Study On Combustion Cyclic Variation Characteristics In A Lean-burn Natural Gas Egnine Based On Nonlinear Dynamics Theory

Lean-burn technology is very effectiveto achieve high efficiency and low emission of ship natural gas engines,but the combustion cyclic variations(CCVs)caused by incomplete combustion or misfire of the too lean mixture become the main obstacle to the expansion of lean combustion boundary and performance improvement of the spark ignition(SI)engine.If the CCVsare eliminated,the power output of the engine can be significantlyincreased in the case of the same fuel consumption.Therefore,the issues of identification,prediction and control of the CCVs have always been the hot topics.In this paper,the characteristics of the CCVs of spark-ignition natural gas engines are systematically identified by using the nonlinear dynamics theory and multifractal analysis method based on the wavelet transform modulus maximum.The statistical characteristics of IMEP time series and the generation mechanism of CCVs were studied by using the classical statistical analysis methods and 3D CFD simulation,and the results showed that the fluctuationsof IMEP time series and their statistical characteristics were more sensitive to the changes of gas injection timing(GIT)and spark advance angle(SAA)under the conditions of low load and lean mixture.With the increase of GIT,(?) gradually decreases,skewness increases monotonously,and the transformation from negative skewness to positive skewness is realized.A higher value of Kurtosis indicates that the CCVs have obvious characteristics of supergaussian distribution.Different GIT will lead to different degree of mixture concentration stratification in the cylinder for the manifold injection natural gas engine.When spark advanced angles(SAA)are the middle value,the fluctuation range of IMEP time seriesare small,the (?) value are larger,the values of Skewness and Kurtosisare higher.At this time,the engine has a strong ability forpower output.Based on the nonlinear dynamics theory,the phase space reconstruction,recurrence plot(RP)and recurrence quantification analysis(RQA)methods are adopted to systematically identifythe nonlinear dynamic characteristics of combustion system,and the results show that the combustion stabilities and the power output are not related to whether the structure of the attractor has circular,nabla or bifurcate structure,and depending on the distribution intensity,location,and the number of discrete state points ofthe combustion system attractor state points in phase space.A method for determing recurrencethreshold based on weighting factor optimization is proposed,and the recurrence characteristic of combustion cyclic variations were identifiedfrom the diagonal and vertical direction respectively,the results show that for different mixture concentrationunder the low and middle load,with the increase of GIT and SAA,CCVs experience mode transitions: quasiperiodic-chaos-quasiperiodic-chaos,chaos-quasiperiodic-chaos,and the more stable combustion causes the high system complexity and the lower layer state.In order to further identify the dynamic behaviors of combustion system in the aspects of time domain and frequency domain,based on wavelet transform and fractal theory,wavelet energy spectrum and the singular spectrum of IMEP time series are studied,the results show that for all loads,ignition and gas injection conditions in this study,CCVs have properties of multi-period and intermittent fluctuations,changes of GIT and SAA lead to the frequent changes of WPS periodic ban distributions and the persistent duration.The fractal characteristics of CCVs are quantitatively analyzed by using the width of singular spectrum ?h.At a engine low load,with the increase of GIT,?h first increases and then decreases,while for medium engine load,?h first decreases and then increases.For the medium and low loads,with increase of SAA,?h increases first and then decreases,The values of h corresponding to the peaks of singular spectrum curve are less than 0.5,indicating that the CCVsare anticorrelated processes.In general,lower CCVs lead to increased multifractal complexity.The results of the study have important theoretical and realistic significance to understand the complex nonlinear dynamic behaviors of CCVs in a natural gas engine,reveal fluctuation pattern conversion and the relevance of the intermittent combustion characteristics,and to provide basic research data for further achieving the combustion prediction and closed-loop control.