Research On Fuel Air Ratio Control Of Si Gasoline Engines

The fuel air ratio(FAR) control is to maintain the ratio between the fuel to air mass flow into cylinders within a narrow range, so that the three-way catalyst performance can be maximized to reduce emissions as much as possible. Because of the great theoretical and engineering significance, FAR control has attracted increasing attentions all over the world.This dissertation, to further improve FAR control performance, mainly deals with the theoretical issues of FAR control in the presence ofvarying time delay, which are also closely related to engineering applications, e.g., sensor aging and wall-wetting effects. The dissertation is divided into five chapters. The state of the art of FAR control is presented, and followed by the in-depth analysis of factors influencing FAR control performance. A time-delay-dependent gain-scheduled memory controller is presented to mitigate the influence of varying time delay. Considering the non-ignorable wall-wetting effect during transients, an observer-based time-delay-dependent controller is proposed for FAR control. Finally, a robust FAR controller is designed to deal with uncertainties and varying time delay caused by the aged oxygen sensor.Chapter 2 presents the modeling for FAR of SI gasoline engine following the work fluid. The change rate of fuel film is modeled regarding the existence of wall-wetting effect. Besides, with respect to the evaporative emission system and estimation errors for air mass flow, the essential reasons for time delay are analyzed. A state space representation, being capable to mimic the behaviors of wall-wetting effect, evaporative emission and varying time delay, is addressed for the SI gasoline engine accordingly. To investigate accuracy of the proposed model, as well as the influencing factors on it, an improved simulation system based on en DYNA is built. In addition, simulations and comparisons are conducted under various conditions, providing that the influence of evaporative emission on FAR mainly arises from the fuel in vapor, the estimate error of air mass flow affects FAR without relation to operation conditions, the influence of wall-wetting effect becomes obvious during transients.Chapter 3 mainly addressed a novel controller regarding varying time delay, which is usually ignored or fixed at its maximum in existing studies. This varying time delay,modeled as a function of engine speed and air mass flow into cylinders, is reserved in the FAR system to achieve improved control accuracy and less conservation. A timedelay-dependent gain-scheduled controller is then designed for such a system. Validation results, obtained based on the above mentioned simulation system, illustrate the promising performance of the proposed controller for disturbance rejection and FAR tracking control.Chapter 4 gives an observer-based controller design method, considering both varying time delay and wall-wetting effect.As stated above, the wall-wetting effect mainly affects FAR control during transients, e.g., fast opening and closing conditions of throttles,and becomes non-ignorable. To obtain the immeasurable states of wall-wetting behavior,a gain-scheduled observer is designed. The sufficient conditions for global asymptotic stability of the observer error dynamics is systematically derived. An observer-based controller is presented for FAR control accordingly. Simulations demonstrate the outstanding performance of the proposed controller during both transients and static conditionsChapter 5 concerns with the controller design issue in the presence of oxygen sensor aging, which results in parameter changes anddeteriorates the FAR control performance significantly. For convenience, an experimental model, identified based on input-output responses, is employed. In further consideration of the sensor aging phenomenon, the model can be regarded as a first-order system with unknown bounded time delay and parameter uncertainties. For this system, a robust controller is designed, of which the gain can be adjusted with engine operation conditions. Simulations, conducted with aged oxygen sensors, show the superior performance of the proposed controller for both FAR tracking and disturbance performances.