Surrogate Models for Heat Transfer in Oscillating Flow with a Local Heat Source

Abstract

Simulative optimization methods often build on an iterative scheme, where a simulation model is solved in each iteration. To reduce the time needed for an optimization, finding the right balance between simulation model quality, and simulation time is essential. This is especially true for transient problems, such as fluid flow within a hydromechanical system. Therefore, we present an approach to building steady-state surrogate models for oscillating flow in a pipe with a local heat source. The main aspect is to model the fluid as a solid with an orthotropic heat transfer coefficient. The values of this coefficient are fitted to reproduce the temperature distribution of the transient case by parametric optimization. It is shown that the presented approach is feasible for different sets of parameters and creates suitable surrogate models for oscillating flow within a pipe with a local heat source. In future works, the presented approach will be transferred from the simplified geometry under investigation to industrial problems.