Powerspex Instrumentation B.V., Powerspex Simulator GmbH and Landscape AI are launching a study on an innovative solution for managing complex district heating networks. This is a project subsidized by the province of Overijssel. The aim: a framework that will be used by numerous district heating networks.
"The generation and distribution of district heating is currently facing major changes," Peter Lasch on behalf of Powerspex Simulator GmbH explains the importance of the project. "Operators are faced with extensive conversions and the integration of new technologies. Think of geothermal, solar thermal, river water heat pumps, waste heat from data centers, heat storage and other decentralized technologies in their networks. The control of these networks is becoming increasingly complex and is often no longer manageable with conventional control methods."
Collaboration three innovative companies
It resulted in a constructive collaboration between the companies Powerspex Instrumentation B.V., Powerspex Simulator GmbH and AI specialist Landscape. They are joining forces to create a solution for an innovative approach to district heating networks.
For this purpose, they combine the core activities of the three companies: process automation, dynamic simulations and artificial intelligence.
Simulation trained by AI model
In het door de provincie Overijssel gesubsidieerde project gaan de drie bedrijven het komende jaar onderzoek doen naar een innovatieve oplossing voor het beheer van complexe stadsverwarmingsnetwerken.
Daartoe maken ze een simulatie van het stadsverwarmingsnetwerk en trainen ze met behulp van zogenaamd Reinforcement Learning een AI-model dat de kenmerken van het netwerk, de producenten en verbruikers van het netwerk en de economische en ecologische randvoorwaarden leert kennen en deze volgens vooraf gedefinieerde criteria optimaliseert.
Lasch said, "The long-term goal of this research project is to obtain a framework that can be transferred to and used by numerous district heating networks. By linking a simulation to an AI model, it can be continuously updated as adjustments and changes are made to the network without having to fall back on historical data."
Funding MIT R&D
De projectaanvraag werd ingediend als onderdeel van het MIT R&D financieringsprogramma van de provincie Overijssel en werd beloond met 83,5 van de 100 punten. De aanvraag is begeleid door Ward Meijers en Lars Smit van PNO Consultants en wordt ondersteund door verschillende partijen waaronder Warmtestad Groningen.
Lasch concludes, "We would like to thank our supporters very much for their support and interest. We all strive for an innovative solution that district heating networks can benefit from in the future."