Abstract:

A gasoline engine torque prediction model was designed based on Sigmoid-ARX algorithm to
improve the accuracy of real-time estimates in torque control strategy. Model parameters were estimated by
using the Akaike's Final Prediction Error (FPE) method and the normalized mean square error (NMSE) method
through identification data set, such as valve opening degree and engine speed. A model predictive controller
(MPC) was designed based on the torque model, tested, and verified. The results show that the relative error
is less than 5% within the full range of operating conditions having a precise control with a torque overshoot
of about 6%, and the expected torque output response is 150 Nm within 3 s. Therefore, the control strategy is
effective to realize accurate control for gasoline engine torque.

Key words: gasoline engines, control strategies, torque estimation, model predictive controller (MPC), Sigmoid-ARX algorithm