THE DESIGN OF COLLIERY INFORMATION AND CONTROL SYSTEMS
Dr. R. K. Holmes / Dr. E. F. Wolstenholme
System Dynamics Research Group
University of Bradford Management Centre
Emm Lane
Bradford, BD9 43L
ABSTRACT
The paper is concerned with describing an investigation of
information usage in the control of colliery operations. The
premise of the work is that to make the most of new information
retrieval technology currently being installed in collieries res-
earch is needed to provide compatible advances in methods of
information usage. The approach adopted was to construct a cont-
inuous simulation model using system dynamics capable of provid-
ing a laboratory assessment of alternative managerial control
policies based on alternative sources and levels of aggregation
of information.
The model developed represents a typical colliery situation
composed of three working coalfaces and incorporating planning
production, development and manpower sectors. The face sectors
transform coal reserves to mined coal output, under manpower con-
straints and geological shocks, and these are all interlinked by
means of allocation policies for manpower and shifts.
A range of policies for the exercise of control through
these allocations are considered subject to a range of shocks.
It is concluded that, although there are difficulties in design-
ing single policies which are universally best, there are clear
advantages associated with fully integrated colliery policies
based on information inputs from all aspects of the operations.
1. INTRODUCTION
During the mid 1970's, as a consequence of the rise in the
price of oil, demand for coal rose and the coal mining industry
in the U.K. was revitalised after a long period of decline (1)
(2) (3). As a consequence of this a new strategy to increase coal
220
output was developed. This involved the exploitation of new
reserves by large new collieries such as at Selby, major extens-
ions to existing collieries with significant existing reserves,
and the increased use of modern technology, both for better and
more efficient mining equipment (Advanced Technology), and for
improved information retrieval (the computer based system MINOS
(4). This paper concerns research into a need generated by the
latter but with the practical potential to have a significant
impact on all aspects of the revitalisation.
With the advent of mini and, in the late 1970's, micro
computers, the technological advances in the speed of operation
and capacity of information retrieval methods has been unpreced-
ented. The true purpose of collecting information is, of course,
to enhance the quality of managerial decision making and through
this to improve performance. Hence the potential of new inform-
ation retrieval technology will not be fully attained until
comparable advances are made concerning information usage. There
is clearly a need therefore to develop methods which are capable
of assessing the effects on performance of alternative informat-
ioninputs to managerial policy making and of alternative policy
formulations. The work described in this paper is therefore con-
cerned with methods for designing managerial control. In fact
it can be argued then that until usage of information has been
examined in such ways it is not possible to make rational decis-
ions concerning information retrieval.
The methodology used in this research to develop a frame-
work for the design of control in collieries was System Dynamics.
This was chosen because it implicitly provided a basis for rep-
resenting a colliery as a complex information feedback system
capable of facilitating a simulation analysis of alternative
operating policies, and their robustness under a wide set of exo-
genous effects such as geological risk and manpower availability.
The representation of a colliery as a feedback system
will first be outlined and subsequently experiments with and
results from the application of the ensueing model will be
presented. The model to be described can be considered as a
trial development to ascertain the feasibility of the approach
for the purpose defined. It represents a typical colliery sit-
uation where the three coalfaces and their associated replace-
ment developments are operated with the object of attaining a
target output given geological and manpower availability fluct-
uations.
2, HE COLLIERY AS A SYSTEM
In the most general sense a colliery can be described as
a system for converting coal reserves buried underground to
mined coal on the surface (which may be used for electricity gen-
eration; industrial; and domestic use). This being said, the
boundaries of the system are easy to identify and relate to the
physical characteristics of the colliery as a unit.
Thus the simplest model of a colliery as a coal conversion
system is shown in fig. 1. This identifies the main states in
Exogenous
constraints! =p( Manpower
variation
FIGURE 1,
Pit room
development
\ccessed
serves
development |q= = = =
E===p Production ‘
y
COLLIERY PRODUCTION PROCESS
ane —
221
Exogenous variation
{geology etc.)
Exogenous constraints!
variation (shaft capacity,
geology, haulage
capacity, ect)
which coal exists and the processes which transfer the coal bet-
ween the states (such as ‘coal face development’). The répres-
entation of a colliery in this way is the first stage in the dev-
elopment of a system dynamics model. It defines the physical
stocks of coal which can be considered as levels or state vari~
ables which are an integral part of this modelling philosophy.
