Accurate predictions of part quality while the process is still live
Tomi Villilä is one of the 'plastics supremos' at Sartorius in Helsinki. As Development Manager Injection Molding, he shoulders responsibility for the quality of the injection molding processes for as many as 50 individual parts that make up one pipette. In response to ongoing developments, he and his strong team recently began using Kistler's ComoNeo process monitoring system with the ComoNeoPREDICT function. Online quality prediction allows forecasting of part quality during the injection molding process, based on numerous learned process parameters and related part-specific quality criteria. Villilä has garnered over ten years of experience in the plastics processing industry, and he is no stranger to the use of cavity pressure sensor technology: 'I've known about Kistler for a very long time, and I'm well aware of the excellent performance their sensors deliver. In fact, I focused my degree thesis on cavity pressure sensors made by Kistler! And since then, I've often been involved with Kistler products throughout my career, so I'm very familiar with them.'
Although ComoNeo only went into operation just recently, users at Sartorius are already convinced of the system's benefits:
"We started out with a fairly non-critical part – a fixture – so we could get to understand the system, practice using it and improve our own process. But even after a short time, the results have been so good that we're planning to deploy six to eight additional ComoNeo units for other critical components as the next step."
Tomi Villilä, Development Manager Injection Molding at Sartorius
ComoNeoPREDICT is especially suitable for processes where the quality requirements are very demanding. The first step is to use the associated Stasa QC software to generate a testing plan – known as Design of Experiments, or DoE – that includes all the parameters needed to determine the process. Which dimensions are to be achieved? How must the machine be set to achieve them? Might it be possible to reduce the cycle time? Comprehensive and precise answers to many questions such as these are provided by the Stasa QC PC software and the ComoNeoPREDICT process monitoring feature. For convenience, the DoE can be generated on a PC and implemented digitally in ComoNeo at a later stage. The Kistler sensor technology installed in the mold supplies the basis for transparent process management of all material-related and part-specific attributes – the part's 'fingerprint', so to speak. Based on the measured cavity pressure and the specific part characteristics measured in relation to it, Kistler's system delivers a prediction of part quality while the injection molding process is still in progress. The benefit: intelligent optimization is possible with no need for time-consuming quality measurements. It took Sartorius only two or three days' work with ComoNeo to set up the specific DoE for modeling of the injection molding process for the fixture; then, following optimization with Stasa QC and ComoNeoPREDICT, very precise results were obtained. Because such large numbers of machine parameters are considered, users gain a clear understanding of the limits of the injection molding process and the relevant part characteristics – paving the way for intelligent process monitoring. It takes very little time to understand what is happening in the injection mold, and to gain knowledge about the entire process. Villilä notes enthusiastically: 'Even colleagues without lengthy experience can achieve outstanding results in next to no time!'
Fast setup and intuitive operation
Based on the prediction model that is generated, ComoNeo can significantly reduce the percentage of bad parts, and they can be segregated automatically if desired. Here as in other respects, simple operation and integration are the keynotes: 'This system is really simple to handle – almost like a smartphone. You only need five or ten minutes to understand how it works. And setting up a DoE poses no problems either, thanks to the software that comes with the product. You don't have to be a mathematician to understand the whole system – which is not the case with other programs. Once you've generated the DoE, you can simply upload it into ComoNeo and then you already get accurate feedback for process optimization in the test phase,' Villilä explains.
He already has ambitious plans for the future: 'I'm certain that the benefits we can achieve with ComoNeo will also be of interest to other Sartorius units – especially because this system is so flexible. I already gave a presentation to our colleagues about our experiences with ComoNeo at the company's gathering of experts in London. If it were up to me, all Sartorius technical parts that require high precision would be produced with the Kistler system in the future.' With this idea in mind, a workshop will soon take place in Helsinki with a practical demonstration of ComoNeo, ComoNeoPREDICT and the online quality prediction feature for interested staff members from the entire Sartorius group.