Multi-objective optimization of performance characteristic of diesel particulate filter for a diesel engine by RSM-MOPSO during soot loading

Abstract

Diesel particulate filter (DPF) is one of the effective technologies for controlling vehicle particulate matter (PM) emissions. In this research, a hybrid multi-objective optimization approach of FGRA-RSM-MOPSO was developed for DPF, with optimization objectives including maximum initial filtration efficiency and minimum pressure drop. Decision variables include diameter, length, wall thickness, porosity, and pore diameter. Firstly, a computational fluid dynamics (CFD) model for DPF was established, and sensitivity analysis of DPF structural parameters was conducted through fuzzy grey relational analysis (FGRA) to screen out key factors affecting DPF trap performance. Then, response surface methodology (RSM) was used to establish the mathematical relationship between key structural parameters, initial filtration efficiency, and pressure drop. Finally, multi-objective particle swarm optimization (MOPSO) is used to optimize the target and select the optimal solution from the Pareto front. Compared with the original DPF, the optimized DPF under standard operating conditions increased the initial filtration efficiency by 46.85% and reduced the pressure drop by 34.88%. The optimization effect is more pronounced under high load conditions.