Digital Twins ain’t no Doppelgängers

Demystifying Digital Twins, proceeding as if Models are legal documents

By Anand Stephen, PE

Dictionaries define doppelgangers as an apparition or a lookalike of a living persons. We understand the presence of doppelgängers as a rare phenomenon. So much so that its occurrence is commonly understood in folklore as a paranormal spectacle. While we do not view twins as paranormal phenomena, we still consider them out of the ordinary. So, what does an article on Digital Twins have to do with Twins and Doppelgangers?

Depending on your perspective, even the existence of doppelgängers is more common than our imagination. There is a one in 135 chance of an exact pair of doppelgängers in our world of eight billion people. Crucially, statistician David Aldous makes a very astute point about understanding doppelgangers: “it depends whether we mean ‘lookalike to a human’ or ‘lookalike to facial recognition software.”

In the recent past, Digital Twins have paradoxically become popular and misunderstood. In concert with the rise in popularity of the Digital Twins, so too have the attendant misunderstandings. The views surrounding them range from simplistic to almost paranormal.

So, what is a Digital Twin?

A simple and realistic meaning that I prefer is the definition from the Digital Twin Consortium. They define a Digital Twin as “a virtual representation of real-world entities and processes, synchronized at a specified frequency and fidelity.” 

Let us pause to think of physical infrastructure; we can think of it as parts that make up roads, bridges, buildings, and, importantly, associated processes involved in engineering, constructing, operating, and maintaining them. Therefore, a Digital Twin as a composition of assets and processes accurately defines most architecture, engineering, construction, operations, and maintenance processes. Further, it is helpful to reflect on Aldous’s comment about Doppelgangers; it is essential to consider the end-user, in his case, the human or software analyzing the accuracy of doppelgangers.

With engineers in mind, let us further demystify Digital Twins by considering simple yet crucial workflows that are part of their daily lives. As engineers, we adhere to a Quality Management System (QMS) that validates our work products. Engineering work products, plans, reports, and calculations are legal documents backed by a sign and seal. Moreover, as engineers, we are bound by our engineering ethics to produce work products for the public’s greater good. However, to this day, most engineering work products are analog artifacts; even when they are digital, they do not meet the rigor of the Quality Control of analog products.

Currently, the practice of requiring Models as Legal Documents (MALD) is minimal. A National Academy of Sciences report noted that only 13 percent of the Department of Transportation (DOT) Agencies required MALD (See Figure 1.0 below; fifty-one percent of the DOTs use models for information only, and the remaining thirty-seven percent do not use 3D Models.

In any design project, there are two critical components. The first is designing new or improving existing infrastructure. However, adhering to QMS processes is equally important. Quality Control and Quality Assurance are essential parts of any QMS. The first is the act of performing predefined checks. While Quality Assurance provides confidence that appropriate Quality Control checks have been completed. Quality Assurance contains artifacts to prove that Quality Control was completed.

Robust QMS exists within the engineering community when delivering plans; an overview of the typical process is shown in Figure 2.0. A detailed process of quality control and assurances is depicted in Figure 3.0. The acts of performing checks, fixes, and verification constitute Quality Control. The artifacts, or proofs, generated from the Quality Control assure that agreed-upon quality control was performed. Both Quality Control and Quality Assurance together form the QMS³.

Even when Models are used and exchanged in projects, it is for information purposes only. The by-products of Models, paper plans (or the pdf variant) are upheld as legal documents. As seen in Figure 1.0, more than fifty percent of the DOTs fall in this category. Critically, in this scenario, when models are shared but not as a legal document, often there is a gap in the QMS.

In the scenario described above, often, there is a critical gap in the QMS, as seen in Figure 4.0. While the analog variants — paper or pdf — will triumph over Models in a legal battle, there is potential for contractors to consume incorrect data. Legal costs and higher construction costs notwithstanding, the gap in QMS of digital models open the possibility of building deficient infrastructure.

