Front-loading: Virtual Validation and Calibration on LABCAR

The advent of BS6 coupled with RDE emission norms has increased the development efforts and costs due to the shear amount of testing and validation on real engines and vehicles which are necessitated by these stringed norms. Front-loading of tasks by moving actual vehicle and engine tasks on to virtual setup, will reduce the development efforts and costs significantly. This front-loading of tasks on to a LABCAR would need real time and highly accurate plant models, tools to parameterize these plant models and accurate data driven models to predict dynamic parameters like emissions. In this collaborative work between MSIL and ETAS India, ETAS VVTB + ICE plant models were parameterized with the data generated on engine test with ASCMO Global DoE test plan by using ASCMO MOCA. These plants models were optimized to an accuracy level of more than 95%. The ASCMO Global test plan also ensures the coverage of data points across the entire engine operating space. The optimized plant model parameters are integrated on ETAS LABCAR and necessary fine tuning is carried out on the LABCAR. The results show, for MIDC cycle the accuracy of simulation torque and airflow results are greater than 95% and fuel consumption simulation has an error less than 6%. The coast down simulations when compared with actual has deviation of less than 1%. The ASCMO dynamic emission models also had good accuracy levels, for e.g. NOx for MIDC cycle has R2 of 0.95.