Research On Fault Diagnosis And Health Evaluation Method Of Fuel System Of Marine Common Rail Diesel Engine

Diesel engines are widely used in marine power because of their high thermal efficiency and reliability.With the continuous improvement of the requirements for energy saving and emission reduction of ship power,diesel engines generally use high pressure common rail fuel systems.Although the marine common rail diesel engine improves the economy and emission performance,its safety and reliability are reduced due to its complex system structure and harsh working environment.As the subsystem with the highest failure rate in the common rail diesel engine,the fuel system directly determines whether the common rail diesel engine can operate normally and efficiently.In order to improve the safety and reliability of marine common rail diesel engines and reduce operation and maintenance management costs,it is very necessary to carry out fault diagnosis and health assessment of marine common rail diesel engine fuel systems.Although domestic and foreign experts have achieved certain results in the research on fault diagnosis and health assessment of marine common rail diesel fuel system,there are still some problems to be solved,mainly including the problems of incomplete method system,weak pertinence of typical fault diagnosis methods,low accuracy of weak fault diagnosis algorithms,and low efficiency of health assessment algorithms.In view of the existing problems,this paper has carried out research from four aspects:the technical scheme of health management of marine common rail diesel engine fuel system,the extraction method of typical fault features,the method of weak fault diagnosis,and the method of health assessment.The main research contents include:(1)In view of the lack of a complete method system,this paper proposes a health management technical scheme and a detailed technical route suitable for the marine common rail diesel engine fuel system according to the connotation of fault prediction and health management technology.According to the potential failure mode and consequence analysis,this paper sorts out the typical failure types of the common rail fuel system.In order to obtain the fault source signal of the fuel system of the common rail diesel engine and realize the state monitoring of the fuel system of the common rail diesel engine,a fault test platform and a fault simulation model platform of the fuel system of the common rail diesel engine are built.(2)In order to extract the sensitive fault features of marine common rail diesel engine fuel system,a hierarchical dispersion entropy method is proposed,and a variational mode decomposition and hierarchical dispersion entropy optimization based on improved fruit fly algorithm are established for typical faults of common rail fuel system injectors and high-pressure fuel pumps diagnosis method.The paper applies the proposed method to test and simulation data.The test analysis results show that for typical faults of fuel injectors,the fault identification accuracy is increased by 8.9%,and the algorithm running time is reduced by 7.7s;for typical faults of high-pressure fuel pumps,the fault identification accuracy is It is increased by 7.8%,and the running time of the algorithm is reduced by 6.3s,indicating that the typical fault diagnosis method proposed in the paper can effectively identify typical faults,and has higher accuracy and faster computing efficiency.(3)Since weak faults are easily submerged by noise,in order to enhance the weak fault characteristics,this paper proposes a weak fault diagnosis method for common rail fuel system injectors and high-pressure fuel pumps based on adaptive multi-point optimal minimum entropy deconvolution and hierarchical weighted permutation entropy optimized by the improved locust algorithm.First,this paper uses the multi-point kurtosis spectrum to determine the fault period T,and then adaptively obtains the optimal filter length L of the multi-point optimal minimum entropy deconvolution by improving the locust optimization algorithm,and uses the obtained optimal filter complete the filtering process.Then,the hierarchical weighted permutation entropy is used to measure the complexity of the filtered tubing vibration signal,and the weak fault features are extracted.Finally,the fault feature vector is input into the least squares support vector machine multi-classifier to realize the weak fault diagnosis of the common rail fuel system injector and high pressure fuel pump.The proposed method is applied to the fault signal,and the results show that the proposed method can effectively identify the weak fault state of the common rail fuel system injector and high-pressure fuel pump.Compared with existing methods,the weak fault diagnosis accuracy of fuel injector and high-pressure fuel pump is improved by 6.7% and 10%,respectively.(4)Accurate health assessment technology is a key step to realize the health management of marine common rail diesel engine fuel system.Therefore,this paper proposes a high-pressure fuel pump and injector of the common rail diesel engine fuel system health assessment method based on improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)and Discriminant Correlation Analysis(DCA)feature fusion entropy.First,the proposed method is used to process the fault source signal,and the effective fault information of the common rail fuel system fault source signal is obtained.Then,through the DCA fusion algorithm,the two information entropy features are fused to obtain the DCA feature fusion entropy.Finally,the DCA feature fusion entropy is used as the feature vector and input into the least squares support vector machine multi-classifier to realize the high-pressure fuel pump and injector of the common rail diesel engine fuel system health assessment.Through experimental verification and analysis,it is shown that the number of incorrectly classified samples of the injector’s health state has decreased from 9 to0,and the number of incorrectly classified samples of the health state of the high-pressure fuel pump has decreased from 7 to 1.The state recognition accuracy is improved by 15% and10%,respectively.