Predicting Oxygen Concentration In Lead-bismuth Eutectic And Critical Heat Flux In Ocean Condition By Machine Learning Methods

Machine learning is a study on how to make machines acquire new knowledge and new skills and identify existing knowledge.With its powerful fitting ability and robustness,machine learning has been widely used in economics,engineering and other fields.It is often used in data mining,computer vision,natural language processing,medical diagnosis,etc.Because strong prediction and classification functions,in order to study the thermal hydraulic characteristics of the critical heat flux(CHF)of the nuclear power plant under the ocean conditions,and establish the oxygen concentration prediction mechanism in the liquid lead-bismuth alloy,based on the existing experimental data,machine learning had been used.In this paper,Artificial Neural Network Neural Network(ANN),Support Vector Machine(SVM),Random Forest(RF),Deep Learning(DL)and other models have been established to predict CHF under ocean conditions and the oxygen concentration in liquid lead-bismuth alloy.Furthermore,the marine conditions in this study are simulated by computer,to oscillating waveform,and then,construct the original training data through the Look up table program based on experimental data.It has been found that with the increase of data volume and data dimensions,the machine learning algorithms can predict more precisely,in the oxygen concentration in liquid metal and the critical heat flux density under ocean conditions.Especially through algorithm integration technology,multiple simple algorithms are integrated,and the integrated algorithm is used for data prediction,which can get better results.This work provides a new idea for solid-state oxygen control of lead bismuth eutectic of primary loop in accelerator driven sub-critical systems(ADS)and safe operation of reactors under ocean conditions,and made a little exploratory work for the way of artificial intelligence control in nuclear power plants and the application of nuclear reactor in the future.