FPT Software - From Linear to Circular – Greening Manufacturing with the Power of AI
Key players insights
From Linear to Circular – Greening Manufacturing with the Power of AI
Author: FPT Software
Manufacturing will be circular. The take-make-waste approach is ecologically harmful and exposes enterprises to regulation violations and reputation loss. To tackle climate change, firms should switch to circular production. AI is intended to ease the protracted transition.
From linear to circular
Circular production serves as an alternative to the linear model based on several criteria, such as efficient use of energy, materials, and resources, eliminating waste and pollution, and minimizing the use of raw and non-renewable resources. The general goal of circular production is to keep products and their components at their highest value for a longer time, optimizing their use not only for one more time but for as many as possible. Circular production, which eventually leads to the circular economy, has been greatly endorsed by governments as part of the attempt to achieve their goals of reduced carbon or carbon neutrality. For instance, the EU adopted the new circular economy action plan, wherein legislative and non-legislative measures were introduced concerning the entire life-cycle of a product. New proposals include the introduction of Ecodesign, which sets out a list of requirements for some product groups, aiming at improving their circularity, energy performance, and sustainability . Similar measures have also been announced by other governments. The UK introduced the Extended Producer Responsibility framework, requiring producers of certain goods to pay for the full costs of their product life-cycle, including collection and recycling costs, which are currently paid wholly or partially by local governments .
Closing the loop of production not only benefits manufacturers in terms of regulatory compliance and sustainability but also financial incentives. Minimizing resource consumption and maximizing material values means cost savings and new growth streams. Indeed, McKinsey & Company estimated that the circular consumer goods market brings about a €400 – 650 billion opportunity for European companies. The growth comes from new segments of the circular economy, such as recycling, resale and rental services, and refurbishing products . Furthermore, circular models are more appealing to consumers. A report by Capgemini Research Institute shows that 72% of the surveyed consumers are willing to adopt circular practices, including purchasing products with higher durability and repairing products to increase their shelf life .
Closing the loop with AI
Changing from a linear to a circular model involves a complete corporate change. AI can smooth the transition. AI applications for circular manufacturing include:
Optimize product designs
Product designs play a critical role in enabling circular production – an optimized design not only minimizes material and energy consumption but also maximizes durability and reusability. Indeed, up to 80% of a product’s total emissions are determined by the decisions made during the design phase . Traditional product design relies upon the expertise and experience of human engineers, who are prone to errors and bias. AI, on the other hand, offers a more accurate and faster approach to product optimization. Inputs (product data) and outputs (desired performance) are determined by human engineers, which are then used to run simulations on different product designs. It may seem like a conventional product optimization approach but what the AI-powered systems differ is that the initial simulation is used as training. Once training is complete, the AI-powered system can learn, adapt and generate results for other inputs, outputs, and designs other than the initial variables. This means that the AI system can provide results significantly faster than the traditional human-based approach, or even other digital product optimization systems. Not only so, the AI system can actively learn and improve accuracy rates as it operates.
McKinsey & Company offers a real-life example of such an application. The company in their study used AI to optimize turbine designs for hydroelectric plants. With the experimentation, the company successfully cut design time by 50% and the end-to-end design process by 25%. But most importantly, the new design improved efficiency by 0.4 percentage points compared to the traditional approach, which translates into 13.5 gigawatt-hours (GWh) of annual energy production, enough to light over 3,000 homes.
Improve e-waste recycling and reuse
With the booming demand for electronic devices, the world is facing a tsunami of subsequent electronic waste (e-waste). It is estimated that 50 million tons of e-waste are produced annually, which can double to 120 million by 2050 if left unchecked. Sadly, only 20% of them are recycled, while the remaining are incinerated or end up in landfills . Although electronic waste contains valuable materials, some are even scarce and expensive; what holds back e-waste recycling is the extreme difficulty in dissembling them from used devices. For smaller items such as smartphones and tablets, the recycling process is even more challenging as they consist of much smaller, more delicate components that require the touch of humans. As it is time-consuming and costly to recycle, mobile devices often have their battery removed and the rest shredded. Precious materials are lost in the process, and new materials and energy are required all over again to produce similar components for a totally different device.
AI offers a potential solution for e-waste recycling problems. The Idaho National Laboratory is currently studying AI applications in improving recycling efficiency for portable devices. The researchers develop an AI-powered system to automatically identify types of smartphones, remove their batteries, and collect other high-value components. What’s ground-breaking about the system is its ability to identify device brands and models (e.g., iPhone 14 Pro Max) to locate and separate its battery and components. To do so, the researchers have developed a database of 2D and 3D scans of different device models and applied a machine learning-based approach to train the system on identifying devices and locating desired components. Despite being in its early phases, the project has received high interest from the public as a potential game-changer in recycling smart devices. In December 2021, the project was awarded US$445,000 from the US Department of Energy .
 European Commission. Ecodesign for sustainable products.
 McKinsey & Company. Playing offense on circularity can net European consumer goods companies €500 billion
 Capgemini Research Institute. Circular economy for a sustainable future.
 McKinsey & Company. Product sustainability: Back to the drawing board
 World Economic Forum. The world’s e-waste is a huge problem. It’s also a golden opportunity.