Back to home
Intelligent fault diagnosis for nuclear power plant based on deep belief network and support vector machine   ZHOU Wen1, LIU Yong-kuo2, XIA Hong3, and LIU Peng-fei4   1. College of Nuclear Science and Technology, Harbin Engineering University, Nantong Street No.145, 150001, Harbin, China (1028678419@qq.com) 2. College of Nuclear Science and Technology, Harbin Engineering University, Nantong Street No.145, 150001, Harbin, China (lyk08@126.com) 3. College of Nuclear Science and Technology, Harbin Engineering University, Nantong Street No.145, 150001, Harbin, China (xiahong@brbeu.edu.cn) 4. College of Nuclear Science and Technology, Harbin Engineering University, Nantong Street No.145, 150001, Harbin, China (c0938mjbn@163.com)   Abstract:For the fault diagnosis of nuclear power plant, using the deep learning and support vector machine technology, a novel intelligent diagnosis method was proposed, which combined the deep learning feature extraction and support vector machine pattern recognition. The characteristics of fault data was extracted adaptively with deep belief network, then support vector machine classification model was used for diagnosing the fault of nuclear power plant. The normal and fault data from Qinshan II nuclear power plant was normalized. The deepbelief network was given unsupervised training with training samples consisted of normalized data. The output of deep belief network was used for training support vector machine classification model. The result indicates that the method can diagnose fault correctly and the accuracy can be above 90%. Keyword:nuclear safety; fault diagnosis; deep belief network; support vector machine
	Click here for viewing full text.