Integrated planning is a key part of the Transformation journey that ArcelorMittal Mining Canada has undertaken. Having the ability to accurately and reliably forecast production and material movements throughout the value chain is an important requirement to achieve our business plan.

It involves planning at the different areas of the value chain while taking into consideration historical performance and gaps to target.

To reliably plan the output of the value chain, the variability of every individual process has to be accounted for. This variability makes daily stock forecasting unreliable over long periods of time, having downstream impacts on shipping and can result in high demurrage costs, which are fees payable to the chartered ship owner if  ships are not loaded or discharged within the time agreed. Sales’ adjustments cannot be done in the short term and require a 2+ month line-of-sight to adjust the vessel lineup.

A vessel lineup aligned with inventory availability has the potential to identify opportunities or extra sales and minimise demurrage costs. Order of magnitude of the impact of this CAPEX-free project is measured in million $ per year, that is a ten-fold return of investment in the first year.

"A timely accurate forecast is the first step towards a completely integrated plan."

Ann Marie Rushworth, director of logistics

To mitigate the uncertainty along the mine value chain, ArcelorMittal Mines Canada (Port Management and Integrated Remote Operations Center teams) and Global R&D (KiN team) launched a collaboration to develop Ironman: a mathematical model to forecast inventory evolution, considering transports and process uncertainties, and then estimate plan fulfillment and potential stock ruptures.  

Ironman learns stochastic models from past data to project the current production and loading plan into the future, assessing the potential outcome scenarios and their likelihood, and summarizes these projections into probabilistic KPIs that allow the identification of risks and opportunities related to inventory levels, vessel and port scheduling, and operations of the concentrator and pellet plant. This provides valuable and actionable information e.g. to reschedule vessels that are likely to wait for loading or to offer additional material for sale if excess is expected, as well as shining the spotlight on the potential causes of the problems.

Ironman has been validated, although key performance indicators are being analyzed, which could lead to minor adaptations of the model.