果蝇优化算法融合SVM的风机叶片损伤识别研究

Study on Identification of Damage for Wind Turbine Blade by Fusing FOA and SVM

为提高风机叶片裂纹损伤和边缘损伤识别的准确率,提出使用果蝇优化(FOA)算法和支持向量机(SVM)相结合的方法。使用硬件系统采集两类损伤故障的声发射信号,然后对信号进行小波处理,提取能量特征,根据能量特征信息,建立支持向量机模型,测试其准确率;采用果蝇优化算法优化支持向量机参数,使模型损伤识别更准确,并将优化后模型识别结果与粒子群优化(PSO)算法优化支持向量机后的识别结果相比较。仿真结果表明,使用果蝇优化算法优化后的支持向量机模型的识别精度更高,能够准确地实现对风机叶片损伤的识别。

In order to improve the identification accuracy of crack damage and edge damage for wind turbine blade,the method combining fruit fly optimization algorithm（ FOA） and support vector machine（ SVM） is proposed. The acoustic emission signals of two types of damage fault are collected using hardware system; then processed with wavelet method and the energy features are extracted. In accordance with the information of energy features,the support vector machine model is established,the accuracy is tested. The SVM parameters are optimized with FOA,to get more accurate damage identification. The model identification result is compared with the result by using particle swarm optimization（ PSO） and SVM. The simulation result indicates that the identification accuracy of FOA ＋ SVM is more accurate. Precise identification of blade damage can be implemented by the method proposed.