The ‘digital twin’ may sound like a term from a sci-fi movie, but it refers to a technology that brings together real-time data, simulation and CAD modeling to allow manufacturers to gain real value by merging data from real-world assets, products or processes and turning it into a matching digital model used for analysis, simulation, virtual commissioning, product lifecycle planning and many more applications.
Since it’s invention in 1957, computer-aided design (CAD) has transformed manufacturing by providing many of the benefits of prototyping and technical drawing for communication, design development and optimization with better accuracy and lower cost than traditional, manual drafting and modeling techniques. In much the same way, early computer simulations provided a way to study the behavior and performance of physical objects in cases where real-world experimentation was too expensive and mathematical analysis was too complicated.
Digital twin brings these two technologies together to provide benefits in time savings, design optimization and product lifecycle management. However, a digital twin is more than a model. Data collection through Internet-of-Things (IoT) connected sensors is what brings the digital twin to life, enabling insights based on real data.
Digital Twin Applications in Product Design and Production
Have you ever wished there was an “undo” button in real life? That’s the advantage of a digital twin in product design. A simulated, digital model of the whole product and its manufacturing process enables cross-functional teams to bring contradictory requirements together, such as high strength and low weight, by using the digital twin as a single source of truth. In addition, a digital twin enables virtual prototyping and testing, saving time and costs while improving time-to-market. Data from previous versions of a product can be brought into the digital twin to solve existing problems and make improvements. In this design phase, the digital twin makes product development faster while keeping costs low.
From there, the digital twin of the product can be used to enhance production. Digital traceability of parts through the production process can improve quality and reduce defects. In addition, production data can be analyzed to improve design for manufacturing, reducing scrap, improving throughput and making production more ergonomic for workers.
For some products, the digital twin can remain connected to products during their entire life cycle. This provides insight to engineers and designers on wear, performance and usability. It also enables software updates to be developed and pushed to products in use, delivering a better user experience. Real-world data can also inform the next generation of products, including information about use cases, new features and more.
How is Digital Twin Implemented?
Creating a digital twin, whether it’s for a product, a piece of equipment or an entire facility, requires specialized software. Today, IoT technology vendors such as Siemens, PTC and Ansys offer software packages for plug-and-play twin building. These multi-technology packages include the modeling, simulation and IoT platform features needed to connect a model to the real data. However, digital twins can also be created using existing CAD, simulation, PLM, ERP and IoT platform software, connected via custom development or API, for example. Many manufacturers will hire system integrators to implement digital twins.
At a high level, the most important strategic consideration when implementing a digital twin is the data source. For example, a factory seeking to implement a digital twin of a CNC milling machine to enable predictive maintenance should consider which machine to equip with sensors, and when to collect data. Manufacturers should try to collect the right granularity and frequency of data for the application. More data is more expensive, but can produce more accurate results. For new implementations, it’s recommended to try a simpler use case that can provide ROI before delving into more complex solutions.
Computing Power: On-premise or In the Cloud?
While cloud computing from vendors such as AWS or Microsoft Azure is quickly becoming a ubiquitous service in many industries, some manufacturers may choose to use on-premise computers and server clusters instead of the cloud for a variety of reasons, including:
- Data privacy and security: Manufacturers may have concerns about the security and privacy of their data in the cloud. With on-premise computers, they can have more control over their data and ensure it is protected.
- Connectivity and reliability: In some manufacturing facilities, internet connectivity may be limited or unreliable. In these cases, on-premise computers provide a more stable and dependable computing environment.
- Regulatory compliance: Some industries have strict regulatory requirements around data storage and management. On-premise computers provide a better solution for meeting these requirements.
- Customization and control: Manufacturers may want more customization and control over their computing environment than is possible with cloud solutions. On-premise computers allow them to tailor their computing infrastructure to their specific needs.
- Cost: Depending on the size and complexity of their computing needs, it may be more cost-effective for manufacturers to use on-premise computers rather than pay for cloud services.
High-Performance Computing for Industrial Digital Twin
The advanced big data analysis and simulation required to power digital twins requires high-performance computers. The main advantage of on-premise computers for manufacturing is complete control over the hardware and configuration of the systems, enabling your IT department to build the optimal computing solutions for your application. In addition, IoT data sent to an on-premise server for processing does not require a reliable, high-bandwidth internet connection, reducing some costs.
The main advantage of cloud computing is scalability, reliability, ease of use and cost. For many businesses, plugging into a cloud service vendor is easier, faster and less expensive than trying to build an on-premise server solution, considering the costs of IT talent, hardware and software maintenance and security. Cloud services are highly scalable, allowing customers to buy more processing power or storage as needed. Cloud services provide a level of reliability, with secure, massive facilities and multiple redundancies. It’s difficult for on-premise solutions in many industries to compete, especially on cost, with the capabilities of cloud computing.
However, in specialized high-performance applications in manufacturing, building and maintaining a custom solution may be more cost-effective than trusting a cloud service provider.
On-premise solutions can even be air-gapped from the internet, providing ultimate protection from cybersecurity threats. These solutions can be more cost-effective in the long run, depending on the size and complexity of the computing needs or when regulatory compliance or geographical location makes internet connectivity less reliable. This ensures your digital twin technology operates smoothly and that data remains synchronized with the physical asset.
Lastly, control over hardware and software configuration allows IT teams to construct the optimal solution for computing needs, such as increased graphics processing power.
In summary, using on-premise high performance computers and servers for digital twin technology can provide manufacturers with greater control over their computing infrastructure, improved security and data privacy, potential cost savings and more reliable connectivity.
This story is one in a series underwritten by AMD and produced independently by the editors of engineering.com. Subscribe here to receive informative infographics, handy fact sheets, technology recommendations and more in AMD’s data center insights newsletter.