Calibration and assessment of the GT-Power “DIPulse” predictive model for a light duty compression ignition engine

The objective of this master thesis is to create the virtual test rig for a light duty compression ignition engine by means of the zero/mono dimensional numerical modeling software GT-Suite. In this paper the predictive combustion model called “DiPulse” was assessed. Taking into account afford mentioned points, one dimensional simulation tools have proven to be essential in the process of confiding with strict regulations that determine the engine building and calibration operation. The engine under investigation in this paper is light duty Diesel, equipped with single stage variable geometry turbocharger, displacement of 1.6 liter, compression ratio of 16:1 and common rail fuel injection system. In order to perform the calibration of previously mentioned engine the data set of 41 operating points was provided. The calibration procedure starts with a preliminary calibration using non-predictive combustion sub-model called Cylinder Pressure Only Analysis (CPOA). CPOA required the measured pressure traces which are then used to obtain the fuel burn rate. Imposed burn rates from CPOA are then treated as an input in the DiPulse calibration. DiPulse calibration is carried for smaller number of operating points that represent the engine map in case of levels brake mean effective pressure (bmep) as well as exhaust gas recirculation (EGR). Calibration is composed of the optimization process during which parameters influencing the combustion process are investigated in order to minimize the factors such as Improved Burn Rate RMS error and IMEP% error. Having reach the satisfactory level of accuracy the NOx emission calibration was performed in order to match the emission levels provided with the data set. Once the calibration was completed, the validation process was performed in which operating points not considered in the DiPulse were evaluated to asses the predictive capabilities of the combustion and emission models. The final model can be implemented in subsequent studies which may include the influence of alternative fuels or fuel blends on the represented engine. In conclusion the DIPulse multizone predictive combustion model developed by Gamma Technologies was vastly employed to perform engine calibration process. The resulting virtual test rig model can be employed to asses the influence of different engine configurations and hardware components hence shortening the experimental activity required to accomplish projects goals.