Research On Application Of Predictive Control To Air-Fuel Ratio Control Of CNG Engine Based On DSPACE

As a kind of new clean source of power, Compressed Natural Gas (CNG) is anideal alternative fuel of engine. Due to Air-Fuel Ratio (AFR) has a closerelationship with engine drive-ability, fuel economy and exhaust emission level,AFR control plays an important role in the CNG engine control.Because the oxygen sensors are mounted in the exhaust pipe, as well as thedelay of injection systems and the engine cycle, measuring of AFR signal isdelayed in the CNG engine. In the mean time, Engine is highly nonlinear andmultivariable, engine parameters will change with engine speed and load.Predictive control is a new type of computer control method. As the multi-stepprediction, feedback adjustment and rolling optimization are used in this algorithm,the control effect of AFR in different engine operating mode will be better thantraditional feedback control.The experiment platform of CNG engine control system is firstly built basedon the MicroAutoBox, and then the control strategy of the CNG engine isestablished using Matlab/Simulink, including injection control, ignition control andelectronic throttle control.Aiming at the nonlinearities of CNG engine and the transmission delay of theair-fuel ratio, a generalized predictive control (GPC) method based on ControlledAuto Regressive Integrated Moving Average model (CARIMA) and a neuralnetwork predictive algorithm based on neural-network model to air-fuel ratio aredesigned in this paper. These two control algorithm were built under theMatlab/Simulink environment and implemented on NQ150N CNG engine withMicroAutoBox hardware of dSPACE. The actual experiments results showed thatthe predictive control algorithms have better steady performance and better controleffect under transient conditions compared with a typical PID control algorithm.