"We are ready for the use of AI in industrial cleaning."
The use of pulsed laser beams in industrial cleaning is an innovative and sustainable cleaning method. Pieter Cretskens, managing partner at Netalux, based in Diest, Belgium, talks about new innovations in this application.
Industrial cleaning with laser technology already exists for some time. Yet Netalux saw room for improvements.
"Definitely. First of all, we focused on user-friendliness. This is reflected in the ergonomics when using the machine. Many operators active in industrial cleaning will use our devices manually. Then it is important to maximize the ergonomics of the device. We achieve better ease of operation by applying a unique system that features a rotating head. In addition, we offer the operator an integrated system of lenses, so that he can choose on the head itself which lens distance he wants to keep to the surface to be cleaned. Apart from that, we also made sure that the operation of the device is done via a touch panel on the machine, and in a very intuitive and ergonomic way."
Are those devices heavy if you want to operate them manually?
"The weight of the handset depends a bit on the model. But you can safely say that they weigh between over two kilograms and just under four kilograms. By the way, you can also apply a balancer, or work with an exoskeleton."
If you're cleaning and a residue arises, how do you dispose of it?
"That is extracted directly at the surface, with an industrial vacuum cleaner. That has a pre-filter as standard and in addition, in many cases, an activated carbon filter. That dust extraction can possibly be attached to the head of the machine. But when operated manually, it can also be easier if a second operator operates the extractor."
Thus, all in all, important ergonomic strokes are made.
"Certainly, and that's the first story we want to tell. The other story is that we already want to capitalize on the opportunities offered by Industry 4.0. For example, we are preparing industrial cleaning for Artificial Intelligence. We actually built a full-fledged computer and developed the electronics and software in-house. The whole machine is basically enterprise-ready to handle a high degree of automation."
How would that work in the field?
"Suppose one of our partners operating as an industrial cleaner wanted to clean a storage tank, in the future the system could determine, based on the scanned gray level of the steel, for example, which way the cleaning will be most effective and efficient. We would then let the algorithm decide which laser patterns to select, for example, with what frequency or luminosity the laser gives the light pulses. Of course, there is still a long way to go for that. We now want to continue developing the algorithm with industrial cleaners and the asset owners, but the hardware is already fully integrated into our devices. Both hand-held devices and laser heads that we mount on a robot will be able to evolve in this way in the future."
Do you see robots in this sector as a serious development?
"Definitely. With several industrial asset owners, we are already looking at this extensively. It is a development that is going quite smoothly. The platform we built is completely prepared for this way of working. We can communicate with any robot we want.”
This allows the industry to respond to the fact that personnel are hard to find.
"That is rather a side benefit. Many asset owners are making the move toward no-man-entry, they no longer want personnel for work to be in a tank for safety reasons. In addition, indeed, it takes some of the pressure off for finding personnel willing to do this work.”
If you deploy artificial intelligence then in a sense it replaces human intelligence. What human intelligence would you then no longer need?
"If you do a laser cleaning, you can set a lot of parameters. You can set the frequency of the number of light pulses, you can set the light intensity per light pulse, you can set the overall power, and you can vary the pattern you clean with in many ways. With a certain amount of repetitive work, in the future the machine will be able to use algorithms to some extent to calculate for itself the best way to clean in a given situation. Not only based on the contamination, but also based on the extent to which actual cleaning is required - i.e., the desired end result.Now, the experience of the operator is the determining factor. Over time, this gives a greater margin of error and in many cases costs more time, if the operator has to make corrections or adjustments to the settings while working. Our platform on the computer is actually now ready to start interpreting practice. What we need now is a practical situation where we can run the automated application combined with a software developer who can then start writing the algorithm for this application. But the platform, the hardware, is ready now."
And when the algorithm is ready, the system continues to learn in practice?
"Yes, based on machine learning and artificial intelligence, the algorithm will continuously improve itself."
Looking to the future, how fast will this develop?
"That depends a bit on the partners that come forward for this. There is a group of customers that bought these machines or rented from us from the rental fleet and work with them. We would like to bring that group together with the asset owners in the big industry. Together you would then want to shape the development of the algorithm in practice. That has to be done in a special project, because the industry deploys cleaning primarily in a particular, critical situation. Then there is no time and opportunity for a test project. A setup that is available for a while would be ideal to effectively work there with the machine, the algorithm and a software developer in a practical environment. So you would have to include it in an innovation or study project."
To be successful with this form of cleaning, you have to convince both industrial cleaners and asset owners.
"In fact, we mainly need to get the asset owners on board; they will then work with cleaners who offer this technique. The advantage of laser cleaning is that it is quiet and much less polluting than sandblasting, for example. You also have no secondary waste, reducing the overall operational burden. With AI and the use of robots, you could clean full time. You can't get that done with human operation."
Published by https://www.europoortkringen.nl/netalux/