Prediction of Combustion Characteristic Parameters of High Efficiency and Low Emission Hydrogen-Enriched Compressed Natural Gas Engine Under Closed-Loop Control

Closed-loop combustion control is highly beneficial for improving the efficiency and reducing the emissions of spark ignition internal combustion engines. In this paper, the key parameter (CA50) of closed-loop combustion control and its effect on the combustion and emissions were explored experimentally in a six-cylinder hydrogen enriched compressed natural gas (HCNG) engine. Moreover, the particle swarm optimization (PSO) back propagation neural network (BPNN) algorithm improved by various hybrid strategies was employed for CA50 prediction. The experimental results reveal that CA50 has a significant impact on the combustion characteristics and emissions of the HCNG engine. Meanwhile, statistical analysis illustrates that CA50 follows a normal distribution and has no self-correlation. Considering the one-to-one correspondence between CA50 and the spark timing, it is suitable to select CA50 as the feedback parameter. The simulation results indicate that the CA50 prediction model established by the PSO-BPNN method has high prediction performance and excellent generalization ability, with an average mean absolute error (MAE) of 0.25°CA and correlation coefficient (R) of more than 0.997. To further enhance the model’s performance, the PSO-BPNN models optimized by various hybrid strategies were compared, concluding that the hybrid strategies can significantly improve the convergence speed without sacrificing prediction accuracy. Among them, the NaPSO-BPNN method has the fastest convergence speed, and its CPU running time is 73.02% less than that of the PSO-BPNN model.