Virtual Manufacturing Enhances Reality

https://commons.wikimedia.org/wiki/File:Robotic_simulation_using_Robcad_software.jpg

APICS Magazine September 2018
Elizabeth Rennie

Simple meaning

When dealing with a subject as complex as digital twins, it’s always best to begin with a definition. Unfortunately, depending on who you ask in business or academic circles, a digital twin can be explained in seemingly infinite ways: an integrated model of a product intended to reflect manufacturing defects; a critical building block of advanced manufacturing; a sensor-enabled digital model that simulates a physical object in a live setting; the digital version of the physical world, based on historical, current and predicted future information — the options are endless.

Urnauer tells the story of a colleague who spoke with four different companies about their definitions, applications and business cases behind digital twins: “Each of the companies had a different understanding. Of course, clarifying the definition would bring a better focus into the discussion.”

While true digital twin clarity is yet unrealized, understanding the technology as a virtual representation that is the exact counterpart of something across its entire life cycle at least provides a starting point. Then, picture real-time data from connected sensors on the physical item that are mapped to the virtual one to enable dynamic recalibration and improved decision-making. Put as simply as possible: Observe a digital twin, and you’ll know exactly how its physical counterpart is functioning out in the real world.

Virtual playground

Gamota adds that establishing a flexible framework upfront is the key to unlocking digital twin benefits and recommends paying close attention to the discovery phase, as it has the potential to be the most valuable step of the process. As such, supply chain management stakeholders must be engaged in digital twin implementations early. “Their global points of view, unique supplier and sourcing perspectives, and carefully honed troubleshooting skills play a pivotal role in defining and refining digital twin architectures,” Gamota says.

Joseph J. Salvo, executive director of digital industrial ecosystems at General Electric (GE) Global Research, who also led GE’s digital twin development team, agrees that it is critical to create value from the beginning. “Most supply chain organizations are conservative,” he says. “Without proof of the value of the technology, they are unlikely to invest in a digital twin transformation.” After showing the initial value, Salvo says scale and collaboration with a broader digital supply chain team are essential.

Urnauer recounts his experience working with simulation models of production plants. He says critical preliminary steps are tool selection and definition of scope. Along the way, he asked himself the following questions:

• What are the core functions that must be modeled by the digital twin?
• Which assumptions and simplifications can, or need to, be made?
• What is the input for both real and simulation data?
• What output will be monitored?

“This way, I could define the scope and make sure that I didn’t get lost in details that are not of importance,” he explains. “When I started to build the model, I asked myself the same questions again.”

Verification and validation, while not done at the beginning of the project, also are vital phases for ensuring successful application of a digital twin and increasing its reliability and acceptance. Validation requires assessing whether the digital twin serves the purpose for which it was intended; verification involves confirming if model components and functions are working properly. Urnauer says both can be tested with experiments in which the outcome of the real system is compared with the outcome of the digital twin when using the same inputs.

Gamota describes this type of experimentation as a “virtual playground,” in which product designers and supply chain management professionals exercise their collective ingenuity and innovation. Together, they can work through a list of hypotheses to be resolved, with each project phase leading to tremendous insights. “The digital twin is a living entity that learns and grows intellectually as each new experiment is conducted and different variables are introduced, understood and manipulated,” Gamota says. “Achieving the right balance of predictability and variability produces the best possible digital twin.”

Piloting the pair

Fundamental to the development and application of digital twins is creating a digital version of supply chain flows. Furthermore, piloting digital twins that effectively address real opportunities for business outcome improvement typically will require data from multiple systems. Salvo says this is no small undertaking and adds that, although most organizations have some form of digitization today, it often is incomplete and difficult to access.

Another obstacle to digital twin pilot success lies in current information technology systems and applications. Resistance to change is difficult to overcome due to a focus on different metrics, such as cost of ownership of the technology, rather than business value and innovation. To address these issues, Salvo advises looking at the highest-value areas for implementation of a digital twin pilot program. Then, he says, as the approach is proven out, it can be scaled to other areas and become part of how the supply chain is being transformed.

Digital twin pilot projects at Gamota’s company begin with a framework that guides users through the discovery, development, industrialization and deployment phases. Next, a significant amount of preparation is performed to establish project goals, functionality requirements and value propositions. Once in place, it’s possible to assess how much and which type of data must be ingested, parsed, structured and analyzed using artificial intelligence, machine learning and different algorithms to build the digital twin. “These early alignments, learnings and outputs will prove invaluable during proof-of-concepts, pilots, risk assessment and beta testing,” he says.

Finally, the Deloitte report states that a major challenge of a pilot program is determining the optimal level of digital twin detail. “While an overly simplistic model may not yield the value a digital twin promises, taking too fast and broad an approach can almost guarantee getting lost in the complexity of millions of sensors, the hundreds of millions of signals the sensors produce, and the massive amount of technology to make sense of the model.”

Optimized Supply Chains

Organizations hoping to capitalize on digital twins have an array of opportunities available to them, including

• product quality improvements from greater durability, extended performance and fewer defects
• more efficient operations because of better engineering change execution, enhanced manufacturing equipment performance and less variability
• service savings through greater efficiency, a clearer understanding of the configuration of equipment in the field and more proactive handling of claims.

Additionally, digital twins can dramatically speed product realization by decreasing or even eliminating the most time-consuming aspects of commercializing a product in the real world. “This leads to increased profit margins because manufacturing efficiency grows quickly when variables that could impede the process are identified and removed,” Gamota says. “Ultimately, there’s an important monetization benefit to be realized if companies can use a digital twin to demonstrate a product’s value proposition before it’s built.”

Salvo believes supply chain optimization is another key advantage because the digital version of the historical, current and predicted future is much less expensive to manage and optimize than the physical. In this way, digital twins can dramatically improve the planning and execution of the global supply chain through improved on-time delivery performance, reduced sourcing costs, heightened manufacturing efficiency, optimized inventory and streamlined new product introduction. “We have developed and implemented digital twin supply chain systems within GE that have helped reduce inventory from 30 to 50 percent while improving on-time delivery and capacity utilization,” he says.

A final benefit to note is the opportunity digital twins provide to link organizational tools, skills and knowledge bases from the assimilation of cross-functional teams. “It’s becoming increasingly apparent that different manufacturing mindsets and methodologies are needed to support project success,” Gamota says.