MODAL IDENTIFICATION OF POWER TRANSFORMER WINDING FOR FAULT DIAGNOSIS

摘要

The mechanical damage in and change in vibration modal parameters of power transformer windings are related. Thus such change could be used as failure indicator. Practically, the challenging task is to extract the changes of the modal parameters of windings from transformer tank vibration. In this paper, the Ibrahim Time Domain (ITD) method is utilized to identify windings' modal parameters based on the transient vibration of the tank structure excited by de-energize events. The analysis of winding vibration shows that the modal frequencies identified by the ITD method agree with the results obtains from the impact test. To further test the modal identification method on non-invasively winding modal parameter extraction from the vibration of transformer tank, one common winding fault, looseness of clamping force, was also introduced into the transformer. The experimental results demonstrate that ITD method is able to track the frequency shift of certain winding modes due to the fault based on tank free-decay vibration and appropriate model order. This paper gives some evidence that a combined method with investigation of modal frequencies of failed winding and ITD method can provide is a powerful tool for winding condition monitoring and fault diagnosis.