Transient NOx Estimation using Artificial Neural Networks

The reduction of exhaust emissions from any kind of vehicle is one of the key challenges for the automotive industry. Therefore a lot of technologies have been developed to reduce the engine out emissions as well as the tailpipe emissions by exhaust aftertreatment. For the accurate control of engine and exhaust aftertreatment, a lot of sensors have been introduced to the latest vehicles which increased the cost accordingly. This paper presents the development of a virtual NOx sensor to estimate the engine out NOx emission of a diesel engine based on a neural network. This information can be used for both, the control of NOx aftertreatment systems as well as the combustion control of the engine. The validation of the virtual sensor has been done under dynamic conditions using different test cycles.