Abstract
The calibration and optimisation of direct injection gasoline engines is very complex, with five to seven control variables being responsible for the development targets of fuel economy, emissions and driveability. The conventional approach to EMS mapping was to divide the calibration problem into many sub-tasks, which required a high communication effort between the teams. In addition, a great number of measurements for each parameter combination had to be made. The approach described in this paper shows how online design of experiments and modern engine modelling techniques can be used to simplify the EMS calibration and optimisation process.