Optimization of diesel fuel injection strategies through applications of cooperative particle swarm optimization and artificial bee colony algorithms

Abstract

New engine hardware and injection strategies allow modern engines to meet stringent emissions regulations but can require extensive engine testing to identify optimum operating points. Swarm intelligence algorithms, which do not require knowledge of the search space gradient, can provide a short cut in finding optimum operating parameters in reduced experimental time than a traditional design of experiments study. In this paper, a modified artificial bee colony (ABC) algorithm and a cooperative particle swarm optimization (CPSO) algorithm are applied to triple and quadruple injection routines. The optimization was applied directly to the operation of a four-cylinder turbocharged production diesel engine operating at a high exhaust gas recirculation rate and medium load. Six and eight variable optimizations were carried out for triple and quadruple injection schedules, respectively. In experimental testing, the cooperative particle swarm optimizer significantly reduced soot and carbon monoxide emissions in only 84 engine tests when using a pilot-pilot-main-post schedule. The modified artificial bee colony algorithm took 176 engine tests to optimize a pilot-main-post triple-injection schedule and was not applied to the four-injection routine.

Adaptive Neural Network Iterative Sliding Mode Course Tracking Control for Unmanned Surface Vessels

“…Particle swarm optimization (PSO) [45] is proposed in 1995. e PSO algorithm is applied to search for the optimal solution in M dimensional search space. e population X � (X 1 , X 2 , .…”

Deep reinforcement learning to discover multi-fuel injection strategies for compression ignition engines

“…Many researchers in the literature have used machine learning (ML) techniques for various aspects of engine research such as simulation, 19 modeling, [20][21][22][23][24][25] optimization, [26][27][28][29][30][31][32] and control. [33][34][35][36][37][38] Traditional reinforcement learn-ing (RL) has been used for hard-coal combustion processes in a power plant, 39,40 for spark ignition and injection timing, 41,42 for energy management strategies for hybrid-electric vehicles, [43][44][45] and for the control of the air-fuel ratio 46 and spark engine exhaust gas recirculation (EGR) operation.…”

Deep reinforcement learning for dynamic control of fuel injection timing in multi-pulse compression ignition engines

“…25 Review articles and books on the subject include. 7,[26][27][28] Machine learning techniques have been deployed for applications in automotive engines, including simulation, 29 modeling, [30][31][32][33][34][35] optimization, [36][37][38][39][40][41][42] and control. [43][44][45][46][47][48] Recent work on spiking neural networks has been proposed for dilute combustion with EGR, 49 leveraging previous work in control of such systems.…”