The diagram also identifies major shocks to the system from exog-
enous factors and captures the main planning and control elements.
Fig. 1, although playing a key role in outlining the main
characteristics present in the colliery system is too abstract
to help in the detailed delineation of the system dynamics model,
which, if it is to be at all valid must be an adequate represent-
ation of the real physical system. Indeed, the detailed physical
structure of the system is an important determinant of the model
structure as will be seen. Thus it is necessary to examine the
colliery in more detail to ascertain how it operates.
The central aspect of colliery operations is the develop-
ment and use of production capacity, and this is illustrated in
fig. 2.
The accessing of reserves is carried out initially by driv-
ing the major roadways and these open up large areas of the coal
seam being worked. This is referred to as pit room development and
as it is frequently carried out by outside contractors it is
considered external to the model and only the more detailed
face room development works as shown in fig. 2, is considered.
This consists of driving two parallel face access roadways for
COLLIERY PRODUCTION PROCESS (plan view)
FIGURE 2.
Planned face capacity
222
6
9983 (20D,
ro----4
I I
I I
I I
I I
I 13]
s
I 13 2
I 12 Hf
I 18 8
! 1g az
i tg
as
-t oe HF
83
t ss
A bs < '
2 Hill |/83 a
wae $8 a s35 88
zua $8 a Ore fe
aS 33 g zoe §s
aQ £3 9 523 83
ues g use 33
£8 8 HEE Gs
J
sAempeos s0leW q
each coalface and the face line which links them at their extr-
emities, to create an area of coal to be extracted known as the
planned face capacity. Such a coalface in mid development is
shown at the top of fig. 2.
There are a variety of ways in which this face room dev-
elopment subsystem could be modelled. For the purposes here
attention will be restricted to the most commonly used form of
development where each operating face has a designated replace-
ment face associated with it; the development of which is started
when the stocks on the production face fall to a pre-determined
level. Before a developed face can produce coal it must be
equipped with the necessary coal extraction machinery and the
final act of development is to install this equipment. This
creates a fully developed face as shown in the centre of fig. 2.
Once a coalface has been developed production can take
place. Production consists of cutting slices of coal from the
face and the whole face line moves forward with the roof being
allowed to collapse behind it. In fig. 2 this is shown in the
bottom coalface with the production face line moving back to-
wards the main roadway. This is technically known as retreat
working on a longwall face (5).
It is obvious from figs. 1 and 2 that developed capacity
is generated by a development rate, and that this developed
capacity is converted to production capacity once the whole face
has been developed. The production capacity is then depleted
by the production rate and the cycle repeated. These two rates
are, to a certain extent, dependent on the number of men assigned
to these functions and this allocation of men must be the major
agency for the control of these rates. These rates are influ-
enced by externa] factors such as changing geology.
In real life a colliery system is complicated by the fact
that a number of coal faces may exist at any one time, of both
development and production types. The rates of production and
development on these faces mist be co-ordinated, and this may
impose conflicting demands on the available resources (primar-
ily men).
The need to control face development and production
implies a plan. That is a desired state of affairs at any point
in time, divergence from which will instigate corrective control
actions. This is expressed in the colliery situation by the
Action Plan which specifies the required present and future
state of the colliery in detail up to a period of 18 months in
the future. The Action Plan provides that the collieries’
operations are integrated within the Area and that long term
business objectives are met. The Plan, which is updated theor-
etically every 3 months, sets targets against which actual per-
formance can be measured for operational control.
Colliery management is held accountable for meeting prod-
uction targets on an annual basis (with quarterly checks),
based on a yearly budget setting procedure. Obviously the over-
all state of the colliery is considered together with the
financial performance, but primary index of performance is
the tonnage of coal produced.
Clearly from the above description of basic colliery op-
erations, a colliery can be conceived of as a dynamic feedback
system as defined by Coyle (6). There is obvious dynamic beh-
aviour in terms of production and development rates. Although
these are subject to some uncontrollable elements they are
largely controllable via manpower development policies. Their
control is in fact a key element in colliery management. Control
is in principle applied by the definition of target states for
cumulative production and development. Information is fed back
from the system states for the exercise of this control and due
to the complexity of the system in terms of multiple production
and development faces which will generate conflicting priorities,
there may be a variety of control policies which can be invest-
igated to improve overall performance.
