Research On Evaluation Method Of In-cylinder Combustion State Of Marine Medium-speed Diesel Engine

With the rapid development of automation control and computer technology,the study of diesel engine combustion state assessment technology has become one of the necessary ways to realize intelligent ship.Diesel engine in-cylinder combustion is affected by many factors,but also affects the operation of the whole diesel engine,therefore,the evaluation of combustion in cylinder can not only detect faults in time.That is,effective in-cylinder combustion data monitoring and combustion state evaluation technology is one of the main means to improve the in-cylinder combustion state of diesel engines and improve the overall performance of diesel engines.The in-cylinder pressure curve is the best signal source for evaluating in-cylinder combustion.However,due to the high cost of the cylinder pressure sensor and its high thermal load,the reliability and service life of the cylinder pressure sensor in long-term operation are limited.The cylinder pressure sensor is used in the evaluation of the combustion state,so this paper uses the torque sensor to calculate the in-cylinder pressure signal,which can save cost and obtain more combustion information in the cylinder.For the common combustion state evaluation,a single signal is used to judge,but it is difficult to judge the corresponding combustion state only by a single signal.A method for comprehensively evaluating the in-cylinder combustion state from torque sensor measurement data.First of all,this paper studies the main influencing factors of the in-cylinder combustion process,and determines the evaluation scheme of the in-cylinder combustion state.The in-cylinder pressure signal is calculated by the torque sensor measurement signal,and the main combustion characteristic parameters are extracted by calculating the in-cylinder pressure,and the error of the main combustion characteristic parameters is about 5%;The monitoring signals of the temperature sensor and the oxygen concentration sensor are used to extract auxiliary combustion characteristic parameters.Secondly,GT-Power simulation software was used to build a single cylinder simulation model of Marine medium-speed diesel engine.Under 100% working conditions,for normal combustion,fuel injection in advance,fuel injection lag,nozzle wear state,lack of inlet pressure,inlet temperature increase state,fuel injection quantity reduce,the increased amount of oil increase state and cut-out state nine,to simulate the combustion status in different degree,and extract the different combustion characteristic parameters of the combustion state changes.Finally,the combustion state evaluation model combined with the extracted combustion characteristic parameters and D-S evidence theory is used to avoid the defects caused by the subjectivity of artificial evaluation.At the same time,the change of combustion characteristic parameters can also make up for the "black box" characteristic of previous intelligent diagnosis methods,namely the opacity of judgment behavior.Evaluation results show that most of the combustion characteristic parameter is in a reasonable range,the output will tend to be normal combustion or normal combustion state probability is bigger,but as the combustion state degree deepening,beyond the reasonable range of the combustion characteristic parameters increase,based on D-S evidence theory of assessment model of the assessment result accuracy is higher.