D5.2 Tool for Predictive Maintenance and outputs for Digital Twin

The purpose of this report is to present the predictive maintenance tool called iPREDICT, that is one of the software technologies that is brought into the SPHERE ecosystem. The tool follows a data-driven approach and it is based on AI and machine learning. The use of predictive maintenance instead of conductive or preventive maintenance enables to minimize downtimes due to unexpected breakdowns and reduce energy waste due to malfunctioning or misuse of equipment. Initially designed for heating and air conditioning and in water heating systems, it can be extended to other equipment or building elements and target structural instability, leakage and other safety risks that may derive in unforeseen construction work and use of resources, energy among them.