Digital Twins and Data Spaces: Feeding the Virtual Replica with the Right Data

A digital twin offers something highly appealing to anyone responsible for infrastructure or public services: the ability to simulate what might happen before it actually does. Testing the impact of a new sewage network, anticipating a neighbourhood’s energy consumption or planning maintenance for a water distribution network without taking any risks in the real world. The problem is that most digital twins deployed today represent only a small part of reality because they rely exclusively on the data available to a single organization.

A building does not operate in isolation from the electricity supply that powers it, the water network that serves it or the weather data that influences its energy demand. The same is even more true of an entire territory. When a digital twin is built using only the internal data of the organization developing it, the simulation is inherently partial, and any decisions based on it inherit the same blind spots.

The solution is to connect the digital twin to a data space involving the different organizations that generate relevant information about that territory, building or infrastructure network. This article explains why this connection is the missing piece in many digital twin projects and how it can be implemented in practice.

What is a digital twin, and how does it differ from a static simulation?

A digital twin is a virtual representation of an asset, system or territory that remains synchronized with its physical counterpart through updated data, typically from IoT sensors and management systems. Its key difference from a traditional simulation lies in this continuous synchronization: while a static simulation runs once using fixed parameters, a digital twin is constantly updated, making it possible to anticipate scenarios based on the current real-world situation rather than a snapshot taken months earlier.

This continuous updating is what makes a digital twin valuable for tasks such as:

• Detecting operational anomalies before they develop into failures.
• Simulating the impact of an intervention (construction work, a route change or a capacity expansion) before it is carried out.
• Planning maintenance based on the actual condition of assets rather than generic schedules.
• Anticipating how a system will behave under extreme conditions, such as demand peaks or adverse weather events.

However, all these capabilities depend on the quality and breadth of the data feeding the model. A digital twin is only as good as the data that powers it, and this is where the need to look beyond an organization’s own data comes in.

The limitations of digital twins built in isolation

When an organization builds a digital twin exclusively using its own data, it encounters structural limitations that go beyond the volume of information available:

• Lack of external context: a building’s digital twin cannot accurately anticipate its energy demand without incorporating up-to-date, high-quality weather data from external sources.
• Artificial boundaries: a water distribution network does not end at a municipality’s administrative boundaries, yet consumption, pressure and water quality data are often fragmented across different operators.
• Duplicated effort: if each organization builds its own digital twin without sharing data with others, they end up repeating data collection and cleaning work, even when the information already exists at a neighbouring organization.
• Incomplete simulations: planning the expansion of an industrial estate without data on the electricity grid and water supply produces a model that overlooks precisely the interdependencies that often cause real-world problems.

A data space addresses these limitations by providing a governed channel through which multiple organizations — public administrations, utility providers and technology centres — can share the data needed to power a digital twin without losing control over their own information.

Digital twins of territories: when the model spans multiple organizations

An urban digital twin, or one covering a wider area such as a district in Castile and León, needs to integrate data that is rarely generated by a single entity: land registry and urban planning data (local authorities), electricity supply networks (energy distributors), water supply and sanitation networks (water utilities), environmental quality data (regional authorities) and socioeconomic data (public administrations and statistical bodies).

Imagine a municipal consortium in Castile and León, the Ribera del Duero Municipal Consortium, that wants to build a digital twin to plan the expansion of industrial areas without overloading existing networks. Through a data space, the consortium can access electricity grid capacity data published by the distributor, available water flow data published by the water utility and cadastral data from the relevant local authority — all subject to usage policies that each organization controls independently.

The resulting digital twin makes it possible to simulate different growth scenarios and identify, before any project is approved, where the electricity grid or water supply would begin to come under strain.

Without access to this external data, the consortium would have to estimate these variables using generic assumptions, and the digital twin would cease to be a reliable planning tool and become little more than an illustrative exercise.

Buildings and networks: joint energy and water simulations

At a smaller scale, the same principle applies to buildings and infrastructure networks. A digital twin of an industrial estate seeking to optimize the combined energy consumption of its businesses needs consumption data from each facility, generation data where renewable self-consumption is available, and information about the distribution grid connecting them. No individual company has the complete picture needed to carry out this optimization on its own.

The same applies to water networks. A digital twin designed to reduce water losses in a supply network needs to combine pressure, flow and quality data from different points in the network, often managed by different concessionaires or public authorities along the water’s route. Sharing this data through a data space, with clear usage policies specifying who can access what and for which purpose, makes it possible to build a shared model without any party relinquishing control over its own infrastructure.

Simulation and territorial planning using shared data

The added value of connecting a digital twin to a data space is not simply having more data. It is the ability to run scenario simulations that combine variables from different domains. Planning the location of a new public facility, for example, can benefit from combining demographic data, energy availability and environmental impact information, all sourced from different organizations.

This ability to run joint simulations is particularly relevant to territorial planning. Decisions that were previously made using partial data and planning horizons spanning several years can now be supported by models that are continuously updated and incorporate the perspectives of all relevant stakeholders across the territory, rather than just the organization leading the project.

For this to work in practice, shared data must be described consistently in a federated catalogue, access policies must be clear for every participating organization, and an audit mechanism must be in place to verify that data is being used as agreed. A data space that integrates these components is what turns a collection of scattered IoT sensors into a digital twin that can genuinely support decision-making.

Build your digital twin on shared, governed data

If your organization manages infrastructure, networks or territory and wants to develop a digital twin that can leverage data from other entities with the necessary safeguards, a modern data space provides the federated catalogue and access policies needed to make it possible. Look for a technology partner with experience in data spaces to explore how this approach could be applied to your project.