A Fault Diagnosis Method in BLDC Motor Drive Systems Using Moving Average Filter for Back Electromotive Force Signal Processing

BLDC has the advantage of obtaining good performance at a low unit price without a brush. A typical BLDC generates a field flux using a permanent magnet. However, the magnet has a disadvantage that the properties of the magnet may deteriorate when exposed to excessive stress. The deterioration of the magnet may cause an accident due to a decrease in the characteristics of the BLDC or an overcurrent and a control failure. Therefore, it is necessary to implement a failure diagnosis technique of the BLDC. In this paper, we present a fault diagnosis technique using the non-exciting phase voltage of BLDC. The control algorithm of BLDC is based on the 120-degree current flow method, so a non-excited section occurs even when driving at a phase voltage. It is possible to monitor the state of the magnet by measuring the back-electromotive force due to the bridging magnetic flux of the motor generated in the non-excitation section. In this paper, a fault diagnosis technique using a non-excited phase voltage is verified through simulation and experiment, and a signal processing technique and a diagnosis algorithm are presented to implement it.