基于KPCR的发电机组参数预测与估计

Parameter prediction and estimation for turbine generator based on KPCR

为了解决机组运行过程中参数失效和优化过程中参数计算的问题,提出了基于核回归(KPCR)的发电机组参数预测和估计方法。首先用正常数据建立机组参数的预测和估计模型,确定各变量之间的回归关系,然后将其用于参数的在线预测与估计。该方法可以有效地捕捉变量间的非线性关系,参数预测和估计效果明显好于偏最小二乘回归(PLS)和主元回归(PCR)等线性回归方法。某电厂1000 MW发电机组烟气含氧量历史特征数据集仿真试验证明了该方法的有效性。

For the parameter failure during generation unit operation and the parameter calculation of online unit optimization,the parameter prediction and estimation based on KPCR（Kernel Principal Component Regression） is proposed. The prediction and estimation model is established with normal data and the regression relationship among variables is analyzed,according to which,the parameters are online predicted and estimated. The proposed approach can effectively capture the nonlinear relationship among variables,better than PLS and PCR methods based on linear regression. Simulation based on the historical data set of the oxygen content in flue gas of a 1000 MW generation unit verifies the effectiveness of the scheme.