UC San Diego

Homogeneous charge compression ignition (HCCI) engines have the benefit of high efficiency with low emissions of nitrogen oxides and particulates. These benefits are due to the autoignition process of the dilute mixture of fuel and air during compression. However, because there is no direct ignition trigger, control of ignition is inherently more difficult than in standard internal combustion engines. This difficulty necessitates that a feedback controller be used to keep the engine at a desired (efficient) setpoint in the face of disturbances. Because of the nonlinear autoignition process, the sensitivity of ignition changes with the operating point. Thus, gain scheduling is required to cover the entire operating range of the engine. Controller tuning can therefore be a time intensive process. With the goal of reducing the time to tune the controller, we use extremum seeking (ES) to tune the parameters of various forms of combustion timing controllers. Additionally, in this dissertation we demonstrate how ES can be used for the determination of an optimal combustion timing setpoint of an experimental HCCI engine. The use of ES has the benefit of achieving both optimal setpoint (for maximizing the engine efficiency) and controller parameter tuning tasks quickly. The lack of a direct combustion trigger makes control of combustion timing during transients especially challenging. To aid in HCCI engine control during transients, we have developed a model that can be used to derive a controller for a thermally-managed, gasoline and natural gas fueled HCCI engine. The model uses an ignition threshold derived from detailed chemical kinetic simulations of HCCI engine combustion to provide an estimate for the combustion timing. The ignition threshold is a function of both temperature and pressure. An estimate of the residual gas fraction from the previous cycle can also be obtained, which is essential information due to the strong temperature sensitivity of HCCI ignition. This model allows the synthesis of nonlinear control laws, which can be utilized for control of an HCCI engine during transients