An optimal convective heat transfer coefficient calculation method in thermal analysis of spindle system

Abstract

Accurate calculation of convective heat transfer coefficients (CHTCs) is the premise to obtain accurate results of finite element thermal analysis (FETA). This paper presents an optimization method of CHTCs in steady state FETA of a spindle system using genetic algorithm (GA), which is especially suitable for complicated FETA with many CHTCs, such as 10 CHTCs. Firstly, the temperature values at thermal critical points, the temperature field, and the thermal fluctuations of a spindle system were obtained by experiment. Secondly, based on similarity theory, the initial values of CHTCs, which are inaccurate but close to real ones, were calculated by empirical formulas, and then, the FETA of the spindle system was conducted. The inaccurate initial CHTCs lead to inaccurate FETA results, so the optimal CHTCs closer to real ones were searched by GA automatically. The CHTCs were regarded as the target values of interest, and the root-mean-square error (RMSE) between the experimental temperature values and the simulation temperature values was regarded as the fitness function. After 200 generations, the optimal CHTCs were searched with which the RMSE was the lowest. Finally, by conducting the FETA of the spindle system with the searched optimal CHTCs and comparing with the experimental results, the effectiveness of the proposed method was proved.