Planificacin Minera Estratgica Integral

Strategic and Integrated Mine Planning

Hugh Thompson
Runge Latin America Ltda.
Gerente General
Telfono : 56 2 333 03 04
Fax

: 56 2 333 02 99

email: [email protected]

Page 1

Runge Latin America Ltda :: Saturday, 09 de July de

Abstract
Planning in the short, medium or long term is understood to be the definition of the
activities and the resources (physical and/ or human) required to fulfil the stated
goals. The benefit of defining a plan which is as close as possible to what can be
achieved in reality translates into significant savings of resources for the company.
In mining today, planning is of paramount importance, even more so when one
considers that the business margins are reducing, requiring the evaluation of
diverse alternatives for the optimal use of scarce resources. Additionally the capital
required to approve some projects demands the best planning possible. Modern
planning requires that all of the aspects of the business be represented, and not just
the planning of the extraction in an isolated form. Planning should therefore define
processes that clearly identify the business, with the business goals as the stated
planning targets. These processes could be defined in the following manner:
Extraction of the Material, Treatment, Obtaining of the Final Product, and Transport
of the Product. Each one of these processes models a flow of activities in order to
get the product with associated characteristics, for the following process. This
integrated planning must interrelate these separate processes of the mining
business, with an ultimate focus of strategic character. It should respond to the
following questions for each process of the business:
1. How will they carry out the activities?
2. When will they be carried out?
3. What human and physical resources are to be employed?
4. What constraints are imposed by the preceding or subsequent process?
5. What will be the budget guides for these activities?
In order to carry out integrated strategic planning in the mining business one needs
to have implemented a culture within the organisation that allows the business to be
understood as a interelated system of processes. These processes must be coordinated so that they interact properly between each other and so define an
integrated plan that optimises the available resources and maximises the benefit to
the company. This defined culture should consider the important variables of each
activity within the processes, and its relevance to the strategic plan. For example
maintenance needs the projected operating hours of equipment. Finally, the
planner needs to have adequate computer tools, and access to all parts of the
process, in order to facilitate the creation and modification of the plan in an efficient
manner. One of the only true knowns in mine planning is that we do not know
everything. Which indicates a need for iterative mine planning, rather than the
application of true mathematical optimisation techniques.
This paper shows an example of computer tools used in order to reach the
described objectives. Within this example some key points are illustrated which all
mining professionals should always maintain:

Page 2

Runge Latin America Ltda :: Saturday, 09 de July de

1. Time should use in analysis and not in managing data and complex techniques.
2. The difference between "Data" and "Information" is its communication.
3. Strategic Planning is different to scheduling.
4. Strategic planning can no longer be seen as a once a year activity.
5. Mining is a business.

Page 3

Runge Latin America Ltda :: Saturday, 09 de July de

Background and Introduction

The drive to optimise comes from the fact that mining businesses have significant
turnover and any improvement in the overall business can provide a large return.
Strategically this is why it occurs. Continuous improvement, which embraces
optimisation, is core to strategic direction of business.
Significant effort is put into mine planning. As planning is labour intensive and time
consuming, often the first plan that meets the corporate criteria is the one that is
used. At best a few scenarios are run. Planners know that, given more time
improvements can be made. This results in the feeling that they should provide the
best plan.
So how do we find this plan AND how do we know it is the best anyway?
The most obvious way to do this is to make sure that the Best plan is the first one
derived and we use tools which are proven (mathematically) to give optimum
solutions. Generally techniques like linear programming and mixed integer
programming are used. These include various attempts to reflect various
components of cost structures, and hence hope to produce a result that satisfies all.
In this way business process integration is attempted.
Much effort and expense has been put into development of optimisation business
planning tools in the mining industry. This should not be confused with the tools
developed for determining the optimum pit or stope size limits. The optimisation
planning tools have generally failed, for one of the following reasons:
1. The problem is very complex to model, therefore gross simplifications are used.
This echoes the garbage in garbage out adage that all understand.
Essentially the mining system must be expressed in well-defined mathematical
models. Often common sense will provide a better solution.
2. Prejudgements of the solution are imposed as rules the solution is the answer
it was told to give.
3. The labour and time cost of constructing a valid model is so large that the model
is obsolete by the time a solution is derived.
4. The skills required to run the model on an on-going basis do not exist within the
mining business. Effectively the available professionals who understand the
processes in the real world cant drive the planning tools as users.
5. When results come from a Black Box optimisation can they be believed?
Basically mining is complex and many of the cost-revenue relationships are only
partially understood at best. However all agree that it is economics which must
drive the planning. Add to this an inherently dynamically changing environment
on both the production and market side - and the result is a planning puzzle that
does not fit well with the technique used.

