ Current research assumes ore as a dry and cohesionless material; incorporation of
cohesion into the DEM simulation would enable the code for consideration of
cohesive type of hang-ups also.
♦ All the current DEM codes somehow overestimate the magnitude of dynamic loads.
Incorporation of rock failures during impacts may provide a significant contribution
to realistic design of ore passes and their gate assemblies.
♦ Incorporation of a comprehensive classification of ore shape/size distributions after
blasting will ensure a more accurate and efficient simulations of ore passes.
♦ A three-dimensional cluster code should help to the better understanding of gravity
flow of ore in ore passes.
♦ Performing more DEM simulations on results from experiments and active ore passes
can provide better assessment of shortcomings in current ore pass design procedures.
This effort will help to reduce the number of injuries and fatalities associated with ore
pass operations.
|
Colorado School of Mines
|
ABSTRACT
The importance of sulfate-reducing bacteria for metal precipitation in anaerobic
passive treatment systems for remediation of acid mine drainage has been established;
however, conditions leading to decline of sulfate-reducing activity and failure of passive
treatment systems are not well understood. Previous research has focused primarily on
the activity of sulfate-reducing bacteria in anaerobic passive treatment systems, while
little research has focused on understanding the biological processes and carbon flow in
these systems. Other microbial groups degrade complex organic material to provide the
simple organic compounds required by sulfate reducers and are essential for long-term
sustainability of passive mine drainage treatment systems. This research tested the
hypothesis that sulfate reduction in passive treatment systems is limited by one or more
upstream microbial activities that function as rate-limiting steps in generating substrates
for sulfate-reducing bacteria.
The major objectives of this research were to (1) develop a method for assessing
microbial activities in anaerobic passive treatment systems, and ( ) apply the method to a
2
column system for the purpose of discerning the rate-limiting step(s) in the degradation
of cellulose-based organic material as they influence sulfate reduction. The final
approach involved the use of a long-term column study in conjunction with short-term
batch studies. During the batch studies, five substrate supplements of central importance
in cellulose degradation and sulfate reduction, cellulose, cellobiose, glucose, lactate, and
acetate, were added to the organic material from sacrificed columns as a way to probe
important microbial activities. The substrate supplements each targeted a distinct
microbial function at a specific step in the anaerobic degradation of cellulose, which is
typically a predominant component of the substrate mixtures used in passive treatment
systems.
|
Colorado School of Mines
|
1
CHAPTER 1: INTRODUCTION
1.1 Acid Mine Drainage
Problem Description. Thousands of abandoned mines that generate acidic metal-laden
drainage exist throughout the western United States, many in remote locations. There are
over 51,700 abandoned mine sites in EPA Region (Western Governor’s Association
8
1998). An estimated 5,000 to 10,000 miles of streams in the western United States are
impacted by acid mine drainage (Benner et al. 1997).
Generation of Acid Mine Drainage. Acid mine drainage is caused by oxidation of sulfide
minerals in ore bodies found in mines and mine waste. Mining processes expose
otherwise stable sulfide minerals to oxygen and water through tunnels, pits, and other
disturbances or by bringing them to the surface in tailing piles. The most common
sulfide mineral is pyrite, FeS]. Oxidation of pyrite results in the formation of ferrous
iron, Fe2+, and sulfuric acid. Ferrous iron is oxidized to ferric iron, Fe3+, which is itself
capable of further pyrite oxidation. Sulfuric acid leaches other minerals, further
contributing to formation of acid mine drainage. The oxidation of sulfide minerals thus
results in a release of soluble metals, sulfate, and hydrogen ions, which can be
transported into an aquifer by precipitation and drainage pathways in mine tunnels.
Important reactions associated with generation of acid mine drainage are summarized
below.
Pyrite reacts with oxygen to form sulfuric acid and ferrous iron:
(1) FeS2(s) + 7/2 02(g) + H 20(l) -> Fe(aq)2+ + 2S04(aq)2" + 2H(aq)+
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.