Frequency Response Analysis Using High Frequency Transformer Model

Abstract

The concept of Frequency Response Analysis (FRA) has been successfully used as a diagnostic technique to detect winding deformation, core and clamping structure for power transformers. The main problem about FRA techniques is to interpret the observed evolution of the frequency response in order to identify failures. Transformer high frequency computer modeling is proposed in this work to be used with the practical FRA measurements. The physical meaning of the model parameters allows the identification of the problem inside the transformer. Two high frequency transformer models based on lumped and distributed parameters approaches are investigated. A comparison of both models is conducted using their transfer function plots, and hence based on the amount of information revealed from the plots, a distributed model is chosen for further analysis. The model validation is carried out through the comparison of the simulation and field results. Mechanical and short circuit faults are simulated into the model to compare the differences in the frequency response of healthy and deformed transformer signatures. The advantage of this technique is that the FRA measurements can be obtained from meaningful parameters that can aid interpretation and classification of FRA signatures.