Investigations On Nonlinear Dynamic Characteristics And Chaotic Predictive Control Of Combustion In Natural Gas Engine

Developing efficient and clean natural gas engine technology is an effective way to optimize energy consumption structure and to meet the increasingly stringent emission regulations.And it is also one of the mainstream directions of marine engine in the future.Applying lean-burn technology can further improve the thermal efficiency and reduce emissions of natural gas engine.Because of its compression ignition difficulty,low flame propagation velocity and narrow optimum operating range,the combustion instability of natural gas engine aggravates under lean-burn conditions.Especially near the engine operating limits,misfire,incomplete combustion and knock will deteriorate the engine stability and make it difficult to control.So far,the internal dynamic evolution mechanism of the combustion instability caused by the interaction of multi-factor is not completely clear,making it difficult to effectively control the combustion of natural gas engine and to ensure its safety,stability and high efficiency combustion.Therefore,revealing the essence of combustion instability and the inherent law of deterministic dynamics evolution,then carrying out characteristics identification,dynamic modeling and nonlinear prediction,finally realizing the chaotic control based on combustion are systematic researches to achieve combustion optimization and solve combustion instability fundamentally.It is of great academic value and engineering application value to expand the operating limits of natural gas engine,and further improve its power and economy as well as reduce emissions.In this thesis,an electronically controlled multi-point injection natural gas engine is developed on the basis of a marine diesel engine.The engine test system is established.The control system and combustion closed-loop system based on in-cylinder pressure are designed.All of these works provide hardware conditions for the research of this thesis.The main contents are carried out as following:?1?The experiment system of natural gas engine is set up.Considering the characteristics of the natural gas engine,the engine structures,including the combustion chamber,the intake system,the gas supply system and the ignition system are optimized on the basis of a marine diesel engine.The engine test systems,such as the torque test system,the air-fuel ratio test system,the in-cylinder pressure acquisition system,the emission test system,the monitoring and calibration system are established.The controller of natural gas engine is developed based on the modular design.The driving controllers of the key executive components,such as the electromagnetic valve,the electronic throttle and the ignition,are designed and developed.And the key modules such as air-fuel ratio and detonation are controlled.The combustion closed-loop control system of natural gas engine is established.All of these works provide hardware conditions for the subsequent combustion characteristics analysis,nonlinear dynamics characteristics analysis,chaotic prediction and control of natural gas engine.?2?Experimental studies are carried out on the combustion characteristics of natural gas engine.The effects of?,SIT and GIT on the combustion characteristics of natural gas engines are studied under typical conditions.The regular combustion characteristic parameters,especially in-cylinder pressure,HRR,exhaust temperature,CCV_IMEPMEP and average IMEP are analyzed to provide theoretical supports for the analysis of nonlinear dynamic characteristics and chaotic prediction.?3?The nonlinear dynamic characteristics for combustion process in natural gas engine are studied.Focusing on the problem of lacking quantitative analysis and multi-factor correlation analysis in the previous researches,the dynamic patterns of combustion under different?,SIT and GIT values are investigated,and compared with the dynamic characteristics of typical periodic and chaotic systems from the views of multi-factor correlation and quantitative analysis.The hidden dynamic law in the combustion process and the correlation between the combustion cycles are fully displayed.?4?The on-line real-time chaotic prediction model is established.In view of lacking on-line real-time chaotic prediction model in the previous researches,the UWT-LLE-RBF neural network chaotic prediction model is established combining with the combustion characteristics and the nonlinear dynamic characteristics of natural gas engine.The“Windowing Prediction Method”is put forward to ensure the real-time performance of on-line prediction and control.Predictions for the interaction between phase points of combustion process,and the randomness as well as determinism co-existing between combustion cycles are realized.The average prediction error is within 5%.?5?The chaotic predictive control strategy is proposed and studied.On the basis of theoretical analysis and experimental researches,in view of the difficulty in combustion control of natural gas engine,the joint closed-loop control strategy and chaotic predictive control strategy are proposed,and the chaotic predictive controller with multiple feedback parameters is designed.The IMEP,p_maxax and?_pmaxmax are comprehensively considered during the control process and the chaotic quantization LLE is introduced as one of the evaluation criteria.The outputs of the control system coordinated control the SIT and the injected natural gas quality of each cycle.The results find that the combustion stability for?>1.4 conditions is effectively improved by using the chaotic predictive control strategy.Compared with the open-loop control,CCV_IMEPMEP decreases by 74.37%for the lean limit at?=1.9 condition.