Automated Vehicle Dismantling – Infeasible or Inevitable?

FUTUR: Ms. Hohlfeld, your company, LRP Autorecycling, is at the forefront as a recycler. What challenges do you see in vehicle recycling as it currently operates?

Hohlfeld:

The challenges begin right at the vehicle procurement stage. In Germany, it is difficult to even obtain end-of-life vehicles due to unimplemented legislation, although things are currently changing here. But even once the vehicle has arrived at our facility, things get complicated: We have very little data available that comes directly from the manufacturer or suppliers. We purchase all the information we obtain about the vehicle through databases. Some of these databases are no longer updated, are incomplete, or are not always 100 percent accurate.

This makes our work very difficult. Our employees spend a lot of time painstakingly gathering data – such as OE numbers for parts – just to create a digital record of a vehicle in our systems. Yet our goal is to identify the best R-strategy for each component and for the entire vehicle. Currently, this relies heavily on our employees’ experiential knowledge. A robust database would be of immense help in determining: Is a component suitable for reuse? Can it undergo remanufacturing? Or does it need to be recycled as material? Just because the vehicle as a whole has reached the end of its life doesn’t mean that every single component has also reached the end of its service life.

For material recycling, we would also need information about the material composition. This is important for future recycling rates and the quality of the recycled materials. We can only separate materials as effectively as we know which materials are used. Especially with plastics, we often have to dismantle a component just to look at the stamp and identify which material was used. This is very time-consuming and hardly feasible in practice.

FUTUR: So time and money could be saved through better data availability, primarily through digitalization? Mr. Uhlig, in your view, what are the advantages and pitfalls of digital recycling?

Uhlig:

Currently, all the data we link to a vehicle serves the purpose of selling used spare parts. Beyond that, dismantling facilities document certain material groups, which are grouped under a waste code number and disposed of. If we look to the future, this will change significantly. We will have to become much more granular when it comes to raw materials. As a software provider, we face the challenge that we need not only data for the sale of spare parts, but also material data.

At the same time, this creates the opportunity to dismantle vehicles much more thoroughly and sort them into differentiated material groups right from the dismantling stage. Intermediate steps such as shredding large mixtures of materials could thus be reduced or rendered obsolete. With manual disassembly, we could theoretically measure and weigh the amount of coolant from every vehicle. However, that would involve many additional steps and be very labor-intensive. In an automated, standardized process with an appropriate database, this would be much simpler. Certain components, such as bumpers, in which the trim and the support structure are made of different materials, could be dismantled more thoroughly than we do today. This would provide us with fine-grained data on material flows. Given the current market situation, this is not economically feasible with manual dismantling. I see a major opportunity here for automated dismantling.

The challenges lie in the differences between production and disassembly. When connecting two new parts with new screws in production, their condition is clearly defined. In an end-of-life vehicle that is perhaps 15 to 20 years old, this screw connection may no longer be non-destructively removable due to rust and corrosion. Add to this the modifications made during the vehicle’s service life that no longer correspond to the condition in which the vehicle rolled off the assembly line. This flexibility where an experienced employee might use a welding torch instead of a wrench to cut the joint is difficult to achieve in automated disassembly. These are the major challenges that need to be solved.

FUTUR: Dr. Kuschan, how can the automation solutions developed by Fraunhofer IPK learn to handle such challenges?

Kuschan:

That will indeed keep us busy for quite some time. But I see many possible approaches. For example, if we know where the screw positions are, different strategies can be applied. The robot first tries to loosen the screw with a torque screwdriver. If that doesn’t work, another strategy is applied, such as a destructive method like welding. The key factor here is the information: What kind of part is it? Do I want to use it later? Are there any problem areas when working with robots there?

We rely on data and information to be able to work safely, efficiently, and effectively. The main advantage of humans is that they can adapt quickly to new circumstances. Production is actually a wonderful environment where the same thing happens over and over – and yet we still have many manual steps there. In disassembly, we have a new, unique case every time. The robot’s strength lies in repeatable work, and that is precisely the strength we don’t actually have in disassembly. That is why digitization and information sharing are so important. Building on this, we can then implement different strategies directly on the robot.

