关键词: 误差反传神经网络, 预诊, 性能评价, 剩余寿命预测, 汽轮机

Abstract: Intelligent prognostic methodology framework was proposed from the perspective of realizing system function of intelligent prognostic to predict and prevent failures. Performance assessment and residual life prediction models based on BackPropagation (BP) neural network were established. The effectiveness and prediction error of BP models were studied in particular. From the perspective of application, updating and dynamic prediction of model were realized. With the increasing of collection data, the prediction model was adjusted. And the dynamic assessment value was attained based on the adjusted model. The proposed method was implemented in a blade material fatigue analysis system at Harbin Steam Turbine Company. The feasibility and effectiveness of the proposed method were verified by blade material performance analysis and residual life prediction.

Key words: backpropagation neural network, prognostic, performance evaluation, remaining life prediction, steam turbine, blade