基于PCA与LSTM的核电站凝给水系统运行状态预测

打开文本图片集

中图分类号：TL353 文献标志码：A doi：10.3969/j.issn.1006-0316.2026.01.004

文章编号：1006-0316（2026）01-0021-07

Abstract ： Aiming at the requirements for operational stability and energy efficiency improvement of the condensate and fedwater system in nuclear power plants，this paper employs data mining techniques to research operational state identification and prediction.First， data preprocesing is performed onthe datasetcollected from condensate and feedwater sensors.Subsequently，a Long Short-Term Memory （LSTM） neural network is utilized to construct a system model for predicting future operational states.Experimental results demonstrate that the LSTM network significantly outperforms traditional Backpropagation （BP） neural networks and Recurent Neural Networks （RNN） in terms of prediction accuracy. The comparative performance analysis of the BP， RNN，and LSTM models applied to the condensateand feedwater system further validates the efectiveness of LSTM in handling such time-seriesdata.

Key words ：condensate and feedwater system iprincipal component analysis ilong short-term memory network ; stateprediction;multi-sourcedata

在核电站发电过程中，凝给水系统是实现水循环利用的关键环节，该系统稳定高效地提供符合多工况的给水需求[1]。（剩余7197字）