Stimulating and detecting changes in the insulation health state of inverter fed ac machines

In modern traction drive applications condition monitoring is getting more and more important. Insulation health state monitoring can be done by evaluating the high frequency current response to voltage step excitation by the signal analyses of the current sensors, which are also used for the control of the machine. This reduces the costs and number of necessary system components. A change in the current response is, amongst others, correlated to a change through insulation degradation. The dominant influencing parameters of the high frequency current response of the machine are parasitic winding capacitances as well as insulation resistances, which are slowly changing over time if the insulation is deteriorated through different stresses, e.g. thermal or electrical. These slow changes can be monitored over long time period and give evidence if additional maintenance steps are required. Thus an unforeseeable outage of the traction drive system caused by damage of the machine can be prevented. Before the insulation state can be evaluated certain signal processing steps are necessary to increase the sensitivity. In this work the method of stimulation as well as detection of changes in the dominant parameters is verified on a 1.4MW induction machine.