AI: A Game Changer for Plastic Packaging

Plastics are light­weight, dur­able and easy to pro­cess, mak­ing them truly ver­sat­ile mater­i­als for almost any applic­a­tion. This makes it no sur­prise that more than 60 per­cent of pack­aged goods in Ger­many are wrapped, bagged or boxed in plastic. The goal of the plastics industry is to sup­port the cir­cu­lar eco­nomy through ideal pack­aging design (design for recyc­ling), min­im­ize envir­on­ment­al impact and at the same time provide optim­um pro­tec­tion for pack­aged goods. Ideally, pack­aging will con­sist of as much recycled mater­i­al as pos­sible. New EU reg­u­la­tions also require plastic pack­aging to be made from up to 35 per­cent recycled mater­i­al at a min­im­um by 2030.

Enter the research being done by KIOp­tiPack, one of two innov­a­tion labs that form part of the AI Applic­a­tion Hub on Plastic Pack­aging (see below, ​“The AI Applic­a­tion Hub on Plastic Pack­aging”), which is receiv­ing fund­ing from the Ger­man Fed­er­al Min­istry of Research, Tech­no­logy and Space. Togeth­er with their pro­ject part­ners, research­ers at Fraunhofer IVV are using arti­fi­cial intel­li­gence as a game changer in recyc­ling and for pro­cessing recycled mater­i­al into pack­aging to meet the tech­no­lo­gic­al chal­lenges and high stand­ards involved in pack­aging pro­duc­tion. The goal is to reduce plastic waste, cre­ate more sus­tain­able pack­aging and lower demand for raw mater­i­als through reuse as recycled mater­i­al, thereby increas­ing the reli­ab­il­ity of the sup­ply in Europe through a focus on cir­cu­lar­ity. To achieve this, the research­ers at Fraunhofer IVV have developed AI-based optim­iz­a­tion tools and con­nec­ted them through a data space. This approach makes it pos­sible to effi­ciently take account of the many require­ments that plastic pack­aging has to meet. In par­tic­u­lar, it also unlocks ways to recycle these pack­ages and optim­ize them over­all. Along­side envir­on­ment­al com­pat­ib­il­ity and impact, the aspects of func­tion­al­ity, design, cus­tom­er accept­ance and man­u­fac­tur­ing feas­ib­il­ity are also factored in. After all the data is fully integ­rated, the AI tools sug­gest the best pos­sible pack­aging designs for a cer­tain product while also min­im­iz­ing mater­i­al use. The data space brings togeth­er all of the digit­al data on mater­i­al prop­er­ties, ori­gins and des­tin­a­tions to do this and cross-links the inform­a­tion. In this way, it provides a shared infra­struc­ture for exchan­ging data past the fact­ory gates. The data mod­el under­ly­ing the full pack­age, which enables the shar­ing of data among the vari­ous stake­hold­ers in the first place, was also developed at Fraunhofer IVV.

Digit­al­iz­a­tion of mater­i­al flows to pro­tect consumers

“To improve the incor­por­a­tion of recycled mater­i­als into pack­aging solu­tions, a shared data space is cru­cial,” says Andrea Büttner, the insti­tute dir­ect­or at Fraunhofer IVV. ​“Recycled plastic is a valu­able mater­i­al. But the qual­ity of the recyc­lable plastic has to be right. If the recycled mater­i­al is con­tam­in­ated with undesired sub­stances such as oth­er kinds of plastic, print­ing inks or degrad­a­tion products, it can severely impair qual­ity and even make the mater­i­al unus­able. Once plastics are recycled, they can con­tain con­tam­in­ants in prin­ciple.” And that poses a par­tic­u­lar chal­lenge for the industry. This means that across the entire industry, from pro­du­cers of recycled mater­i­als and com­pan­ies that man­u­fac­ture pack­aging films to the food sec­tor, sys­tem­at­ic data shar­ing is needed, along with a single lan­guage for deal­ing with plastics. The mater­i­als involved are mainly poly­ol­efins, which rep­res­ent the biggest group of plastics and can change sub­stan­tially dur­ing processing.

AI helps with choos­ing the right recyc­ling method

“AI-driv­en tools are a must when it comes to high-qual­ity recyc­ling of plastic mater­i­als. The mod­u­lar soft­ware solu­tions we devel­op at Fraunhofer IVV help with aspects like char­ac­ter­iz­ing and ana­lyz­ing mater­i­al prop­er­ties, and they link dif­fer­ent inform­a­tion flows togeth­er in such a way that a suit­able field of applic­a­tion can be iden­ti­fied for recycled mater­i­als with fluc­tu­at­ing prop­er­ties,” says Mat­thi­as Reinelt, head of the Shelf Life and Pack­aging Mod­el­ing group at Fraunhofer IVV. ​“Our AI tool for pack­aging pro­duc­tion uses as much inform­a­tion as pos­sible about the recycled mater­i­al to determ­ine the optim­um pro­cessing chain so that what ends up on store shelves is a sus­tain­ably man­u­fac­tured yogurt cup with homo­gen­eous wall thick­ness and the desired shape, for example.” The research­ers also ana­lyze the qual­ity of the recycled mater­i­als using chro­ma­to­graphy, which in turn also relies on AI tools for improved sub­stance iden­ti­fic­a­tion. The goal here is to pre­vent unsuit­able or con­tam­in­ated recycled mater­i­als from enter­ing the cycle, which can occur even with good sorting.

The AI-based optim­iz­a­tion tools and inter­con­nec­ted data space provided by Fraunhofer IVV and its pro­ject part­ners are even dir­ectly incor­por­ated into man­u­fac­tur­ers’ pro­duc­tion pro­cesses in some cases. This makes real-time sug­ges­tions for fur­ther pro­cessing pos­sible, depend­ing on the recycled mater­i­als being used on the machines on-site.

AI tools address recycled mater­i­al gap

Use of recycled mater­i­als in plastic products is rising stead­ily. How­ever, even des­pite this trend, the volume of recycled plastics will be insuf­fi­cient to reach the levels of recycled mater­i­als required by law. This will lead to an ever-widen­ing gap between sup­ply and demand in the case of recycled plastics over the next few years. By 2030, demand for recycled mater­i­als is pro­jec­ted to sur­pass sup­ply by 30 per­cent. ​“Isol­ated solu­tions won’t be enough to bridge the recyc­ling gap in the plastics sec­tor in the long term. We are address­ing this issue with our AI-driv­en optim­iz­a­tion tools, which inter­link mater­i­al and inform­a­tion flows. We’re bring­ing the stake­hold­ers from a wide vari­ety of dif­fer­ent indus­tries togeth­er and try­ing to respond more effect­ively to resource short­ages,” Büttner says.

Experts from the research sec­tor, industry, gov­ern­ment and asso­ci­ations from all across Ger­many will be gath­er­ing at the Zukun­fts­for­um Kun­st­stoffkre­is­lauf event, to be held in Ber­lin from July 1 to 3, to dis­cuss a full range of top­ics related to plastics, pack­aging and recyc­ling of both. The goal is to bring people togeth­er to dis­cuss innov­at­ive approaches to the trans­form­a­tion to a resource-effi­cient cir­cu­lar eco­nomy and trans­late con­crete solu­tions into real-world practice.

The AI Applic­a­tion Hub on Plastic Packaging

To close the cycle for plastic pack­aging as much as pos­sible, 51 part­ners from industry, the research sec­tor and civil soci­ety are work­ing togeth­er in two innov­a­tion labs: KIOp­tiPack, which focuses on design and pro­duc­tion, and K3I-Cyc­ling, which deals with recyc­ling and sort­ing of mater­i­als. The labs were cre­ated to enable cross-lab shar­ing of data and ensure that all rel­ev­ant find­ings are taken into account across the entire value chain.