Page 161 - Welding Robots Technology, System Issues, and Applications

Robotic Welding: Application Examples 149

1. Actual robot manipulators have closed controllers, not allowing real time
position correction by costumer programming.
2. Most of the actual robot controllers do not allow remote control from an
external computer.
3. It is very difficult to attach guidance sensors with good performance,
because most of the robot controllers are not prepared to do it, or restrict
that possibility, not providing fast interfaces freely accessible by advanced
users.
4. Robot programming environments are not powerful enough to handle
tasks that require complex control techniques (learning, supervisory,
adaptive, etc).

With the system introduced here several of these limitations are reduced as
follows:

1. A robot control system that allows position correction commanded by a
remote computer is used. That’s not a standard feature, but was added to
the system by using other features available from the controller.
2. A distributed and client-server based software architecture, that enables
remote control using Ethernet networks was developed.
3. The necessary sensors may be attached to the computer that controls the
robot and not to the robot controller itself. This excludes the seam
tracking sensor which requires a low-level interface with the robot
controller.
4. A personal computer is used as programming environment, taking
advantage of the huge amount of programming and analysis tools
available for those platforms.

The basic setup is composed of an industrial robot (model ABB IRB1400 M98), a
robot controller (model ABB S4C+ M98), a MIG/MAG welding power source
(ESAB LUA 315R) and a computer running Windows XP (any other DCOM [4]
based version of the Microsoft Windows operating system could have been used).
The welding power source is connected to the robot controller IO subsystem, being
in this way completely controlled by the robot controller. The robot controller is
equipped with an Ethernet board being accessible from the local area network
(Figure 5.1) by any computer also connected to the network.

Achieving automatic parameter selection also means using sensing devices not
only for guiding the welding torch (joint tracking), but also for real-time
acquisition of the welding geometry. Simple torch guidance and joint tracking can
be obtained using mechanical, electrical or optical sensors: current sensors (to
sense the arc current during torch weaving), laser beams, etc., (see Chapter 3).
Monitoring the welding geometry requires more sophisticated sensors, like the
laser 3D cameras available for example from Meta Systems – UK and Servo-Robot
– Canada (see Chapters 3 and 4). By joint geometry we mean information about
the gap, cross section area, mismatch and type of joint (fillet, corner, lap, V-
groove, butt, etc.). With that information, and with the appropriate software and