As part of an effort to use automation to reduce costs, improve
productivity, and increase profitability, the Ordnance Div, FMC Corp’s Defense Systems Group, San Jose, CA, tested the capability
of a robotic arc-welding system developed by Advanced Robotics Corp,
Columbus, OH. Called Cyro Vision, it is an advanced 3-D laser-based,
real-time adaptive system.
The welding of aluminum ball-port sections for the Bradley Fighting
Vehicle served as a beta test site for the system. Based on its
performance to date, we believe it offers significant opportunities for
improving weld quality and reducing production costs, while
significantly expanding the range of practical robotic arc-welding
Why adaptive control
Our Ordnance Div produces a variety of armored vehicles with
aluminum hulls and structures. We began investigating automated welding
five years ago when robotic technology had advanced to where it made
possible high deposition rates with fewer torch passes on jobs requiring
larger welding wire. Real-time adaptive control of process parameters
and weld path was required to make this a reality.
There are thousands of parts used in the manufacture of armored
vehicles. Even slight variations in part tolerance can create fitup
inaccuracies that require the torch and process to adapt. Prior to
CyroVision, we lacked this capability.
The harsh welding environment had rendered earlier vision systems
useless. We prefer using 3/32″ wire for these applications because
it yields the best weld. However, this requires over 450 amps, and
creates extreme heat, light, and smoke.
We evaluated several through-the-arc seam trackers, but none could
respond adequately to the electrical resistivity of the aluminum, and
this precluded precise control. The intense arc-light emission and high
reflectivity of aluminum result in a signal-to-noise ratio too low for
most vision systems to perform satisfactorily.
The system we needed had to be capable of adapting in real time to
variations in joint geometry and position. After an exhaustive search,
we have concluded that CyroVision is the only system that meets our
requirements of real-time process control, seam finding and tracking,
uses large-diameter wire, and allows changes in process parameters.
How it works
This system is a joint development of Advanced Robotics and
Oldelft, a European supplier of optics and laser-related products. It
makes 2500 depth measurements per second and reproduces the seam profile
During welding, the vision system precedes the torch, gathering and
processing data on both the weld path and the seam profile. By
correcting for weld path and joint volume variances in real time, it
enables a robotic work cell to produce quality welds in applications
previously considered impractical for robotics. Weld parameters can be
adjusted in direct proportion to joint deviations without affecting weld
quality. The system identifies tack welds according to their individual
volumes. The compact camera is air purged and water cooled.
The advantages of this system over manual welding became apparent
at once. The number of welding passes were reduced. The extreme amount
of light and heat generated by high-density welding make manual
operations impossible. Manual welding with low-density wire required as
many as five passes to complete the weld. With robotics, we can use
3/32″ wire to complete the weld in a single pass, even for parts
3″ thick. Thus, we achieve dramatic savings in production time.
Secondly, robotic welding eliminates many of the quality problems
normally encountered with manual welding in this situation, it
eliminates brushing, bubbles, and grinding, while providing excellent
consistency in fill height.
Thirdly, because it accommodates seam-location tolerances of up to
[plus-or-minus]2, it allows us to re-evaluate our tooling requirements
and our need to know the precise location of the part. Adjustments for
part tolerance are made without affecting weld quality. When working
manually, a welder normally must accommodate fitup inaccuracies and
joint changes by adjusting torch travel speed. Now, adjustments are
made in voltage and wire feed so a constant torch travel speed is
maintained, resulting in greater consistency in weldment quality.
The systems proved to be very reliable. Productivity increased
dramatically due to increased arc on-time, faster torch travel speeds,
and higher deposition rates. Also, the need for finish grinding has
been reduced or eliminated. Arc on-time averages better than 85
percent, and we expect to increase this to 90 to 95 percent as we gain
more experience with this equipment. Torch travel speeds are 60 to 80
percent faster, and deposition rates are up more than 200 percent over
We anticipate a significant increase in the number of units
produced per day, based on production levels achieved during beta-site
testing so far. Translated to direct cost savings, we expect to save
$400,000 per robot per year.
By 1986, FMC expects to rely exclusively on robotics for welding
the entire outside of the Bradley vehicles. We look forward to the
time–very soon–when robotic arc welding can demonstrate its full
potential by performing all the welding operations required in
During testing, we found that intensive employee education is
necessary to successfully implement any stage of welding automation. Our
educational program includes seminars, training sessions, and lectures
for management and engineering staffs as well as machine operators.
These programs emphasize that people are essential in robotic arc
welding. We stress the fact that there will always be a need for
skilled people with the experience and expertise to properly evaluate
weld quality. Our program also stresses to employees how important
robotics will be to help FMC stay competitive in the future.
For more information on CyroVision arc-welding systems from
Advanced Robotics, circle E62.