论文部分内容阅读
为实现水电机组轴系运行常见故障的快速实时诊断,提出了一种基于支持向量机的故障诊断及预测方法。该方法应用支持向量机分类的基本原理,提取机组振动信号的频谱能量作为学习样本,通过训练建立基于水电机组轴系运行常见故障的分类模型,进行故障类型识别。同时,结合状态监测系统的实时采集数据,应用时间加权因子和支持向量机回归模型,实现特征数据的实时预测。经实验分析验证,该诊断方法具有较高的准确性,其回归预测方法有效可行,能满足实时故障诊断的要求。
In order to realize the rapid real-time diagnosis of the common failures of hydropower unit’s shaft operation, a fault diagnosis and prediction method based on support vector machine is proposed. The method uses the basic principle of SVM classification, extracts the spectral energy of the vibration signal of the unit as a learning sample, and establishes a classification model based on the common faults of hydropower unit shaft operation through training to identify the fault type. At the same time, real-time prediction of feature data is realized by combining the real-time data collected by the condition monitoring system with the time-weighted factor and the support vector machine regression model. The experimental analysis shows that the method has high accuracy and the regression prediction method is effective and feasible, which can meet the requirements of real-time fault diagnosis.