3. A SYSTEM DYNAMICS COLLIERY MODEL
3.1 Model Structure
The first part of the system dynamics modelling process
is to conceptualise the system under investigation in
terms of level and rate variables, and to express these
in diagramatic form to illustrate their interconnective-
ness. Such an Influence Diagram is shown in fig. 3.
This illustrates the simplest possible feedback inter-
pretation of the colliery system described in the previous
FIGURE 3.
(Version 1)
Initial Colliery Planning and Control Model
Face ¥®., opment
Cumulati Om room
development rate
CFRDR
y
Face room
development rate
®
©
completion rate
FRDCR
Bg
28
gge
eee
S20
Bere
x kSS
28>ak
gy ee
zee
evs
= EO
= §
Be
s
+
HP eiom
Face room capacity
+
&
2
ira
of
Time to
FRC
o/
Production rate
correct
errors 13
©
‘Average production
©
Production
|
a
rate APR.
Averaging time 12
PR
cover T1
224
i
section of the paper. The development rate is accumul-
ated into the development capacity, which when equivalent
to a complete face is instantaneously. transferred to face
xoom (production) capacity, which is then depleted by
the production rate. In this situation the production
rate drives the system with the development rate being
dependent on the production rate and current state of the
achieved development capacity.
In order to account for the complexity of the real
world colliery system this basic structure was extended
through a number of iterative steps. This evolutionary
approach in which each added factor was tested and valid-
ated resulted in a validated final model of a hypothet-
ical yet typical colliery.
The use of a hypothetical as opposed to actual
colliery situation was chosen so that the results obtained
would have sufficient generality to be widely applicable.
A specific situation may have such characteristics as to
make it unique and so invalidate any generality as to
the conclusions drawn from it.
The general structure of the model created is shown
in fig. 4 which clearly illustrates the separate sectors.
There are three coal face sectors as this situation is
not uncommon in reality and gives sufficient complexity
to illustrate the merits of different control policies.
Working
Population
-—— >
<«-+
12
i Y 2 tz
& wo )z5
c tol =
a 855|GE5
e65r 4 god| 265
2u 209| 260
go 1 23%) $9e
s serepenBel SA 1 OS
T ge
1 w,
FACE SECTOR
FACE SECTOR
FACE SECTOR
A
Loe
ee
KEY
=== Flow of coal (physical)
&
oO
Q
2
a
— — — Flow of men (Physical)
ACTION
PROGRAMME
** Flow of information
—-==- Flow of control
225
COLLIERY SECTORS AND THEIR RELATIONSHIPS
FIGURE 4.
13
These face sectors translate accessed coal reserves into
coal output.
The operation of these face sectors is dependent
on the manpower available, the manpower being generated
within its own subsector. The allocation of the manpower
in terms of men to machine shifts and machine shifts to
faces is carried out subject to the policies adopted by
the colliery management in order to meet the specified
plan. These policies are defined in a further policy
sector.
An individual:coal face sector is illustrated in
fig. 5. This shows the physical flows of coal and men
through the actual development and actual production
subsectors associated with the coalface and identifies
the major variables. The schematic in fig. 5 also shows
that two planning subsectors are associated with each
coalface. ‘These planning subsectors essentially mirror
the 'actual' subsectors and generate the two variables
‘planned development capacity’ and ‘cumulative planned
production rate’ against which the actual performance
is compared. Discrepancies are thus generated which
are used as controlling variables in the managerial
control policies. Information feedback on these discrep-
ancies occurs with the discrepancies being fed back from
each coalface sector to the management control policy
sector.
OT
Machine Shift Allocation
‘Machine Shift Allocation
14
Manpower Allocetion
Manpower Allocation
Development Start Trigger
Equipping Start Trigger
226
A TYPICAL FACE SECTOR.
FIGURE 5.
15
This very briefly describes the major aspects of
the model. The three face sectors are simply replications
of the basic face submodel with parameter changes to
represent the individuality of each face.
3.2 Managerial Control. Policies
It is obvious from the foregoing that the core of
the model is the managerial control policy sector. The
basic information links into this sector from the face
sectors have been described in the previous section,
however. dependent on the sophistication of the policy
to be implemented further information may be required.
