Effective multi-objective optimization of Stirling engine systems

In this article we demonstrate the supremacy of the Non-dominated Sorting Genetic Algorithm-II with Simulated Binary Crossover and Polynomial Mutation operators for the multi-objective optimization of Stirling engine systems by providing three examples, viz., (i) finite time thermodynamic model, (ii) Stirling engine thermal model with associated irreversibility and (iii) polytropic finite speed based thermodynamics. The finite time thermodynamic model involves seven decision variables and consists of three objectives: output power, thermal efficiency and rate of entropy generation. In comparison to literature, it was observed that the used strategy provides a better Pareto front and leads to improvements of up to 29%. The performance is also evaluated on a Stirling engine thermal model which considers the associated irreversibility of the cycle and consists of three objectives involving eleven decision variables. The supremacy of the suggested strategy is also demonstrated on the experimentally validated polytropic finite speed thermodynamics based Stirling engine model for optimization involving two objectives and ten decision variables.