Broadband excitation signal techniques for electric machine diagnostics

Three-phase induction machines are widely used in industry and they are known as reliable and cheap components of drive systems. However, even if these components are reliable, they can be submitted to external stresses coming from the system and degradation can occur even in the electrical and/or mechanical parts. Fault detection in induction machines has been investigated largely for the last twenty years with success and progress is still being made. The objective of this study is to present a new method for early detection of anomalies, in particular broken bars and porosities which can occur in the fabrication process. The method presented is general and may be applied to the induction machine whilst in rotation or at standstill. It´s based on broadband excitation signals which are able to excite the faulty modes of devices under tests. Experimental measurements are made both on healthy and faulty machines and the comparison gives a difference in the flux and current machine signatures which is basically the most commonly used method of fault detection. This fact is very important for industrial suppliers of induction machines, because in this way manufactures will have a way of early detection of defects before selling machines.