Multi-objective constructal design for quadrilateral heat generation body based on thermal-entransy theory and NSGA-II

Abstract

Based on thermal-entransy theory, the multi-objective constructal design of quadrilateral heat generation body (HGB) with similar shapes of leaves is studied further. The relationship between the aspect ratio of quadrilateral HGB and average temperature difference based on thermal-entransy dissipation is compared with that between the aspect ratio of quadrilateral HGB and the maximum temperature difference (MTD). The relationship between a composite function, consisting of linear weighting sum of the average temperature difference and MTD, and aspect ratio is obtained, and the optimal aspect ratios under minimum composite function with different weighting coefficients are obtained. Using the NSGA-II algorithm, the Pareto frontier containing a series of compromise results of average temperature difference and MTD is obtained, and optimization results are compared using the deviation index. There is no aspect ratio to make both MTD and average temperature difference reach the minimum, and the optimal aspect ratio under the minimum MTD is smaller than that under the minimum average temperature difference. The optimal aspect ratio is obtained by making the composite function reach the minimum, and the optimal aspect ratios obtained by minimizing the composite function with different weighting coefficients are different. Compared with the construct of the initial design, the value of the composite function with optimal construct decreases by 1.9%, and the aspect ratio of the quadrilateral HGB decreases by 9.1%. The average temperature difference with the optimal construct increases by 2.1%, and the MTD with the optimal construct decreases by 5.6%. The deviation index under multi-objective optimization is smaller than that under single-objective optimization, and the obtained construct has better comprehensive thermal conductivity. Compared with TOPSIS and LINMAP decision-making methods, the average temperature difference with composite function optimization increases by 0.55% and 0.62% respectively, but the MTD with composite function optimization decreases by 0.84% and 0.96%.