Evaluation of Fit Algorithms for NOx Emission Prediction for Efficient DeNOx Control of Transient Diesel Engine Exhaust Gas

Abstract

To adequately control the reductant flow for the catalytic reduction of NOx in diesel exhaust, a tool is required that is capable of accurately and quickly predicting NOx emissions from the diesel engine's operating variables. In this paper three algorithms for nonlinear modeling are evaluated: neural networks, the split and fit algorithm of Bakker et al., and a polynomial NARX model, which is linear in its parameters. Measurements were carried out on a transient automotive diesel engine. Each algorithm was able to make excellent predictions, combined with a short computation time (0.3 ms). This makes them very promising tools in automotive NOx emission control. When the model training times are compared, the split and fit algorithm is favored (50 s). It is proven that the algorithms are much more accurate than the frequently used engine map and the linear fit model.