The polyurethane industry has been improving its processes since the first PUs were invented in the 1930s. EuroPUR rounded up the state of play in this process at its recent meeting in Amsterdam, Simon Robinsonreports.
In business, doing things better is often not as commercially interesting as doing things in the most cost-effective way. In fact, trying to do things too well can seriously reduce profits – and a company’s ability to meet its customers’ orders at a price they are willing to pay.
There is always a tension between investing in new technology or processes, and the need to deliver good returns to the owners and shareholders. The PU industry has taken process improvement to its heart, however, and the machinery employed today is a far cry from the first prototype impingement mixers based around fuel injectors.
Digitalisation is now becoming more important. A recent poll at the EuroPUR meeting in Amsterdam in November found that respondents agreed that ‘Digitalisation and AI are not just buzzwords, we are on the verge of something bigger than just an improvement in processes which could fundamentally change the way the foaming industry works in the coming decades.’ The meeting pulled together six speakers from the machinery side of the business to discuss how they believe digitalisation and AI will affect the industry in the near future.
One challenge that the industry faces is that, according to the session’s chair, EuroPUR president Bart Ten Brink: ‘This industry makes machines with a long lifespan, longer than their economic life times.’ In other words, machines continue to be useful long after they have been written off by accountants. Consequently, many machines in use in the PU foam industry were built either before there was demand for more than process control or, if they are designed to export data about processes, the connections might not be suitable, or the data could be in the wrong format for modern data processing.
‘Some people have thrown out all the old machines and replaced them, some are doing it gradually,’ he said. ‘The big question is: which approach do you take?’ There was a lively debate on just how open different companies are likely to be with their data. If raw materials suppliers knew the state of supplies at their customers, for example, then they might be able to allocate resources more accurately. Or that foamer’s customers could pass their orders directly into their upstream suppliers’ production planning software. While either of these scenarios might enable foam to be made more efficiently, how likely is either of them?
Christoph Moisel, an expert in industrial IoT and artificial intelligence, is all for sharing data between players in the value chain. ‘Working together, we can develop holistic end-of-life solutions that cover the range, without a lot of interfaces,’ he said. ‘There are many industries where there are definitions for data exchange. We need standards for this in the flexible foam sector. If we do that, then there is a good chance of success. There could be a new business model to help manage risks and help companies to become more agile.
Ten Brink, who was previously managing director of FoamPartner, said: ‘In my past, I heard that this is a nice idea. But many companies give the raw material supplier access to the gate, and no further. They don’t want them to see what they are doing.’
Torsten Heinemann, head of group innovation at Covestro Deutschland, believes that compliance and data security concerns get in the way of greater openness. ‘Looking at North America , this seamless interaction is much more advanced,’ he said. ‘From our perspective, looking at process optimisation at the customer, it is already working. In APAC, we can see that customers are really willing to experiment and to fail. That is the global perspective. What is the best of all the worlds? We are not fully exploiting these opportunities in Europe. I wouldn’t say we are closing the doors, but there are opportunities which are exploited more in other regions.’
Moisel said: ‘A foamer does not have to share all their secrets, but the upcoming regulations and the need to have end of life solutions, I think that we will find that customers want to have a form of transparency to share more data.’
But perhaps the attitude of Euroepan flexible foamers was summed up best by Angel Vinas of IPF. He said: ‘In my experience, we all like to talk about data sharing and data management, but there comes a point when we talk to a customer about specifics and even when we show the benefits of sharing different levels of knowledge, companies are reluctant. It is really difficult for me to see data sharing between companies.’
Newer companies are more open to sharing, he said. ‘Older, more traditional companies want to do more things inside. I see companies willing to improve, to open the doors especially if they are newer, younger companies. For them it is a big gain.’
Hennecke’s Joachim Berthold, who is the company’s senior director for sales of refrigerated appliances, moulded foam and slabstock, added: ‘The larger companies can probably use data between plants… they will be the pioneers and they will learn from this. In a competitive situation, no way.’
So while it is currently difficult to see companies sharing the data they have with their supply chain partners, how far could it get within the company itself ? Michael Tillmann, technical vice presiden, at Fecken Kirfel, helped the meeting with some definitions before looking at how to collect and use data. ‘First of all, the machines in a work group need to network with each other and know what the other machines are doing,’ he said. ‘These need to be networked to a server system or a cluster of plants. This is vertical networking. Each machine can send data to the cloud.’
He said that Fecken Kirfel machines have been fitted with IPC controls for 35 years now. This enables them to be connected to intelligent software solutions, and offers networking possibilities. ‘There is advanced digital technology in the nesting and routing functions of our cutting machines,’ he said. ‘There are also superior machine controls and planning systems, and connections are available to the cloud for production data analysis or production monitoring or predictive maintenance. Looking to the future, AI could autonomously plan production, check part quality and self-optimise the process.’
