Research On Key Technologies Of Combustion Timing Estimation For Hcci Engine With Mixed Fuel Under Complex Operating Conditions

Homogeneous charge compression ignition(HCCI)engine has become a research hotspot in the field of internal combustion engines for its advantages of low pollution gas emissions and high fuel utilization rate.However,the combustion of homogeneous mixture gas in the cylinder during HCCI combustion depends entirely on the chemical dynamics of the gas itself.As a result,it is difficult to determine the combustion timing,which is an indicator of its performance,and thus the combustion timing cannot be controlled.Therefore,accurately estimate of combustion timing is an important prerequisite for it’s accurately control.In recent years,the emergence of neural networks has made it possible to estimate the combustion timing using black-box models,and some achievements have been made.However,there are still problems that the estimation is not accurate due to the complexity of the actual operating conditions of HCCI engines,the influence of fuel octane number on the combustion timing when mixed fuels are used,and the coupling relationship between combustion cycles.Therefore,in order to solve the above problems,this thesis uses Long Short-Term Memory(LSTM)neural network to build a black box model to estimate the combustion timing,and uses vibration acceleration signal to correct the above model,thus ensuring the accuracy of the estimation results.The main work of this thesis is summarized as follows:1.Study on the influence variables of combustion timingIn this thesis,the HCCI engine simulation model is built in the engine simulation software GT-POWER,which is combined with MATLAB/SIMULINK environment to build a joint simulation platform.On this basis,the effects of various variables on the combustion timing of the hybrid HCCI engine are studied,and the causes are discussed and analyzed,providing data and theoretical support for the subsequent model building.2.Research on combustion timing estimation of HCCI engine with mixed fuel undercomplex operating conditions based on LSTMAiming at the problem that the existing models do not consider the complexity of the actual working conditions of the HCCI engine with mixed fuel,the influence of the octane number of the mixed fuel on the combustion timing,and the coupling relationship between the engine combustion cycles,which makes it impossible to accurately estimate the combustion timing,this thesis adds the engine speed and the octane number of the mixed fuel into the input variables,and uses LSTM neural network to build a black box model to estimate them.The comparative experimental results under Federal Test Procedure(FTP)show that the model proposed in this thesis can accurately estimate the combustion timing of HCCI engine with mixed fuel under complex conditions.3.Optimization of LSTM black box model based on vibration acceleration signalIn order to solve the problem that the combustion timing cannot be measured in reality,and the combustion timing obtained by other methods may not be accurate enough due to interference or calculation errors,thus the estimation results of LSTM neural network black box model cannot be guaranteed,this thesis uses the correlation between vibration acceleration signal and combustion timing to establish a Back Propagation Neural Network(BPNN)and Support Vector Regerssion(SVR)models identify vibration acceleration signals,and the loss in the identification process is used to correct the LSTM black box model.The experimental results show that the model established in this thesis can ensure the validity and stability of the estimation results.