II. UNDERLYING PRINCIPLE
Signals, whose frequency contents do not change in time, are called stationary signals. All frequency components that exist in the stationary signal exist throughout the entire dura-tion of the signal. Nonetheless, if the frequency components change with time in the signal, then it is called nonstationary signal. One of the downfalls of the FFT is that it does not give any information on the time at which a frequency component occurs. This is not a problem for stationary signals but does leave room for improvement when nonstationary signals are involved.
Voltage and current waveforms in the transformer are treated as nonstationary signals in the transformer. The pro-duction of particular frequencies depends on the physical pa-rameters and conditions of the system, which may change in time. It is essential to be able to detect these changes. There are many tools that can consider time frequency analysis, and among them, Short Time Fourier Transform (STFT) is pop-ular and easy to implement. Time-frequency analysis of nonstationary signals indicates the time instants at which different frequency components of the signal come into calcu-lation. One direct consequences of a treatment will be the pos-sibility to accurately locate in time all sudden changes in the signal and estimate their frequency components.
The sensitivity of fault detection can be significantly im-proved by using the STFT method for the evaluation of im-pulse tests on transformers. Relative changes will help distin-guish the various types of failure and also provide an indica-tion for test repeatability. In the previous literature, the re-search aimed to quantify the relative change in amplitude (DA) and in resonant frequency location (Df). The relative changes are computed as: