Applying Modelling & Simulation, Machine Learning to Address Packaging Challenges: A Snapshot of Ongoing Research and Development

Introduction:

In the packaging industry, the product's life cycle involves a constant interplay between showcasing, safeguarding, and maintaining product quality. While traditional build and test approaches have proven successful over time, the surge of e-commerce and direct-to-consumer (D2C) deliveries presents new hurdles. These novel challenges, coupled with a growing commitment to sustainability, demand an innovative approach, as conventional strategies find it difficult to adapt to the evolving requirements of modern packaging.

The unprecedented rate of digital progress, embodied in developments such as modelling & simulation technologies and machine learning, holds considerable promise for revolutionizing packaging. The goal of this piece is to offer a glimpse into current research and development endeavours harnessing these technologies to navigate the changing landscape of packaging.

Modelling and Simulation:

Several companies have leveraged advanced modelling to tackle challenges in primary packaging design and evaluation using digital tools. However, simulating transit scenarios remains a significant hurdle due to the intricate nature of:

Material characterization of paper products and corrugated boards, which are sensitive to changes in temperature and humidity affecting their strength.

Understanding the loads packages endure due to unpredictable vibrations in transit and potential mishandling.

Virtual prototyping enables swift iterations and cost-effective testing, reducing the dependency on physical prototypes. By simulating transit environments, packaging flaws can be detected and corrected, enhancing safety, longevity, and efficacy.

We have invested significant time and resources in creating robust virtual transit tests for traditional palletized shipments and new e-Commerce/D2C shipments. Such efforts are critical as brands strive to meet sustainability targets while addressing the challenges e-commerce presents to packaging.

When inputs - including geometric definitions, material characterizations, and definitions of loads - are meticulously considered and implemented, the simulation of the complete supply chain becomes possible. Modelling the entire supply chain allows for early identification and mitigation of potential risks and bottlenecks, ensuring seamless operations that cater to the specific challenges of product distribution.

Machine Learning:

Machine learning (ML) involves the use of data and algorithms to improve systems. It has multiple applications within the packaging industry, with several examples illustrating its impact:

Brands often possess a wealth of legacy data, including packaging specifications, lab reports, and field performance data. Extracting insights from this data can drive package simplification, standardization, and new design initiatives. Specific use-cases include:

1.1 Designing rigid packages with improved protection performance and sustainability: ML algorithms are built using meta data and performance data from numerous historical bottle designs . This tool helps optimize shape and thickness, considering factors such as type of plastic, recycled content, energy consumption, and recyclability potential.

1.2 Automating carton design: ML-powered software can utilize historic transit test results and corresponding metadata to automate the design and transit performance evaluation process for snack and cereal cartons.

1.3 Detecting packaging faults: Computer vision and ML algorithms can identify and predict defects during production, thereby enhancing quality and reducing waste.

1.4 Predictive analytics: ML models can predict results of long-duration tests ahead of time, predict results of a weight loss test typically done over six months using only three weeks of data.

Summary:

The integration of modelling, simulation, and machine learning is driving a significant shift in the packaging industry, particularly as the industry responds to sustainability challenges. Despite the complexities of modern distribution, these technologies offer potent solutions to sustainable packaging demands. Research and development will continue to evolve, resulting in even more innovative solutions that will redefine the packaging industry, guiding it towards a more sustainable and smarter future.

#PackagingInnovation #DigitalTransformation #MachineLearning #SimulationTechnology #EcommerceChallenges #SustainablePackaging #AIinPackaging