基于能量熵重构与支持向量回归的水电机组状态趋势预测

State Tendency Prediction Research of Hydropower Unit Based on Energy Entropy Reconstruction and Support Vector Regression

针对水电机组运行状态趋势预测的问题,提出了一种基于能量熵重构(EER)与支持向量回归(SVR)的混合预测模型。先针对复杂非平稳监测信号,利用快速集成经验模态分解(FEEMD)算法将其分解为多个本征模态函数(IMFs)分量和单个残余分量;然后基于能量熵(EE)理论对各分量进行重构,以有效降低分量的复杂度;最后,将生成的重构本征模态函数(RIMFs)作为SVR的输入,训练模型参数得到最优的SVR,用于预测机组状态发展趋势。与实例对比分析表明,所提混合预测模型具有较高的预测精度,为机组运维策略的制定提供了一定的指导。

To solve the problem of state tendency prediction of hydropower unit,a hybrid forecasting model based on energy entropy reconstruction(EER)and support vector regression(SVR)was proposed in this paper.Firstly,for the complex non-stationary monitoring signal,fast ensemble empirical mode decomposition(FEEMD)was used to decompose the original signal into a series of intrinsic mode functions(IMFs)and a residue component.Secondly,the energy entropy theory was used to reconstruct these components for reducing the complexity effectively.Finally,the refactored intrinsic mode functions(RIMFs)after mode reconstruction were served as the inputs of SVR model to train the model parameters and obtain the optimal SVR,which was used to forecast the state tendency of unit.Compared with other methods,the results of a case study show that the proposed hybrid prediction model has the higher forecasting accuracy,which provides guidance for the maintenance strategy of hydropower unit.