Good sense welding

Citation

Loughlin, C. (2002), "Good sense welding", Industrial Robot, Vol. 29 No. 2. https://doi.org/10.1108/ir.2002.04929baa.001

Good sense welding

Welding has always been one of the leading applications for robots. Originally this was just for spot welding but continuous path MIG and TIG welding soon followed. People can make excellent welders but robots win hands down when it comes to consistency and productivity.

If you want to weld hundreds of the same items in a production environment then you will almost certainly have considered using robots. However if your unit volumes are small then the cost of the robot itself and all the associated fixtures and fittings mean that the majority of companies will still be employing people.

I spend quite a bit of my time in welding workshops and know that a very large number of welding tasks are simply impractical to robotize. The main reasons for this are the relatively low cost of human labour, the problems of access and the imperfect setup. The human welder may not be as good or as fast at welding as the robot, but they can get the job done, and be ready for the next, in the time it takes to pick up a Teach Pendant. And the job is done well enough for the application concerned.

Market forces are constantly adding customisation and reducing the unit volumes that manufacturers are being asked to produce. Add to this the requirements for low stock holding and Just-In-Time delivery and it is no wonder that the majority of welding is still done by people. However, at the same time the push for higher levels of quality and consistency will be making the promise of shiny robot welds a tempting proposition.

This means that there is a very big market waiting for the company that can give welding robots the skill and flexibility of their human counterparts. So what are our metallic friends missing? The term lights on no one at home comes to mind. Most welding robots are pretty dumb and are only able to weld at all thanks to the sophistication of today's welding power supplies. These monitor weld current, voltage and wire and gas feed rates to control the weld parameters and the deposition of metal. All the robot needs to do is move from A to B at a fairly constant speed, and you can be pretty sure of a good weld provided that the pieces to be joined are in the right place at the start and do not distort as the welding progresses.

Before they can take over many of the low volume welding tasks currently undertaken by people, robots must acquire both the ability to sense how the weld is progressing and the intelligence to know what to do when the inevitable distortions and positional tolerances occur.

A most impressive amount of work has gone on in this area (I have colleagues who have been at it for at least 20 years) and the results are now very good indeed. However, manufacturers are still being slow to embrace this new level of technology, and as a result the rate of further progress is decreasing.

This issue includes many examples of advanced robotic welding systems. Manufacturers should give these developments very careful consideration or be prepared to lose market share to those who do.

Clive Loughlin

Call for papersThe theme for Volume 29 no 6 will be "Automotive Industry". If you would like to submit a paper for consideration in this issue, please refer to the notes for contributors at the back of this issue, and ensure that copy is submitted by 22 June 2002.