Eventos Anais de eventos
ENCIT 2016
16th Brazilian Congress of Thermal Sciences and Engineering
ANALYSIS OF MULTI-OBJECTIVE OPTIMIZATION ALGORITHMS APPLIED TO INTERNAL COMBUSTION ENGINES
Submission Author:
Germano Menzel , PR
Co-Authors:
Luís Mauro Moura, Stephan Hennings Och
Presenter: Luís Mauro Moura
doi://10.26678/ABCM.ENCIT2016.CIT2016-0691
Abstract
Thermal efficiency is used to measure the effectiveness of an engine’s combustion process and the volumetric efficiency is used to measure the effectiveness of an engine’s induction process, and is highly influenced by the intake system. Maximizing these variables would optimize the whole operation of the engine as well, highlighting their importance on power, torque and emissions. The aim of this paper is to optimize these efficiencies by altering the valves’ timings through two multi-objective optimization methods. A mono-cylindrical engine with a zero-dimensional, two-zone spark ignition combustion model was described for the combustion chamber due to the complexity of the simulation and the high amount of computational usage required for the iterations on the optimizations. A direct search model called “Random Search Approach” will be compared to a more complex method, an elitist multi-objective evolutionary algorithm called “Non-dominated Sorting Genetic Algorithm”. When a large initial population size is applied on the former, it only reproduces a small amount of Pareto Front solutions. On the other hand, the latter, although requiring a higher computational usage, not only results in a more crowded Pareto Front but it generates better results. Both efficiencies, or decision variables, rise from 1000 to 3000 rpm then decrease at 4000 rpm, meaning that the optimal engine speed should be between 2000 and 4000 rpm. There is a tendency of high efficiencies in higher exhaust valve opening angle values due to the pressure pulsing in the exhaust process, and a rise on the thermal efficiency with the increase of the inlet valve closing angle is due to the fluid’s inertia or RAM effect. The results confirm the early assumptions that the second algorithm has better effectiveness (higher efficiency values and more solutions), and that it should be used for the simulation of an engine, where a small gain on each efficiency has large effects on other operational variables, such as torque, power and emissions.
Keywords
Multi-Objective Optimization, Internal combustion engine, Volumetric Efficiency, Thermal Efficiency, valve timing, Multi-Objective Optimization, Internal combustion engine, Volumetric Efficiency, Thermal Efficiency, valve timing
