WIND TURBINE FAULT MODELING AND CLASSIFICATION USING CUCKOO-OPTIMIZED MODULAR NEURAL NETWORKS

Abstract

The wind turbine is rapidly becoming one of the world's most significant renewable energy sources. Wind turbines must be massive to increase amounts of electrical energy. The blades of a wind turbine are commonly made of fiber materials due to their low cost and low weight properties. However, blades are affected by gusts of wind, poor climate factors, uncertain wind loads, lightning storms, and gravity loads, resulting in a surface crack of the blade. As a result, it is important to monitor the state of each wind turbine and its location fault condition. In this research, a cuckoo-optimized modular neural network (COMNN) is proposed for detecting and classifying a crack in the blades of a wind turbine. The method is created using a piezoelectric accelerometer to calculate the blade vibration response when it is energized. Cuckoo optimization is applied to initialize and adjust the weight vector of the Modular Neural Network. The experimental result shows the COMNN accurately detects and classifies faults in an acceptable time. The proposed approaches classify the fault type with an accuracy 98.1 % higher than the existing techniques, such as convolutional neural networks (CNN), recurrent neural networks (RNN), and artificial neural network ANN + support vector machine (SVM) algorithms.