Twin reality for greater autonomy: the next frontier in digital manufacturing

The manufacturing industry requires significant change, driven by the rapid evolution of AI and its adoption in multiple processes — from the design and engineering of greenfield plants to operational efficiencies like waste reduction and quality improvements. The imperative to act is now, with an impetus, for example, to rebuild the US manufacturing base and start an entire ecosystem of chip manufacturing for AI factories.

Wouldn’t it be something if the handling issues and challenges associated with these ongoing capital projects could be a thing of the past, with the advent of simulated digital twins?

Optimizing capital spend in manufacturing, supply chain and operations

Traditional approaches to managing the often-spiraling costs of capital projects have included leveraging various discrete event simulation (DES) tools, which at best and many times are simply a base-level design tested with actual physical build out, leading to delays and costly rework. While DES tools can help with optimization and assist many trade-off decisions around automation and flow, the nature of issues faced by manufacturers today requires dynamic simulations, bi-directional operational connectivity and scenario planning.

Many of the challenges in these programs apply to both new greenfield builds and brownfield modernization projects. These projects, while substantial, often face challenges such as delays and budget overruns due to limited collaboration among internal functions (from design and engineering to production and testing), right through external suppliers of equipment/ parts.

Especially in an existing facility, a project for a process automation — like a robotic cell addition for example – may cause adverse disruptions to the existing production schedule. Further delays in execution may even render the program obsolete.

These workflows infused with AI simulations can enable greater accuracy and collaboration. Bi-directional connectivity on a single platform has the potential to substantially reduce capital spend — resulting in on-spec delivery of equipment and achieved targeted production/ throughput rates.

Evolution of digital twins

At the core of this revolution is the digital twin, a virtual replica of a physical system or process, that provides a dynamic and real-time view of an asset’s health, performance and other key metrics. Within manufacturing, they can be created for a singular machine, complete production lines or entire facilities. These digital models are continually refreshed with live data from sensors, so manufacturers can oversee and scrutinize their operations in real-time, pinpointing areas for enhancement.

The allure of digital twins lies in their potential to yield significant capex (capital expenditure) as well as opex (operational expenditure) savings. For example, layout scenario planning and pinpointing operational inefficiencies, choke points and areas of excess. Digital twins empower manufacturers to streamline their operations and avoid unnecessary costs.

Yet, there can be challenges associated with the realization of this vision. The startup cost of implementing and then maintaining digital systems can be a barrier, particularly for smaller manufacturers with more nimble supply chains. A shortage of skilled talent with the expertise to operate and manage advanced technologies presents another challenge. Digital technologies often require big changes to current processes and workflows, which can be difficult for employees to keep up with, resulting in a slower rate of technology adoption for the enterprise overall. Siloed digital twins, for example, lead to fragmented data, causing inconsistent decision-making across the factory lifecycle and when digital twins are disconnected, teams struggle to communicate effectively, hindering problem-solving and innovation.  Finally, concerns about cybersecurity and data privacy must be addressed to ensure the responsible use of data, AI and generative AI, within the community.

Creating the next frontiers in industrial digitization and physical AI with NVIDIA

Accenture and NVIDIA are collaborating to address many of these manufacturing industry challenges.. At the heart of this is the NVIDIA Omniverse platform, which Accenture is leveraging to develop digital twin and physical AI solutions that let manufacturers create precise simulations of their physical production lines and warehousing operations in real-time. These solutions enable these enterprises to capture behavioral models and precisely accurate details of operating equipment, process conditions and human interactions/ flow. Think of them as a digital brain in physical space.

What distinguishes the solutions built by Accenture with NVIDIA libraries is their tailored industry approach. By combining the latest AI and simulation technologies with technical know-how and manufacturing domain expertise, enterprises can streamline and enhance operational efficiencies. The digital twin and physical AI solutions take Accenture's deep roots and understanding of industry practices, augmented by NVIDIA’s advanced technologies and libraries, to deliver comprehensive analytics and performance insights. Moreover, Accenture acts as a conduit to facilitate adoption, enabling manufacturers to integrate automation with minimal disruption. This harmonious blending of technology and deep sector expertise delivers successful and streamlined processes.

