ANN metamodel assisted Particle Swarm Optimization of the performance-emission trade-off characteristics of a single cylinder CRDI engine under CNG dual-fuel operation

摘要: Abstract The encouraging potential of a CRDI coupled CNG application in providing superior PM–NO x –BSFC trade-off footprint as compared to the BDO single cylinder diesel engine, provided strong motivation undertake an in-depth exploration parametric design space investigate possibility obtaining emission-performance characteristics experimental endeavour. To obviate infeasibility full factorial based approach intended exploration, offline metamodel assisted optimization study within same range input/control variables was considered viable strategy. FIP, CES and engine load were chosen control for study. An Adaptive Merit Function (AMF) constituted objective function be optimized encapsulating simultaneous necessity meet EPA Tier 4 PM–NHC emission mandates lower BSFC expectations consumer. ANN developed correlate with while Latin-Hypercube DoE scheduler provide initial population sample iteration sequence. PSO algorithm due its inherent simplicity efficiency yielding ensured convergence at significantly less computational cost. Validation obtained optimal solutions steps 25%, 50%, 75% yielded 9.9%,16.4%, 10.5%, 2.7% higher AMF than that by best CRDI–CNG operation corresponding respectively during actual experimentation. all such values simultaneously honoured PM NHC mandates.