A STUDY OF WAVELET ENTROPY THEORYAND ITS APPLICATION IN ELECTRIC POWER SYSTEM FAULT DETECTION

The wavelet transform result data of electrical transient signals were abundant, and there was no quantitative method, so to mine and fuse one or a series of universal applicable quantities to detect system fault and stability was essential. Combining wavelet analysis with entropy theory by exploiting the time-frequency localization ability of wavelet analysis and the ability of entropy to token system state, three wavelet entropy concepts, i.e. wavelet energy entropy(WEE), wavelet time entropy(WTE), wavelet singularity entropy(WSE) were defined, and corresponding algorithms were put forward. The mechanism that these three wavelet entropies can token system fault were disclosed. Simulation results of two theoretic signals and transmission line PSCAD/EMTDC simulation signals indicate that wavelet entropy can reflect the system change, and they can eliminate the disturbance of noise and be applied to power system fault detection.