Architecture for a BIM Digital Twin

How to Leverage BIM Models of Buildings and Infrastructures to serve as the Basis for a Digital Twin? This question arises frequently because the technologies available on the market have not yet provided a definitive answer. 

 

Most consider that the digital models feeding the Digital Twins are static and there is no need to worry about changes. However, buildings, telecom towers, substations, bridges, pumping stations, pipe racks, and most civil infrastructures change and update at incredible rates, sometimes in unpredictable ways that are difficult to track and document.

 

For example, a government client with a campus of four buildings in eastern Spain explained to us how their Facilities team handles an average of 1,000 intervention requests per year, from changing a Wi-Fi access point to moving an entire department. A thousand interventions to be planned and documented (or not) in the BIM model. In the same way, a large retail client who constantly renovates their centers to keep them attractive for sales campaigns underscored how difficult it is to maintain up-to-date overall documentation. Dozens of changes, some managed from the design center and others decided directly at the sites, are at best schematically recorded in the plans.

 

Poorly documented changes are not exclusive to buildings. Consider remote installations like telecom towers, substations, or water plants, where planning a change often starts from incomplete documentation because it becomes costly to send a technician to create an accurate as-built or validate that the installed or changed elements match the planned ones. And we don’t take into account changes due to unforeseen needs or breakdowns that are resolved on-site and are rarely documented anywhere.

 

BIM is not a magic solution, but it does help to have a single, up-to-date picture of the asset at any given time, and that is certainly an advantage. If a change needs to be made, doing it in the BIM model ensures that the entire asset and all its graphic documentation are updated.

Here, we need to introduce the distinction between the BIM project information model, the asset model, and the digital twin.

 

The BIM project model is the one used for planning, decision-making, and documenting the changes that are intended to be made to an asset. It is the model that is edited and worked on to produce a project or to monitor a construction or installation. The project model may contain information relevant to the operation of the building or infrastructure, but its main objective is to document changes, support the study of alternatives, and contain information relevant to the design and construction phases, much of which loses its value once the project is completed.

 

The asset model is the representation of the building or infrastructure, along with all the elements that are specific to it and are of interest for its operation, whether for controlling depreciation, maintenance, or being an essential part of the business characteristics needed by the owner or operator. An asset model, for example, might exclude reinforcement details, support structures, or budget-related information if the operation of the asset does not require that information. Counterintuitively, the level of information needed for the asset (LOD500) can be substantially less than the level of information needed for construction (LOD400). An asset model is the map on which the digital twin can be built.

 

The Digital Twin results from the integration of the asset model with data from other information sources to enable the organization’s use cases. The Digital Twin allows simulating, analyzing, and optimizing the asset’s performance and operation in a safe digital environment. It is a dynamic, evolving tool that facilitates informed decision-making and continuous improvement of the asset or asset portfolio.

 

On the other hand, it is important not to confuse a Common Data Environment (CDE) with a Digital Twin. The primary purpose of a CDE is to serve as a collaborative space for project development and information exchange, whereas a Digital Twin is what the organization needs it to be, with specific use cases, controls, and integrations. 

Therefore, if we say that buildings and infrastructures change frequently, and that using BIM to describe and document them simplifies achieving an accurate representation, we can say that a digital twin, whose use cases depend on a good representation of the asset, must be more or less permanently connected to the BIM model. Ideally, a change in the asset should be planned in the BIM model and, once executed, the model should be used to update the digital twin.

 

To achieve this, it would be necessary to develop an architecture that connects the BIM models through a common data environment (CDE) or a project data web platform with the organization’s Digital Twin platform. Otherwise, the risk of having a Digital Twin with outdated information increases significantly.

Before detailing the architecture to develop a Digital Twin based on BIM, let’s examine the relationships that must be established between the project and asset models to keep a Digital Twin updated. 

The asset model should always result from the evolution of the project model, but adapted to meet the Digital Twin’s needs.

Our experience shows that the adaptation or conversion of a project model to an asset model can be done automatically, provided certain minimum conditions are met in the project BIM model. The key activities to convert a project model into an asset model are threefold:

• Remove all elements and attributes irrelevant to asset management. For example, eliminate structural reinforcements, cost control, or mounting information.
• Organize elements differently from the project using as a basis existing attributes or relationships with other elements. For example, group terminals by the room they occupy or equipment by process instead of discipline.
• Add information to project elements using existing coding. For example, add warranty information or installation dates from an identification code.

