基于偏最小二乘算法的热耗率回归分析

Heat Rate Regression Analysis Based on Partial Least Squares Algorithm

针对热耗率影响因素众多且呈现高度多重相关的特征,提出了偏最小二乘(Partial Least Square PLS)算法建立热耗率回归分析模型。在数据预处理方面对机组热耗率的主要因素和主要参数做了相关性分析,进而更合理地确定了偏最小二乘回归分析的数据表,有效建立了热耗率预测模型。预测模型的检验方式采用交叉有效性检验,选定对模型有显著改善的PLS主成分个数。通过实例验证了偏最小二乘方法能够有效解决自变量集合高度相关的问题。

For the characteristics of heat rate that there＇re the multiple influencing factors and they have shown high correlation,a novel approach of partial least square regression algorithm is proposed to analyze heat rate.For data pre-processing,the correlation of unit heat rate are is analyzed.The data tables are determined for the partial least squares regression analysis.The partial least square model is established to predict the heat rate effectively.The effectiveness of cross-examination methods of the model test is used to select the number of PLS principal component,which can improve the mode significantly.Example analysis proves that the partial least squares regression method can effectively solve the problem of high correlation in the variable collection.