Rail & Bulk Supply Chain Modelling

Clients bring questions ranging from the capacity of an existing operation to the configuration of a proposed supply chain. Some studies investigate infrastructure or fleet investment; others support recurring capacity reviews, annual planning or maintenance programmes.

Scoping begins with client discussions led by InterDynamics, drawing on our experience to examine how the operation works and what the study needs to establish. I generally worked as the sole simulation developer, often alongside a general manager responsible for project management. My role commonly extended from scoping and design reviews through to delivery, training and technical support.

The required outputs guide the level of detail. Data availability also sets a practical limit: a detailed representation needs evidence to support its assumptions. Model design establishes the data requirements, but the available data can in turn change what is reasonable to model.

Across these projects, I developed reusable simulation components and the system used to maintain and place them within a model. The main benefit is faster network construction, alongside consistent behaviour between models and less effort when the scope changes.

The Coles Myer project in 2004 was my first supply-chain model built entirely from input tables driving generic network components. The network could be regenerated as designs changed, allowing the same modelling logic to represent different configurations.

CBG subsequently used an early reusable rail-location component. By the Bowen Rail project in 2023, reusable rail-location components were used throughout the mine-to-port model, with their configuration defined by internal tables. This approach has made initial proof-of-concept models quicker to build.

Internal model construction and client configuration are separate decisions. Some models expose the complete network definition to users; others provide controls for agreed operating and infrastructure alternatives, with further structural changes remaining development work.

Existing operations provide the strongest starting point for checking model logic. Cycle times and achieved capacity can be compared with actual performance before introducing proposed changes. For Coles Myer’s redesigned network, the configurable model could first represent the existing network using the same underlying logic.

Where practical, model outputs follow the data structures clients already use. That makes comparison with operational records more direct. Charts provide an overview, while train graphs are particularly useful to rail clients reviewing modelled movements.

Reviews often expose mismatches or incomplete assumptions in the input data. A mine cycle that appears too short may indicate a missing process, or time that belongs within an existing pre-load or post-load activity. The task is to identify what explains the difference and represent it at an appropriate level of detail.

Delivery depends on how the model will be used. For some assignments, I operate it within the project team and provide analysis and reporting. For others, clients receive a model they can configure and run themselves, supported by training and subsequent technical assistance.

Deciding what clients can change is part of the model design. Clients need access to the assumptions and alternatives relevant to their planning work, while changes beyond that boundary require the model to be extended. Continued support allows the simulation to develop alongside new operating questions and project scope.