Eventos Anais de eventos
COBEM 2017
24th ABCM International Congress of Mechanical Engineering
MULTI-OBJECTIVE DIFFERENTIAL EVOLUTION AND NSGA-II OPTIMIZATION ALGORITHMS COMPARISON WHEN APPLIED TO A SINGLE-CYLINDER INTERNAL COMBUSTION ENGINE
Submission Author:
Germano Menzel , PR
Co-Authors:
Stephan Hennings Och, Luís Mauro Moura, Germano Menzel
Presenter: Germano Menzel
doi://10.26678/ABCM.COBEM2017.COB17-2079
Abstract
In this work, two multi-objective optimization methods were applied to a single-cylinder compression ignition internal combustion engine. The methods are NSGA-II (Non-dominated Sorting Genetic Algorithm - II) and MODE (Multi-objective Differential Evolution). The exhaust and inlet valves opening and closing angles (valve timing) were chosen as the design variables, and the thermal and volumetric efficiencies were chosen as the decision variables. The simulation and optimizations were performed for different engine speeds, to better understand the process as a whole. The zero-dimensional model is used for the combustion in the chamber, solved by the fourth order Runge-Kutta method, and the rate of energy release is given by Wiebe’s function. The valves’ lift curve is described by a parabolic model. A predictive one-dimensional model is used for the flow in ducts, whose fluid is considered a perfect compressible gas, and the flow 1D, unsteady and non-isentropic. The physical laws of mass, momentum and energy conservation are applied and this hyperbolic system is solved by two step Lax-Wendroff method with TVD flow control. The boundary conditions are applied on the method of characteristics on duct’s open ends and flowing through valves. Both optimization methods performed well and according to recommendations from the literature, their results were very similar, the main difference is the computational usage being a little lower for the NSGA-II. The decision variable that most affects the efficiencies is the exhaust valve opening angle. Engines at lower speeds require no valve timing optimization for their efficiencies are not highly affected by them.
Keywords
Multi-Objective Optimization, Internal combustion engine, Volumetric Efficiency, Thermal Efficiency, valve timing
