Anomaly detection in heat pumps: experimental setup, testing, and data analysis

This thesis investigates anomaly detection in heat pump systems, with a focus on evaporator fouling, to enhance reliability and efficiency. Conducted at Carel Industries S.p.A., the research combines experimental setups, testing, and data analysis to examine the effects of fouling on heat pump performance. The study begins with a comprehensive literature review, highlighting the current state of research in anomaly detection techniques, including machine learning approaches and numerical simulations. Following this, the fundamental principles and critical components of heat pumps are detailed, such as the compressor, condenser, expansion device, and evaporator. The methodology involves a carefully designed experimental setup to simulate various fouling conditions. Tests were conducted under different ambient temperatures and load conditions, using both poorly-distributed and well-distributed coverings to mimic real-world scenarios. Data collected from these experiments were analyzed to understand the impact of fouling on the heat pump’s performance, demonstrating significant reductions in efficiency and increased energy consumption. The findings underscore the importance of continuous monitoring and timely maintenance to detect anomalies early. By implementing real-time data analysis and exergy analysis, the study effectively identifies and characterizes faults, providing a proactive approach to maintenance. This approach can reduce downtime, lower operational costs, and extend the lifespan of heat pump systems. Key findings include the detrimental impact of both poorly-distributed and well-distributed fouling on heat pump efficiency, and the effectiveness of real-time data and exergy analysis in fault detection. The research contributes to improving HVAC technologies for better energy management and sustainability in heat pump operations.