Molten salt energy storage technology is widely used in solar thermal power generation due to its high thermal capacity and good thermal stability. To optimize the influence of key operating parameters on energy storage efficiency

numerical simulation methods are used to analyze the mechanism of input velocity

initial temperature

temperature difference and other parameters on the formation of thermocline and heat storage efficiency at different horizontal positions. The results show that

increasing the temperature difference and the input speed can significantly promote the development of the thermocline

and increase the heat storage efficiency by more than 10%. The parameter optimization algorithm based on response surface methodology identifies an optimized parameter combination

which improves the heat storage efficiency by a maximum of 16.3 percentage points compared to the previous simulations. At the same time

to quickly and accurately predict the operating temperature of the system

three machine learning models are compared

and it finds out that the random forest model has the best prediction with an accuracy rate of 98.78%. The research results provide theoretical basis and application reference for the optimization design of molten salt energy storage systems.