Research On Fault Diagnosis Of 6-DOF Manipulator Based On Model

With the development of control theory and intelligent technology,the manipulator is used not only in industrial production,but also in fields such as aerospace,deep-sea exploration,medical services and so on.Nowadays,the manipulator has become an important part of the core technologies needed for human development.A manipulator is a complex device that integrated a lot of techniques such as electrical technology,mechanical design,control technology,etc,and unexpected faults may occur during the operation of control system,drive system,detection system and execution system.The losses caused by the manipulator faults are often huge.Therefore,the fault diagnosis of the manipulator is deeply studied based on the detailed analysis of the kinematics and dynamics model of the 6-DOF manipulator.The research work of this paper mainly includes:Firstly,the background of fault diagnosis of manipulator and the main classification and research status of fault diagnosis technology are introduced.On this basis,the kinematic model and dynamic model of 6-DOF manipulator are studied by D-H method and Lagrange method respectively,and the fault types of the manipulator system are described in detail.Secondly,for the faults of the actuator of the manipulator,a neural network observer is designed based on the approximation characteristics of the neural network and the idea of state feedback.The detailed proof process of convergence and stability is given,and the effectiveness of the observer is verified by simulation analysis.Then,a sliding mode observer for fault diagnosis of manipulator is proposed to meet the requirement of system fault reconfiguration,and a detailed fault diagnosis strategy is given.The fault diagnosis effect of sliding mode observer is verified by simulation analysis.Finally,an experimental platform is built to verify the proposed fault diagnosis method based on the sliding mode observer.Through the analysis of the diagnosis results,it is found that the sliding mode observer can detect the time of the fault and reconstruct the fault.