In other words, the conversion from project model to asset model should only involve automatable activities. 

While developing a BIM project model involves manual work, to use this model as the basis for a Digital Twin, it must be convertible to an asset model automatically and without human intervention.

In this way, any modification to the asset must be made through the project model, and the asset representation is created using a model that is generated automatically.

However, the organization must adjust this approach to its assets’ characteristics and available resources. Organizations with a few large assets might consider a manual conversion between project and asset models, but Utilities or Telecom companies with thousands of service points have no alternative but to automate it.

Since 2020, we have been developing Digital Twin BIM projects under this approach. Below, we describe the main components of the architecture we use.

 

 

0. Sovereignty Over the Digital Asset Model

Maintaining control and access to the 3D models forming the basis of a Digital Twin is essential for any organization. 

By having direct access and control over these models, the company ensures adaptability to changes and new needs, preserving its technological independence and avoiding third-party dependencies. 

Moreover, possessing this control facilitates integration with other systems, ensuring data management consistency. Lastly, 3D models are not just representations; they are strategic assets crucial for planning and decision-making in today’s world.

The digital twin architecture must align with the digital asset model’s sovereignty, ensuring access to model information and structuring architectural components so they can be changed without affecting the rest. Excessive dependence on a format or technology can lead to operational problems and significant additional costs.

 

1. BIM Project Model

The BIM project model should be highly oriented towards editing. For this purpose, commercial alternatives (such as Autodesk Revit, Archicad, etc.) that use proprietary formats still seem the best. The model should be maintained with the highest quality and health possible, including the minimum number of information containers, lists, annotations, and plans necessary.

 

2. Common Data Environment (CDE)

To properly control exchanges, publications, and issues specific to the design and construction stages, it is advisable to host and manage the BIM project model in a modern Common Data Environment. 

Besides storing the models, these applications integrate their geometry and properties with the platform’s functionalities: plans are managed together with the models they derive from, issues can be created in specific model locations, and elements and properties can be extracted. The CDE can also be used to store the asset model.

 

3. BIM Asset Model

To maintain sovereignty over asset data and given that, as proposed, the asset model is a snapshot of the project model, enriched but static, using an open standard like IFC (Industry Foundation Classes), a schema promoted by the Building Smart foundation to describe objects in the built environment, is appropriate.

 

4. Digital Twin Platform

The final element of our architecture is the platform that automates model conversions and integrates the common environment with the Digital Twin. This system must serve two purposes:

On one hand, it should orchestrate model exchanges, automatically convert the project model to the asset model, convert and publish it for use by the Digital Twin, and manage related permissions. Additionally, it can add capabilities to the CDE specific to the organization, such as library management, integration of CDE issues with tickets in other systems, or data extraction from models.

On the other hand, the same platform should be used to develop Digital Twin use cases. Organizations that prefer to develop their own Digital Twin environment can extend this platform to manage necessary functionalities, connect with data sources and devices, and ensure proper integration with the asset’s 3D models natively.

There are multiple alternatives for this platform, but all involve custom development for the organization and its users. Whether developed on Azure Digital Twins, AWS TwinMaker, or on top of Bentley’s iTwin or Dassault’s 3D Experience environment, achieving the described vision and level of integration will require considerable programming and customization effort.

At Modelical, we are developing Modelical.io, our Digital Twin Platform to connect the worlds of asset and project, perfect for creating digital twins based on BIM models. It is a fully adaptable solution that includes visualization, model management, automation, data extraction, and user management tools. Modelical.io is a toolkit serving the most demanding organizations with their use cases, ensuring the quality and availability of their asset information, and designed for complete project integration.

Thus, it is possible and highly recommended to leverage BIM models of buildings and infrastructures to serve as the basis for a Digital Twin. As we have addressed in this article, the keys to this process are:

• Accepting that assets are always changing.
• Having specific BIM models for projects and assets, where the asset model is an automatic result of the project model.
• Automating conversions and their maintenance.
• Providing a platform that is sufficiently powerful and versatile to integrate the asset model with external data sources and enable the realization of desired use cases.

At Modelical, we can support you in developing the Digital Twin for your organization. We offer comprehensive BIM services, Digital Twin development, and digital transformation. We combine our expertise in digital asset management and technology implementation to add value to our clients’ projects and businesses. For over 10 years, we have been working with leading and demanding sectors including energy, infrastructure and transportation, oil and gas, logistics, real estate, and AEC, among others.