Research Of Diesel Engine Fault Diagnosis Based On The Signal Feature Analysis

As the power support for industrial production,diesel engines have been used as core components in vehicles,ships,electricity and other fields,which are important support to promote the development of the society.The working environment of a diesel engine is often complicated and changeable.When a kind of fault occurs during operation,the work process will be disturbed and the efficiency will be reduced.In the worst case,the unit may be damaged,and even on-site staff may be injured.The diagnosis of the machine should be made quickly for the follow-up work when a fault occurs,ensuring the smooth progress of production,and maintaining safety and reliability with good and stable operation during the work.Therefore,it is of great significance to study the fault diagnosis of diesel engines.In order to improve the diagnosis accuracy,this thesis proposes effective methods for fault diagnosis of diesel engines based on the study of common faults.The diagnosis results are proved with relevant fault simulation experiments and the practical applications.The contents of this thesis are as follows:Firstly,this thesis proposes research on characteristics extraction methods of fault diagnosis test and training data based on the distribution of data in the changeable operating conditions.A fault diagnosis method based on the domain adversarial network is proposed,and a domain adversarial network diagnosis model is established.The analysis and processing ability of training and test data are improved which are not affected by the changes in the operating environment,and the impact of data sources on the diagnosis results is reduced.The fault diagnosis process is realized under the condition that the distance between the fault feature and the training set sample is large,and is verified by fault simulation experiments.Secondly,this thesis proposes research on diesel engine fault probabilistic inference method combining the principle of probabilistic graphical model and fault diagnosis theory.Aiming at the common fault types of diesel engines,a multi-layer network fault probability diagnosis model is established.The probabilistic reasoning about the possibility of potential faults is realized based on the signal processing and analysis of unit component life status.The diagnosis network framework is designed,and the corresponding relationship between the fault type and the signal characteristics of the network nodes is determined.The proposed method is verified through simulation experiments.Finally,this thesis proposes research on the diesel engine fault diagnosis method of misfire in the cylinder based on the study of the fault characteristics and the engine structure.A kind of fault diagnosis method for diesel engines based on multi-signal features is proposed,and a misfire fault diagnosis network is established.The multi-state correspondence between the cylinder node and the signal component node is determined,and the complexity of parameter setting of the diagnosis model is reduced to realize the auxiliary reasoning.The stability of the proposed method is verified by multi-operation misfire experiments.This paper proposes a new method of fault diagnosis with the study of the fault diagnosis of diesel engines,and broadens the research ideas of diesel engine fault diagnosis.It provides a guarantee for the good working operation of diesel engines,and has important significance in practical engineering applications.