readhome (Всё сделанное в учебнике (переведены тексты и сделаны упражнения))

Automated robotic deburring produces quality components.

All machined parts need some type of deburring. Many are done only by hand today. An automated means to generate robot deburring programs from CAD models of parts yields benefits in cost, part-assembly fit, and final product performance. The use of a combination of sensors helps monitor the process and control it to achieve consistent quality.

The process begins with interpreting a CAD model of a machined part, fixturing, and tool. Next, a path is generated based upon the workspace processing. The path is grossly checked for collision in the approach phase and finely checked in the deburring phase. The path is then transferred to the workstation controller where it is used to direct the robot with real time adjustments based upon acoustic-emission-feedback signals corresponding to the depth of cut.

Interactive process. Automated-robotic deburring uses a computer-aided design system (CAD), CATIA, and a computerized workstation that provides quality precision machined parts. The workstation is linked to the CAD system for automated robot path definition,

which aids the development designer to quickly make prototype parts.

The workstation is self-calibrating for correction to part machining tolerances; and for varying position of the part in the tooling due to nonprecision placement or wear of the tooling points. The deburring process is an interactive process with computer control that uses an acoustic emission sensor to produce a high-quality chamfer.

Traditional methods of deburring include: tumbling, chemical, sand blasting, and hand deburring. Robotic deburring is a selective means of removing burrs from machined parts, whereas tumbling, sand blasting, and others are a global deburring process. Manual deburring is a local method, but is inconsistent, slow, and costly, as well as dangerous to the operator when handling the part before and during burr removal.

Description. The required geometric data for path planning are extracted from the CAD (CATIA) database and the data are preprocessed. The work-piece, deburring tool, fixturing, and the robot are created in CATIA. The edges that are to be deburred are specified in the CATIA workpiece model with blinking line code. The curved edges are represented as straight edge segments, and the faces of the workpiece reoriented for outward pointing normals.

Path planning and collision-checking software have been demonstrated. Paths are planned for the approach phase, and the deburring phase. The approach phase involves large motions of the robot links and the tool as the robot moves the tool toward the initial contact point with the workpiece. Collision checking is performed for both the robot and the tool to detect and avoid collisions.

For the deburring phase, when the tool follows the contour of the workpiece, the movement is made by very fine motions with the end joints of the robot and tool. Therefore, collision checking is performed only for the tool and the robot-end links. Every step of the path generation process is also examined for acceptable kinematic constraints of the robot, and optimum orientation of the tool relative to the workpiece.

The robot-coordinate points, defining the path, are downloaded into the robot controller, a PC/AT. A robot control progкam in the PC/NT operates on an interrupt basis while robot motion is synchronized by a pulse from the robot. The robot position is initialized with a predefined home position. The robot-path-point matrix is transferred from the CAD database postprocessing to the workstation controller. The matrix is simply a file of coordinate points, and end-effector velocity used to direct the robot.

Realtime-feedback control is developed to calibrate the workstation, and conduct the shallow chamfer deburring with quality control by virtue of an acoustic emission transducer. The robot is calibrated in the workspace with a systematic contact procedure by triangulation using the force-torque wrist transducer. The calibration accounts for the uncertainty in the workpiece location due to fabrication tolerances of

the workpiece, and inaccuracies in the robot, end-effector, tooling, and fixture holding the workpiece.

Force-torque measurements are made for positional control between the robot end-effector and the workpiece. Force-torque was not effective for the deburring phase because of the shallow tool engagement with the workpiece, resulting in low-chip load and low forces. At the point of tool engagement with the workpiece, the acoustic emission (AE) transducer transmits a signal representing contact at a sensitivity level less than the force-torque transducer. From contact through complete translation along the workpiece, the AE provides the workhorse signal for engagement control.

Tool engagement to the workpiece is monitored and controlled with a root-mean-square (rms), acoustic-emission signal. The AE signal provides an excellent means to control material removal. The global deburring methods control edge finishing by the robustness of the process and exposure time. AE provides an immediate response, which is a measure of the edge quality.

This is then a controlled response in an automated process. Force-torque is monitored while the cutting tool is engaged with the workpiece; but sensitivity of force to small depths of engagement or fine burrs on the workpiece do not provide a correlation to the workpiece chamfer. On the other hand, the AE signal is responsive to contact and sensitive to depth of engagement. Additionally, a metal-removal-rate model is developed for devising control strategies, and designing s robot-end-effector system.

An x-y table is used to maintain a constant depth of engagement based upon the AE signal during deburring. A laboratory apparatus has proven the x-y table can track a skew of the workpiece. Additionally, the x-y table is capable of compensating for robot deviations. Thus, the robot provides for large continuous movements and the x-y table can compensate for positional errors.

An Asea robot is being used in the deburring workstation. The robot has 6° of freedom, thus providing horizontal and vertical planner motion, and rotation of the end joint. The Asea IRB6/2 robot has a repeatability of ± 0.008 in. and a minimum incremental move of ± 0.004 in. The robot can provide large continuous movements, but its accuracy limits edge precision.

Two principal limitations exist with the robot. One is the joint flexibility and compliance, and the other is the sawtooth movement inherent from the arm design and the control algorithm. The combined positional error can be overcome by using a micro positioning x-y table. The x-y table provides the positional accuracy needed to achieve the objective of 0.003 to 0.007 in. chamfer. The AE signal provides the means to control the x-y table for edge following and deburring engagement.