Multi-objective optimization of a thermoacoustic regenerator

Abstract

This paper presents a new mathematical approach for optimizing the geometry of a thermoacoustic regenerator, aimed at producing efficient thermoacoustic engines. Optimal set of parameters describing the device are computed for a chosen thermoacoustic couple to illustrate this approach. Hence, a non-linear multiobjective problem is formulated in GAMS and solved using Lindoglobal solver. Lexicographic optimization is presented as an alternative optimization technique to the common used weighting methods. This approach establishes a hierarchical order among all the optimization objectives instead of giving them a specific (and most of the time, arbitrary) weight. In this work, the optimization criteria are chosen as work output, viscous resistance as well as thermal losses that are typically disregarded when modeling the device. A practical example is given, in a hypothetical scenario, showing how the proposed optimization technique may help thermoacoustic regenerator designers to identify Pareto optimal solutions when dealing with geometric parameters.