Journal of Korean Institute of Architectural Sustainable Environment and Building Systems ISSN:1976-6483(Print) 2586-0666(Online) 한국건축친환경설비학회논문집

This study proposes an analysis method to determine the performance of physical sensors required to achieve the target performance of the virtual sensor-based fault detection in building energy systems. The proposed method was applied through the building energy simulation with fault modeling. Simulation was performed using the Designbuilder and EnergyPlus tools. The simulated datasets consist of physical sensors and energy consumptions for the normal and faulty system conditions. The virtual sensor-based fault detection method was constructed for fouling faults in the target system, using the simulated datasets. Then, the required performance of the key physical sensors was analyzed by the proposed method under the different levels of random errors of temperature sensors in the virtual sensor-based fault detection of the chiller fouling. According to the case study, the accuracy of physical sensors has a great impact on the fault detection performance. In order to achieve the target performance of 90% detection for the fouling level of below 10%, the physical temperature sensors should have accuracy within ±0.5°C (random errors). This study shows the proposed method can be effectively used in the fault detection design stage, thus proving reliable fault detection performance under the physical sensor uncertainties in real operational systems.