Page 4

Runge Latin America Ltda :: Saturday, 09 de July de

Whereas if the objective is simply producing a better plan, and this is not limited by
theoretical optimisation, three things happen:
1. You still get improvements in the bottom line.
2. You are no longer required to express reality in complex mathematical formula.
Therefore eliminating the black Box issue, so that results are tangible.
3. You can deal with real world rules and complex trade-offs particularly those
that were not apparent when the planning cycle began.
This planning philosophy where you produce one plan, assess it, then produce a
better plan is called iterative mine planning. Essentially this process allows the
planners to know the economic possibilities of mine better, and in turn to drive the
planning.
Characteristics of a planning model
There are many considerations to be made when discussing a planning model.
This paper discusses the relevant points of two of the aspects of this. These are
the basic model design, and the role of data. These are topics that are continually
overlooked in planning.
Planning Model Design
Firstly there needs to be a conceptual design stage for any model used in planning.
This is the time when needs and technical considerations are contemplated. This
can be done in a many ways, however the following always need to happen,
although not necessarily in this order:
1. Decide what are the key results indicators required for evaluating each iteration
of the plan. This should be seen as changeable, in so far as it may only be with
iteration #4 that the need for key indicator gamma is obvious. Eg. tonne
kilometre per hour, or % duty cycle for the engines are usually better indicators
of required haulage effort than tonne * kilometres. But this may not be obvious
at the start of the plan when there is a mix of different haul types.
2. Understanding the individual processes of the business. This involves
understanding the difference between a process and an activity which is
contained in the process, and the relationship between the processes.
3. In turn this means understanding how these should be correctly planned and
modelled. Usually this involves defining a quantity for each process (eg.
Tonnes, metres, ounces), and deciding on either a productivity, or a method to
calculate a productivity. Some processes, eg. transport, may need to be
modelled as a series of discreet packets, not as a continuous action.
4. This will lead to the setting of targets, and the initial analysis of capacity of the
orebody and resources to deliver these targets.

Page 5

Runge Latin America Ltda :: Saturday, 09 de July de

5. Further analysis of the targets and the deposit should produce a set of rules and
guides for the planning staff. Often these can be expressed in terms of how the
target should be made up, and is the initial manner in which a mine plan
becomes
integrated
across
disciplines.
Then planning can begin which values the imposition of successive rules, which
may be assigned priorities. And so the plans are changed and improved
(hopefully)
with
each
iteration.
6. The rules, data, targets and objectives all need to be hosted in an environment
where they are available and verifiable. This is the mine planning database.
7. The long term, or strategic planning model should be vertically and horizontally
integrated with other planning models in the mine. Vertical integration in this
sense may mean that the long term and short term models are properly aligned.
Horizontal integration can mean that the planning model is capable of delivering
different styles and types of reports for different users. Eg. management
accountants need different data to process plant people, which is different again
to what maintenance requirements are.

For example, the base plan is to deliver to the mill 10,000 tonnes per day. Next
comes the imposition of the rules shown in Table 1
Table 1 :: Integrated Planning Rules
Expression

Source

A maximum of 1,000 tonnes of Arsenic per month can be produced.

Oxide % of mill feed to be between 10% and 25%, as long as is possible.
Each truck can have only 500 operating hours per month.
Recovered head grade, rather than feed grade should be maximised.
Quality controls need to be met on a 21 day cycle, to fit in with shipping
Only 10% of production in any given period from below the water table
Three separate ore sources need to be in production at the same time

Environment
Plant
Maintenance
Process
Marketing
Mining (& others)
Risk control

These rules may have only been generated after initial plans have been published
and distributed for comment. In an ideal world they would be known beforehand.
The interrelationship of the processes and their combined effect on economics is
evident. For example environmental constraints stopping a dump location; which
can impact on truck requirements, waste removal, ore availability and revenue
security.

8. Plan for change. It is expected, and desired, that the planners learn from the
results. Therefore there will be a need for modifications to the planning model in

Page 6

Runge Latin America Ltda :: Saturday, 09 de July de

the future. It should not be a static system. Obviously some model

development should be deferred, for example do not introduce radical new
methodology whilst trying to produce the final Life-Of-Mine plan to be published
tomorrow. However changing the recovery calculations so that they match the
current best practice on the mine should be done immediately. Fundamentally
the system needs to be capable of change. In turn this implies two often
forgotten concepts.
i.

