A Study on Fault Diagnosis Technology of Nuclear Power Plant Based on Decision Tree

The technology of real-time fault diagnosis for NPP has great significance to improve the safety and economy of reactor. At present, expert system, artificial neural network (ANN) and support vector machine (SVM) algorithms are most widely used in the field of NPP fault diagnosis. According to the shortcomings of expert systems, ANN and SVM, the decision tree algorithm is applied in the field of NPP fault diagnosis in this paper. ID3 and C4.5 are applied separately to learn from training samples which are the typical faults of NPP, and diagnose using the acquired knowledge. Then the diagnostic results are compared with the results of SVM method. The results show that: comparing with SVM, decision tree has the advantages of much faster training speed and a little higher accuracy. Furthermore, decision tree can obtain rules from the sample set, so it has good explanatory ability for the diagnostic results.