This presentation was made at the NAFEMS Americas Seminar "Model-Based Engineering: What is it & How Will It Impact Engineering Simulation" held on the 1st of October 2019 in Columbus Ohio

Resource Abstract

Digital Twins are a hot market trend right now – and for good reason. They offer companies the tantalizing promise of further-optimized product performance and extended life. This fits well with (and is enabled by) both computer-aided engineering (CAE) and the rapid emergence of data intelligence.



The value of model-based engineering and virtual prototyping has evolved and improved over decades. Now, due to the availability of better information about how products are actually operating, physics-based digital twins can be more tightly coupled with data-driven digital twins thanks to real-time data collection, communication/transmission, analysis, and to machine learning.



In this presentation, we will share several real-world cases to illustrate that the promise afforded by physics-based digital twins and data-based digital twins is real and available today.



The examples presented will showcase different industries and pertain to companies large and small. They will show how a complete and open digital twin platform gives every company the flexibility needed to accommodate its own unique workflows and tools.



Our examples will show interesting and valuable combinations of the use of two or more of the following building blocks for this platform: 3D CAE; OD and 1D system-level modeling & simulation leveraging open standards such as Modelica; reduced-order modeling (ROM); the ability to include 3rd-party tools through the Functional Mock-up Interface (FMI); data intelligence; machine learning; and Internet of Things (IoT) technology.



These building blocks can be categorized into those which help companies to Develop better future-generation products, to better Operate current-generation products, and to tightly Connect the virtual and real worlds for optimized results and continuous improvement. Model-Based Systems Engineering serves as the heart and soul of this important, enterprise-wide product lifecycle activity.



Examples will include:



• Physics-driven digital twin combined with data-driven digital twin for a wind turbine, as used to predict remaining useful life and time to failure.



• Reduced Order Modeling (ROM) coupled with physical testing, as used to optimize the “flight” (path) controller of undersea drones.



• Automatic code generation for embedded systems and real-time data communication via Internet of Things (IoT), as used to automatically update 19,000 rooftop cooling units in need of new controller software.



• Condition monitoring and machine learning, as used to enable predictive maintenance on belt-driven systems while reducing the number of sensors required.



• Pattern recognition, as used to help significantly reduce robot breakdowns.



Not surprisingly, companies have leveraged one or more Digital Twin building blocks in different combinations at different times as they matured their process. This has allowed each to start wherever they would like, to use whatever tools they believe to be best-in-class for their purposes, and to methodically progress their work step-by-step in ways that best support their corporate vision and goals for digital twins.



These examples hit home the reality that there is no one-size-fits-all approach. As each company defines its own standardized processes and best practices for its digital twins, a complete and open digital twin platform will help them get there faster without hitting dead-ends or wasting precious time and energy.