【摘要】 基于神经网络“能量―故障”映射关系,提出将小波频带分析与灰色预测理论相结合进行水电机组故障预测。运用小波分解提取各频带能量,应用预测理论建立水电机组故障特征量的预测模型,预测各频率成分能量的变化,重构由预测各频带能量成分组成的特征向量,应用于水电机组故障预测分析。以水轮机主轴摆度信号为例,应用该方法进行了特征信息提取和预测,表明将小波能量提取与灰色预测理论相结合进行振动特征信息的预测比较有效,为故障预测提供了新思路。更多还原

【Abstract】 Based on the mapping of “energy-fault” in neural network, it is proposed to combine the wavelet transform based frequency spectrum analysis with grey theory to forecast the fault of hydroelectric generating sets. In this method the energy in different frequency bands is extracted by wavelet decomposition and by use of forecasting theory the forecasting model of hydroelectric generating set’s characteristic quantities is established to forecast the energy changes of different frequency components, then the reconstructed characteristic vectors consisting of the forecasted energy components in different frequency bands are applied to the fault forecasting of hydroelectric generating set. Taking the main shaft of hydraulic turbine for example, the extraction of characteristic information and fault forecasting are carried out by use of the proposed method, the results show that it is more effective to combine the wavelet transform based energy extraction with grey theory to forecast the characteristic information of vibration.更多还原