Experimental investigation and measurement of the bearing current in a multilevel inverter-fed induction motor drive

The development of Space Vector Modulation (SVM) based AC adjustable speed drives (ASD´s) has increased the inverter output AC voltage, performance, efficiency, and controllability in speed controlled induction motor applications. However due to fast switching of the MOSFETs/IGBT´s, introduces the high frequency transients at the output of the inverter. In addition, due to non-sinusoidal voltage output at the inverter, there exists the voltage at the star point of the Induction Motor (IM) with respect to the common ground (gnd). Due to this, the rotor shaft voltage is induced, resulting in bearing current which is harmful to the bearing. The grease film in the bearing acts as a capacitor and due to the voltage transitions in the star point of the stator winding it gets charged. The breakdown of the film causes transient current that induces electrical discharge machining (EDM) and due to that pitting occurs in the inner surface of the bearing, thereby reducing the bearing life. Significant efforts were made by the researchers in this direction to minimize the problem. This paper presents the experimental investigation and measurement of bearing current for the modified 3-phase squirrel cage IM connected to a 3-level SVM based inverter bridge circuit. PIC Microcontroller and other associated electronic interface circuits control the inverter bridge circuit. Standard current probe, LISN and a high frequency 4-channel Mixed Signal Oscilloscope (MSO) and digital storage oscilloscope (DSO) with differential probes were used to measure the bearing current in terms of voltage from the current probe and other parameters. For comparing the results with Federal Communications Commission (FCC) and Special Committee on Radio Interference (CISPR) standards the graphs are plotted with Frequency vs dBμV and the bearing current in dBμA for the 3-level inverter (SVM method) using signal analysis software. For conclusion the SVM method results are compared with SPWM- - method published results.