Fuzzy Neural Inference System for Pattern Recognition of Power Quality Events Using Rule Generation

This paper presents pattern recognition of time series data and subsequent temporal data mining of power signal disturbance events that occur frequently in power distribution networks using multiresolution S-transform and Fuzzy neural inference system . This system yields relevant features, which are used in a Fuzzy expert system to separate the transient time series data and steady state short-term duration time series data including various harmonic time series. The transient time series data is then passed through the Fuzzy MLP along with inference system to yield a set of rules required for recognition of various transient disturbance patterns (power quality events).