What Are Digital Twins? A Primer on Virtual Models

What Are the Benefits of Digital Twins to the Semiconductor Industry?

According to a Cap Gemini research study of 1,000 companies in segments including life sciences, consumer, energy/utilities, and discrete manufacturing, 80% were employing digital twins. The reasons were varied: new product development, operational efficiency improvements, enhancements to worker safety, and better sustainability.

The Digital Twin Consortium, which promotes awareness, adoption, interoperability, and development of digital twin technology, notes that these virtual representations are beneficial because they accelerate a holistic understanding of the product, while enhancing decision-making and effective action. Real-time and historical data allow digital twins to represent the past and present and simulate the future.

Historically, digital twins have largely been used for mechanical parts and components. But as electronic systems continue to feature more complex compute platforms and more software, digital twins have become more common over the last several years. Used at the hardware and software levels, digital twins provide valuable insights that allow designers to examine performance issues, test-drive new features, and optimize different aspects of their product throughout the design and manufacturing processes. They can be applied to virtually any component, from the chip level through sub-systems and all the way up to the system level (like, say, a vehicle).

Silicon chip manufacturers are taking a cue from other manufacturers and using digital twins to accelerate production timelines. For example, Bosch’s 300mm wafer fab in Dresden, its first AIoT factory leveraging AI and IoT technologies, uses digital twins to simulate process optimization plans and renovation work without disrupting operations. Intel has transformed its microprocessor fabs with digital twins and has made its technology available to other manufacturers.

Electronic design automation (EDA) vendors have been offering solutions for design, verification and implementation of SoCs and multi-die systems by semiconductor companies. Synopsys is now expanding its solutions to address the needs of system companies by creating electronics digital twin solutions, providing a digital replica of an electronics system, hardware, software, and environment, used throughout the product lifecycle for software bring-up, power analysis, and software/hardware validation. With these digital twins, semiconductor and system companies can collaborate more closely, together ensuring that designs will work as intended in context of the full system.

What Are Some Examples of Digital Twins?

The integration of AI with digital twins provides a level of intelligence that generates valuable, actionable insights from data produced by modeling or collected by sensors. Guided by machine learning, companies can take advantage of predictive insights, for example, or get alerts of anything abnormal that is captured by the sensors. Using this intelligence, they can perform “what if?” analysis to facilitate better decision making or take corrective action earlier on.

Here are some more examples of where digital twins, often in combination with AI, are making an impact:

• Automotive industry: With increasing electrification and reliance on software and electrical/electronic (E/E) architectures, digital twins can provide a productivity and quality boost. They can help validate over-the-air software updates or vehicle subsystems, for instance. Replacing physical electronic control units (ECUs) and physical benches for system validation with digital twins and virtualization-based testing with virtual ECUs (vECUs) provides an efficient path toward functionally correct, robust, safe, and secure automotive software. One automaker that creates a digital twin for each of its cars is Tesla, which applies AI to analyze data collected by the cars to understand whether the vehicle needs maintenance or is working as intended.
• Aerospace, defense, and government sector: Digital twins can provide a virtual representation of a semiconductor subsystem within a larger system or of a larger system, such as an aircraft, saving time and budget for testing, validating, and verifying hardware and software under real-world conditions. An example of this, presented during SNUG Silicon Valley 2023, is Niobium using analog mixed-signal (AMS) emulation for an aircraft digital beamformer sensor design. They can also be used during program reviews with government entities to demonstrate how systems would work.
• Healthcare: Experts anticipate that someday, humans will have digital twins that can be used to develop tailored treatments for each person’s ailments and to predict the effects of different drugs. Already, the Neurotwin project, funded by the European Union, aims to use digital twins of the human brain to study the use of non-invasive brain stimulation for conditions such as Alzheimer’s and epilepsy.
• Transportation and infrastructure: Digital twins can influence how infrastructure and transportation systems are designed. As part of its digital roads strategy in England, National Highways has outlined a plan to create digital twins of all major roads in the country to enhance how the roads are designed, built, operated, and used. Amsterdam is home to the first pedestrian bridge constructed entirely using 3D printing; the 12m steel bridge has its own digital twin for analysis of performance and safety during daily use. In Los Angeles, the Department of Transportation, working with the Open Mobility Foundation, is creating a digital twin of the city’s transport infrastructure to better understand everything from shared-use bicycles and scooters to ride-sharing services and, eventually, autonomous taxi drones.
• Urban planning: Digital twins are being created of cities such as Shanghai, New York, Singapore, and Helsinki to study and manage everything from traffic patterns to energy consumption. In Shanghai, its digital twin was used in addressing the COVID-19 pandemic and can also simulate the impact of natural disasters to aid in response planning. Las Vegas is modeling future energy use, emissions, mobility, and emergency management.
• Buildings: Digital twins of buildings provide facilities managers a way to handle predictive maintenance for ventilation systems, manage the use of their spaces, and provide digital navigation systems to guide visitors. Builders are designing greener buildings using virtual models to better understand and forecast energy consumption, resource usage, and other environmental impacts. 

The Future of Digital Twins

It’s easy to see why digital twins are so attractive in a myriad of industries and why we can expect nearly every system in the future to have a digital twin. They provide an efficient, cost-effective way to validate product functionality, test-drive new features, and project the impact of a variety of enhancements on factors such as budget, productivity, and energy use.

Considering the cost, time, and effort involved in developing any kind of physical product, using digital twins to streamline the process simply makes good business sense. The addition of AI in the digital twin toolbox only takes the technology to the next level. As businesses strive to create the next unique product, there’s a good chance that digital twins will facilitate the journey.