Prediction and Numerical Study of Thermal Performance of Gradient Porous Structures Based on Voronoi Tessellation...

Porous materials are a new type of engineering material with both functional and structural properties. Compared with regular porous structures and random porous structures, a gradient porous structure is a porous structure with a spatial variation mechanism, which can adjust the layout of the structure by changing its own load and boundary conditions according to different situations, thus obtaining better performance. In this paper, three spatial Voronoi structures with different spatial gradients are designed using the spatial Voronoi tessellation method. The differences in thermal protection performances between the Voronoi spatial gradient structure and the regular structure and the effects of porosity, gradient direction and heat flow density on the three-dimensional Voronoi stochastic gradient structure were investigated via data simulation. The results show that the effective thermal conductivity of the Voronoi spatial gradient structure is lower than that of the regular structure. The effective thermal conductivity of the structure gradually decreases with increasing porosity. Taking the gradient Voronoi structure consisting of 3 × 3 × 3 units as an example, when the porosity increases from 83% to 94.98%, its effective thermal conductivity decreases from 0.586 to 0.149 Wm-1K-1. The anisotropy of the random structure leads to effective thermal conductivity errors of more than 5% in all three gradient directions. In addition, according to the principle of thermal resistance superposition, we designed a battery pack set for calculating the effective thermal conductivities of pillar-based porous materials, including three-dimensional Voronoi gradient random porous materials on the Grasshopper platform. In this way, the effective thermal conductivity of a pillar-based porous material can be predicted more accurately. The predicted calculation results and the simulation results basically agree with each other, and the relative errors of both are within 10%.