Advance Fuzzy Neural Network for the Detection and Diagnosis of Faults in the VAV Systems

Abstract

Variable air volume (VAV) fault detection and diagnosis is essential for energy consumption and stable operation of the Air Handling Units (AHU). This article introduces the Advance Fuzzy Neural Network (AFNN) model for detecting and diagnosing of the VAV systems. As the initial step, fault id detected based on Fuzzy Logic (FL) method, then the type of the faults is identified based on Neural Network (NN) classification approach. Our proposed model is tested with simulated data. It also tested on six different VAV systems from two levels of the Land Custody and Development (LCDA) building in Sarawak, Malaysia. The results show that proposed AFNN model can detect more faults and accurately classify faults from different size of VAV systems.