Evaluation on Aging State of NPP Canned Motor Pump Based on SVM

A new method for aging assessment of canned motor pump in nuclear power plants was given in the paper.In practice,it is difficult to do the aging assessment for this kind of pump for its complicated structure,lack of samples and aging data.In order to solve these problems,the Support Vector Machine（SVM） which had its own advantages of solving the classification and evaluation in the case of limited examples was applied in this new method.The assessment indices including two performance parameters,and two status parameters were selected.And then the indices grade model and code coding rules were established and used to evaluation the aging condition on SVM with different kernel function.The results show that the application of the SVM method in aging assessment of canned motor pump is effective and reliable,especially with RBF kernel function that can get faster calculation speed,high generalization capability and more exact result.