RESEARCH ENERGY CONSUMPTION REDUCTION THROUGH ARTIFICIAL INTELLIGENCE (AI)
Armac Industrial Automation (Armac) and Radboud University, partner Think East Netherlands, are collaborating on an AI research project to examine energy usage by residents of a defined area north of Amsterdam. The goal is to come up with an algorithm for energy usage through the use of artificial intelligence (AI).
The cooperative venture between Amrac and Radboud University has been in place since the middle of 2020. Together with the cooperation of a district heating company they will analyse the energy usage in the earlier-mentioned area. Through this method, they hope to create a predictive algorithm. The resulting data should allow the possibility to predict future heating patterns from local power plants, predict energy usage, and more efficiently regulate future energy streams so that the plants aren’t generating heat when there’s no demand. Armac has already been working for 10 years with this district heating company. They were trend-setters in that they implemented a cloud-based system to monitor energy usage. Over the years, Armac has become increasingly specialised in artificial intelligence. For example, they’ve learned how to process stored data into useful applications. Armac’s software engineers have been learning how to use AI to make these processes even more effective and efficient.
The venture that includes Radboud University goes even further. In this project, data is analyzed, which helps Armac develop even more into a knowledge partner. The district heating company gains insight into its systems and how to improve them. The goal is to deliver energy at lower costs, more efficiently, and with better overall quality.
The first results are already being measured and show that data analysis results in large efficiency gains. One result is a 20% reduction in losses. A complete model to make city warming more efficient through the application of AI is being planned for a later phase of the project. This should be achieved over the coming two years.
One of the next steps is to integrate external data, which should help improve the predictive qualities of the algorithm.