Abstract Considering the various irreversibility conditions caused by heat transfer and working processes in a dual cycle, the power density performance is optimized by applying finite-time thermodynamics theory, and multi-objective optimization is performed by using NSGA-II. The effects of cut-off ratio, maximum cycle temperature ratio, and various losses by heat transfer and working processes on the relationships between the power density and the compression ratio and between the power density and the thermal efficiency are analyzed. The thermal efficiency and engine size obtained under the conditions of maximum power output and power density are discussed. The results show that for a dual cycle, the heat engine has a smaller size and higher thermal efficiency under the condition of maximum power density. The cycle compression ratio and cut-off ratio are selected as decision variables, and the dimensionless power output, thermal efficiency, dimensionless ecological function, and dimensionless power density are selected as objective functions. Multi-objective optimization is performed with different objective combinations. The deviation indexes under the LINMAP, TOPSIS, and Shannon entropy approaches are discussed, and the number of generations when the genetic algorithm reaches convergence are obtained. The results show that the genetic algorithm converges at the 341st generation for the quadru-objective optimization, at the 488th generation for the tri-objective optimization, and at the 399th generation for the bi-objective optimization. When the bi-objective optimization is performed with dimensionless power output and dimensionless ecological function as the objective functions, the deviation index obtained based on the LINMAP approach is 0.1400, which is better than those obtained for other single- and multi-objective optimizations.
Read full abstract