Neuron, for example, is a groundbreaking solution developed by Accenture, built on NVIDIA Omniverse libraries, designed to revolutionize the way manufacturing plants, and their components are designed, engineered, assembled and commissioned. A cloud-based solution that combines NVIDIA Omniverse technologies with generative AI — including Accenture’s AI Refinery platform, also built on NVIDIA libraries — to build Digital Twin Factory Simulations. These factory simulations can help improve the speed of collaboration, construction and testing of production equipment, deliver capex cost optimization and program risk mitigation, targeting both greenfield and brownfield programs.

Additionally, opex-focused initiatives provide flexibility for ongoing enhancements. Unlike capex projects, which often entail prolonged commitments, opex improvements can be swiftly materialized and fine-tuned based on real—time insights and feedback. This agility empowers manufacturers to quickly adapt to market shifts, embrace emerging technologies and adopt procedural enhancements without the constraints of long-term investments.

Let us illustrate this with some practical examples. Picture a manufacturing site that's brought to life with real-time monitoring of its processes through the use of sensors and Internet of Things (IoT) devices. The data streaming in from these sensors is like a digital mirror, a twin, reflecting the facility's every move. Engineers can then, from a distance, spot bottlenecks and fine-tune the production flow.

Case Study: Reinventing Operations with KION

KION GROUP AG, working with Accenture and NVIDIA, is reinventing supply chain and warehouse operations using artificial intelligence (AI) and digital twins. This project builds on the Mega NVIDIA Blueprint, to develop digital twins of industrial environments, enabling the simulation and optimization of operational scenarios such as layout planning, robot interactions and workforce management.

These digital twins act as a virtual testing environment where operations of autonomous warehouse robots can be simulated, validated and adjusted according to changes in demand and inventory, enhancing operational flexibility. By testing various configurations in physically accurate simulations, KION identifies the most efficient strategies before implementing them in real-world facilities, improving performance metrics like throughput and task completion times.

The use of AI and digital twins is leading to more autonomous and efficient warehouse operations, reducing the need for manual intervention and allowing for quicker adaptations to operational changes. This approach has streamlined warehouse management, minimized disruptions, and increased overall efficiency, demonstrating a practical application of operational excellence principles in a manufacturing environment.

Case Study: Accelerating operational excellence with autonomous factories for Schaeffler

Accenture is working with Schaeffler AG to reinvent industrial automation with physical AI and robotics. The companies developed a proof-of-concept (PoC) demonstrating the benefits AI-powered simulations can bring to Schaeffler’s factories and distribution centers on three levels:

• Planning ideal facilities using digital twins to optimize layouts and automation levels, reducing commissioning times.
  
  
• Paving the way for physical AI and general-purpose humanoid robots: Accenture and Schaeffler are starting to adopt the Mega NVIDIA Blueprint to test robot fleets, including general-purpose humanoid robots, in industrial digital twins of factories and warehouses.
  
  
• Optimizing live robotics operations: Accenture and Avanade, its joint venture with Microsoft, are collaborating with Schaeffler to use simulated data to optimize the physical performance of robots in the warehouse and on the shopfloor — for example, to avoid congestion. Data from various simulated scenarios is fed into Microsoft Fabric, a unified, AI-powered data platform. Site managers and workers responsible for operations can then compare key performance indicators, such as availability, utilization and overall equipment effectiveness for these scenarios to identify potential issues in time to act.

The potential for digital twins to transform the manufacturing landscape is vast, and companies are seeing results now. Neuron can help reduce capex and operational expenses through efficient design and simulation and ensure project success with detailed risk assessments and collaborative problem-solving.

Don’t be outpaced. Accelerate innovation by leveraging AI and advanced visualization to push the boundaries of what’s possible in your manufacturing processes, today.