Research On Robot Grinding System And Control Method Based On Floating Platform

In order to meet the market demands for small batch size,low cost,high efficiency,high consistency and high surface quality during the grinding process,many companies have adopted advanced robotic grinding technology to gain a place in the fierce market competition.However,because parts' dimensional and shape tolerances,fixture and robotic grinding equipment' position errors and accuracy errors exist in grinding process which often cause deformation and vibration in grinding process,making uneven material removal and poor product surface quality and causing normal grinding operations stopped.The force control problem between the grinding tool and the workpiece is a prominent technical difficulty.It has become a bottleneck restricting the promotion and application of robot grinding technology.To this end,this paper introduces a new developed six-degree-of-freedom industrial robot constant force grinding system based on floating platform,adopting the adaptive intelligent control algorithm to solve the problem of normal grinding force control in the robot grinding process.The grinding system has the advantages of low cost,high grinding efficiency,wide range of machined parts,and easy control.The project has been funded by the National High-end CNC Machine Tool and Basic Manufacturing Equipment Science and Technology Major Project(2015ZX04005006),Guangdong Science and Technology Plan Project(2014B090921004),and Zhongshan Science and Technology Major Special Project(2016F2FC0006).In this paper,the stiffness model of the six-degree-of-freedom robot is studied.which provides a design for the robot pose.Based on the theoretical basis,the mathematical model of the robot is established.The main factors affecting the grinding force and material removal during the grinding process are analyzed.The mathematical model of normal grinding force and tangential grinding force was established;the mathematical model of the developed floating platform is analyzed.Combined with the mathematical model of the robot and normal grinding force,the mathematical model of the new developed constant grinding force system based on floating platform is proposed.This paper designs neural network kalman filter based on robot grinding system,which is composed of BP neural network that receives deviation and its variation based on the residual theory and kalman filter that receives current measured normal grinding force,and the former adjusts the system parameters of the kalman filter in real time,while the latter obtains the optimal estimation value of the normal grinding force according to the kalman filter principle.The neural network kalman filter has the advantages of high filtering precision and good real-time performance.This paper designs an adaptive intelligent control algorithm which combines hierarchical integration,advance differential,fuzzy control,particle swarm optimization and immune feedback.It has advantages of fast response,small steady-state error and Strong adaptability;the robot grinding comparison experiment based on the adaptive intelligent control algorithm and other control algorithms are performed,and the surface quality of the workpiece polished by the adaptive intelligent control algorithm is the best.At present the three countries have applied for two patents for inventions are accepted,utility models of two patents are accepted.The results of this paper include: utility models of two patents are accepted,two patents for inventions are accepted,and one of the articles is to be published in the Chinese core magazine called "Mechanical Design and Manufacturing".Research results of this paper can resolve the problem of force control in the grinding process of robots,and promote the development of automated grinding and the improvement of processing efficiency.