Predicting the Class of Salt Contamination for Wind Turbine Blade under a Lightning Strike Using Fuzzy Inference System and Probabilistic Neural Network

This study predicts the class of salt contamination for wind turbine blades that undergo a lightning strike using a fuzzy inference system (FIS) and a probabilistic neural network (PNN). The experiment uses four classes of equivalent salt deposit density to determine the salt contamination status for a wind turbine blade. The greatest degree of salt contamination is 338.9 µg/cm². The rotational speeds of the blades are set as 0, 50, 100, and 150 rpm. Many lightning strike surge experiments are conducted for different types of salt-contaminated blades. The lightning current signals in the grounding line are measured using a high-frequency current transformer. This study uses an experimental method to determine the effect of salt contamination on wind turbine blades during lightning strikes. In terms of prediction, FIS and PNN are used to predict the four classes of salt contamination. The peak absolute value and the average power of the measured signals are the input datasets for the FIS and the PNN. The predictions for the ESDD classes are 94.8% for a FIS and 88.8% accurate for a PNN. The research results show that the proposed method accurately predicts the salt contamination class for wind turbine blades.