Journal of Korean Institute of Architectural Sustainable Environment and Building SystemsISSN:1976-6483(Print) 2586-0666(Online)한국건축친환경설비학회

This study proposes performance improvement methods of fault detection of AHU systems with measurement of variable importance and comparison for three fault models based on ASHRAE 1312-RP experimental dataset. The results show that input variables were determined differently for each of the representative three fault types. The fault detection performance be improved more than 29% compared to base input condition using major input variables which measured importance of system variables. Also, the impact of variables and models be differed for each type of fault. The results of this study contribute to failure protection of HVAC systems and exploration of fault location in advance. It is necessary to verify the generalization performance of the proposed model in actual system conditions.