A.K.M. Mohiuddin, A.A.I.S. Ashour, and Y.H. Shin (Malaysia)
Optimization, Valve timing, Multi-objective genetic algorithm, GT-Suite, and modeFrontier
This paper aims to demonstrate the effectiveness of Multi Objective Genetic Algorithm Optimization and its robust practical application on the automobile engine valve timing where the variation of performance parameters required for finest tuning to obtain the optimal engine performances. The primary concern is to acquire the clear picture of the implementation of Multi-Objective Genetic Algorithm and the essential of variable valve timing effects on the engine performances in various engine speeds. Majority of the research works in this project were in CAE software environment and method to implement optimization to 1D engine simulation. The paper conducts design optimization of CAMPRO 1.6L (S4PH) engine valve timing at various engine speeds using multi-objective genetic algorithm (MOGA) for the future variable valve timing (VVT) system research and development. This paper involves engine modeling in 1D software simulation environment, GT-Power. GT-Power is one of the CAE tools available in GT-SUITE offers the only true "virtual engine/power train" tool, capable of integrated simulations of the total engine and power train system. The GT-Power model is run simultaneously with mode Frontier to perform multi-objective optimization. Multi-objective Genetic Algorithms (MOGA) are an extension of Genetic Algorithms (GA) that does not require multiple objectives to be aggregated to one value. Instead of static aggregate such as a weighted sum, MOGA dynamically determine an aggregate of multiple objective values of a solution based on its relative quality in the current population, typically as the degree to which the solution dominates others in the current population. By using the MOGA optimization approach in modeFrontier, multi-objective optimization could be done by coupling with several software environment; this does not require any modification of the original model. Thus, it is able to provide professional, reliable and accurate solution. In preprocessing of optimization, modeFrontier Response Surface Method (RSM) is able to model the behavior of engine performances corresponding to the change of design variables.
Important Links:
Go Back
