基于最小二乘支持向量机和机电综合特征的发电机故障诊断

Generator fault diagnosis using least-squares-based support vector machine and mechatronical features extraction

提出了1种最小二乘支持向量机和机电综合特征相结合的发电机故障诊断模型.用二次损失函数代替传统支持向量机中的不敏感损失函数,将不等式约束条件变为等式约束,从而将二次规划问题转变为线性方程组求解,降低了计算的复杂性.提取发电机故障中的综合特征,即振动信号和电流信号,整理后作为诊断模型的特征值,从而得到了故障的典型特征,提高了诊断的准确率.最后从SDF 9型模拟发电机中实测数据进行分析,结果表明,与常规的方法相比,该模型具有较高的分类速度和较好的故障诊断准确率.

In this study, a generator fault diagnosis model is first developed using least-squares-based support vector machine and mechatronical features. Then, the insensitive loss function regarding traditional support vector machine is replaced by the quadratic loss function. In this way, the inequality constraints are adapted to the equality constraints so as to simplify the quadratic programming problem as the linear equation groups, viz., reduce the computational complexity. Next, the mechatronical features are extracted from the vibration and current signals. Furthermore, these features are organized as eigenvectors, i.e. typical fault features of diagnosis model, so as to improve the diagnosis accuracy. Finally, the testing data, which are sampled from the SDF-9 Simulation Generator, are used for analysis. Therefore, it is found from the experimental results that, compared with traditional approaches, the proposed model possesses such strongholds as classification efficiency and diagnosis accuracy.