Bayesian network analysis of heat transfer deterioration in supercritical water

•The impact of different influencing factors on HTD and the causal relationship between various influencing factors were calculated and analyzed.•A Bayesian network model diagram for predicting HTD is constructed successfully.•The predictability of HTD is obviously improved compared to conventional methods. Using the experimental data of supercritical water heat transfer, a Bayesian network is constructed for various parameters which affect the heat transfer deterioration (HTD)in supercritical water. Through the Bayesian network model, it is concluded that the most direct factors are the actual heat transfer coefficient and the Pr number, and the factors which are most closely related to HTD are thermal conductivity, dynamic viscosity, Reynolds number and heating power. The interaction of various parameters results in the complexity and changeability of the situation. The Bayesian network is capable of showing the effect of various factors under different conditions, and then it comes to make reasonable inferences and analysis on the possibility of HTD in supercritical water with different condition parameters. Then the mechanism analysis of the effect is carried out.