The Application And Research On Reverse Engineering In Robotic Grinding

Blades,aircraft shell and other complex parts are widely used in aviation,aerospace and other industries.Grinding is the key process.The traditional manual grinding is harmful and poor consistency,CNC grinding is expensive,low flexibility and efficiency.In recent years,the robot technology has been gradually applied to the grinding industry to achieve automated production for its features of flexible,easy to expand.At present,the robot grinding firstly finish off-line programming which is based on the workpiece theory model.However,this kind of grinding processing method is not applicable to the workpiece which is unknown or deformed.In this paper,the reverse engineering technique is applied to the robot grinding system for complex parts to make it possible.This paper studies the surface data measurement,data processing,surface reconstruction and trajectory planning.In the aspect of data measurement,several measurement methods are compared and then choose the line laser scanner as the measuring device.A Hand-eye calibration algorithm which bases on virtual sphere is adopted to demarcate scanner coordinate system,and verify it through experiments,the experiment showed that the range of scanning error is between 0.1 mm and 0.2 mm.In the aspect of surface reconstruction,the combined de-noising algorithm and the distance reduction method are used to de-noise and simplify the point cloud.the node parameters and the node vector are obtained by the inverse algorithm,the least squares method is used to obtain the control points,and the reconstructed surface is obtained by NURBS,finally the engine blades are reconstructed,and the reconstructed model and the theoretical model were matched and analyzed in Geomagic software,and the standard deviation of the matching result was 0.231 mm.In the aspect of trajectory planning,this paper use equal chord error method to determine the step size and equal residual height method to determine the spacing by comparing several common trajectory planning methods.The attitude of the robot is determined by coinciding the support vector of grinding head with the normal vector of workpiece in order to improve the grinding accuracy and simplify the calculation,and the rotation matrix is converted into quaternion so that the robot controller can read the generated grinding program.At the end of this paper,a set of self-developed software with the function of point cloud processing,NURBS surface reconstruction and trajectory planning is introduced.The grinding experiments are carried out on the local surface of wind turbine blade,the surface roughness of the blade is taken as the evaluation index,and the average value is 2.532?m which meets the current process requirements.The effectiveness of the proposed method is verified.