Application of Improved Grey-Least Squares Combined Forecasting Model in Gasoline Engine Intake Air Mass Flow

Abstract

Increasing the forecast precision of the intake air mass flow of gasoline engines by reducing the effect of hysteresis in the mass air flow sensor, an improved gray forecasting model is constructed by adding a weakening buffer operator and variable background values based on traditional GM(1,1) model, and then combined with the least squares to form a combined forecasting model. In the weakening buffer operator, the background value weighting coefficient is determined based on collected sample data from the mass air flow sensor, utilizing the particle swarm optimization (PSO) algorithm. Some of test data of single-cylinder gasoline engine acceleration condition, deceleration condition and motorcycle China IV emission are simulated. The results show that the improved grey-least squares combined forecasting model can accurately predict the intake air mass flow as well as come up with a solution of prediction mutation, and the prediction accuracy is improved.