FUTUR: Ms. Mügge, with the Circular Economy Assistant, you are developing exactly such a digitalization solution for disassembly. How does the concept work?

Mügge:

The Circular Economy Assistant which we are developing at Fraunhofer IPK is a central digitalization component in our joint DAADEM project. With it, we utilize data throughout the entire product life cycle that we can make available at the end of life to help recyclers select suitable circular strategies.

We provide data at the vehicle, component, and material levels and factor into our decision-making: What is technically feasible? What do regulations require? And which circular strategy makes the most sense from an environmental perspective? At the same time, we integrate automated image recognition into the concept. With this technology, we can uniquely identify components at the end of their life and assess their quality condition, for example, detecting cracks or corrosion. This data feeds back into the CE Assistant and influences decision-making.

For the disassembly process itself, we use augmented reality technologies to visually guide the process. We can display relevant information such as disassembly instructions, locate components, and specify the order in which fasteners like screws should be loosened and with which tool. With the AR glasses, we also have the advantage of being able to capture data simultaneously. We can see: Which tool is being used? If a screw can no longer be loosened with the screwdriver, what is the next step? We collect all this data to train our robot later and thus take the step toward automated disassembly.

FUTUR: So the AR glasses serve a dual purpose: They support human employees with information while simultaneously collecting data for future automation?

Mügge:

Exactly. They guide the person through the disassembly process and learn from them at the same time, so that this experience can later be transferred to automation.

FUTUR: Ms. Hohlfeld, you mentioned the missing data. Would a solution like the digital product passport help here?

Hohlfeld:

Absolutely. It would be a dream come true to have a perfect digital product passport for a vehicle. For example, if we remove an engine because we’re selling it as a spare part, we could directly assign the removed materials to the product passport. That would be important not only for us, but for the entire cycle. In the future, OEMs will have to account for a recycled content quota and demonstrate how much recycled material is in a new car. How much is actually recovered from end-of-life vehicles is determined operationally by ATFs, dismantling, shredding, and recycling facilities. This data is reported by these entities – not by the OEM.

So we can’t yet map these material flows with that much precision today. Today, recyclers generally do not distinguish between vehicle scrap and other input streams during recycling. The materials are processed together, so that it is no longer possible to clearly determine in the resulting recycled material what proportion comes from end-of-life vehicles. This would be possible with a digitized material or product passport. I see digitalization as a very important and helpful factor for traceability throughout the entire cycle.

But what’s also important to me is this: Even with a perfect product passport and complete geometric data, we still haven’t accounted for the fact that these are end-of-life vehicles. Accident damage, scratches, modifications – none of that is reflected in the original product passport. That’s why I believe we can’t do without visual recognition systems. They must capture the current actual condition, which may differ from the digital target condition. This is also particularly important for choosing the R-strategy. For example, what was the mileage of an engine or a transmission? That can differ from the vehicle’s total mileage. We need such information to decide: Is a component still good enough for reuse or remanufacturing?

Mügge:

To do this, the digital product passport must also be updated throughout the entire life cycle. The mileage of individual components, which parts are actually still installed – this changes compared to the post-production state and significantly influences the R-strategies.

Uhlig:

It is also important that the R-strategy not be based solely on physical properties. It’s pointless to dismantle parts and store them if I can never sell them. There are vehicle models where certain components never break down. Based on a purely physical decision, these would always be identified as spare parts – I could stockpile hundreds of them and never sell a single one. Demand and the market must be taken into account.

