Research Of The EPS Controller's Online Calibration Technology Based On Parameter Identification

With the continuous deterioration of the global environment,it is particularly important to save energy and protect environment in all fields,and the automotive industry is certainly no exception.The electric power steering system(EPS)controls the assisted force of steering mechanical structure via EPS motor.Compared with the electronic hydraulic power steering system(EHPS),EPS is widely used in vehicles because of its energy-saving,environmental protection,maneuverability and stability.The EPS is studied on the basis of the system identification which belongs to the modern control theory,combined with the automotive calibration technology in the thesis.The researching parts are listed below.First,the EPS assisted motor is studied,and its model is built based on the theory mechanics and Kirchhoff laws.The motor's transfer function is obtained with the input of terminal voltage and the output of armature current by the Laplace transform.The characteristics of the direct and indirect identification method in the continuous-time model are compared through the motor's transfer function.Using the CONSTID toolbox of Matlab,the simulation identification is completed in different identification conditions,determining the best identification algorithm and conditions.The motor driver is designed,and the motor identification experiment platform is built to accomplish the motor's resistance and parameter identification experiments by the data acquisition system programmed with LabVIEW.Through these experiments,the following parts are obtained: the relations between resistance and current,duty ratio and terminal voltage,and the mathematic model,parameters of the motor.Then,the EPS is studied,and its model is built by simplifying some complex components.The EPS's transfer function is also obtained with the input of steering wheel's torque and the output of rack's displacement by the Laplace transform.The EPS's characteristic is studied by simulation and experiments.The input and output of experiment identification are not the same as the simulation identification.The torque sensor's main voltage is used as its input,and the load motor's current is used as its output.The EPS's bench is built,and in EPS's assisted mode,the control models of PID and model-based algorithms by Simulink are demonstrated in dSPACE's real-time processor.Then,the two control algorithms' features are analyzed accordingly.Finally,the thesis is focused on the automotive calibration protocol CCP and XCP.The structure of A2 L file is resolved,and the comparison among three general method generating A2 L files is accomplished.The A2 L file is generated by RTW,and some details are integrated,like ECU's variable address,communication interface etc.The EPS's calibration platform is built,using motor's identified parameters to calibrate the controller,and two experiments are done respectively according to an EPS's controller,one using INCA software and ES581.3,the other using USB-CAN and customized calibration tool.