Nuclear power equipment key quality characteristic extraction method based on material visual angle

• Main Authors: YE XUQIAN, FENG LINHAN, WEN PEIHAN, LIU YIYANG, WANG HAO
• Format: Patent
• Language: Chinese; English
• Published: 13.05.2022
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention provides a nuclear power equipment key quality characteristic extraction method based on a material view angle. The method comprises the following steps: firstly, collecting point cloud and energy data of nuclear power equipment key material element atomic clusters, and analyzing and identifying element composition of atomic cluster energy characteristics related to the structural stability of nuclear power equipment; inputting the quality characteristic elements of the atomic cluster into the XGboost prediction model and the LightGBM prediction model respectively for training, then judging whether the energy characteristic values of the atomic cluster output by the XGboost prediction model and the LightGBM prediction model are optimal or not, and if yes, judging whether the energy characteristic values of the atomic cluster output by the XGboost prediction model and the LightGBM prediction model are optimal or not; if yes, inputting quality characteristic elements of the to-be-measured atomic c