Research On Model Based Air-fuel Ratio Control Of Spark-ignition Engines

With the sustained growth of automotive production and sales,countries around the world have enacted increasingly stringent emission regulations to limit the automotive exhaust emissions.The emission control of spark ignition(SI)engines is based on the electronic fuel injection technology in combination of the three way catalytic converter(TWC).The engine control system needs to maintain the air-fuel ratio(AFR)to be close to the stoichiometric value,so as to maximum the TWC efficiency of CO,HC and NOx in the exhaust gas at the same time.As a consequence,accurate AFR control is significant to the limitation of SI engine exhaust emission and it is also indispensable with the development of bio-fuel applications and lean-burn technology in automotive engines.Therefore,the research on AFR control of SI engines is valuable and meaningful.However,the AFR control in SI engines is currently based on the data MAP table accomplished by the arduous engine calibration tests.The AFR control performances should be enhanced and there is no adaptive feature in the AFR control strategy to handle the parameter variations.The inherent nonlinearities,time delay,time-varying characteristics,and uncertainties in the SI engine operating process result the AFR control system design still to be a challenging task.It is reasonable to design the AFR control strategy based on the mathematic modelling of the AFR dynamics in the engine.Meanwhile,the AFR control strategy needs to be implemented by the engine electronic control unit(ECU)which can only be produced by the foreign auto part monopolized companies.So,in this paper the research of model-based AFR control technology was carried out by the discussing of control object,control strategy and controller implementation.The main research achievement is listed as following content:(1)The theoretical analysis on the effect of varying AFR on SI engine performances,fuel economy,emissions and TWC conversion efficiency was discussed and the experimental tests on the SI engine fueled with biofuel blends,such as bioethanol and bio-butanol and ABE(Acetone-butanol-ethanol),was performed.By the investigation of the differences in AFR control using biofuel,a unified control aim and the key implementation steps were proposed to guide the AFR controller design.(2)In order to provide the theoretical references for AFR control strategy design,the AFR dynamical model was proposed based on the mean value engine model(MVEM)which consisted of intake air mass dynamic model,fuel film mass flow dynamic model and crank shaft dynamic model.Then,the enDYNA engine simulation model was improved by adding the AFR dynamics,and the virtual simulation platform was constructed for the AFR control strategy implementation.(3)The model-based AFR control strategy was carried out based on the engine operating conditions and the control aim.During the transient state,the AFR was control by the feedforward fuel film compensator in the open loop.At the steady state,the combination of the feedforward compensator and self-tuning feedback controller were selected to regulate the AFR based on a UEGO sensor measurement.Furthermore,a Kalman filter observer was implemented to estimate the intake air mass in the cylinder which could not be measured by sensors and was crucial for the fuel mass calculation for the AFR control.The simulation and experiment results of applying the proposed method showed the AFR controller exhibited excellent performance.(4)To face the challenge of unknown modeling error,noise,time-varying parameter,delay,disturbance,nonlinearity and uncertainties in the AFR control loop,a new approach of adaptive feedback AFR controller design was proposed using the self-tuning multi-step predictive method.Using the CARIMA model to represent the AFR dynamics,the proposed controller could estimate the model parameter by on-line identification and regulate the AFR output deviation adaptively.Finally,four simulation tests were carried out to evaluate the closed-loop system performances against varying structure parameters,noise disturbance,time-varying model parameters and joint simulation with enDYNA model.The results showed that the output AFR could track the referenced AFR value at different operating conditions,also,the proposed control method has good robustness and adaptivity.(5)The hardware development and hierarchical software design of the ECU were completed by adopting the V-shaped development process.After the application layer control logic modeling and simulation tests,the controller software integration process was realized by the script file for automatic code generation and compiling.Also,in order to implement hardware-in-loop(HIL)tests,the HIL simulation platform was constructed based on the xPC-Target and enDYNA simulation tool and a signal interface matching circuit which could effectively reduce the computation was proposed.(6)Based on the analysis of the CAN Calibration Protocol(CCP),the CCP modularization driver was designed for ECU.Meanwhile,the calibration test system was set up for on-line calibration and parameter monitoring using the automatic generated A2 L file.Experiments were conducted on a SGMW-B15 engine test bench to evaluate the proposed ECU hardware and AFR control strategy.The results showed that the AFR controller had a considerable performance compared with the OEM ECU and was effective to track the given AFR reference value under the disturbance caused by the noise,delay,time-varying parameter and uncertainties.