Machine learning-based multi-objective optimization and thermodynamic evaluation of organic Rankine cycle (ORC) system for vehicle engine under road condition

Abstract

Organic Rankine cycle (ORC) cooperative multi-objective optimization under dynamic driving cycle is the key to obtain the potential of waste heat recovery of internal combustion (IC) engine. The exhaust waste heat energy of IC engine has the characteristics of complex change and strong fluctuation under actual road conditions. The analysis and optimization of the performance of ORC system under single engine operation has some limitations in evaluating the actual waste heat recovery characteristics. Based on Japan10-15 road condition, this paper builds ORC system for vehicle engine model oriented to driving environment. The interaction between operating parameters of ORC system and thermodynamic and thermal economic performance is analyzed. Based on the operating characteristics of the ORC system, the thermal efficiency and power output per unit heat transfer area (POPA) data driven models are established, and the structural parameters of the network are optimized. Based on the multi-objective optimization algorithm, the thermodynamic performance of ORC system under road condition is optimized. The results show that the ORC system has a maximum thermal efficiency of 2.23% and a maximum POPA of 0.55 kW/m2 under Japan10-15 road condition. In this paper, the performance evaluation and multi-objective optimization results of ORC system under road condition can provide a certain reference for the practical application potential of ORC system.