In general terms control can be considered in a
number of dimensions. Firstly the key variables by which
control is implemented must be defined; secondly the
frequency of application of the control must be spec-
ified; and thirdly the type of control (that is which
information is to be used and how) must also be defined.
The key control variables are the number of machine
shifts to allocate to each face on each day (0, 1, 2 or 3),
and the manpower to allocate to each machine shift.
The frequency of application of control depends
very much in colliery management on the level of manage~
ment at which the control action originates. A colliery
manager may take action himself at any time (i.e. carry
out continuous control), but is ultimately accountable
16 227
and held in check by the planning/budgeting system. Hence higher
levels of management in conjunction with the planning department
may influence or impose changes on the colliery at discrete points
in time (reyiew point control). The control policies which have
been tested using the model fall into two general types: the first
category concerns what we shall refer to as semi-integrated (SI)
policies, which are based on only a subset of the total information
available from the total state of the colliery and which are current
feasible and commonly applied. ‘The second category concerns what
we shall refer to as fully-integrated (FI) policies which should
be feasible to apply given the information retrieval methods bec-
oming available. The fully integrated manpower allocation polic-
ies are based on the work of Wolstenholme (7) who has developed
and applied similar algorithms to the management problems of contro-
lling bunker discharge rates in conveyor belt systems. This is not -
to imply that these algorithmic policies are programmable in the
sense of automating managerial decision making. They can, however,
form a logical basis for such decision making and obviously the
manager must also take into account a whole variety of other fact-
ors when making such an allocation decision.
Fig.6 illustrates the control policies tested and it will
be noted that consideration of review point control will be rest-
ricted, as in practice to overall guidance on machine shift all-
ocations and only through semi-integrated control methods. Each
policy will be briefly described below:
a) Manpower Allocation Policies
In an ideal situation there will always be sufficient
men available to give full manning on all shifs on all
FIGURE 6.
v7
CONTROL POLICY DELINEATION
| Key variable
used for
policy Review point control Continuous Control
implement-
ation
Machine
shifts Semi - Sem! - Fully-
(per face integrated integrated integrated
per day)
Manpower per
machine Semi - Fully-
shift integrated integrated
b)
faces at all times. Howeyer, when there are insuff-
icient men they must be deployed on some sort of rational
basis. Manpower deployment is a common problem (see
for example Teal (8) and can only be done when the men
actually turn up for work (approximately continuously).
The continuous semi-integrated policy defined in Fig.
6 assumes that no information is available or required
on the state of the coal faces as men are allocated to
production in preference to development. ‘his is done
such that development at least continues with minimum
manning if possible and in the extreme case of the faces
not being capable of minimal manning then a development
shift is dropped.
In contrast the fully integrated policy uses the
face discrepancy as the basis for manpower allocation,
and compares discrepancies between faces such that the
face which is furthest behind receives most men (whether
it be a production or a development face).
Machine Shift Allocation Policies
In the foregoing policies for allocation of men to
machine shifts it was assumed that the allocation of
machine shifts was fixed. In reality, shift patterns
are maintained for as long as possible as changes can
cause a great deal of disruption (not least in industrial
relations). However, if necessary, machine shift alloc-
ations can change and these policies delineate several
19
rational bases for such changes.
Changes at discrete points in time will be con-
sidered first. This is analogous to the situation where
the machine shift pattern is reviewed when the Action
Programme is updated. This occurs every three months
and on the basis of the actual status of the system at
this review point the planned allocation pattern for
machine shifts may be revised. It is assumed that the
manager will then work to this pattern in the actual
allocation of machine shifts.
The basis for changing the planned machine shifts
is defined here to be on the basis of discrepancies. The
production faces are considered in isolation and if they
are behind schedule to such an extent that they are
‘eritical' then their shift allocation is increased.
This increase may be catered for in two ways: simply by
absorbing a built in productive capacity (i.e. spare
men); or by sacrificing a development shift for each
extra production shift allocated. The first of these
recognises that spare capacity may be built in to cater
for anticipated future geological effects. The second
is where an unanticipated exogenous shock may occur.
Given that the planned number of shifts may be
set at a review point the manager may wish to change
the actual allocation between these points to cater
for unpredictable changes in the situation and this
20 229
is the basis for continuous control.
