Digital Twins for Buildings and Infrastructure

Information only — not legal advice.

A digital twin connects a digital representation with information about a real asset to support a defined purpose. The connection, update process and decision use matter more than the visual sophistication of the model.

A static model can be valuable without being called a digital twin. Before adding live data or complex simulation, identify a decision that better information could improve.

Start with one operational question

Useful questions might include whether a ventilation system is performing as intended, which assets need investigation or how an operational change affects energy use. Define the user, decision frequency, required accuracy and action that follows an observation.

Then identify the necessary sources: asset identifiers, location and system relationships, sensors, maintenance records, occupancy information or an analytical model. Different purposes need different update frequencies; not every useful connection must operate in real time.

Example: investigating abnormal energy use

A fictional building team links meter readings to the correct equipment and spaces, alongside operating schedules and relevant weather information. It flags an unexpected pattern for investigation.

Before concluding that equipment is faulty, the team checks timestamps, units, missing readings, sensor quality and whether the operating schedule changed. A technician investigates and records the result. The digital representation supports a decision; it does not make every observed correlation a proven cause.

The team measures whether the workflow finds actionable issues with an acceptable false-alarm rate and reasonable investigation effort. A dashboard's visual appeal is not the acceptance criterion.

Establish trustworthy connections

Use stable identifiers to connect physical assets, models and operational records. Define who maintains those mappings after equipment replacement. A sensor assigned to the wrong unit can produce convincing but misleading conclusions.

Record source, timestamp, units, expected update interval and data-quality state. Make stale or missing data visible. Decide what the system does when an input is unavailable rather than displaying an old value as current.

Separate observation from control

A read-only analytical service and a system that can command physical equipment have different risks. Any control capability needs suitable engineering, authorisation, safety and cybersecurity assessment, with clear responsibility and fallback behaviour.

Limit access to sensitive asset and occupancy information. Review integrations and credentials, and avoid collecting personal data simply because a sensor or platform makes it possible.

Plan for the whole life

Account for sensor maintenance, calibration where relevant, integration support, model updates, cybersecurity, staff time and eventual migration. Identify who owns the operational information and what can be exported if a service ends.

Pilot one use case with a clear baseline and success measure. Expand only after the team demonstrates that the information remains dependable and improves the intended decision.

The Centre for Digital Built Britain's Gemini Papers discuss connected digital twins and the importance of purpose, trust and effective functioning. They are a useful conceptual reference, not proof that any particular product delivers a measured benefit. Continue with asset information handover, data ownership and vendor management.

Important information

This article provides general information and does not constitute legal advice or replace a qualified professional's assessment of your circumstances. Laws, official guidance, standards and product features can change. Check current primary sources and the requirements applicable to your jurisdiction, contract and project before making decisions. Examples illustrate the discussion; they are not official rulings or instructions for a particular case. For legal decisions, consult a qualified lawyer; for technical or other specialist matters, consult the responsible professional.