A concrete example is e-motors from electric vehicles. These consist essentially of rotating parts, and their service life will be significantly longer than the vehicle’s useful life. If there are no longer any vehicles on the market that use such electric motors, I may not need to keep these parts in stock as spare parts – even if they are still in good physical condition. Every vehicle type has a product life cycle. If electric motors are only compatible with a specific vehicle and that vehicle is no longer available in sufficient quantities on the market, I don’t need spare parts for it. Especially since electric motors, due to their design, experience significantly less wear and tear than internal combustion engines. They break down less often; the demand for used parts simply isn’t there.

FUTUR: What role do regulations on data exchange and standardization play in the project?

Mügge:

For us, regulations play a huge role. For our digitalization solutions, we rely on having data throughout the entire product life cycle in a standardized and interoperable format. With the multitude of new regulations, this data provision is legally required, and we can use this data.

Relevant examples here include: the digital product passport from the Ecodesign for Sustainable Products Regulation, the Circularity Vehicle Passport from the new End-of-Life Regulation, the Environmental Vehicle Passport from Euro 7, or the Product Circularity Data Sheet from ISO. All these regulations share the common goal of presenting product data in a structured manner and making it usable to implement circularity strategies.

Regarding data exchange itself, a number of DIN standards are being developed, for example on data carriers, unique identifiers, data storage, interface descriptions, and exchange protocols. In addition to traditional regulations, we also consider industry initiatives such as Catena-X or the IDTA regarding data modeling and cross-company data exchange.

Uhlig:

For software providers like us, it is crucial that we map the processes in the dismantling facilities. These facilities contribute significantly to meeting quotas, which is already not entirely straightforward based on current regulations. If requirements increase, we may need to collect vehicle-specific data. Currently, we work with average values. We track vehicle inflows and material outflows via transfer slips. For mass flows within a company – that is, changes in inventory for spare parts, materials, or vehicles – we use average values. For example, 800 to 900 kilograms of body scrap from each vehicle goes to the shredder.

In the future, we may need to track things in much greater detail: How much oil do I remove from a vehicle? How much brake fluid? How much glass, how much plastic, and of what type? Employees can no longer handle some of this today. Identifying the different plastics, removing them, weighing them, and assigning them to the vehicle is a massive undertaking. I imagine automated disassembly, supported by vehicle data and process steps that weigh every material removed, as a huge relief.

FUTUR: Let’s conclude by taking a look into the future. Where will vehicle recycling be in 20 years – in a perfect world?

Hohlfeld:

In my ideal scenario, we have a highly digitalized, automated dismantling plant with robot-assisted processes. Visual recognition systems identify the vehicle. We know which materials and components are installed. When I dismantle a component, a digital product passport is precisely assigned to record what I’ve removed and how much material is associated with it. Data exchange between stakeholders works seamlessly. Recyclers know the material composition of a residual car body and can recycle accordingly. This produces high-quality secondary material, and the cycles function effectively. We have taken the first step toward a true circular economy, moving away from downcycling..

Uhlig:

I fully agree with that. For me, this involves a utopian state in which all stakeholders work together cooperatively – 
from manufacturers to vehicle users and operators to recycling facilities. Data exchange becomes significantly easier. A first step in that direction is a platform like Catena-X. There are still a few mental barriers that need to be broken down. But if the common goal is to keep materials within our circular economy and not export them as we have in past decades, then everyone must pull together. Then it must be possible for a dismantling facility to obtain the necessary data for dismantling without having to spend large sums of money to do so.

Kuschan:

I would be very pleased if, in 20 years, we had robots in the dismantling plant and the data we need for automation were available directly within the company, or if we could send a request to the OEMs and receive the necessary information for each individual dismantling step in return. This would ensure data separation, and the OEM wouldn’t have to disclose all relevant company knowledge to outside parties.

I foresee a very high degree of autonomy for robots. We’re currently seeing rapid progress in robot commissioning, which has a positive impact on dismantling. Draining vehicles – that is, removing fluids – can already be automated very quickly. I don’t see any major challenges there, and automation can quickly help separate materials from one another. Proper disassembly will take a little longer. But I’m optimistic that we’ll achieve this through end-to-end digitalization and a digital product passport.