In continuous control a semi-integrated policy
such as was described for review point control may be
applied. However, it is also possible to consider the
application of a policy which takes into account the
situation on all faces in determining the changes to
be applied on one face (that is the status of both
production and development faces). In essence when a
discrepancy on any face becomes critical the shifts
allocated are increased and this increase is compensated
by a reduction in other allocations proportional to the
shifts previously allocated. This allocative algorithm
also has the capability of allocating 'spare' shifts
(that is shift capacity released when a development is
completed, for example). This means that any ‘spare’ men
are fully utilised.
The basis on which the consequent shift changes
are made here is relatively simplistic being simply in
proportion to the prior allocation, however the principle
could be extended to other bases such as the magnitude
of the discrepancy between planned and actual perform-
ance (relative to the total discrepancy across the
colliery).
The above range of policies demonstrates the ability of
the model to cater for the evaluation of widely differing manag-
erial control policies.
21
4. CONTROL POLICY EXPERIMENTS AND RESULTS
In order to test the effectiveness of managerial control
policies of the type described in the previous section a series
of experiments was designed and carried out. The experiments
catered for varying situations which tested policy robustness
by imposing several exogenous shocks. These shocks represented
firstly a deteriorating geological situation causing a reduct~
ion in face output, and secondly a deteriorating manpower
availability. ‘The shocks were replicated in each series of
tests and each policy so as to obtain comparability of results.
The results fell into two types: qualitative graphical
output; and quantitative output using a variety of performance
indices. The simulation model being written using the DYSMAP
(9) language.
The qualitative displays illustrate face capacity fluct-
uations and manpower resource utilisation. ‘The reference mode
for face pzhaviour is illustrated in Fig. 7, and Fig. 8 ill-
ustrates this equilibrium behaviour generated by the model. By
simplifying the number of variables displayed and combining the
faces, the positions of the face states relative to each other
may be displayed as shown in Fig. 9. Aggregated variables of
interest may also be displayed as shown in Fig. 10 which displays
overall production discrepancies, and manpower availability
and allocation.
DYNAMIC REPRESENTATION OF FACE PERFORMANCE
FIGURE 7.
Face room
capacity
Tonnes
PFRDC
Production face
22
\.
e
a
—
3
8
a
a
@
c
aaa
planned
time to
equip face
TIME
Planned time to
develop the face
230
8
=
=
|
100. ans
Hi ed
(|!
Vt
1
50. [
ti. | ards \ nemcee a
1
il |
o, + } 4 4 } i yi re i }
18 6.0 20. 40, 60.80. = 100,120,140, 160. 180.- «200. 220. 240,
bs | | 1 TIME @ SIMULATION TIME
j | Leeca a FACE ROOM CAPACITY ~ FACE A
| —-CFADR (CUMULATIVE FACE A DEVELOPMENT RATE
+ -—-SFADC m SPARE FACE A DEVELOPED CAPACITY
_.—SDAT on QUANTITY IN FACE A EQUIPING DELAY
EXPT-2B, 1.1
FIGURE 8. MODEL GENERATED FACE DYNAMICS
FILE CPC9E2B :
300.
T
i
250.
aH
200. 4
i
pall
I
100.
i
so.
1
|
wi
i
50, I
|
oe —_
jf er eerie eee eer en FR career romani
1
150. -
"
an bees pt
300 20. 40. 60. 80.100. 120. 140." 160. 180. 200. 220. 240.
#10 li TIME SIMULATION TIME
iL erca (FACE ROOM CAPAEIT
| Lo FRCB a FACE POOM CAPACITY
ma FACE ROOM CAPACITY - ‘FACE C
a” SUITCH TO INDICATE FACE 4 DEVELOPHENT CAN PROCEDE
af SWITCH INDICATING FACE 8 DEVELOPMENT CAN PROCEDE
aM SVITCH TO INDICATE FACE C DEVELOPNENT CAN PROCEDE
FIGURE 9.
RUN NO. 1.1.1.1, NO SHOCK
COAL FACE STATES
MANPOWER POLICY EXPTS,
~ SUBSET 1 (A)
rd
Tez
25 26 232
s
— With regard to quantitative output this must of course
s be indicative of the performance of the system as represented
48
by the model. To this end a number of performance indices
a were defined to represent various aspects of colliery oper-
ations such as cumulative production, developed faces awaiting
3
‘a 2 production starts (spare capacity), downtime due to late
s completion of a development, and downtime due to insufficient
gS
+ . men to run a face.
