The Research On Fault Diagnosis Of Integrated System Of EPS And SAS Based On Information Fusion

Worked as the main components of control system,integrated system's performance affects the overall condition of the vehicle.Reliability and fault diagnosis technology of the integrated system is one of the key technologies by which the integrated system of EPS and SAS is realized into production.It is very important to diagnose the fault of integrated system of EPS and SAS in time and accurately in decreasing the risk of vehicle and reducing the economic loss.Based on characteristics of integrated system and the breakdown principles,with the combination of information fusion about fault diagnose method,the fault diagnosis technology of the integrated system was researched.Firstly,the fault diagnosis technology and diagnosis method for automotive electronic control system were described.The concept of artificial neural network,the structure of neuron,learning process of artificial neural network and the principle and concept of information fusion were narrated. And then the failure mode of the integrated system on EPS and SAS was ascertained,the parts of the system and the working principle about the system were analyzed in detail.The main failure components were determined and also the breakdown phenomenons such as torque sensor,ECU,vehicle speed sensor,power steering motor,accleration sensor were carried on the breakdown mechanism analysis.In terms of the failure mode and effects analysis theory,the potential failures of the integrated system were classified.Secondly,a strategy of the fault diagnosis about the integrated system on EPS and SAS based on the artificial neural networks(ANN) was stated.After that,through researching the characters of fault about the integrated system of EPS and SAS,with a three-layers BP neural network and improved BP algorithm,a fault diagnosis system of integrated system of EPS and SAS was developed,and the simulation were made to confirm it.In order to improve the rate of accuracy and realize the survival of the fittest of the two algorithms, information fusion fault diagnosis of neural network and evidence theory integration was stated.At last,the Data Acquisition System was designed,and the output signal of the Hall electric current sensor and the torque sensor was gathered and processed.And the experiment data in different working conditions of the integrated system of EPS and SAS was classed.The fault diagnose network was trained with the training samples,and tested with the test samples.Two neural networks were built to do a preliminary diagnosis.A 3-13-6 network structure was used with E1 network,and a 12-13-6 network structure with E2 network.And then the outputs of two networks were as evidence of D-S evidence theory.The result proved that,the experimental confirmation result and the fact consisted basically,the success ratio of failure diagnosis achieved 86.6%,the failure diagnosis effect was remarkable.Therefore,the validity and the feasibility of information fusion method to the failure diagnosis integrative system were confirmed,and it can be used to diagnose the fault of the integrated system of EPS and SAS.