PREMA-HEAPS – Predictive Maintenance and Fault Detection for Efficient Heat Pumps

This proposal is for funding of a Swedish contribution to IEA HPT Annex on Digital Services for HPs through a joint project with RISE, CLIMACHECK, and Halmstad University (HH). The project aims to develop advanced predictive maintenance and fault detection capabilities for commercial HP systems.

The main goals are to implement machine learning algorithms to detect sub-optimal operation, predict failures, and perform automated root cause analysis. The project will also investigate measurement uncertainty in HP data and its impact on fault detection algorithms. The methodology involves conducting state-of-the-art analysis, developing AI/ML algorithms, validating them through real-world data, quantifying measurement uncertainty, and providing guidelines to minimize uncertainty. Expected outcomes include innovative predictive maintenance methods, fault detection algorithms, measurement uncertainty, and industry adoption facilitated through knowledge dissemination activities.