4 Stroke Gasoline Engine Performance Optimization Using Statistical Techniques

The engine designer has to find novel methods to optimize the engine efficiency faster as the engine development cycle is getting shortened due to the continuous growing market demands. Engine optimization involves fine tuning of the various engine parameters and conducting a large number of tests on actual engine test bed.

In this paper, modern techniques that have been used to optimize a small 4stroke air-cooled engine performance have been described. The engine has been modelled using one-dimensional thermodynamic engine modelling software (AVL-BOOST). Design of experiments (DoE) tools have been used to optimize the engine variables. The input parameters form an orthogonal array of L27 matrix and the out put characteristics of the engine (responses) have been predicted by using BOOST software. This design matrix has been used to study and optimize thirteen factors in three levels (3¹³). The results have been analysed to find out optimum factor/level combination for the desired performance characteristics and limited number of parts have been built and tested to validate the predicted results. This methodology significantly reduced the engine development time, and the designer had a valuable insight into the engine tuning parameters which reduced the design iterations.