An artificial intelligence strategy for multi-objective optimization of Urea-SCR for vehicle diesel engine by RSM-VIKOR

Urea selective catalytic reduction (SCR) is one of the effective technologies for controlling automotive nitrogen oxide (NOx) emissions. In this paper, a three-dimensional computational fluid dynamics (CFD) simulation model of a urea injection system including a four-channel cyclone mixer was developed. Based on the CFD database, the relationship between the decision variables and the optimization objectives was analyzed. Then, the mathematical relationship between the decision variables and the optimization objectives was established using the response surface method (RSM), and the optimal combination of solution sets for the decision variables was obtained. The entropy weighting method was used to assign weights to the optimization objectives, and the VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) method was used to find the combined optimal solution among multiple objective solutions. Compared with the original parameters, the NH3 uniformity of the optimized SCR system increased by 8.4 %, the NH3 concentration increased by 81.74 % and the wall film mass decreased by 75.2 %.