Research on the interaction relationship between common rail diesel engine injection parameters based on neural network

High-altitude calibration of CA6DL2-35E3R common rail diesel engine was finished on engine high altitude simulating environment test bed. The interactions between injection parameters and its impact on engine high altitude performance were studied. Results show that the fuel delivery per cycle per cylinder increases with the adwance of injection timing and the increase of common rail pressure at the calibration sweep as a whole; fuel delivery per cycle per cylinder increased about 0.1-0.3 mg with advancing 1° CA injection timing and it would increase 0.07-0.1 mg with increasingl MPa common rail pressure averagely. The fitting error of constructed radial-based function neural network model was below 10^-12 and the predictive error of which was between 1.5%, which can fulfill common rail diesel engine characteristics modeling demand. The model can help alleviate the influence of the injection timing and common rail pressure on fuel delivery per cycle per cylinder, and achieve the impact of one single injection parameter on engine performance, which can help increase the understanding of common rail diesel engine injection characteristics.