Research On Key Technology Of Robotic Blade Grinding

With the development of the industrial technology, turbine fluid machinery to become a kind of widely used machinery equipment. In order to obtain better performance, the requirements of dimensional accuracy and surface quality of the turbine blade are very high. At present, for high precision requirements blades, the usual manufacturing process is the use of casting, high precision grinding and milling as well as repeated manual, off-line measurement method of combining. The precision machining procedure of the turbine blade almost the manual finishing processes, which not only difficult to ensure consistent product quality, but poor working environment for workers has significant health hazard. In order to obtain better processing quality, the use of industrial robots to solve the turbine blade machining is an effective means of processing problems.Based on the current operations of China's industrial robot grinding applications and recent development needs of the grinding robot related technology that should attach importance to and increase investment in research and development of new industrial automation technologies. As the manufacturing industry to the flexible manufacturing requirements to further improve, resulting the needs of surface grinding of complex parts processing technology in simpler and more efficient for the integration of industrial robots. In this paper, use industrial robots as the research object for blade grinding applications, carry out the key technology research of the robotic blade grinding, develop a set of function in rich, highly integrated, interactive intelligent robot grinding blade automated machining system. Development of flexible automation technology for the blade grinding to increase processing capacity of China's high-precision grinding, promoting the state's aerospace-related technology development, promotion industrial manufacturing equipment upgrading take on the of national significance.Practical production shows that in the grinding process, the use of industrial robot system as the processing equipment can provide an even broader flexibility and a richer means of control. For turbine blade and other complex surface grinding process of workpiece characteristics, this paper set up the robot belt grinding system based on reverse engineering for blades and other complex surface grinding workpieces. The system is equivalent to the actual production and processing of hand-processing, using a robotic grasping blades to replace the human hand, coupled with a belt grinding machine, using the same materials and processing techniques, using the same manual operation granularity of the belt in order to obtain better results than the hand-grinding machining.The robot belt grinding system based on reverse engineering mainly consists of six major components, namely abrasive belt grinding machine, grinding machine control cabinet, laser scanners, industrial robots, robot control cabinet and the control computer. By system functions, the entire system can be divided into sub- of robotic reverse engineering measurement subsystem and the robotic abrasive belt grinding subsystem. Robot reverse engineering measurement subsystem main features include measurement path planning, robot workpiece shape of non-contact measurement, point cloud data pre-processing operations and the workpiece model of reverse remodeling and so on. Robotic abrasive belt grinding subsystem main functions include the grinding robot path planning, off-line programming, force feedback control and workpiece belt grinding processing.In this paper, robot grasping the workpiece, fixed grinding tools for processing. Tool coordinate system relative to the world coordinate system is fixed relative position of the workpiece coordinate system based on robotic grasping of the workpiece, relative to the robot's relative position is fixed. The use of high-precision displacement sensor, complete the accurate calibration of the tool coordinate system.Using the CAD models of blade grinding to plan the robot path can improve the system automation and blade processing productivity. However, in some cases, such as damaged blades and blade distortion occurs, as well as blade grinding process, there is no CAD model of blades and the information of blade surface at this time is also unknown. As most of the blade surface are freeform surfaces, through a simple measurement to obtain the precise blade CAD model is not possible, so need to reverse remodeling the robot blade model by robotic measurement.In robot measurement system, by building a mathematical model of the scanner to complete the scan data of the calculations to determine the scanner and the relative position between the robot and scanner calibration. Robot measurement system for object measurement is done by planning the robotic scanning path. A scan path is a set of laser line collection, a complete scan path along the scanning path will be scanned objects completely enveloping them. Based on the characteristics of an object designed two sets of scan mode. Two kinds of scan mode for each scan path contains the path number, start scanning position and the termination of scanning position, the entire scanning process by execute the path in order of implementation of the robot scans to complete.Rotating scanning measurement method calculate the rotation angle of robot scan path and the total number of robot scan times through the chord length of the cylinder boundary. Four view scanning measurement method divided each piece into four view, scan workpiece region of each view. For the the workpiece of blade, direct measurement the cross-section line of blade or use the blade point cloud slices to carry out the blade model for reverse reconstruct.After acquired point cloud data, it is need to pre-process for point cloud data. The process of pre-process of point cloud through a multi-view combination, noise processing and data reduction such as multi-step operation to complete. Noise processing to achieve by smooth filter and data reduction by isometric reduction. Point Cloud Data triangular mesh process by connect the point cloud data into