Investigation of a multi-objective optimization tool for engine calibration

Abstract

Since multi-objective optimizations are becoming more important in engine calibration, the paper investigates multi-objective genetic algorithms in the application of engine optimization. Although several multi-objective genetic algorithms have been developed and some have been applied successfully in the automotive industry, it is difficult to determine which multi-objective genetic algorithm outperforms others in finding the set of optimal solutions or Pareto-optimal front in a practical multi-objective optimization problem. Based on some widely used multi-objective genetic algorithms, the paper proposes a combined scheme to deal with the difficulties in finding the optimal solution set during the engine calibration process. In the proposed approach the real-coded representation is employed in the genetic algorithm and the elitist strategy is applied for each multi-objective genetic algorithm used. To assess the proposed approach, two computational examples are given to minimize the brake specific fuel consumption and to maximize the output power torque simultaneously. The results show that the proposed approach is well suited to multi-objective optimization in engine calibration.

Study on the efficiency of a spark-ignition variable displacement and compression ratio engine

“…Moreover, optimisation tools can be combined with simulations. The use of optimisation schemes, such as direct search, numerical methods, and explorative methods, 16 allows multi-objective optimisation 22,23 and the reduction of computational efforts; extensive cases do not need to be simulated when an optimal path is identified.…”

Advantages and applications of transforming empirical model input space with dimensional models

“…Several areas of academic enquiry have emerged in the last decade or so, such as the optimum methods of collecting data for empirical models to be used for calibration, 12–19 online calibration in the test cell, 20 adaptive calibration, 21 fast data acquisition for calibration, 22 transient calibration, 3–5,23,24 calibration of integrated engine-aftertreatment systems, 25,26 empirical modeling methods 27–39 and the computational optimization process. 40–44 Books on model-based calibration addressing the last two areas have started appearing, see, for example, works by Edwards et al, 45 Alberer et al 46 and Röpke et al 47 There is a growing body of literature addressing the application of model-based calibration and DoE techniques for regular engine development, for example, Suzuki et al 48 and Jiang et al, 49 and for unique applications such as noise reduction 50 and dual-fuel calibration 51 and calibration for gas-to-liquid (GTL) fuel vehicle conversion. 52 There even exists literature on the teaching of established engine calibration technology.…”