| With the development of modern ships towards large scale and high-power density,marine diesel engine is treated as the main power unit of ships,and is easy to cause rough operation and serious combustion noise radiation when it operates for a long time under high power conditions.Notably,the in-cylinder pressure resonance caused by knock combustion is regarded as the significant source of combustion noise and structural vibration.However,the investigations of in-cylinder pressure resonance were almost based on automotive engines,and there is little investigation related to marine diesel engines.Considering the differences of structural size and combustion environment between automotive and marine diesel engine,the characteristics and optimizations of in-cylinder pressure resonance must be different.Therefore,this thesis focuses on the analysis and coordinated optimization of in-cylinder pressure resonance characteristics in marine diesel engine.Based on the numerical simulation and signal processing technology,the influence mechanism of in-cylinder pressure resonance can be revealed,and the oscillations responses of resonance were evaluated.Besides,intelligent and machine learning algorithms were employed for exploring the optimal design strategies of incylinder pressure resonance and other multi-objective characteristics.The main research work includes the following aspects:(1)The numerical model of marine diesel engine was established by CFD simulation,which was used to analyze the generation and evolution of in cylinder pressure resonance during combustion,and the coupling mechanism of pressure wave was also revealed.In addition,the spatial distributions of pressure oscillations caused by resonance in the combustion chamber under different loads were compared to explore the influence of combustion conditions on the response of resonance pressure oscillations.The results show that the superposition of pressure waves generated by fuel injection and ignition causes cavity resonance in the combustion chamber.During the combustion stages,the shape of the resonant mode gradually stabilizes but its position rotates significantly.At the same time,the pressure waves excited by combustion will be coupled,which generates the local cancellation and superposition region in resonance mode.In addition,the generation and evolution of resonance phenomena are greatly affected by cylinder volume and combustion conditions.Also,the pressure waves will be not only transmitted to cylinder head,but also aggregated on the clearance area and pit area of bowl shape combustion chamber.(2)Considering the time-varying evolution characteristics of in-cylinder pressure resonance during combustion process,the signal decomposition and feature extraction algorithm based on TVF-EMD and PD-CPD were proposed,and the relationship between in-cylinder pressure resonance signal and combustion process in time domain was systematically investigated on the basis of multi-condition experimental designs.According to the results of signal processing,the combustion characteristic indexes that have great influences on in-cylinder pressure resonance were selected,and multiple regression method was used to predict the in-cylinder pressure resonance energy.The results show that the whole combustion process contributes to the resonant pressure oscillation,and the intensity of the resonant pressure oscillation excited by different combustion phases varies greatly in the time domain.It is also found that maximum pressure rise rate,maximum derivative of pressure rise rate and maximum temperature can reflect the characteristics of in-cylinder pressure resonance,which could be used to accurately predict the energy of in-cylinder pressure resonance signal.(3)One-dimensional simulation surrogate model of marine diesel engine was established and calibrated based on experimental test data.Combining Latin hypercube sampling and the response surface method,the sensitivities between the related engine control parameters and in-cylinder pressure resonance energy,fuel consumption and NOx emissions were investigated.Also,significant control parameters that greatly influence the multi-objective characteristics were selected.The results show that the control parameters,including injection timing,intake valve closing,intake pressure/ intake temperature and intake cam phasing,were found to have the greatest influences on in-cylinder pressure resonance energy and performances,which were utilized for multi-objective optimization.(4)In order to achieve the coordinated optimization of in-cylinder pressure resonance energy,fuel consumption and NOx emission in marine diesel engine,a GA-ICSO intelligent optimization algorithm was proposed.The proposed algorithm not only incorporates the advantages of swarm intelligence algorithm and evolutionary algorithm,but also combines with the enhanced levy flight and adaptive learning factor.The proposed algorithm was applied to GT-Power simulation surrogate model to explore the optimal designs of injection and intake parameters.The results show that GA-ICSO algorithm has the strongest optimization ability and the best stability compared with other optimization algorithms.Meanwhile,GA-ICSO could provide the best control strategies of injection and intake parameters under different working conditions.(5)Considering the problem of high computation cost when intelligent optimization algorithm is applied to the simulation surrogate model of marine diesel engine,a SQPHKRVM-EMOPSO hybrid machine learning model was proposed,which combines advanced machine learning model SQP-HKRVM with fast multi-objective intelligent optimization algorithm EMOPSO.The proposed algorithm could achieve the optimal design of five significant control parameters using small sample data.Based on the comparison validation in GT-Power software,it is found that the proposed hybrid optimization algorithm can make the computation time of surrogate model reduced 65.2% while the optimization accuracy sacrifices only 8.2%.As such,the proposed algorithm shows great potential in on-line optimization of marine diesel engine. |