Hennecke’s Joachim Berthold said that his company is now offering a completely new control and data acquisition system. He added that Foamware controls all of its continuous slabstock lines and enables visualisations. ‘This allows the data to be displayed to make the process more transparent to the user,’ he said. ‘We can show the machine status – how the pump is running, for example. There are systems for preventive and predictive maintenance, and we also offer remote support, troubleshooting, service manuals and also a spare part shop.’
Hennecke systems can also interface with other third-party systems, but it will be some time before truly smart foaming machines are developed, Berthold believes. ‘The foaming process is very complicated, so we cannot give automatic recommendations to improve quality,’ he said. ‘We are some way away from automatic or semi-automatic advice. Machine operation relies strongly on the expertise of the operators to get this experience into data is very difficult.’
However, Berthold said, an expert system with intelligence might one day be able to automatically adjust machine settings and optimise product quality, reduce waste and consider ambient conditions like humidity or temperature and give advice. ‘It’s a long and tough way to overcome the hurdles, but we understand the benefits of further digitalisation,’ he said. ‘It’s going to happen; the question is how fast, and when. We cannot stop it.’ Fecken Kirfel’s Tillman outlined some of the barriers that have to be surmounted before machinery become truly smart. ‘AI can help as an assistant to optimise machines, and it could help new operators learn the machine and avoid mistakes,’ he said. ‘Should machines be fully controlled by AI? We say no, because a deterministic system must keep the machine safe. This is why the EU machinery directive excludes software as a way to keep machinery safe.’
He added that it is important to bring staff along on the digitalisation journey. ‘People fear that they may lose their jobs, or that they are being monitored,’ he said. But these qualms are not restricted to the shopfloor. ‘Companies have reservations about the security aspects of cloud systems; we need to explain what we are doing when we use AI and digitalisation. There is a big chance that machine learning and AI can help to make our daily lives easier, to share information and improve processes.’
Moisel said that interoperability poses a further challenge. ‘If you are in a world of new smart machines, you can do a lot, but you may have many sites with old machines from many different suppliers,’ he said. ‘The question is how to join them and capture the right data, what data architecture do you need, and how to apply it. About 80% of digital is this groundwork, standardising processes, getting the right connections to machines. If you get these questions right, you can achieve a lot. You need a strategy.’
Specialist suppliers should be able to assist. IPF’s Vinas explained that because his company helps to build factories for flexible foamers, he feels it is in a good position to help the industry achieve digitalisation success with its MyFoamPlant software.
Neveon’s lean and digitalisation expert Christan Auer rounded the discussion up with some fundamental points. ‘There are many new technologies, and it is important to choose the right tools,’ he said. ‘Secondly, data is the basis of digitalisation and AI. Its quality is very important; it must be reliable and current. Digitalisation is not just about improving processes.’
Making the transformation
Neveon has put a lot of effort and time into digitalisation, its lean & digitalisation expert Christian Auer told the meeting. ‘The process involved interviews and workshops across the whole company to find the needs and potential of the business and merge it with new ideas to create a destination,’ he said.
He said that Neveon plans to become the leading digitalised business company in the core business and to use digitalisation to realise growth along the value chain, add value and improve the customer experience. ‘We want to optimise our internal processes and make it better for our customers to deal with us,’ he said.
This process has led to Neveon creating 20 focus areas which the company ranks across two groups of metrics: impact and effort; and strategic and economic relevance. On the operations side, the company wants to combine lean processes and digitalisation. ‘We need processes to be lean, and to digitise them,’ he said. ‘Software architecture is important, but we want to look at the big picture.’
The company is looking at each digitalisation tool to see how they fit within the overall landscape, and how they would work with its existing tools. ‘If there is a fit, then the interface between them needs to be worked out as well,’ he said.
This process is leading to two concrete things. First, the company is developing a consistent planning process to make things clearer for clients. ‘It will also help make us faster and to make better decisions, and to optimise our processes,’ he said.
Secondly, Neveon is working on robotic process automation, using digital robots that can imitate the behaviour of people to automate processes that are time-consuming or which have a high risk if failures happen, and which are strongly rule based. These should be easy to automate, he believes. ‘It is important to use the data which we have and to use it to generate more transparency, to help us make faster and better decisions and optimise our processes,’ he said. He cited specific projects, including a consistent planning process. ‘This is very important for our clients,’ he said. ‘This will look at processes from our customer through production and our supply chain.’
Neveon was created by Austrian conglomerate Greiner, and employs 3500 people from six former Greiner companies and a joint venture. It has 61 locations in 18 countries, includings 10 foaming lines. It has sales of about €500m/year across its three divisions, comfort and care, mobility, and speciality businesses.