基于KPCA和灰色模型的凝汽器故障预测

汽轮机凝汽器的故障预测为其故障自愈的研究提供了理论依据.提出一种基于核主元分析和灰色预测模型的汽轮机凝汽器故障预测方法,首次将灰色预测理论应用于凝汽器的故障预测.采用核主元分析法对故障特征数据进行分析和处理,提取反映故障的主要特征量,以降低特征变量之间的非线性相关性,同时减少灰色预测模型的预测参数的数目.然后应用灰色预测理论建立故障特征的预测模型,对每一个主要特征量的趋势值进行预测,重构故障特征向量,用于汽轮机凝汽器故障的预测分析.

Condenser Fault Prediction Based on KPCA and Gray Prediction Model

Condenser fault prediction and its fault self-recovery problems are discussed. A method based on the kernel principle component analysis (KPCA) and gray prediction theory is presented for condenser fault prediction. The gray prediction theory is applied to condenser fault prediction. Fault feature data is processed to extract main features by KPC A. As a result, the nonlinear relativity a-mong feature data and the number of features to be predicted are reduced. Based on gray prediction theory, the predicting model of fault feature is established to reconstruct the feature vectors used for the fault identification of condenser.