triangular facets to check and repair of point cloud data with large gap between the point cloud, as well as deletion and scanning point cloud formed by interference holes.Patches reconstruction modeling method directly on the measured data points by fitting surface patches, after the transition of the surface film, mixed, connected to form the final surface model. The use of B-spline surface reconstruction model as a mathematical expression, namely through the point cloud model reconstruction data inverse caculate B-spline surface, inverse calculation process is constructed a k * l B-spline surface interpolation was given topology rectangular array of data points, by the expression for the tensor product of the reverse process of calculating the surface, thus the surface of the inverse problem into a two-stage curve of inverse calculate problem. Loft the surface by reconstruct a family of curve-fitting to a surface, this family curves are cross section of the blade point cloud.Off-line programming is a way of planning blades off-line processing of the path, generate robot simulation processing program, after simulate the correctness and effectiveness of the verification program, then automatic generate the final robot processing program. Robot programs are uploaded to the robot controller, running the program by robot controller to complete the automation process.Path planning according to the workpiece shape data, according to belt grinding on the requirements of the position and orientation, transform it into a robotic grinding tool of its path, and then through the coordinate transformation, to acquire the position and orientation paths of the robotic grasping the workpiece point of contact with the contact wheel. After determine the path processing method, according to the processing line spacing to determine the number of grinding path. Computing blade section curve or direct use of isoparametric parameters as the blade grinding path curve. Processing path through a given number of points, or two adjacent path points given the distance between the discretization of the path curves to generate the robot path points processed. Processing step to determine the number of processing points, processing path to the minimum number of points required to meet the Douglas-Peucker algorithm. Robot's gesture requires grinding belt grinding machine always keep the workpiece surface with the blades fit and keep in touch. The processing gesture of each processing path points is maintain the Z axis of contact wheel coincide with the normal vector of the workpiece surface at the point.Using robot programming language to program the robot grinding program, in the whole machine process, mainly the use of line motion commands to perform the robotic grinding, robot grinding path is of the linear interpolation. According to grinding technical to plan the velocity of each path points. The velocity of approach point and retract point of grinding path is slightly higher than the robot's velocity in grinding. Robot in not processing stages to meet the security and processing efficiency requirements, using a high velocity. In addition, the processing technology identified by the abrasive belt grinding machine machining parameters, mainly including belt velocity of the abrasive belt grinding machine, belt tensioning force values, the pressure value by providing the contact force for the grinding machine, whether to use the cooling water and so on.System security, non-collision is the basic safe requirements of the entire processing system. Robot motion simulation processing validation rules to ensure the joint angle were within the normal scope of work, validated whether the joint angles of all axes can be reached through the robot kinematics.Grinding path optimization algorithm calculated by analyzing the curvature of the path curve characteristics, depending on the curvature of the path curve will be characterized by segmentation, through the generation of discrete path points rules, different sub-accompanied by a discrete distance from a different generation to meet the requirements of the processing path points . In addition, the processing path by calculating the curvature of characteristic curves to select the grinding wheel diameter.The optimization principle of selected the grinding path overlap region width of the workpiece is the two adjacent paths overlap exactly the width of the region will cover the edge arising from the two paths side by side processing to eliminate the processing edges. By calculating the edge width of the machining path, to determine the optimal spacing of selection process.In this paper, including industrial robots accuracy, robot positioning repeatability accuracy, robot zero calibration accuracy, displacement sensors accuracy, tooling fix positioning accuracy, grinding machines precision accuracy, model reconstruction and path planning accuracy for precision grinding process, the grinding system error affect factors are analyzed and evaluated. To the impact of complex factors and the actual processing of applications, design and development the blade grinding method based on virtual profiling, the impact of factors on the processing error in the processing system for online calibration and compensation, in order to achieve the robotic blade grinding.The blade grinding method based on virtual profile of using a standard blade workpiece as the reference for other workpieces to be processed, on-line calibration errors affecting the processing, compensation, and then revised the initial robot standard parts program in order to get the standard blade shape of the robot grinding. The program of other parts to be processed are modeled on the standard grinding program with necessary adjustments to process.The blade grinding method based on the virtual profile grinding tests on the blades, grinding test results show that robot machining system to achieve high machining accuracy, the accuracy of blade surface is less than 0.3mm fully meet the surface precision. By grinding a variety of different types of blades, grinding results show that the processing quality and processing efficiency is much higher than workers in manual work.