Learn the Path from Systems Model to Digital Twin

Siemens Digital Industries Software has sponsored this post.

Engineers are in the midst of a digital revolution, and if you haven’t woken up to it, consider this your alarm. The industry is racing to find new ways of working in a world that is getting more connected and complex. As a result, it can feel like engineers are drowning in a sea of digital tools. Those that can’t keep up are going to fall behind.

But don’t panic. It’s not too late to start thinking about digital transformation. Better yet, even small jumps into these technologies can yield large improvements. However, without proper planning, it’s hard to pull off. There is a risk of unintentionally implementing disparate and siloed systems that act no better than passing paper between cubicles.

With digital transformation, it’s best to start small and grow the system naturally. As that system grows, it becomes more complex until it is eventually a digital representation of the product, processes and operations of an organization—or in other words, a comprehensive digital twin.

This systematic, lifecycle approach is why model-based system engineering (MBSE) is a good starting point to digital transformation. The methodology and technology act as a foundation that guides the creation of a digital twin by connecting the development, engineering and manufacturing processes to products in the field.

So, how does MBSE become the digital twin backbone?

It Starts With System Level Thinking

MBSE can become the skeletal structure of a digital transformation because it forces those using it to develop system-level thinking skills. Nand Kochhar, Vice President of Automotive and Transportation Industry at Siemens Digital Industries Software defines system-level thinking as “an approach that seeks to maintain a holistic view of the problem which needs to be solved.”

This view not only includes the product itself but the entire development, manufacturing, use and environmentally-aware disposal of the product and all the subsystems it contains. Connecting this all together is not an easy feat because the subsystems will span various domains of expertise (from mechanical to electrical and software) and they contain lots of models, simulations and data.

“MBSE is a methodology that helps to manage the production and process of digital twins,” says Kochhar. “It ensures that the information is accessible to all stakeholders around the company while maintaining an authoritative source of truth. That information can be applied across the entire organization. It builds on the solid product lifecycle management core of the systems.”

In digital transformation, this concept of data continuity is called the digital thread. The idea is that throughout the lifecycle of a product, from its earliest conception to the time it is recycled, data are continually captured in one single source of truth. MBSE, as well as product lifecycle management (PLM), are good examples of how this digital thread can be maintained.

How to Model a System

The idea of producing a system-level model sounds daunting. But it has a simple first step: ask fundamental questions. For example, if you’re designing a vehicle, you might ask:

• Will my product shuttle people across nearby cities or freight across continents?
• Will it operate over land, air or sea?
• Will it be battery-powered, hydrogen-powered or use an internal combustion engine?

As the answers to these questions become clearer, engineers will have a better idea of what the full system should look like. This can then be mapped out using MBSE tools.

This version of the MBSE model starts off as little more than a flow chart that, at best, contains simple models of a few systems and subsystems. By mapping them all together in one place, engineers can start to expand the size of the system model or cascade into the specifications of each subsystem.

“MBSE is the fundamental basis of the [digital twin] methodology,” says Kochhar. “It starts with the requirements and then you form the architecture at a high level. The next step is concept modeling. In the concept state, you don’t have all the details, you have the basic information. But you are still able to leverage simulation, artificial intelligence, machine learning and other technologies to select the appropriate architecture. Then you cascade that downstream and cross-domains to the mechanical, electrical and software groups within your organization so they can deliver the design details [for each subsystem] that meets the objectives.”

As the design teams begin to produce high fidelity models of each part and subsystem, the crude system model becomes more deeply defined. These models will be informed and often produced using tools like CAD, CAE and CFD. To keep the system-level thinking alive, MBSE must be a constant presence in that design process. The high-level perspective can’t be forgotten as low-level design details are decided. As the system model grows upon itself, it eventually becomes a digital twin.

It is here that the engineers can begin to see the trade-offs reality will force upon the design. For an electric vehicle, for example, how do we find the sweet spot between battery capacity, weight, performance and cost? We need a full system-level picture to find the answer.

MBSE in the Validation Process

As each subsystem becomes better defined, everyone working on the project, from the system modeler to a CAD engineer, can verify and validate that each subsystem continues to fit and work together.  This real-time collaboration breaks down engineering silos and streamlines communication, uncovering problems or design issues early, saving time and costly downstream changes.

For instance, an engineer testing the sensor systems within a robotic arm could use the MBSE model to produce a system-wide simulation of the arm and its surroundings. This would ensure that the sensors not only work with the control system, but that the arm’s geometry doesn’t block any sensors.

These system-level simulations can then feed information back in the other direction by providing boundary conditions that a particular subsystem might experience—such as the stresses put onto the arm’s gears and braking systems when a worker abruptly walks into the range of a sensor. By plugging those boundary conditions into a finite-element model, the engineer can assess if the gears and brakes are mechanically up to the task. Similarly, if the gear isn’t designed yet, engineers could use those boundary conditions as inputs into generative design tools that can suggest workable geometries. Moving back and forth allows engineers to continually design, validate and verify a system at every level.

“MBSE is a methodology,” says Kochhar. “At the end of it, you have a product and a virtual representation of a product and its performance. That's when you can bring together your virtual to the physical product, so you can correlate the two.”

In other words, after the validation process is completed and the product is in the actual world, the MBSE and the digital twin can act as a predictive tool. As the real-world sensors on the arm detect its performance, the data can be fed back into the digital twin to produce continual performance improvements or to predict when a maintenance crew should be called in.