2-stage PET bottles from one line
How Krones is integrating the Netstal plants
The acquisition of Netstal by Krones has been the subject of much discussion in the industry. Time to ask around in Neutraubling to find out how the new arrival is getting on. We spoke to Jochen Hirdina, Head of the Plastics Technology department. In passing, the conversation drifted towards AI.
PETplanet: Mr Hirdina, Krones has consolidated Netstal in the “Filling and Packaging Technology” segment since March 28, 2024. How did this come about?
Mr Hirdina: In our approach of offering systems for the entire PET bottle cycle, we lacked an injection moulding machine. With MHT, we had added an outstanding preform mould manufacturer to our portfolio years ago, so the acquisition of Netstal was a logical consequence.
PETplanet: Netstal had developed the side extraction as a new range. Would you have taken over Netstal if there had only been the range with top removal?
Mr Hirdina: Strategically, I think so. However, side extraction is more favourable for our system concepts.
PETplanet: With Netstal, you have acquired a plant manufacturer that sells mainly to converters and less to bottlers. Does Krones now also want to address the converters as a target group?
Mr Hirdina: Yes and no. When we acquire companies, we leave them largely independent. Examples of this are Kosme, Gernep or MHT. Nevertheless, we see ourselves as a system provider rather than an individual machine supplier. For our target group, the bottlers, we therefore offer an integrated injection moulding stretch blow moulding system.
PETplanet: How does it all fit together?
Mr Hirdina: We call our new system Prefero. It combines our line concept with the plastic cycle. The line starts with PET material preparation. The preforms are produced in the Prefero machine (built by Netstal) with an MHT mould, which are transferred by the Prenexo Link directly to the stretch blow moulding machine where they are processed into bottles. The performance of the stretch blow-moulding and injection moulding machines is coordinated. Depending on the capacity, the clamping force is 300 to 500 t and moulds with up to 192 cavities can be used.
PETplanet: In other words, in contrast to a 1- or 1.5-stage process, the preforms cool down completely and are then specifically reheated in the stretch bubble?
Mr Hirdina: Yes, this makes sense for high outputs. The advantage of direct transfer to the stretch blow moulding machine (SBM) is that the preforms are fed in the order in which they were injection moulded. This allows the SBM to constantly adjust to fluctuating material qualities, which is particularly important with rPET. If preforms are processed in the wrong order, one batch could result in clear bottles, but the next batch could result in milky or yellowish bottles. No SBM, however good, can compensate for this. With direct transfer, we also reduce the influence of fluctuations in humidity, temperature or preform weight, as these factors only change slowly and the SBM control system can react accordingly.
PETplanet: Where does AI fit into the scheme of things? You already have experience with such control systems, for example with the Linatronic AI empty bottle inspection systems.
Mr Hirdina: You’re right, with AI we can significantly reduce sorting errors. However, I think we will have to wait a little longer when it comes to controlling systems or lines.
PETplanet: What is the difference between a classic control system and AI?
Mr Hirdina: A control system consists of control loops: An input value influences an output value, and if this changes, the input value is adjusted to achieve the desired output value. AI, on the other hand, works differently: It analyses a wide range of data and measured values and automatically determines correlations in order to find optimum solutions. This already works well with Linatronic AI, for example it can distinguish a drop of water on a bottle from an actual particle.
PETplanet: Where does the AI get this amount of data from? From all systems installed worldwide?
Mr Hirdina: This is a sensitive topic. Nobody gives out their data voluntarily. If a company participates, we guarantee anonymisation of the data. We have implemented many protection mechanisms so that the data cannot be traced.
PETplanet: Is this comparable to the development of bottle designs, where designers draw on projects that have already been realised? But not in the minds of the designers, but in the IT?
Mr Hirdina: Yes, something like that. A new bottle design is rarely developed from scratch. We hope that this acceptance will also increase in AI projects. Our initial successes confirm this: with Linatronic AI, we can reduce reject rates. But it will probably be a while before everyone puts aside their reservations.
PETplanet: Thank you very much, Mr Hirdina.