Skills to make changes need to be available locally (ideally the mine

planners themselves), and

ii. the planning model as a whole need to be documented and readily

understood by other people as well as those who created it.
9. Know that a variety of different results, in different formats will be required.
These should all be identifiable as coming from the one base plan, and should
be readily produced. Whilst planning may be done in a specialised planning
system, that is a rarely where the results need to reside. Usually this needs to
be passed to other, simpler packages, such as MS Excel, MS Word or MS
PowerPoint for communication to others. The ease and range of output is a
serious consideration. It is the interfacing that will provide the best solution for
each unique situation.
10. Strategic planning is often used interchangeably with long term planning.
However this disguises the fact that the short term plan is in fact strategic in
nature. Therefore the planning model does need to be able to properly direct
and relate to short term planning considerations.
The reverse is also occasionally true in that planning for the whole life of the
reserve may not in fact be strategic. This can often be the case where a mine
has vast reserves, and decisions on year 25 have limited impact on the strategic
decisions that need to be made within the next 2 to 5 year horizon.

Inputs and Data

In the definition of the productivity, or the quantities in developing amine plan,
planners need to access and use data. Nowadays there is no shortage of potential
inputs of data into the planning process. Some straightforward rules regards input
data are:
1. Data sources and collection methods need to be understood for the resulting
data to be of any use. For example, using raw data from a mine production
recording system to assess operational delays, without understanding why and
how certain delay codes are recorded, may cause problems. Another example
is the use of historical cost reports for generating unit rates. Quite often these
are published from accounting systems in place, which may not be a true activity
based system.

Page 7

Runge Latin America Ltda :: Saturday, 09 de July de

2. Any data that is being considered for use also needs to be validated. This data
checking may be brief, if a high level of confidence is placed in the data, but it
always needs to be done. Production databases used to forecast future
equipment productivitys are typical of these. In turn this indicates that some
skills in trend analysis need to be present.
3. Data from geological databases needs to be checked, however it must be said
that this is less of a problem today than in years gone by.
4. Most critical data should be cross checked with other sources. This may be
tedious, but it needs to be done. Even if the check is done on a global reality
check level. For example do the hours logged in the production recording
system match the down hours in the maintenance history database?
5. Structures for holding data are important, and do change with the data type.
How the data is held inside the planning model depends in part on how it is to
be used and this needs to be understood. Recoveries for example may only be
known at a defined number of test points, on a curve and the historic process
plant reconciliation data. Depending on the ore characteristics and the process
plant design, it may be reasonable therefore to use lineal interpolation between
the known points. However hourly operating costs for equipment usually come
in ranges, which are step functions, rather than continuous lines between known
points.
6. Always confirm units used in measuring and reporting data.
7. The meaning of null values, zeroes and expected ranges of variables should be
clear. For example grade estimates that have -99 have a different significance
to those that have 0.
8. Understanding exactly what data is available for use in mine planning is
important. Traditionally mining engineers have had little knowledge or
experience with maintenance planning techniques or data sets. However it is
the maintenance planning that is key to short to medium term cost control in
most modern open pit mines. It is a continuing surprise that people separated
by 20 meters of floor space do not share information, only because they have
not been made aware that they are all part of the same process. As it is usually
mining engineers working as mine planners who do the strategic overall
planning, it is they who must go and find what data is available.

Example Planning System

XPAC has been developed to cater for this type of data validation, and for planning
the mining business in an integrated manner. This has been developed over 20
years, and is focussed on providing a systematic and integrated planning
environment. One of the fundamental concepts of XPAC is that there is a need to
develop a database for planning the mine. Typically this starts with physical

Page 8

Runge Latin America Ltda :: Saturday, 09 de July de

reserves, and then further attributes for each block are developed. A typical mine,
as modelled inside XPAC will have the following types of information;
Geological Reserves :: usually on a block basis. This block may be the same size
as the geological model blocks, or it may be a larger unit, such as a blast or a bench
or a stope. Or possibly a combination of the two. This is held in a hierarchical
structure allowing the user to quickly work with the data. These reserves may have
to be massaged into a as mined condition by the application of loss and dilution
criteria

if
not
applied
previously.
A similar structure is created to hold the dump destination data. It is becoming
increasingly obvious to planners that the development of the waste dumping
destinations needs to be included in a more fundamental manner in the planning
process. In this way it can be. As costs associated with waste placement are
usually a significant proportion of the mine costs, this is sensible.
Revenue Calculation basis :: This will usually be made up of recuperation data
and pricing information. The recuperation values typically come from one of the
following sources, a) imported on a block by block basis, b) defined using a
global constant c) derived using an equation or d) referenced from a lookup table
or curve. All of these can be used, and depends on the level and style of data.
Pricing data is usually a global constant for the plan, or time dependant where.
prices vary from period to period.
Equipment factors :: Usually used to derive productivities, and hence operating
hours required for a given block. Therefore when the block is placed in a position in
a planned sequence, the resources required to mine this block, the equipment
hours, explosives tonnes, are also known and accumulated.
Base Costing Data :: Costing can be developed using a number of techniques.
Generic costing may be applied, or unit costs applied to the blocks operating
requirements can be used. This way the true cost of the block can be derived. Its
location may affect costs (eg. deep benches, or distant pits and stopes), as well as
material characteristics (eg. oxide and sulphide ores have different processing
costs).
Working Hours assumptions :: Just when resources are available for these
activities needs to considered. In part this is dedicated equipment availabilities
and/or utilisations, however there may be other implications. For example, ore
mining may happen on day shift only, but the equipment used during the day on
ore, is used on waste at night. Or shaft and ore pass capacity is limited, although
there are enough production LHDs to handle requirements.
Physical access type mining rules :: To be mined, material must be physically
accessible. Any plan that can not guarantee that this rule(s) is honoured needs to
be doubted. Different rules may need to be applied in different places of the mine.
This is particularly so when coordinating between activities, such as development
and production in an underground mine.

