Capacity / Utilization Predictive Models

For an organization whose business is dependent on meeting demands for service, the key elements in any capacity and utilisation modelling are the:

• Consequences of failing to meet these capacity requirements and the business cases needed to evaluate their impact
• Accuracy of the growth in demand scenarios
• Lead times required to ensure that the augmentation is available at the time desired, taking into account the approval process and the construction timeframe.

The relationship between customer expectations and the organization's capability to deliver the service must be well defined and communicated. Strategies in service provision depend on trade-offs between:

• Affordability (is the customer willing to pay the real cost?)
• Expectations (what the community wants or is willing to accept).

Customers will be satisfied if they are provided with a level of service that they expect, at a cost that they can afford to pay.

Specialised modelling applications that quantify and balance these factors need to be employed to ensure that the demand for service is matched to current and future demands. Some of the functional considerations are:

• Inputs that clearly indicate the need for the asset
• Utilisation being made of the assets
• Quality of service required, eg uninterrupted
• Pattern of demand for services - may be seasonal or fluctuate during the day, as in the case of electricity demand.
• Operational scenario to accommodate demand, eg re-route, load share or operate the asset in an alternative mode to provide additional capacity
• Environmental issues, eg noise level restrictions during the night.

Capacity modelling of networked infrastructure is highly sophisticated, with many applications meeting the needs of different asset groups, including water supply reticulation and drainage modelling systems.

For mature networked assets, the modelling requirements are split into:

• Macro models: deal with the key elements of the system or the main arteries or distribution systems which will require regular analysis and checks to be made to ensure that the macro system capacity and planned augmentation will meet the demands for growth etc.
• Micro Systems: desirable for the micro distribution or networked reticulation assets. These models do not tend to require as regular an update, as the systems tend to be more stable with less growth following their initial utilisation phase. In later times these models will be used to ensure that appropriate demands are made on the macro model system and to accommodate any significant changes that may occur in the area, due to dual occupancy dwellings, shopping centre redevelopments etc.
• Medium models: for large service authorities medium system elements may need to be modelled separately, as a break between the major macro distribution systems and the actual reticulation or minor (micro) systems.

These models are shown below:

Functionality

For all capacity models the inputs are very similar and are:

• The physical characteristics of the assets
• The demand model characteristics
• Specialist modelling parameters.

The physical asset characteristics fall into two categories:

• Static characteristics involving the type, size, material type
• Dynamic characteristics such as the current condition, level of service or other indicator necessary to model the system, i.e. condition or friction coefficient for water mains.

All of these parameters can generally be found in asset registers or associated condition modules.

The demand characteristics or utilisation trends will generally be available in system capacity monitoring models or SCADA systems. This information needs to be fed into the capacity analysis model separately, to suit the model's input requirements.

Other modelling parameters, such as operational requirements and unusual physical characteristics for complex facility elements, will generally need to be fed in manually as a separate input. These key inputs are shown in the following figure.

If the electronic asset register is set up appropriately, then all physical and dynamic data should be automatically transferred from the asset register to the modelling application.

In some cases the linkages or nodes can be recorded in a GIS overlay and the necessary data transfer software can download the appropriate physical parameters to the model.

These factors can significantly effect the overall cost of setting up, updating and running the models. The organization should make sure that these factors are taken into account in their system design.