水电机组状态退化评估与非线性预测

Condition Degradation Assessment and Nonlinear Prediction of Hydropower Unit

提出了基于Shepard曲面、经验模态分解法(ensembleempirical mode decomposition,EEMD)、混沌理论和灰色理论的水电机组状态退化评估与趋势预测模型。该方法首先用Shepard曲面建立综合考虑有功功率、工作水头作用的水电机组状态退化趋势模型。然后将水电机组状态退化趋势进行EEMD分解,得到若干个相对平稳的固有模态函数(intrinsic mode function,IMF)分量和一个余项分量,对每个IMF分量进行特性识别,根据其不同属性,选用混沌预测模型或灰色模型预测,同时对余项分量进行灰色预测。最后将所有分量的预测结果进行重构,获得最终预测结果。实例分析表明,该方法能有效地评估水电机组状态的退化过程,且能提高退化趋势的预测精度。

Based on Shepard surface, ensemble empirical mode decomposition （EEMD）, chaos theory and grey theory, a model to assess and predict condition degradation trend of hydropower unit is proposed. Firstly, utilizing Shepard surface a condition degradation trend model of hydropower unit, in which the actions of active power and working head are synthetically considered, is established; secondly the condition degradation trend ofhydropower unit is decomposed by EEMD to obtain several relatively steady intrinsic mode function （IMF） components and a remainder term component, and then each IMF component is identified and according to their different attributes the chaotic prediction modeI or grey model are chosen to perform prediction, meanwhile the grey prediction of remainder term component is carried out; finally prediction results of all components are reconstructed to obtain final prediction result. Case study results show that using the proposed method the condition degradation process of hydropower unit can be effectively assessed and the prediction accuracy of degradation trend can be improved.