Design And Implementation Of Auxiliary Analysis And Diagnosis System For Automobile Engine Faults

As the heart of the car,automobile engine is one of the most important components,which is of complex structure.Many automobile faults are derived from the engine,at present however,most maintenance companies are still using manual diagnosis methods to detect the failures of automotive engine.Due to the uneven technical level of maintenance workers,the time and quality of fault diagnosis varies severely.Moreover,the time of diagnosis is much longer than the maintenance time.Therefore,a system that can quickly assist in the analysis and diagnosis of failures can not only save the diagnosis time,improve the accuracy,but also improve the after-sales service quality of the car,thus establishing a decent reputation for a company.Based on the above reasons,this thesis is devoted to design and develop a prototype system which can help with the analysis and diagnosis of the automobile engine.Firstly,in this article we take the workshops which provide the after-sales service as the core to study the diagnosis programmes of the automotive engine fault home and abroad.Besides,we analyze and select important engine parameters from the engine data flow perspective,and integrate the engine life cycle process into the analysis of data flow.From the common engine fault phenomena,we analyze the faults of various engine systems.We analyze the engine fault from multiple angles,study the business needs and determine the system function requirements.Secondly,according to the specific needs of the enterprises and the characteristics of the existing data,we adopt the improved algorithm BA-BP and BA-SVM to do the simulation experiment comparison on the Matlab R2016 a platform,and choose the min-max method to normalize the experimental data.According to the experiment comparison results,the BA-SVM algorithm is confirmed as the diagnosis algorithm of this paper.Using the engine data flow,the data of the grievous extent of the multiple fault phenomena and the combination of BA-SVM algorithm to form a fault diagnosis model,and through the model training and test results to verify the correctness of the model.For different motorcycle type and engines,it is necessary to use models which manage and train different parameters are needed,which form the new models.Finally,the system management module,basic information module,integrated query module,model management module,and fault diagnosis module are implemented according to the functional requirements of the system,and the functional testing and performance testing of the system are completed to validate the correctness and stability of the system.This thesis combines the BA-SVM algorithm model with the actual needs of the enterprises to provide a variety of diagnosis methods such as diagnosis of automobile engine data flow,diagnosis of various fault phenomena,and query-type fault diagnosis,which can provide users with multiple perspectives to analyze the dimensions of engine failure,and suppoertin the later maintenance decisions.