Modeling and Analysis on Emission Characteristics of Light-Duty Diesel Engine After-Treatment System Based on Neural Network

With the increasing complexity of diesel engine after-treatment systems, the explosion of dimensions has made it difficult for traditional bench test research to be competent for multi-parameter analysis. In this paper, based on the experimental samples derived from a light-duty diesel engine bench test, a neural network model for the emission characteristics of the after-treatment system is established. After correctly training, the R values of the neural networks for gaseous and particulate emission are all close to 1. The influence of the precious metal parameters and structure of diesel oxidation catalyst (DOC) and catalytic diesel particulate filter (CDPF) on diesel engine gaseous and particulate emission reduction efficiency is analyzed through the neural network. The result shows that with the increase of the engine speed under the external characteristic, the particulate emission increases while the gaseous emission reaches a peak at medium speed; as the DOC precious metal loading and the DOC precious metal ratio (Pt/Pd) increase, the reduction efficiency of the gaseous emission increases and the former has a greater impact; as the CDPF precious metal loading increases, the reduction efficiency of the gaseous and particulate emission both increase; as the CDPF precious metal ratio (Pt/Pd) increases, the reduction efficiency of PN decreases; increasing the length-diameter ratio of the DOC improves the reduction efficiency of gaseous emission, while the length-diameter ratio of the CDPF is more sensitive to particulate emission.