A method of bad data identification based on wavelet analysis in power system

Historic load data are so distorted by kind of influential factors that results in the false analyzed results of the EMS and DMS advanced application software. In fact, bad data are regarded as singularity points or anomalous sharp parts in load data curve. Discrete dyadic wavelet transform can be used to detect positions and characters of the local singularity points in the noisy surroundings. In this paper a method based on the wavelet singularity detection and the wavelet de-noising scheme is presented for bad data identification in power system. It uses modulus maxima value to identify the local singularity of signal, and its process is simpler than complicated bad data identification of state estimation. The validity of the algorithm is proved by real data analysis.