; 8
é z ny
1s 3 x A set of four experimental runs were then designed in
4 Ss
. BES & terms of the magnitude of geological and manpower shocks to the
s 3 _
aft pose =] < system and the previously defined managerial control policies
w gf =
of ic tested against these.
42 8 R888 26
, 3 ggs® ant 2 It is not proposed to present the results in detail as
“Ig & Ska8 ~ wo they are available elsewhere (10) but Figs.11 and 12 give an
SE.u2 ~ ot
a288 indication of the dynamics of continuous machine shift and
; a8 2 oie
3 3
43 8 © Ss g manpower allocation policies under the geological shock comp-
w oo 88 Ss ared with the base case of no feedback control.
= ess z 8
ho In a manpower shortage situation, . the continuous
od
we 2 allocation of manpower is an ideal mechanism by which relat- -
x
SLES g o ively small discrepancies between the actual and desired
1
—1| 28 8 states of coalfaces may be corrected. Given this need for
* og as
% | | ff 2 an allocative mechanism the policy which most effectively
2 ! ao =
¥ | q achieves this objective is of the fully integrated type
a
J ig ge taking account of information from all production and devel-
a eee q opment face states.
és “6 ¢ 4 6 © ¢s ¥ ¢ B15 3
& 8 8 €£ &@ 8 8§ K&S Bis
& 8 gf 2 8 ge Ble &
& =
i
3)
#
g
Fy
==dete=famieeta es
g 8
Sa eee
8
5
Ti
i
Es 0
i
a ad}.
i {i
a H 80.
i
100,
{i
0
i
ti
— — — 7
. oo 20. $0. to $0, —60. 100.120. ta. 160. —te0. 300. 220.200.
ae lt TIME swmnarion re
iL geca mn FACE ROOM CAPACITY = FACE A
[LrFece (1) PACE ROOM CAPACITY — FACE 8
: tm "pace nan ciency - re
an firm “Sorrow 0 MOTCATE FACE A OevOPENT can PRocDe
fia —_SviTen rearing Pace cevtomen? conocer
fiz “Seta 0 moleare Face € DEVELOENT Cm PROCEE
250. Tl 1000. ! m7) 1000.
1 750. i] 750.
a
zoe oa
2 § z s
ae See eee ee ee]
3
a eee eee eee]
ex
=
50. 50. +
-250. -250.
a \ a4 i
! | i i
y 500. 4 500. 4
® a a a OO EL OE 9. a ee OO Ce Ce
Ip wet TIME a sinsarion rie [| rt TIME @° smxarion re
| {Legpag ay cuenta acne ren a cet i 'Legeae my caxeenr mon. o1scercr rao wacer
| Li epeoR H_ -COLLTeey paca. aIScREDANCY FeDR PLAN | 1) COLLIERY PROD, DISCREPANCY FRY PLAN
iia feo “oni sa oe ce ev atuoae fm “tora naar Face TEV AUTRE
feo “Torn a “ev Miser
yD TOTAL Na. OF PEM ALLOCATED
FIGURE 11. RUN 38.2.1, GEQLOGY 2nOCK FIGURE 12,
MACHINE Shit ® POLIS
RUN NO._1.2.1.1, GEOLOGY SHOCK
MANPOVER POLICY EXPTS. - SUBSET 1 (A)
29
The experiments also emphasised the need for a
structural link between development and production so as to
synchronise the development and production phases on a face.
The link tested was relatively crude in-so far as the product-
ion face state only triggered the development start. The
planned development rate through which this’ control trigger
was exercised depended in part on the historical production
performance only. ‘The production face state after the trigger
point was not fed back to aid development control and thus
major changes in the production rate could not be recognised
and compensated for in the development, causing on occasion time
delays between exhaustion and replacement of the production
face. This was of greater importance when continuous machine
shift control was exercised. It is not unreasonable to conc-
lude that the continuous feedback of information on production
face states so as to redefine continuously the planned devel-
opment rate would improve this situation, particularly where
integrated control over men and machine shift allocation is
exercised.
