Simulation Investigation On Auto-ignition And Combustion Mode Of Mixture In A Confined Vessel

Knock and super knock are the main bottlenecks to further improve the thermal efficiency of spark ignition engines.The occurrence of knock is related to the auto-ignition of the mixture.The mechanism of auto-ignition and knock in a confined vessel is complex,and different combustion modes may lead to different knock intensities.However,most of the current studies on auto-ignition and combustion mode focus on one-dimensional simple conditions,and influencing factors such as temperature stratification and turbulence have not been fully considered.The mechanism of auto-ignition and combustion mode is still unclear,and there is also a lack of effective prediction methods.Based on numerical simulation methods and machine learning methods,this paper studied the effects of temperature stratification and turbulence intensity on auto-ignition and combustion mode,the prediction of auto-ignition and combustion mode,and the application of rapid auto-ignition in combustion.This paper can provide theoretical guidance for more reasonable organization of engine combustion process and knock control.Firstly,the effects of temperature stratification and turbulence intensity on auto-ignition and combustion mode are explored.The results show that with the change of the initial conditions,different combustion modes of auto-ignition are observed,including supersonic autoignitive deflagration,direct detonation,deflagration to detonation transition and subsonic autoignitive deflagration.Under the condition of temperature stratification,the auto-ignition reaction front acceleration and the transition from deflagration to detonation are more likely to occur in certain regions characterized by intermediate temperature gradient and intermediate ignition delay time.Turbulence promotes the transition of the auto-ignition reaction front from deflagration to detonation.Turbulence increases the propagation speed of subsonic autoignitive deflagration.Turbulence breaks the large intermediate temperature gradient region into several small intermediate temperature gradient regions.Secondly,the auto-ignition and combustion mode were predicted based on machine learning methods.The results show that when the partial information of temperature field,temperature change field and temperature gradient field is used as the input feature of the neural network,the temperature field and the mass fraction field at the next moment can be accurately predicted.Although the heat release rate field and the pressure field have some prediction errors.The prediction error is related to the influence of the gradient region on the auto-ignition reaction front,the interaction between the two relatively close auto-ignition reaction fronts,and the pressure wave disturbance.The support vector machine can get accurate prediction results with fewer training samples,and consider the influence of multiple factors such as hot spot interaction and pressure wave disturbance on the combustion mode.Finally,the application potential of rapid auto-ignition in turbulent jet ignition system is explored.The results show that under normal conditions,the vortex formed in the main chamber can promote the propagation of jet flame.When auto-ignition occurs in the pre-chamber,the flame propagates in the form of spherical flame instead of normal jet flame in the main-chamber due to the strong leading pressure wave.The spherical flame formed under the influence of the strong leading pressure wave has faster propagation speed and higher heat release rate,which is favorable for flame propagation in the lean mixture,but the strong pressure wave caused by auto-ignition may also cause damage to the turbulent jet ignition system.