Compressor production at Atlas Copco's Airtec division in Wilrijk is one of the jewels of our manufacturing industry, driven by a sound approach to innovation in both product and production process. And although Atlas Copco has been around since 1875, the company remains resolutely focused on the future. We met Karl Mast, one of the pillars of the company's Airtec division.


How come Atlas Copco is not a Factory of the Future? Despite a very focused vision of the future, Atlas Copco is not yet a Factory of the Future award winner. Karl Mast, VP Global Operations and member of Agoria's Factories of the Future steering committee, explains why: "From my position on the Factories of the Future steering committee, I often see large companies that are doing wonderful things in line with the spirit of this programme, but whose size and nature means that they cannot meet all the requirements. Atlas Copco is one of them.

A high-tech product requires high-tech machines. In recent years the company has invested in a dozen machines, with a focus on automation.

Human Centred Oganisation

The ‘human-centred’ transformation is a good illustration of the different approaches implemented by companies, according to Karl Mast. "For some time now, our facilities for staff have been as varied as they are extensive: we have our own sports clubs with the necessary infrastructure, we run a programme to prevent burnouts and we try to provide disabled people with the facilities they need to be able to work here. In the factory, we strongly believe in the principle that ‘the person most directly affected by a problem is also the closest to the solution'.

Based on this belief, we have decided to find a way to give our operators more responsibility in their working environment. This initiative has developed into a team meeting, where our operators monitor daily operations and solve problems on their own. Our prototype department, which has 20 employees, has been working in this way for some time now and we are trying to extend this approach systematically to the rest of our production department, which operates continuously even at weekends.

In this context, employees have to further develop qualities such as coordination, organisation and problem-solving skills. We rely heavily on trust and solid structures. Throughout the Covid-19 outbreak, production was maintained - with the exception of a few hours - partly thanks to our employees’ confidence in these structures. They know that we deal with safety issues and that we take their health concerns into account. This whole situation also illustrates our capacity for innovation. When training on the machines became too complicated, we quickly switched to other solutions, such as video via GoPro cameras."

3 main pillars in production

"Our production process is based on three main pillars: the production of the casings (the compressor housing), the production of the rotors and the assembly," explained Karl Mast. "In casing production, castings are machined to create housings. This is done with an accuracy of up to 10 µm. We try to carry out as few machining operations as possible, often limited to 1 or 2 steps at most that we carry out on the same gauge." This also applies to the next stage of production, the production of the rotors.

Mast: "Everything is done in-house. Previously, this was a multi-stage process in which the rotating cylindrical part had to be turned, the profile milled, the cylinder ground and deburred. Thanks to technological developments and innovation, we have managed to reduce this number to 2 or 3 steps for the smaller versions, and to automate it even further. Flexibility is essential here.

For example, we have gradually moved from fixed automation such as gantry systems to more flexible automation such as robotics. Our cobots – there are now 4 - are also part of our automation strategy. Together with the operators, we looked at the points where the workload was high and decided to use a first cobot to take over the repetitive work. This first cobot was like a 'beacon', which gave us the opportunity to transfer difficult or repetitive tasks to other stages of the process". During assembly, the casings, rotors and other associated components are assembled. "Here we have chosen not to use a conveyor belt; everything is done in an assembly cell by a single worker," said Mast. "Due to the increase in the number of variants - I think tenfold in the meantime - we are now working according to the principles of the lean method with a kit system that makes it easier to handle these different models. Assembly is also becoming more and more automated. For the automotive sector, this has been the case for a long time now, but our series are much more limited in size and less repetitive. In addition, our product range has become more diverse in recent years, with the number of variants systematically increasing, making automation more and more indispensable".

Shortage of workers

“In recent years, we have invested heavily in our machines and in technological development. We are talking about a dozen machines, an investment cost that is increasing rapidly (see below), although we are not making these investments solely for the purpose of optimising our production process.

We have to adapt to the shortage of skilled workers. At some point, the market will no longer be supplied with new profiles with the right skills. We need to anticipate this now, because the days of the traditional machine operator are over. Lifelong learning and retraining will be needed to keep pace. Already today, we are hiring according to attitude, rather than skills per se.”

Digital Factory

Since 2018, Atlas Copco Airpower has invested a total of 35 million Euros in new production technologies in Wilrijk - in addition to the annual R&D&I budget (Research & Development & Innovation, 95 million Euros in 2019). "We are one of the few companies that use 'connectivity' in both product and process. We have been connecting our compressors using the latest network technology for the past decade, allowing us to remotely monitor their use and condition and optimise equipment and customer service. As this involves thousands of units worldwide, this is truly Big Data. In addition, our own fleet of machines - a total of around 100 - is gradually being connected.

In both pillars - product and production - we are able to implement large-scale optimisations. The next challenge will be to link these two 'data strands'. The potential here is enormous: we will be able to optimise processes using the information from our machines worldwide. AI and in-depth learning, among other things, will help us to recognise patterns that we had not detected before. On the basis of these patterns, it then become possible to optimize your product and process, resulting in an even lower life cycle cost for the user. In the end, the purchase price is only a fraction of the whole. Lower maintenance costs, reduced consumption and increased efficiency - that is our goal."

Networked Factory

"The creation of knowledge is crucially important," said Karl Mast. "I am active in Agoria’s Advanced Manufacturing working group. As Atlas Copco, we also collaborate with Flanders Make, as well as with our network of suppliers, in order to realize new developments together. For example, some of the design features of our compressors were developed in collaboration with foundries. A common long-term goal ensures that we pursue the same interests. In addition, we are constantly conducting at least 10 knowledge-sharing projects with non-competitive companies, within the framework of ICON or SBO projects or other partnerships with research institutions such as Flanders Make and Sirris, and the education sector.

However, we also collaborate with companies that, at first glance, seem alien to our own ecosystem. For example, we have set up a learning network with a company from another sector, around the labelling and structuring of Big Data. These are completely different applications: for us, it is machine data, for them it is heavy photo files for quality control. We worked together on this project so that our data could be used in our process. Of course, at a certain point we each went our own way, which makes sense because the applications are different. However, the generic basis is the same and allows us to work together on the solution. Would we have found it without them? Maybe, but we would have missed out on the other side's valuable knowledge."