Circling back to our definition of Digital Twins as comprising assets and processes, for design engineers, assets are the models (plans) they develop. QMS procedures are part of the process. So how could we create a “virtual representation” of these “real-world” entities and processes?

One way in the engineering world is to leverage the iTwin platform, Bentley Systems’ implementation of Digital Twins. Using Bentley’s Model Review, we can create Digital Twins that are a virtual representation of the assets we design and the QMS processes that we practice. In addition, with Bentley’s Model Review, we can create a virtual representation of spatial and non-spatial assets. For example, Digital plans and models, which are geospatially located, and snapshots of digital plans, such as pdfs, are examples of spatial data without a link to a digital geospatial reference. In contrast, reports and calculations are examples of non-spatial data.

Virtual representations are created through Federation and Synchronization. These sound like complicated terms. While the technology to implement them is complex, the concepts are relatively simple. The Architecture, Engineering, and Construction (AEC) industry has been practicing Federation and Synchronization since the beginning of the industry when the sector started producing plans.

Let us consider the process of Federation. It entails creating an aggregate data model from multiple disparate data sources. In the analog era, it is the aggregation of plans that meet specific criteria—for example, compiling a set of plans that provide information on all the Drainage features in the project. Synchronization, on the other hand, is the process of establishing consistency between the aggregated data model and multiple data sources. In the analog age, synchronization occurs at an event triggering the compilation of the plans; for instance, printing plans before the milestone submission or a Project Manager calling a team lead requesting a print set.

Fast forward to the current digital era, Federation involves pulling together various models into one aggregate model. Connecting design data on Bentley’s iTwin platform is one mechanism to Federate a model. The data resides in a Common Data Environment (CDE), such as ProjectWise 365. Bentley provides several data connectors which enable the connection of data from disparate data formats. The data are read into the model and viewed as an aggregate via a web browser. In the analog era, you can think of connectors as printing from disparate sources, formats, and locations. The data connectors remove significant friction in the data Federation process.

Using Bentley’s Design Review platform, the consistency between data on the CDE is maintained by setting the synchronization at a predetermined schedule. A word of caution: any automation adds responsibility. There are use cases when the same model must be synchronized at different intervals: one for everyday internal progress and another for stakeholder review monthly, for example. While these two models pull data from the same sources, they are potentially markedly different. Like in the analog world, internal daily check prints are potentially different than milestone prints. In the digital era, it happens behind the scenes, and there is a tendency to assume that the same model synchronized at different intervals conveys the same information.

For any successful implementation of Digital Twins, merely representing assets is not adequate, the mirroring the processes from the “real” to the “digital” world are equally important. Using the Design Review platform, we can map the QMS processes from the analog to the digital world, see Figure 5.0. The inset depicts the process shown in Figure 3.0, while the main graphic shows the equivalent Digital Twin processes mapped within the iTwin platform.

As discussed previously, Quality Control and Quality Assurance are integral components of a QMS system. Using Digital Twins for the Design and Review, we can record and access all the issues and their resolutions from a common issues registry. In addition, Digital Twins also maintains an audit trail of all the problems and solutions, providing adequate assurance that Quality Control has been performed.

There are many use cases of Digital Twins; however, I hope this article demystifies Digital Twins by highlighting an everyday use case for us engineers and designers. Digital Twins are no doppelgängers. They are representations of assets and processes that facilitate users to access pertinent information. Digital Twins are a lot less mystical than doppelgängers with everyday applications.

Anand Stephen, PE, is the Digital Delivery Leader for the Roadway Group at Gannett Fleming. He holds Professional Engineer’s licenses in NJ and CA, and certificates in Bentley and Autodesk Platforms. He has over two decades of experience in engineering, software development, and sociology. Anand has led digital delivery across the US, he mentors engineers to use design technologies. He published articles and taught university-level courses. You can contact him at astephen@gfnet.com.