Prediction of real-time NO based on the in-cylinder pressure in Diesel engines

Abstract

Nitric oxides are one of major limiting factors in developing Diesel engines because of emission regulations. There are two methods to reduce tailpipe NOx emissions: reducing the engine-out NOx emissions and using after-treatment systems. Therefore, the control of both in-cylinder combustion and after-treatment systems are important in reducing the NOx emissions; to accomplish this goal, prediction of engine-out NOx is essential.In this study, a real-time nitric oxide prediction model was developed based on the in-cylinder pressure and on data available from the ECU. As computational fluid dynamics can describe the process of NO formation which is not directly obtainable from experiments on a physical basis, the NO formation model was developed based on both the analysis of CFD results as well as on a physical model. Furthermore, the in-cylinder pressure is used to predict the amount of NO formation under various engine operating conditions as the pressure reflects the change in the combustion characteristics. The prediction model consisted of a simple calculation process; therefore, the model could predict the cycle-by-cycle NO in real-time. The validation results show that the model presented can predict engine-out NO well; thus, this model can be applied to engines and after-treatment systems as a useful tool to control the engine-out NO without the use of an NOx sensor. In addition to being a virtual NO sensor, the prediction model can be applied to 1-D simulations, such as GT-SUITE and AMESIM, and demonstrate improved NO prediction results as the model is able to predict the NO level as same standard as the 3-D CFD simulation.