Correlation and optimisation of thermodynamic model parameters with experimental data to predict emissions from a diesel engine

Abstract

The biggest challenge in engine development is emission regulation today, which gets more stringent with time. Computer aided engineering (CAE) has also increasing importance together with increasing potential of processors for easy and quick calculations and also it is less expensive compared to conducting tests. Therefore it is highly important and challenging to predict emissions by using computer models and simulations precisely. For diesel engines, NOx and particle (soot) emissions have the highest importance. 3D CFD of combustion systems and phenomenological combustion models can provide such information. Phenomenological combustion models have the advantage of much lower runtimes compared to 3D CFD and can be used for wide variety of engine operating points for prediction. Such a phenomenological diesel combustion model is used and correlated with limited amount of test data for the prediction of emissions and is compared to a larger set of emission test data to see its performance. Assessments are made how well the emissions can be predicted by using such a model. The sensitivity of the model against different start of injection and exhaust gas recirculation values is tested. NOx model show good accuracy whereas soot model needs some modification. Both emission models can provide some insight in terms of direction (increase-decrease).