Identification of Damage to Wind Turbine Blade Based on Fruit Fly Optimization Algorithm-support Vector Machine

摘要

In order to improve the correctness of identification for crack damage and edge damage to wind turbine blade, a method combining with Fruit Fly Optimization Algorithm (FOA) and Support Vector Machine (SVM) was proposed. Acoustic emission signals of two kinds of damage failure were collected with hardware system, and then handled with wavelet to extract energy characteristics. According to these characteristics, Support Vector Machine (SVM) will be established, and accuracy rate will be tested. Then, the characteristics of support vector machine were optimized by fruit fly algorithm to make damage identification of model more accurate. Finally, the identification result after optimization was compared with the identification result after optimizing the support vector machine by Particle Swarm Optimization (PSO). The simulation result shows that the identification accuracy of support vector machine model optimized by fruit fly optimization is much higher, and can accurately realize the identification of damage to wind turbine blade.