Robot systems for automation of welding are today a robust technology that combines productivity with quality in a variety of designs of robot stations. The technology is combined with a set of supporting techniques related to welding engineering (different processes, control of the welding process, etc.), sensor technology and simulation and programming. New welding processes such as rapid arc and laser welding will put further demands on control aspects on robot systems as the travel speed increases and process operating parameters become more narrow. The use of new production processes is however linked to product design and the preparation of robust programs for the Robotic Arc Welding station so that the benefits of new processes can be utilized to their optimum.
The present trend in robotic welding system is to further increase the system's flexibility. One such example is the use of robot systems in welding large structures in areas such as ships, bridges and earth moving equipment, etc. Within these areas, the number of work pieces of each type is relatively small, sometimes down to one and high volume products in the range of a thousand. To a large extent, this is a reality today and there are attempts to take this further into one-off production. With the increased use of sophisticated 3D CAD/CAM systems and combining the information with robot simulation systems, this trend will be a reality within a few years to come. Already there are industrial adapted systems that in an economical way produce parts in one-off production, but in such cases the products are defined by parameters and programming is in such cases fairly straightforward. The general case is much more demanding since it has to cover issues related to parameters of the welding process and robot related issues related to joint limits and singularities of the kinematics of the manipulator.
In this context, simulation is an extremely important tool that will develop further in the coming years. In summary,
* More complex work pieces and weld cases can be simulated, programmed and tested as there is no principle restriction in work piece geometry.
* Simulation, testing and programming can be done using a virtual model of the work piece before it exists, providing short lead time in preparing for production.
* Design of weld jigs can be done including testing without the need for a physical mock-up.
* Programs are in general more consistent where part of the programs can be reused. Better documentation and reuse of data gives in general better quality programs. The programming method provides a safe way to produce programs.
This is of importance in complex industrial robotic systems including those with integrated gantry systems, servo tracks and positioners. Manual on-line programming should otherwise force the programmer to work within the system which, especially in complex systems, demand great resources both with respect to personnel safety, and the machine system.
Moreover, simulation systems and online programming provide the necessary tools to address issues related to generation of robust programs. In first phase software tools must be developed to optimize the robot program considering both the robot and the welding process. In the second phase we can expect a coupling between a simulator with the world model of the robot station and the robot controller. Robot status and various sensor information will provide important data to alert the controller if something is going wrong. In such case the concept of a world model can be used for generating programming instructions online as a method for error recovery. Such coupling is illustrated in where an industrial robot is hooked up to a simulation system in a laboratory environment.
This type of connection can be used in several ways and one possibility is, as discussed above, to close the loop concerning sensor data to the simulator which holds information about the world model. Through this coupling, on-line analysis can be made concerning issues related to joint limits, singularities and collisions and error recovery can take place before such problems occur by taking over the control from the Robotic Arc Welding program in the robot controller. Note that such a connection is possible with today's robot controllers by using a "Remote Procedure Call" (rpc) where an external process (control systems) sends robot native instructions in real time. It may not be realistic to have one expensive robot simulator for one robot. Instead, the high level as represented by the simulator is not expected to take control during normal execution and in such cases a high level controller may be connected to tens of robots in a real implementation.
by Bolmsjo, GunnarView Profile; Olsson, Magnus; Cederberg, Per. The Industrial Robot29.2(2002)
No comments:
Post a Comment