The imposition of periodic review point control whilst
necessary as an accountability measure can be seen as causing
a certain amount of disruption if the extra manpower requirements
are not catered for in the initial design of the number of
faces and shifts to be worked. However, even with this greater
sensitivity to manpower review point control can ensure slightly
better performance if it is carried out by temporarily borrow-
ing development shifts for production purposes.
30
When continuous reallocation of machine shifts is
considered this can pre-empt review point control making it
unnecessary. However, considering the continuous machine
shift allocation policies it is apparent that integrative
policies are superior by providing for the better utilisat-
ion of the available men and greatly increasing the total
colliery production. The policy does suffer from the draw-
back of greater sensitivity to manpower availability and the
problem of phasing developments into production faces. This
latter drawback can be overcome by the provision of an
improved production development link as mentioned Previously.
The results from the machine shift allocation experiments also
raise the possibility of using different policies under diff-
erent types of exogenous shocks and constraints. For example,
given a restricted manpower establishment it may prove more
beneficial to use the continuous machine shift allocation
policies under geological shock conditions and either review
point control or no machine shift allocation policies at all
under certain unstable manpower conditions.
5. CONCLUSIONS
The research described in this paper has conclusively
demonstrated a number of valuable points. The most basic of
these, and which now may be seen to be self evident, is that
a colliery can be described as a dynamic system within which
feedback control is exercised. Having demonstrated this fact
through a description of a typical buthypothetical colliery
234
using the system dynamics methodology, the simulation model
produced as a result of this description has proved to have
the capability of elucidating and testing managerial control
policies,
The validated model was used to evaluate a variety of
managerial control policies, these policies operating through
a resource allocation mechanism of machine shifts and men to
these shifts. The control policies included simplistic,
though realistic, policies which treated the faces on an
individual basis, and integrated policies which considered
the interactions between faces.
The identification of 'good' or ‘better’ policies through
this analysis naturally has implications in terms of the feed-
back of information. The policies require certain types of
information at defined times (for example face states exp-
ressed in capacity (tonnage) terms). Although this is by no
means the only operational information required by colliery
management it does define a minimum needed and is based on a
rational analysis. Ad hoc or potentially all embracing
information retrieval systems may include it, but the analysis
previously described not only defines what information is
required, but also how it is to be used.
The model developed may be extended and used in a variety
of ways. It may be used for further control policy analysis
in a theoretical context or, equally importantly, the principles
embodied in the model construction can be used to generate
32 235
simulation models of actual collieries for use as managerial
decision aids. It is additionally felt that the concept of
establishing actual and desired states for each coalface
essentially formalises a subconscious principle employed by
good managers. ‘The exploration of this concept through the
medium of a feedback model could contribute significantly
to management training by establishing the merits of the
principle in all managers and providing a basis for assessing
alternative ways by which discrepancies can be corrected.
33
References
q@)
(2)
(3)
(4)
Gs)
(6)
[oD
(8)
(9)
(10)
National Coal Board; 1974, Plan for Coal.
National Coal Board; Plan 2000.
Taylor, R.j 1979; The Challenge for Coal; Management
Today, Nov. 1979, pp. 74-81.
Hartley, D.; 1978, The Remote and Automatic Control
of Mine Operating Systems; The Mining Engineer, Nov.
1978, pp. 363-370.
Woodruff, S.D.; 1966; Methods of working Coal and Metal
Mines; Vol. 3; Pergammon Press.
Coyle, R.G.; 1977; Management System Dynamics: John
Wiley & Sons.
Wolstenholme, E.F.; 1980; Designing and Assessing the
Benefits of Control Policies for Conveyor Belt systems
in the Underground Mines; Dynamica; Vol.6; Part II;
Winter 1980, pp. 25-35.
Teale, C.W.; 1979; The computer based manpower deploy-
ment system developed and operated by: N.V. Kempense
Steenkolennijnen, Houthalen, Belgium; National Coal Board
Operational Research Executive internal report, No.
OR.923/1/1; March 1979.
Cavana, R.¥., and Coyle, R.G.; 1982; Dysmap User Manual;
System Dynamics Research Group, University of Bradford.
Holmes, R.K.; 1982; The Design of Colliery Information
and Control Systems; Unpublished Ph.D. thesis; The
University of Bradford.
236
