RELAI

Important results from the project

Sigholm is active in the energy and infrastructure area and develops optimization solutions for the Nordic combined heat and power industry. Where advanced algorithms are central. The project aim is to start Sigholms AI-journey by improving these algorithms using AI/ML. The goal was to increase the AI competence at Sigholm by developing improved algorithms that could be implemented in Sigholm´s existing softwares to create good economic and environmental profits for Sigholms customers.

Expected long term effects

The project has provided a broader competence within AI / ML, where the company has been introduced to best practices for successful projects. Furthermore, the core group has gained knowledge and practical experience of applying non-linear AI models for time series forecasts, with increased insights of which AI algorithms are robust and most suitable for integration into Aurora By Sigholm, and also have the potential to provide more cost-effective and energy-efficient roadmaps for our customers.

Approach and implementation

The project was carried out in two parts. As a start, experts from Neurolearn and Chalmers gave a broad introduction to AI / ML for the entire company. Furtheron, the core group from Sigholm iteratively tested different AI models to predict robust electricity price forecasts linked to production planning for heat and powerplants, and evaluated the results and methods together with the experts. Furthermore, architecture for implementation and maintenance of these in cloud services was examined.