基于改进核主元分析法的三电平逆变器故障检测

Fault Detection for Three-Level Inverter Based on Improved Kernel Principal Component Analysis Method

核主元分析法能够有效利用提取的主元信息对系统运行情况进行评价,但实际应用中存在有效数据信息"淹没"的问题。为更好地实现三电平逆变器开关管故障检测,对传统核主元方法进行改进,增加灵敏度分析,分别计算各变量对故障统计量的贡献值,由此引入贡献值权重矩阵对逆变器故障特征进行加权,加权后的数据可以消除由不同量纲和噪声对数据造成的影响,同时可以反映故障数据的主要特征。基于此,计算不同开关管故障情况的权重矩阵,并对故障情况进行检测,降低了系统的数据处理量,提高了故障检测的准确性。实验结果验证了所提方法的有效性。

Kernel principal component analysis（KPCA） can be used to evaluate system operation. It is widely used in nonlinear process monitoring. However, KPCA does not always perform efficiently because useful information may be submerged under retained KPC. To overcome this shortcoming, a weight-based KPCA is proposed for fault detection of three-level inverter. Variable contribution plots are constructed for fault isolation based on sensitivity analysis theory. Improved KPCA is used mainly to set different weighting values on KPCs to highlight useful information. Improved feature eigenvectorscan eliminate influence caused by amplitudes and noises, comprehensively emphasize on fault features.Efficiency of the proposed method is demonstrated. Inverter fault detection results indicate that the proposed method is superior to conventional KPCA.