Application of CFD and ANN in Predicting the Flow Nature of Flue Gas in the Catalytic Converter

Abstract

The main objective of this research is to improve the performance of the catalytic converter by achieving uniform flow distribution of petrol engine exhaust gas using a distributor plate in the upstream section. The flow analysis in the catalytic converter is conducted using computational fluid dynamics (CFD) Solid works 2017 simulation software. Totally, flow simulation is carried out for twelve different cases. It has been observed that a 140 mm diffuser and distributor plate positioned 117 mm from the entrance delivers the optimal flow trajectory. This catalytic converter highly exhibits a non-linear correlation between the input and output variables. Results of such simulation study are then considered as an input for artificial neural network (ANN) model, an artificial intelligence (AI) technique. Results predicted by the ANN model are compared with CFD output. Compared outputs are in good agreement with each other. To sum up, we conclude that by introducing the distributor plate in the flow regime of the conventional catalytic converter, we can accelerate the rate of uniform flow distribution.