Soft fault detection strategies in heat pumps: A case study investigating virtual sensor methodologies for evaporator fouling

Abstract

Heat pumps, known for their energy efficiency, can experience soft faults, such as evaporator fouling, that degrade performance while allowing continued operation, making detection challenging. This study presents a method for detecting evaporator fouling in air-to-water heat pumps using virtual sensors. These sensors estimate key parameters, including condensing and evaporating temperatures, refrigerant mass flow rate, and air enthalpy at the evaporator outlet. The approach relies only on variables typically measured in the system – such as those from the compressor, fan control, and secondary fluids (air and water) – eliminating the need for direct refrigerant measurements. Virtual sensors are calibrated against simulated fault-free performance maps and tested on a heat pump with variable-speed components. The method’s effectiveness is assessed through simulations and experiments with induced evaporator fouling. Simulations show high accuracy, with deviations within ±5%; however, experimental results exhibit higher deviations due to challenges in measuring air dry-bulb temperature at the evaporator outlet, where fouling leads to non-uniform airflow. Despite these limitations, virtual sensors offer a cost-effective and practical alternative to traditional physical sensors for fault detection.

Keywords

• Heat pumps
• Fault detection and diagnosis
• Virtual sensors
• Evaporator fouling
• Soft faults