Fraunhofer Institute for Solar Energy Systems ISE
In many buildings, systems are not running optimally due to incorrect operating times, faulty sensors, defective actuators and inefficient controls. We use digital technologies to proactively improve the condition and efficiency of systems and reduce costs and CO2. Facility managers benefit from targeted maintenance based on the actual condition of the systems.
COMETH framework for fault detection and diagnostics
In order to monitor and predict the energy efficiency of buildings and systems, we need to evaluate numerous signals over time and their interdependence. For optimal monitoring, we have developed the automated fault diagnosis system COMETH Rules and COMETH AI, which is based on both rules and AI. An important component of the tool is a combined AI method with an innovative feedback mechanism that enables the algorithm to learn adaptively and recognize new, previously unknown faults. The program can also identify data points and system topologies, estimate the effects of errors and visualize data points live.
AI-based optimization of technical building systems
We develop advanced, data-driven, adaptive control methods for building technology that learn system models from data and optimize operating strategies without manual calibration. Our approaches outperform conventional controllers in terms of performance and flexibility. Examples include automated heating curves and thermostat controls as well as integrated thermal management systems. Automated forecasts, e.g. of energy consumption, reveal operating deficits and promote load management measures.