一种基于SVDD和PCA的风机故障诊断方法

A Fault Diagnosis Method for Fan Based on SVDD and PCA

针对风力发电机叶片故障诊断中正常与故障数据不均衡问题,提出一种基于SVDD的故障诊断方法。首先,设计高通滤波器消除风噪对检测声信号的影响,提取出多维时频特征量;然后,引入主成分分析法(PCA)优化初始特征向量,有效表征叶片的健康状态,在此基础上训练建立基于SVDD的单分类模型。最后,基于实测数据,实验验证了算法的有效性。

Aiming at the imbalance between normal and fault data in fault diagnosis of wind turbine blades,a fault diagnosis method based on SVDD is proposed.First,a high pass filter is designed to eliminate the influence of wind noise on the sound signal and extract the multidimensional time frequency characteristic.Then,the principal component analysis(PCA)is introduced to optimize the initial eigenvector and effectively characterize the health state of the blade.On this basis,a single classification model based on SVDD is established.Finally,the validity of the algorithm is verified by the measured data.