Optiflex is an AI-based tool reducing energy costs of industrial processes. The tool aims at moving the energy demand of the process to time slots where prices are low and/or local renewable production is high. Not only does synchronising energy demand and local production reduce energy costs, but it also helps the grid by reducing injection (and thus load). This is done by taking into account all constraints of the underlying process, including temperature level for fridges, water volume for reservoirs, minimal state of charge for a battery (among others). To this end, the tool is able to forecast market prices and all variables related to the process’ constraints. Using these forecasts, an optimisation process is used to compute the optimal commands schedule for all controllable assets in the process. Finally, it is important to note that both the schedule and forecasts are continuously updated in real-time to always compute control commands in light of all available information.

Many different industrial processes can benefit from such an approach. As such, the variability of business processes to model can be daunting for the solution. This variability brings many challenges. These will be the focus of the talk, as well as the solutions Haulogy has implemented to tackle them. In particular, the talk will focus on the node-based modelling spirit of the Optiflex tool and on the many advantages it brings for delivering forecasting and optimisation algorithms in production environments.