Development of dissimilar heat transfer promoter by genetic topology optimization

Abstract

From the effective use of energy, it is desired to develop a vortex generator that realizes heat transfer enhancement while suppressing flow resistance, that is, the dissimilar heat transfer promoter. In recent years, various thermal-fluid engineering devices such as heat sinks have been developed in laminar region by topology optimization using adjoint analysis. However, in a strong non-linear flow field, second or higher-order variation cannot be ignored if integration is performed for a long time, in other words, the variational principle breaks down. Hence this optimization method cannot be applied to the development of dissimilar heat transfer promoters. Therefore, in this study, we develop a dissimilar heat transfer promoter by genetic topology optimization called the NG-net method, which is the method without using gradient information of objective functional. The purpose of this study is to find out what shape vortex generators should be installed in order to maximize heat transfer in plane channel with temperature differences on the upper and lower walls. The Reynolds number is set in a range where vortices appear time-dependently. In the poster presentation, the optimum shape, the Reynolds-number and time-integration-dependence of this shape will be shown, and the reason why the obtained shape realizes greater heat transfer will be discussed.