Analysis On Oil Film Characteristic And Study On AFR Control Strategy Under Transient Condition Of Gasoline Engine

Transient operation condition Air-fuel ratio (AFR) affects emission performanceand fuel economy performance of gasoline engine crucially. With AFR closes to thetheoretical value, the air-fuel mixture burns adequately, exhaust contains less pollutiongas, at the same time, the three-way catalytic converter (TWC) works with the highestefficiency. Control of AFR to close to the theoretical value under transient condition issignificant to promote the energy conservation and emission reduction. This thesisstudies AFR control strategies of gasoline engine under transient condition, establishesa hybrid intelligent control system based on fuzzy control theory and RBF neuralnetwork theory.Takes HL495Q as the research object, this thesis establishes the simulation modelof gasoline engine based on Simulink, simulates with different throttle percentagevariations and variation rates, analyses the dynamic oil film characteristics of gasolineengine under transient condition. The simulation results show that throttle percentagesaltation is the main reason for oil film dynamic balance break and AFR deviation,throttle percentage variation rate determines deviation amplitude directly. Afeed-forward compensation controller is built up to compensate fuel oil injection. Thecontrol simulation results show that AFR under transient condition is controlled to betheoretical value accurately. And the compensation results of fuel oil injection are usedto analyse the dynamic oil film characteristics further.According to feed-forward compensation control results, the AFR fuzzy controlrules are formulated and AFR fuzzy control system is established. The fuzzy contrlosimulation results show that, without considering the control flaw because of oxygensensor transmission delay, AFR deviation under transient condition is controlled within±0.06, the maximum deviation ratio is±0.41%. In order to eliminate the oxygen sensortransmission delay defect, creates two RBF neural networks respectively for AFRprediction under acceleration and deceleration condition. After training, underacceleration condition, AFR prediction MAE(Mean Absolute Error) of neural networkis 0.0032 and MRE(Mean Relative Error) is 0.021%; under deceleration condition,MAE is 0.0027 and MRE is 0.019%.Finally, a hybrid intelligent control system based on fuzzy control and RBF neuralnetwork is established to control AFR deviation under transient condition of gasoline engine. It takes throttle percentage as input and AFR as output, RBF neural networktransmits the predictive value of current time AFR to fuzzy controller, fuzzy controllertransmits the compensation value of fuel oil injection to ECU of gasoline engine, AFRdeviation is controlled because of fuel oil injection flow adjustment. Hybrid intelligentcontrol system needn’t identify oil film parameters, and doesn’t exist AFR feedbackdelay flaw. The simulation results show that hybrid intelligent control system controlsAFR deviation within±0.1, the maximum deviation ratio is±0.68%. TWC wouldwork in the most efficient conversion window(φ_a=1±0.01)under whole transientcondition.