Prediction of Critical Heat Flux Based on Gaussian Process Regression

Accurate prediction of critical heat flux(CHF) is very important for reactor safety and operation. Aiming at the shortcomings of existing artificial neural networks(ANNs) prediction methods, a Gaussian process regression(GPR) based CHF prediction method is proposed. Firstly, the obtained CHF data under local conditions are preprocessed, and the data is divided into a training set and a test set. Then, the training data is used to train the GPR model, and the optimal hyper-parameters are obtained. Secondly, the trained GPR model is used to predict the CHF, and the results are compared with the radial basis function neural networks(RBFNN). Simultaneously, the effect of important parameters on CHF is analyzed. The results show that the prediction results of the GPR model have higher prediction accuracy and smaller error compared with RBFNN, and agree well with the experimental data. The parametric trends fit the general understandings.