Design And Research On Fault Diagnosis System For High Pressure Common Rail Electronic Control Injector

The high-pressure common rail fuel injection system of marine diesel engines has a high failure rate and has a bad failure impact.To detect and locate faults in time,it is necessary to carry out fault diagnosis research on the system.Considering that the existing research work lacks a complete fault diagnosis technical route,this paper has researched the fault diagnosis method of the fuel injection system,and on this basis,a fault diagnosis system is developed to realize the intelligent diagnosis of electronic control injectors,thereby ensuring the marine diesel engine reliable operation.First,the failure mode and effect analysis of the high-pressure common rail fuel injection system is carried out.Based on the analysis results,the electronically controlled injector with a higher failure hazard is selected as the research object.To obtain effective fault source signals,an experimental plan was designed,the fault experimental platform was bulit,and the fault signal collection was realized by setting the appropriate console parameters.Then,researched the fault diagnosis method of the high-pressure common rail electronic control injector,mainly including(1)Research on the signal preprocessing method.Aiming at the end effect defects of the Local Mean Decomposition algorithm,an adaptive signal continuation method based on similar signal energy and skewness is proposed,and a threshold function that can adaptively control the degree of denoising according to the noise content of the signal is constructed.Combining the improved Local Mean Decomposition algorithm with the adaptive threshold function,a new signal denoising method is proposed.Compared with the commonly denoising methods,the results show that the proposed method has better denoising performance.(2)Research on feature extraction algorithms.Aiming at the nonlinear characteristics of high-pressure tubing pressure fluctuation signals,a Composite Multi-scale Weighted Permutation Entropy(CMWPE)feature extraction algorithm is introduced.Comparing CMWPE with commonly used feature extraction methods,the results show that compared to Multi-scale Permutation Entropy(MPE)and Multi-scale Weighted Permutation Entropy(MWPE),CMWPE-based Support vector machine(SVM)fault recognition accuracy has been improved by 6.0% and 3.3%,and the standard deviation of CMWPE is smaller,which proves that CMWPE has better inter-class dispersion and stronger robustness.(3)Research on feature selection methods.Considering that the fault feature set extracted from the measurement signal contains a lot of redundant information,a feature selection method based on mutual information is proposed.The method is applied to the feature sets of MPE,MWPE,and CMWPE.The analysis results show that after feature screening,the SVM fault recognition accuracy has been increased by 7.3%,7.3%,and 6.0%respectively,and the total calculation time has been reduced to a certain extent.This proves that the method can effectively reduce the difficulty of pattern recognition tasks and improve the efficiency of the entire fault diagnosis process.(4)Research on pattern recognition algorithm.Aiming at the problem that it is difficult to optimally select the key parameters of SVM,an improved particle swarm optimization(IPSO)algorithm is proposed to search for the optimal parameters of SVM,and the inertia factor adaptive adjustment strategy and particle random mutation mechanism are introduced to prevent the Particle Swarm Optimization(PSO)algorithm from prematurely converging to the local optimal solution.The proposed IPSO-SVM classification algorithm is applied to the MPE feature set and compared with other classification algorithms.The results show that compared with the PSO-SVM algorithm and the SVM algorithm,the fault recognition accuracy of the IPSO-SVM algorithm is increased by 5.3% and 8.0% respectively,which proved the effectiveness and superiority of the IPSO-SVM algorithm.Finally,a new fault diagnosis method for electronically controlled injectors is proposed.Through the fault diagnosis experiment on the measurement signals under the three working conditions of 70 MPa,90MPa and 110 MPa,the results show that the fault identification accuracy rate of the three working conditions has reached more than 99%,which proves that the method has good fault identification ability.Finally,the electronic control injector fault diagnosis system software framework is designed based on the proposed fault diagnosis method,and the high-pressure common rail electronic control injector fault diagnosis system is developed using the Lab VIEW.The function of the diagnosis system was tested with the data obtained by the experimental bench,and the result showed that the fault identification accuracy rate of the system reached 99.3%,which met the actual diagnosis requirements.