PROCESS CONTROL | MACHINERY
seeing leaders in this space invest in empowering front-line teams with systems and processes directly. Although we are also seeing manufacturers building out sophisticated data science teams and data lakes, the winners will be bringing usable and impactful insights directly to the plant teams who can actually solve production issues.
Sustainable trends
Sundblad also sees sustainable manufacturing continuing to be a key focus for many years. “However, with unstable supply chains and rising costs, we will see more experimentation with recycled materials as both a corporate social responsibility initiative and to support the bottom line. However, testing new materials can be risky, so finding a balance and partners to help reduce the risk will be key.”
He identifies this as an area where AI could
make a real difference. “The rapid adoption of AI is an exciting trend for all types of industries, and plastics compounding is no different. Workforce is a continuing challenge, and AI has already been making it possible to augment the workforce and help the teams that exist be more productive and ease the learning curve for new employees. Instead of thinking of AI for employee replacement, it should be thought of as an assistant or productivity amplifying tool. For example, Large Language Models (LLM) can be used to distil large manuals that can take months to memorise into a natural language, accessible library to be queried. Building on that, integrating this new format of manuals to actual production data can remove the billions of data points of ‘noise’, and provide clear predictive recommendations.” Sundblad warns, however, that choosing an AI
partner is a challenge that must be approached with care. “One thing to keep in mind with AI adoption is that there are many new AI focused companies, some of which are helping manufactur- ers,” he says. “However, many of them are not starting with true manufacturing problems, or an understanding of the industry. It is important to evaluate these vendors based on a long-term partnership, rather than a flashy set of tools.” Looking to its own developments, Oden says it has launched both predictive quality and predictive recommendations capabilities recently. For predic- tive quality, it offers a new capability that helps manufacturers, particularly those with compound- ing processes, auto-detect quality issues and other events on the line. This is based on historical, rather than theoretical, runs. It means that staff can be alerted in real-time of quality issues that can then
www.compoundingworld.com
be quickly addressed rather than waiting for an offline quality test and continuing to make a product that will ultimately need to be scrapped. For predictive recommendations, the company
has introduced a new tool that provides the insights needed to maximise run cost efficiencies and/or speed in a few clicks. It predicts what process settings will yield the best results at a line and product level based on historical achievable runs, predicted quality results, and current con- straints. By allowing process engineers to compare the predicted impact of different process settings it is possible to run more effective process improve- ments, more quickly and with a large amount of reliability and little risk. The company says that an early pilot customer achieved a nine times return-on-investment (ROI) through cost reductions and speed improvements. Predictive quality analysis with a real time model and predictive recommendations were used to optimise for cost and line speed, with the result that the customer saw reduced costs of 5% on the one line it was implemented on without reducing quality. The solution also identified opportunities to increase line speed without affecting other con- straints. Oden says the customer plans to scale this solution to other machines to improve margins further. For the future, Oden says it is working on
projects that use Generative AI to provide work instructions directly to operators and other front-line staff. The goal of these developments is to make it easier and exponentially faster for operators to get to the same level of productivity as an experienced operator. The company is also exploring ways to make it
easier for teams to understand what is going on in production. Strategies involving Data Lake and Data Warehouse architecture offer opportu- nities to connect high-quality, contextual- ised process data with traditional enterprise data. Oden says its proposed solution does
IMAGE: LEISTRITZ
Left: New control systems such as LinXX from Leistritz aim to upskill plant operators to lift productivity (page 52)
August 2023 | COMPOUNDING WORLD 49
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62
