Evaluation of Electromechanical Failure of Nodal Systems and Components in Wind Power Generation Systems by Neural Network

The probe is specifically targeting 13 wind power plants in Bulgaria, all of which have an installed capacity that above 9 MW. After observing for a period of five years, a sample was developed that includes the most commonly damaged elements and systems. The defining characteristic of these failures is the percentage at which normal operation ceases. To assess dependability, the outcomes of the observations are employed as input data for a neural network. The importance of each individual measure was evaluated using the self-organizing map technique. The weighting parameters were used to prioritize the individual failures.