Multi-objective optimisation of thermodynamic performance parameters of a Gifford-McMahon refrigerator

Abstract

The maximisation of cooling capacity and thermodynamic efficiency of a Gifford-McMahon refrigerator is a tedious task due to the occurrence of complex thermo-physical mechanisms inside its expander. Thus, in this paper, a hybrid simulation approach is introduced by using a combination of the numerical model, response surface model, and non-sorted genetic algorithm to maximise its cooling capacity and relative Carnot efficiency. The selected design factors are length and diameter of the regenerator, its porosity, stroke length of the displacer, and the delay angle of the displacer motion. A Box–Benhken design approach is adopted to create a robust design matrix by using the aforementioned design variables and responses. The correctness of the model is detected through the analysis of variance. Finally, a non-sorted genetic algorithm is utilised to maximise both responses. The numerical analysis at optimal parameter suggests that a major portion of expansion work loss takes place at the regenerator.