A neuro-fuzzy based system for fault detection and diagnosis of 3-phase PFC rectifier

Fault diagnosis in PFC rectifiers is becoming more and more important in many industrial applications. PFC rectifiers allow high power factor operation in AC/DC conversion. This power converter is normally the first interface between the ac power system network and the electronic equipment. In this way a good diagnosis system can avoid unplanned standstill. Under this context, this work presents a method to detect and identify the open transistor circuit fault. The developed method has two major steps. First, the input line currents are converted into a new pattern representation. Second, a neuro-fuzzy algorithm will be used to identify the fault. Several simulation and experimental results are presented to show the effectiveness of the proposed approach.