Page 9

Runge Latin America Ltda :: Saturday, 09 de July de

Restrictions :: There will usually be reasons why we can not mine some portions
of the reserve, even though they are physically available. For example, fase 4 can
not start until 2005 because the HV electrical line needs to be moved prior to its
commencement. This will take 2 years to do. Similarly the equipment can pose
restrictions, such as I have to use only 20 trucks for the next two years, but after
that I can justify more if needed. Obviously blocks may be affected by such
restrictions today, but be released at some point in the future.
Objectives :: Of the blocks that are physically available, and without restriction,
then the best sequence to comply with stated objectives can be sought. These
objectives can take many forms, such as minimise cost or maximise return
(these are not the same), as well these can be more specific, such Silica grade
must be between X and Y or oxide % of feed is a maximum of 15%.
Output :: XPAC has a wide variety of numeric and graphical outputs available,.
The user configures these to their exact requirements.
In this way XPAC has data and information from the following groups; Geology,
Metallurgy, Marketing, Maintenance, operations, Accounts and senior management.
And so it becomes an integrated planning model.
Similarly the user thinks and defines the rules and restrictions, rather than being
concerned with pre-determining a sequence at the beginning of the planning cycle.
The data structures used inside XPAC to hold this data depend on the nature and
use of the data. There are curves, tables, and hierarchical database available
within the XPAC environment.

Benefits
There are a number of differences in using an automated planning process with
XPAC, when compared to a more traditional system. The key differences are;
1. Rules, restrictions and objectives are documented within the planning model.
2. Key data and variables are easily identified and audited.
3. Easier to determine business drivers (eg for use in sensitivity analysis, shadow
pricing).
4. Configuration - an order of magnitude harder. Users have to think and
understand the planning environment at a deeper level. This may be difficult the
first time, however most users agree that the process required delivers a wider
and deeper understanding of what the business is really about.
5. Operation - is an order of magnitude easier. Base data does not have to be
continually re-handled.

Page 10

Runge Latin America Ltda :: Saturday, 09 de July de

6. Running schedules - 2 to 5 orders of magnitude faster. The flow on from this

is; if the strategic model can be re-run easily, then intervals between the reassessment of the strategic plan can decrease. Hence strategic planning
becomes more than an annual event.
7. Successive iterations can be compared on a like with like basis very easily,
and incremental advance (or not) of successive plans measured.
8. Management criteria that changed at the last moment can be easily handled.
In Figure 1 illustrates the conceptual flow when using XPAC. In a traditional
systems there is no real way of defining the logic which dictates the rules and
restrictions. Therefore a new schedule means a new sequence being generated
(usually in a manual manner ), and that new sequence may or may not comply with
the required rules.

Figure 1 :: Automated Planning with XPAC Concept Flow

Set Corporate Objectives

Multi Discipline
Input
User defines
objectives
& trade offs

No
Solve
Schedule
Report
schedule

User defines
operational
rules

Feasible
Plan ?

Yes

Compare
to corporate
objectives

Meets
objectives ?
Yes

No
Communicate
Plan

Conclusions
Mining is a complex, costly business which operates using high volume units of
production. Therefore large increased benefits due to incremental improvements in
better planning are desired. Due to the complexity of mining, the strategic planning
can not solely focus on the extraction of material, although this is the initial stage.
So planning models that are integrated across disciplines and processes need to be
developed, understood and used. As this results in still higher complexity, users of
the model need to become focussed on understanding the requirements, rules and

Page 11

Runge Latin America Ltda :: Saturday, 09 de July de

restrictions of the various processes rather than just trying alternative sequences
in the hope that this will lead to success.
Iteration rather than optimisation is going to facilitate a successfully integrated
process which can be managed and driven by the mine planners.
Planning systems, such as XPAC, have been developed so that this can happen in
a straightforward and repeatable manner.

References
Runge, March 2002, Great XPACtations, p 1, Runge Pty Ltd Australia,
www.runge.com

Runge, Integrated Mine Planning, training course notes, Runge Pty Ltd Australia

Page 12

Runge Latin America Ltda :: Saturday, 09 de July de

Page 13