The Mercedes-Benz plants in Rastatt (Germany), Kecskemét (Hungary) and Beijing (China) will be producing new MMA (Mercedes Modular Architecture) models from the middle of the decade. For the first time, Mercedes is integrating artificial intelligence, the Mercedes-Benz Operationg System MB.OS and the digital twin based on NVIDIA Omniverse into the digital MO360 ecosystem. With this "digital first" approach, Mercedes aims to leverage efficiency potential and accelerate the ramp-up even before the MMA models are launched in the global production network. With MO360, Mercedes-Benz is able to produce electric, hybrid and combustion models on one line and thus flexibly scale the production of electric vehicles in line with market demand.

Virtual commissioning of an assembly hall

The Mercedes-Benz plant in Rastatt is using the digital twin to prepare for the ramp-up of next-generation vehicles with minimal disruption to assembly in the existing halls. With the digital-first approach, Mercedes-Benz is ensuring that the new production line for models on the MMA platform can be modified, configured and optimized in the existing plant using digital simulation techniques. The exact position of machines, supply routes and production lines can be determined and simulated in advance without interrupting production of the current A-Class, B-Class, GLA and EQA compact cars. In addition, production can be ramped up significantly faster compared to a conventional planning process. Mercedes-Benz is investing an amount in the low three-digit million range to prepare the Rastatt plant for next-generation vehicles. The virtual commissioning in Rastatt serves as a blueprint for the global production network and is also to be rolled out accordingly at the cooperation plant in Beijing.

Digital twin in production

© Mercedes-Benz

Mercedes-Benz is developing new production technologies with NVIDIA. The company is contributing its expertise in the areas of software platforms, data and artificial intelligence (AI). For the plant expansion at the Hungarian site in Kecskemét, Mercedes-Benz has created a fully digital map of the entire hall for the first time. The company is relying on NVIDIA Omniverse, a platform for developing Universal Scene Description (OpenUSD) applications for industrial digitalization, such as the digital twin. The digital twin comprises the entire building and its infrastructure, including break areas and changing rooms for employees. With the digital twin for production, assembly areas can be planned, converted and checked virtually. Among other things, the requirements for short distances, sufficiently wide aisles and appropriate fire protection can also be taken into account. Virtual commissioning enables greater speed, transparency and flexibility in production processes without the need for cost-intensive hardware. Further cost savings can be identified through the simulations. By using Omniverse, Mercedes-Benz can interact directly with its suppliers and halve the coordination processes. The use of a digital twin doubles the speed of converting or building an assembly hall and improves process quality at the same time. With the data collected in the virtual world, assembly processes can be optimized more quickly and potential errors and defects can be identified at an early stage - without any impact on the real world.

Efficiency through AI

© Mercedes-Benz

In addition, the introduction of AI opens up new areas of energy and cost savings. Starting with the Rastatt plant, digital production is being introduced in the paint shop area. AI helped to increase efficiency in the top coat booths, where the vehicle chassis receive their paint and protective coats. Instead of a conventional programmable logic controller, AI took over the monitoring of the relevant sub-processes in the pilot test, resulting in energy savings of 20% compared to the conventional PLC control system. In addition, AI enormously reduced the time required to initiate the sub-processes. The successful application of AI-controlled process technology is an important step in the digitalization of production and is now being rolled out to other Mercedes-Benz plants.

Last May, Mercedes-Benz launched a ChatGPT project within the MO360 digital ecosystem to optimize production processes and accelerate the identification of defects. With the "democratization of data", important production data can be accessed by many employees. This allows them to evaluate processes and data in real time.

MB.OS centralizes software operation

The Mercedes-Benz Operating System (MB.OS) - developed on a chip-to-cloud architecture - will be used in series production for the first time. The necessary software will no longer be controlled by individual measures, but will be used in cyclically timed software packages in production. MB.OS in the vehicle and the technologies derived from it in production enable software updates ten times faster than before while simultaneously increasing the amount of data. Thanks to the direct connection of production to the Mercedes Intelligent Cloud (MIC), all vehicles leave production "always up-to-date". Mercedes-Benz is thus supplementing its over-the-air software strategy for vehicles that have already been delivered. Software optimizations can be rolled out globally within a few hours, which significantly increases the safety and quality of the vehicles and ensures compliance with increased certification requirements. The MIC is directly connected to the MO360 Data Platform within the development, production and fleet operations network. It is responsible for software update packages and installation processes. The two are connected via an intelligent analysis interface so that every employee can read and analyze the current software status of any vehicle in the network via the MO360 Data Platform. This is supported by an app for intelligent fault diagnosis. The app detects potential errors at an early stage and uses an intelligent algorithm to provide instructions for quality assurance during the production process.