Proposal of fault detection and diagnosis system architecture for residential air conditioners based on the Internet of Things

Abstract

This paper aims to present a proposal of architecture for fault detection and diagnosis (FDD) in residential air conditioner systems using the internet of things and being able to detect performance degradation faults such as refrigerant leaks and filter fouling. Hardware for data collection and upload was developed, which takes information from an entire operating air conditioner system and sends it to a cloud environment. These data are used to model the behavior of fault-free equipment. The model is fed with working data and compares expected and provided data to create residuals. When the expected behavior differs from the modeled data, the system indicates that a fault is present. Depending on the residual pattern, a specific failure is revealed, and the user is notified on a mobile phone.