Re-writing the future: The simulation-based digital twin

In our last blog, we looked at how dramatically the digital twin landscape has changed over these last few years.

We also previewed the newest and most exciting class of digital twin yet—the simulation-based twin.

ANSYS is taking the lead here, with the recent news that they are extending Microsoft Azure Digital Twins with physics-based simulations to get a new level of digital validation for new products.

The ANSYS Twin Builder enables companies to reuse existing components and quickly build, validate and deploy complete systems simulations and digital twins for real-time asset performance monitoring and scenarios like predictive maintenance.

Manufacturers can then export and plug this twin into factory floor platforms such as PTC ThingWorx and SAP, where these digital twins can leverage AI to learn from the data streaming from the factory floor, and generate insights for optimizing design, performance, operations, and much more. What is most exciting is that this data can be used to either validate or optimize the design, development, delivery and serviceability of future iterations of digital twins using an intelligent feedback loop. ANSYS has developed a platform by the name Minerva, which can be leveraged by designers and engineers at the OEM to collaborate, build and publish the digital twin models to the customer environment. Data and insights are collected from the asset operations as a feedback loop for the engineers at the OEM.

ANSYS is also working with industrial automation leaders such as Rockwell Automation to help customers realize better business outcomes through simulation-based digital twins.

According to Sameer Kher, Director of Product Development at ANSYS, “ANSYS Twin Builder is the first software solution that offers a packaged approach for simulation-based digital twins — enabling engineers to quickly build, validate and deploy these digital representations of physical products. Once deployed, these digital twins can help reduce maintenance costs by up to 20 percent over the product’s lifetime. With this collaboration with Microsoft, ANSYS Twin Builder runtimes will become easily accessible to large enterprise customers who are looking to build their own IIoT platforms using Microsoft Azure Digital Twins.”

Back to the satellite example from our first blog. This simulation-based twin model is highly interesting for remote asset deployment scenarios like in the aerospace industry, most notably. You cannot launch a satellite and then figure out what went wrong after the fact. Traditional analytical models are limited in their scope of analysis and their ability to predict the product performance once deployed in the physical world.

Adding simulation capabilities to digital twin innovation allows you to understand the consequences of virtual and physical interactions, enabling you to do much of your analysis upfront using simulation and modeling, and then feed the learnings back into the system. Many manufacturers are already on a journey toward digital verification and validation. This fundamentally requires multiple simulation models to work together, exchanging data between them to build a digital model that predicts the behavior of their assets under real-world conditions.

For example, the simulation-based digital twin model of an air compressor might include a Finite Element Analysis (FEA) model for strength and structural integrity of the compressor, a Computational Fluid Dynamics (CFD) model for air flow modeling, a heat exchange or a thermographic model to ensure heat dissipation and validate operating temperatures, and potentially a model to account for vibrations. All of these integrated simulation models can then be leveraged toward a simulation-based digital twin strategy to create serialized digital representations of each of these instances during and after production.

By collaborating with ANSYS to leverage their extensive library of simulation models along with their Twin Builder tool, manufacturers can develop digital twins models that provide them insights using simulation-based approaches. To help make these simulations scale faster and to help manufacturers with a secure and scalable computing platform, we offer specialized capabilities such as Azure High-Performance Computing capabilities that let you run your simulations and twins in the cloud. Or, you can work in a hybrid environment, allowing you to burst into the cloud, leverage Azure Synapse and Azure Time Series Insights for large time-series data analytics, and use Azure Data Sharing to securely exchange data, collaborate and work across the value chain.

With a common data model and compliance with standards like OPC UA, manufacturers can enable efficiencies across the value chain including Tier 1 and Tier 2 suppliers and engineering partners. And then of course, we are investing in advanced technologies like Automation Markup Language, machine learning, artificial intelligence, and a broader Digital Twin definition language (DTDL) to enable communication across different digital twin technologies. All of these layers of technology help manufacturers and partners leverage the very best of technology available to make digital twins interoperable, real and beneficial.

A rich partner ecosystem

We are proud of our partners in the digital twin ecosystem. In addition to collaborating with companies including ANSYS and Altair to leverage HPC for simulation-based twin innovation, we are proud to collaborate with Aras, Eurostep, PTC and Siemens around strong PLM, IoT and simulation capabilities.

For example, PTC offers Windchill, their PLM capability, via a managed service to customers leveraging Azure technology.

PTC is developing interfaces between Windchill's digital definitions and ThingWorx's physical experiences to bring the analytical digital twin to life. This gives engineers access to factory and field data, and vice versa. Not only is PTC enabling the digital twin, but they are also developing applications over it for top use cases like Data-Driven Design, Closed Loop Quality, and Augmented Reality Procedural Guidance. Often, a digital twin is in the design phase of a lifecycle, requiring simulation with realistic field data. PTC, through their partnership with simulation companies like ANSYS, is able to bring real-world products/environments, real-world field data, and realistic simulation for the in-design product. 

PTC extends their capabilities into supply chain and performance twins through aftermarket parts planning and inventory optimization capabilities. Another example of our joint partnership with PTC in terms of delivering value through digital twins is Howden. Howden is a leading manufacturer of industrial machinery, with over 160 years of experience in designing and manufacturing rotating equipment like pumps. Howden has built an interactive mixed reality experience using Vuforia AR solutions for HoloLens to help their customers visualize digital twin data from ThingWorx on Azure and to perform routine maintenance and self-service. They provide an overlay of the IoT-based operational sensor data and telemetry from the specific pump in the mixed reality to help users understand critical operational parameter incorporating alerts and analytics to identify potential failures. Animated sequences built from their CAD models allow users to interact with and explore the components and sub-assemblies and demonstrate the appropriate repair procedure. This enables exciting new levels of visibility into manufacturing.

You also have newer emerging companies like Uber Cloud and Rescale, who bring in the capability and flexibility to simulate seamlessly across multiple software providers. Their series of software versions provide engineers with computations from different solvers. Partners including Cadence, Mentor Graphics and Synopsis offer digital twin capabilities for specialized workloads such as electronic design which are core to the high-tech and semiconductor industries.

The next frontier

This is just a snapshot of where we are today with digital twin innovation and our partners. The landscape is growing significantly and rapidly. Simulation-based twins are an exciting breakthrough. We will see more and more dimensions of twins emerge as the industry continues to innovate and advance. On the horizon are capabilities such as performance-related twins, extending the concept to things like financial modeling of a company considering assets, risks, cash flows and other financial dimensions in the twin. Or, even a bi-directional twin that would allow aerospace companies to optimize aircraft performance 40,000 feet in the air from the touch of a screen or some advanced weather data related computations on the ground.

It is an exciting time. It is an exciting category of manufacturing innovation. Microsoft is pleased to be collaborating with the industry’s top partners to transform how we build, move and empower our world.