A Research Of Automobile Engine Fault Diagnosis Based On Data-Driven

With the rapid development of modern electronic technology,automobile industry has made a breakthrough. However, the automotivemechatronics level continuously improved, which brings the driver moreconvenience, the feeling of high-technology and more security risks at thesame time. Especially the increasingly serious pollution problems makeseverybody re-examine great convenience and improvement along with allkinds of shortcomings brought by this industry. So to speak, fault diagnosisand emission control has gradually become the focus of attention. More andmore research and development tendency have focus on that after the UnitedStates firstly began to implement OBD legislation. Engine as the core part ofthe whole car has got more attention. All kinds of fault detection anddiagnosis methods in the past few decades with the constant progress ofscience and technology are perfected step by step. The methods are fromsimple artificial experience at the first to simple instrument detection, andthen analysing limit value to judge whether the signal is in the normal range.We can say, the probability of misjudgment has been reduced little by littlewith generations of their continuous efforts. Nowadays, fault diagnosis waybased on the data and the model is gradually known by people and widelyused with the control science’s promotion. For engine as the complicatedstructure of mechanical and electrical integration, considering its nonlinearand great inertia, along with the continuous advancement of sensortechnology, the difficulty of modeling and simply obtaining data, all of themmake fault diagnosis method based on data have further promotion.Considering the above analysis and related literature for reference, thispaper proposes one method for fault diagnosis based on data-driven. Firstly,modeling for the engine with Simulink and AMEsim as data source and thenset faults by changing the imput mainly affecting the sensor parameters.Besides, dealing with obtained data of multiple time domain and frequencydomain, in order to better reflect the fault characteristics and facilitatediagnosis classification. Finally detecting and identifying fault by usingalgorithm of support vector machine (SVM). The support vector machine(SVM) algorithm with advantages and reliabilities proved through a lot of simulation experiments can be used for fault diagnosis of engine.