水电机组混合智能故障诊断方法的研究

Research on Hybrid Intelligent Fault Diagnosis Method for Hydroelectric Units

现有支持向量机方法在水电机组故障诊断应用中存在建模能力差、未考虑水电机组故障数据样本不确定性问题,难以保障故障诊断结果的准确性。针对此情况,本文在传统支持向量机方法的基础上引入粗糙集理论,构建粗糙多类支持向量机,并将其应用于水电机组导摆度偏大故障诊断中,进而确认此混合智能故障诊断方法具有较高的准确性。

The existing support vector machine methods have poor modeling ability and fail to consider the uncertainty of hydroelectric unit fault data samples in the application of fault diagnosis,making it difficult to ensure the accuracy of fault diagnosis results.In response to this situation,this paper introduces rough set theory on the basis of traditional support vector machine methods,constructs rough multi class support vector machines,and applies them to the fault diagnosis of large guide swing of water motor sets,thereby confirming the high accuracy of this hybrid intelligent fault diagnosis method.