Audi’s technology factory of the future unveiled

Automaker Audi has revealed how it is building a fully networked, highly efficient and sustainable production system.

The goal is to design a resilient, agile and flexible value chain to successfully meet future challenges. These include an increasing diversity of variants, the transition to electromobility, the increasingly volatile supply situation and political uncertainties.

“We use synergies and look at production as a whole – from the worker to the fully automated cycle,” said Gerd Walker, production and logistics board member. “We digitize specifically as part of a culture of open innovation. We guarantee efficient value creation and make it possible to use resources and capacities flexibly and efficiently.

Modular assembly

Audi says the traditional automotive assembly line is increasingly reaching its limits. Many derivatives and individualization options make the products more and more varied. In a rigid and sequential process, this complexity is increasingly difficult to master.

“Modular assembly is one of our answers to future challenges,” said Gerd Walker. “We are using digital technologies specifically to the benefit of our employees while achieving a more flexible and efficient assembly process.”

In the Audi Production Lab, the team of project manager Wolfgang Kern prepares the modular assembly for mass production.

Audi is initially implementing the concept in the pre-assembly of the interior door panels at the Ingolstadt plant. In the test operation, the job no longer follows a uniform sequence. Instead, it caters to particular needs.

Automated Guided Vehicles (AGV) bring the door panels directly to the station where the components are assembled.

Audi Automated Guided Vehicles (AGV)

“By reducing production time through a focus on value creation and self-guiding, we can increase productivity by up to 20% in some cases,” Kern said. “In addition, we can tie modular assembly to specific production stages. For example, now only one worker can install complete blinds. Previously, this required two or three workers due to predefined processing times in an assembly line.

Virtual Assembly Planning in Smart Production

Virtual assembly planning not only saves hardware resources, but also enables innovative and flexible collaboration between different locations. It eliminates the need to build prototypes in the planning process. A scanning process generates three-dimensional point clouds that can be used to perform virtual reverse engineering of machines and infrastructure. The software is based on artificial intelligence and machine learning. It allows Audi employees to virtually navigate assembly lines. Volkswagen’s Industrial Cloud offers them an efficient tool that allows them, for example, to compare locations and use suitable solutions from other production lines in their planning.

Right now, Audi is working with NavVis to test Spot the robot dog so they can perform the 3D scans as efficiently as possible. About four million square meters and 13 factories have been involved since the digitization of the site began in 2017. The digitization of 100,000 square meters – for example, in the production of the Audi A6 in Neckarsulm – takes around three weeks in one team. Analyzes can only be carried out at night or on weekends. On top of that, structural obstacles like steps and doorways make the scanning job more difficult.

On the other hand, Spot the robot dog can complete this scan in 48 hours and determine its route autonomously. Audi has been testing Spot extensively since December 2021.

Spot the dog
Spot the dog

“The test results are extremely promising and can be updated regularly,” said project manager André Bongartz. “Input data is constantly coming in, and we can use it to plan new car models.” Any range of 3D scans can be integrated into the virtual images, for which Andrés Kohler’s team is responsible. “Merging all of the planning data into our digital twin gave us a holistic view of our future production plans years in advance,” says Kohler. As in a real factory, it includes the workshop, conveyor technology, tools, shelves and containers alongside the new Audi model.

Assembly sequences and logistical aspects are largely designed and optimized by interdisciplinary teams in so-called 3D process workshops. With the digital twin and an in-house VR solution, Audi is harnessing the benefits of digitization and visualization. These include daily updated component data and a view of the different car variants. “Above all, we look at the production based on what it will look like later as a whole,” explained Andrés Kohler. He emphasizes that collaboration remains a central element: “I am always fascinated again as soon as we put on the VR glasses and meet our colleagues as avatars in the virtual world. First, we build our new Audi there or look at a computer-generated avatar and how it applies as a real-time simulation. And if necessary, while we are at it together, we discuss and optimize the sequences and the work environment, such as how to set up the necessary equipment or tools.

Artificial intelligence in production

Artificial intelligence and machine learning are key technologies in Audi’s digital transformation and modern production. An AI algorithm in the press shop in Ingolstadt helps identify component faults. This procedure is supported by software based on an artificial neural network. The software itself identifies the smallest defects and marks them reliably. The solution is based on deep learning, a special type of machine learning that can work with unstructured and high-dimensional volumes of data. The team used several million test models to train the artificial neural network for months. This database includes several terabytes of these images from the presses of Audi sites and several Volkswagen sites.

In another pilot project, Audi is using artificial intelligence to check the quality of spot welds in high-volume production at its Neckarsulm site. It takes around 5,300 welding points to connect the body components of an Audi A6 together. Until now, production personnel used randomized ultrasonic scans to monitor the quality of resistance spot welding (abbreviated WPS in German). As part of the WPS Analytics pilot project, experts are using artificial intelligence (AI) to automatically detect quality anomalies in real time. Currently, the algorithm, dashboard and in-depth quality analysis application are all used to build the body of the A6 and A7. It is a model for other applications in network production.

Edge Cloud 4 production concept

With the Edge Cloud 4 Production local server solution, Audi is initiating a paradigm shift in factory automation. After successful tests in the Audi Production Lab (P-Lab), three local servers will take over worker support in the Böllinger Höfe. In production in Neckarsulm, the Audi e-tron GT quattro1 and the R8 share an assembly line. The small series produced there are particularly well suited for testing P-Lab projects and experimenting with large series. Audi wants to be the first manufacturer in the world to turn to this type of centralized server solution in cycle-dependent production. If the server infrastructure continues to operate reliably, Audi wants to deploy this automation technology – the only one of its kind in the world – for series production across the entire group.

With Edge Cloud 4 Production, a few centralized and local servers will support the work of countless expensive industrial PCs. The server solution allows virtualized clients to be leveled on the total number of them – a much more efficient use of resources. Production will be saved, including software deployments, operating system changes, and IT-related expenses. “What we are doing here is a revolution,” says Gerd Walker, Member of the Board of Management of AUDI AG Production and Logistics. “Before, we had to buy hardware when we wanted to introduce new features. With Edge Cloud 4 Production, we only buy applications in the form of software. This is the crucial step towards computer-based production.


Comments are closed.