Research on performance prediction and parameter optimization of combustion characteristics for opposed single cylinder free piston generator system with direct injection based on BP neural network

As the core index to measure the energy conversion efficiency and low carbonization performance of free piston generators, thermal efficiency directly bears on the power density and carbon reduction potential of the system. In this study, the indicated thermal efficiency of the opposed single-cylinder free piston generator (OSFPG) with direct injection is accurately predicted through the BP neural network model with adaptive error backpropagation mechanism. More, based on the causal analysis derived from orthogonal experiments, the synergistic effect mechanism of compression ratio, operating frequency, and ignition phase on the indicated thermal efficiency of OSFPG was systematically and quantitatively investigated. The results indicate that the ignition phase exerts the most significant influence on thermal efficiency, followed by operating frequency, with the compression ratio having the least impact. Further, the thermal efficiency of OSFPG exhibits a pronounced positive correlation with both the compression ratio and operating frequency, whereas the influence of the ignition phase on the thermal efficiency of OSFPG demonstrates a typical unimodal response. Within the studied change range of parameters, the indicated thermal efficiency reached 44.3 % at an ignition phase of 12^oCA, an operating frequency of 35 Hz and a compression ratio of 10. This research provides a theoretical basis for the dynamic parameter optimization of OSFPGs, as well as an important reference for the research of high efficiency and low carbon performance of OSFPGs.