Anomaly detection in air conditioners using IoT technologies

Anomaly prediction and diagnosis of various machinery using IoT technology have been widely studied. Although air conditioners are commonly used in many facilities, few studies have focused on anomaly diagnosis, especially for small and medium-sized air conditioners. As the installation of necessary IoT sensors is not easy for such air conditioners, the design of the anomaly diagnosis and detection system is not straightforward. In this study, we propose an approach for installing a data collection and diagnosis system for existing air conditioners. Some of the challenges faced in detecting and diagnosing anomalies in air conditioners include noisy environments, cost constraints, and seasonal changes of targets. Therefore, careful coordination of the data collection system and the diagnosis method, and the seasonal tuning of the diagnosis model could realize an effective and inexpensive anomaly detection system for air conditioners.