BMW has standardized on a new technology unveiled by Nvidia, the Omniverse, to simulate every aspect of its manufacturing operations, in an effort to push the envelope on smart manufacturing.

BMW has done this down to work order instructions for factory workers from 31 factories in its production network, reducing production planning time by 30%, the company said.

During Nvidia’s GTC November 2021 Conference, members of BMW’s Digital Solutions for Production Planning and Data Management for Virtual Factories provided an update on how far BMW and Nvidia have progressed in simulating manufacturing operations relying on digital twins. Their presentation, BMW and Omniverse in Production, provides a detailed tour of how the Regensburg factory has a fully functioning, real-time digital twin capable of simulating at scale production and finite scheduling based on constraints down to work order instructions and robotics programming on the shop floor.

Improving product quality, reducing manufacturing costs and unplanned downtime while increasing output, and ensuring worker safety are goals all manufacturers strive for, yet seldom reach consistently. Achieving these goals has much more to do with how fluid and real-time the data from production and process monitoring, product definition, and shop floor scheduling is shared across manufacturing in a comprehensible format each team can use.

Overcoming the challenges of achieving these goals motivates manufacturers to adopt analytics, AI, and digital twin technologies. At the heart of these challenges is the need to accurately decipher the massive amount of data manufacturing operations generate daily. Getting the most value out of data that any given manufacturing operation generates daily is the essence of smart manufacturing.

Defining what a factory of the future is

McKinsey and the World Economic Forum (WEF) are studying what sets exceptional factories apart from all the others. Their initial collaborative research and many subsequent research studies, including the creation of the Shaping the Future of Advanced Manufacturing and Production Platform, reflect how productive the collaborative efforts of McKinsey and the WEF are today. In addition, McKinsey and WEF have set high standards in their definition of what a factory of the future is, as they’re providing ongoing analysis of the select group of manufacturers’ operations for clients.

According to McKinsey and WEF, lighthouse manufacturers scale pilots into integrated production at scale. They’re also known for their scalable technology platforms, strong performance on change management, and adaptability to changing supply chain, market, and customer constraints, while maintaining visibility and cost control across the manufacturing process. BMW Automotive is an inaugural member of the lighthouse manufacturing companies McKinsey and WEF first identified after evaluating over 1,000 companies. The following graphic from McKinsey and WEF’s research provides a geographical view of lighthouse manufacturers’ factory locations globally.

Above: McKinsey and WEF’s ongoing collaboration provides new insights into how manufacturers can continue to adopt new technologies to improve operations, add greater visibility and control across shop floors, and keep costs in check. Source: McKinsey and Company, ‘Lighthouse’ manufacturers lead the way—can the rest of the world keep up?

BMW’s factories of the future blueprint

The four sessions BMW contributed to during Nvidia’s GTC November 2021 Conference together provide a blueprint of how BMW transforms its production centers into factories of the future. Core to their blueprint is getting back-end integration services right, including real-time integration with ProjectWise, BMW internal systems Prisma and MAPP, and Tecnomatix eMS. BMW relies on Omniverse Connectors that support live sync with each application on the front end of their tech stacks. Front-end applications include many leading 2D and 3D computer-aided design (CAD), real-time visualization, product lifecycle management (PLM), and advanced imaging tools. BMW standardized on Nvidia Omniverse as the centralized platform to integrate the various back-end and front-end systems at scale so their tech stack could scale and support analytics, AI, and digital twin simulations across 31 manufacturing plants.

Excel at customizing models in real-time

How BMW deployed Nvidia Omniverse explains why they’re succeeding with their factory of the future initiatives while others fail. BMW recognized early that each system’s different clock speeds or cadences integral to production, from CAD and PLM to ERP, MES, Quality Management, and CRM, needed to be synchronized around a single source of data everyone could understand. Nvidia Omniverse acts as the data orchestrator and provides information every department can interpret and act on. “Global teams can collaborate using different software packages to design and plan the factory in real-time, using the capability to operate in a perfect simulation, which revolutionizes BMWs planning processes,” says Milan Nedeljkovi?, member of the Board of Management of BMW AG.

Product customizations dominate BMW’s product sales and production. They’re currently producing 2.5 million vehicles per year, and 99% of them are custom. BMW says that each production line can be quickly configured to produce any one of ten different cars, each with up to 100 options or more across ten models, giving customers up to 2,100 ways to configure a BMW. In addition, Nvidia Omniverse gives BMW the flexibility to reconfigure its factories quickly to accommodate new big model launches.

Simulating line improvements to save time

BMW succeeds with its product customization strategy because each system essential to production is synchronized on the Nvidia Omniverse platform. As a result, every step in customizing a given model reflects customer requirements and also be shared in real-time with each production team. In addition, BMW says real-time production monitoring data is used for benchmarking digital twin performance. With the digital twins of an entire factory, BMW engineers can quickly identify where and how each specific models’ production sequence can be improved. An example is how BMW uses digital humans and simulation to test new workflows for worker ergonomics and efficiency, training digital humans with data from real associates. They’re also doing the same with the robotics they have in place across plant floors today. Combining real-time production and process monitoring data with simulated results helps BMW’s engineers quickly identify areas for improvement, so quality, cost, and production efficiency goals keep getting achieved.

Above: BMW simulates robotics improvements using Nvidia’s Omniverse first before introducing them into production runs to ensure greater accuracy, product quality, and cost goals are going to be met.

For any manufacturer to succeed with a complex product customization strategy like BMW has, all the systems that manufacturing relies on must be in sync with each other in real-time. There needs to be a common cadence the systems are operating at, providing real-time data and information each team can use to do their specific jobs. BMW is achieving this today, enabling them to plan down to the model-by-model configuration level at scale. They’re also able to test each model configuration in a fully functioning digital twin environment in Nvidia’s Omniverse, and then reconfigure production lines to produce the new models. Real-time production and process monitoring data from existing production lines and digital twins help BMW’s engineering, and production planning teams know where, how, and why to modify digital twins to completely test any new improvement before making it live in production.