Obtaining more accurate convective heat transfer coefficients in thermal analysis of spindle using surrogate assisted differential evolution method

Abstract

Obtaining accurate convective heat transfer coefficients (CHTCs) is crucial to improve the simulation accuracy in finite element thermal analysis (FETA) of spindle. A surrogate assisted differential evolution (SADE) method was proposed in this paper to obtain more accurate CHTCs in FETA of a spindle. Firstly, the temperatures at six key temperature points of a spindle were obtained through thermal experiment and FETA with initial CHTCs calculated using empirical formulae, respectively. And the comparison results demonstrated that there exist notable discrepancies between them. Afterwards, by regarding the CHTCs as the variables to be optimized and the simulation error as the objective function, the optimal CHTCs were searched using the proposed SADE method. Finally, the good effectiveness and high efficiency of the proposed SADE method were verified by comparing with the experimental results and by comparing with the conventional differential evolution (DE) method and genetic algorithm (GA) method, respectively.