Effect of Phase Transition Temperature on Latent Functionally Thermal Fluid and Optimal Selection of Phase Transition Temperature Based on Neural Network

Abstract

At the present research stage, there are fewer studies on the natural convection of latent heat type functional fluids, and the existing models have the problems of excessive computational volume, poor stability and accuracy. In this paper, a single-fluid SRT lattice Boltzmann model of latent heat functional fluid is developed to study the natural convection characteristics of latent heat functional fluid under constant and sinusoidal heat sources at the wall, and the local Nusselt number at its wall is predicted by BP neural network. The results show that the best heat transfer performance can be obtained with phase transition temperatures in the range of 0.3 to 0.7, and that the heat transfer strength of the sinusoidal heat source wall is positively correlated with its wall temperature. Furthermore, when the phase transition temperature is certain, the reduction of the Stefan number (increase in the latent heat of phase transition) greatly enhances the effect of heat transfer inside the square cavity. The error between the BP neural network predicted data and the simulated data is around 5%, which is highly reliable.

Keywords: Latent functionally thermal fluidSRTLattice Boltzmannnatural convectionneural network