The reliance of data-driven fault detection and diagnosis (FDD) approaches on well-labeled datasets poses challenges in real-world applications, where such data may not be readily available. Our work explores domain adaptation as a solution, enabling the use of labeled data from a source domain (e.g., a laboratory testbed) to diagnose faults in an unlabeled target domain (e.g., a real building system). We adapt the contrastive adaptation network (CAN) algorithm, originally developed for image classification, to address this challenge in HVAC applications. The paper details how to transform time-series data into image-like inputs and presents experimental results on air handling unit (AHU) datasets. 

Congratulations to Naghmeh Ghalamsiah on this impressive publication. Special thanks to Dr. Jin Wen for her guidance, and to co-authors Dr. Teressa Wu, Dr. Selcuk Candan, Dr. Zheng O’Neill, and Asra Aghaei for their valuable contributions. 

Read the full paper here: https://authors.elsevier.com/c/1kqmi1M7zHK1Dt