Model Predictive Control considering disturbances in diesel engine air intake systems

Abstract

Model Predictive Control (MPC) is applied in order to improve the transient response in diesel engine air intake systems. In actual engines, a steady state error is generated due to disturbances such as the fluctuation of air intake and fuel injection amounts and the modeling error of the systems. This causes target tracking deterioration. In order to cancel the steady state error, we evaluate three compensation methods in simulations and actual transient response tests: a) a method of incorporating the constant disturbance model, b) a method of explicitly incorporating an integrator, and c) a method using a disturbance observer in motion control. We clarify the configuration of MPC effective for improving target tracking performance.

Model predictive emissions control of a diesel engine airpath: Design and experimental evaluation

“…Prior research on pressure and/or airflow setpoint tracking using linear MPC often uses switching or piecewise linear models to cover the engine operating range 5‐9 . This approach may lead to chattering and poor performance; multiple estimator/controllers running in parallel are often needed to overcome these issues, increasing the computational footprint of the controller.…”

Dynamic feed-forward control aided with active disturbance rejection for boost pressure and mass air flow control of diesel engines

“…Disturbance observer based control [14,15] and rate-based control [16] were adopted as the counter measures. The control-Lyapunov function (CLF) based control [17][18][19] was also proposed, among others, but the complicated construction [20] limits its appeal in practical applications.…”