Virtual Calibration of Injection Parameters for a Heavy-Duty Common Rail Diesel Engine at Different Altitudes

Under high altitude conditions, diesel engine performances are obviously worsened due to combustion deterioration. Injection parameters calibration plays a key role in improving the engine performances at plateau. In order to improve the calibration efficiency and effectively recover the engine power and the fuel economy, a vertical calibration simulation model was established, in which Latin hypercube sampling(LHS)was used to reduce the experiment points from 20 580 to 210. A second-order response surface model that combines the quadratic polynomial function and neural network was used to revise injection parameter maps. An engine bench experiment was conducted according to the revised maps. The results show that, after the virtual calibration, the power output of the engine does not drop at the altitude of 2 km, but drops by 6.8% and 30.9% at the altitude of 3 km and 5 km respectively. The exhaust temperature before turbine, the turbocharger speed and the maximum combustion pressure do not exceed their limits at these altitude operations. This indicates that the virtual calibration built is successful. © 2020, Editorial Office of the Transaction of CSICE. All right reserved.