AAOM Handbook

• random variation in the performance of each process that occurs over time, and • effectiveness of buffering provided by stockpiles. The only way to successfully predict the outcomes of this variation occurring within processes is with an appropriate statistical performance model. A statistical performance model uses Monte Carlo simulation to replicate the impact of random variation occurring within the system being modelled. In setting up to use a statistical model we must realise that the variation seen in the measured performance of a system may not all be random. There are two types of non-random variation that may be present. Firstly, the total variation may be comprised of both independent variation (that arising from factors within a process) and dependent variation (that arising from the influence of a preceding or proceeding process) - note that the purpose of storage is actually to isolate variation in one process from affecting other connected processes. The dependent variation cannot be treated as random in the modelling process and hence needs to be removed (cleansed) from the data fed into the model. Secondly, the remaining variation may still have periods where it was not random, but had data that was linked to a specific, non-random cause – e.g. a process shut-down for maintenance. These events can be identified from a Control Chart of the cleansed data, where they should appear as 'special causes'. These non-random events will need to be treated as a separate input to the model. The statistical performance model, if correctly designed and calibrated, and fed with the cleansed data, should make reliable predictions about the both the dependent variation that will occur, and the overall system performance (see examples in Appendix 1). This performance model can help to identify; • what level of performance is required from each element of the Business Structure if overall Performance Expectations are to be met, • what and where the constraints on performance are - either a capacity 'bottleneck' at a point in the process, excessive variation constraining performance across sections of the process, or ineffective buffering by stockpiles, and hence • what and where the opportunities for performance improvement are. The third area of performance for which targets may need to be set is Sustainability - a measure of the availability of key 'resources' required for a process to continue operating. There may be tree types of key resources: • Physical materials such as ore reserves, in-process stocks, fuel supply, etc. that can be assessed via a measure of stocks divided by consumption rate. This yields a measure of 'time' such as years of reserves, or days of stocks or fuel. • Physical assets such as steelwork, concrete and electrical systems for which the life can be depleted if appropriate sustaining (integrity) service strategies are not implemented. This situation will reduce the functional life of the assets and can be measured via an increase the threat of failure over time.

© McAlear Management Consultants 2006

Operational Planning: Set Performance Targets

Updated: August 2018

Page 6