Model-based Control of Automotive Engines and After-treatment Devices

Abstract

As environmental requirements on automotive vehicle emissions have steadily increased over the last decades, embedded control technology to perform real-time management of the driver's requests has appeared as a key technology and has become ubiquitous. Today, every subsystem that is found under the hood is controlled, most often in closed-loop mode. A key factor in the development of a new engine or a new vehicle is the duration of its design cycle. The design and tuning of all the intricate control loops is very time-consuming. One of the main reasons for this is that tuning procedures are often used to compensate for certain un-modeled effects such as neglected couplings of dynamic variables. In this talk, we explain how models can be used to reduce expensive calibration time significantly. As will be shown in the light of several experiments (involving airpath, fuelpath, and after-treatment systems), simple models bring effective and practical solutions involving mature tools from nonlinear control theory.