Research On Fault Diagnosis Based On Extraction Of Key Points In Diesel Engine Combustion

The diesel engine combustion process is the decisive process for diesel engine power generation.Diesel engine combustion process contains a large amount of diesel engine condition monitoring information.Using this information,it is possible to detect possible combustion failures and induced faults during combustion of diesel engines.And it can help to detect the combustion process of the diesel engine and be able to detect the combustion conditions required for certain faults in time.Thus we can adjusting the diesel engine to eliminate these undesirable combustion conditions.There are three common methods for detecting diesel engine combustion processes: instantaneous speed method,in-cylinder force curve method and vibration signal method.Among them,due to sensor limitations,the instantaneous speed method is difficult to detect high-speed diesel engines.However,the in-cylinder pressure curve method pressure sensor is expensive and difficult to achieve non-intrusive detection,so large-scale industrial applications cannot be realized.The vibration signal method is easy to install and measure,and the vibration signal contains a large amount of combustion process information,which is an effective means to realize the non-disintegration detection of the combustion process,and has important research and practical value.In this paper,the vertical two-cylinder diesel engine is taken as the research object,and the combustion process using vibration signal is studied.First,in order to eliminate the influence of other excitation sources on the analysis of the combustion process,determine the combustion window.That is,the time of combustion.In this paper,the Wigner-Ville spectral moment suppression of other excitations is used to quickly determine the combustion window period and achieve the separation of combustion shock and other components.Secondly,considering that the diesel engine has a second-order cyclostationary characteristic,the spectral correlation density function is used to adaptively find the frequency band related to the combustion process,and the combustion frequency is filtered by the optimization frequency band.After that,the speed signal is obtained by using a numerical integration method on the filtered acceleration signal.Finally,the phase signal is sampled by the tacho signal to correlate the speed signal with the crankshaft angle of the diesel engine.Then use the synchronous angular average to eliminate the influence of random interference on the extraction of the special point of the speed signal(the starting point,the pressure rising point,and the maximum pressure point).After that,several key points in the combustion are associated with the crank angle.By analyzing the crank angle of the key point,the engine combustion hysteresis,pre-ignition,common rail pressure abnormality and other faults are detected,and the measured data is passed.The validity of the method was verified.