Harbour Terminal Digital Twin Review

The passenger terminal had been in service for two years, and the facilities team had inherited a model that looked right but had never been checked against the building that was actually standing in the harbour. The commission was a post-occupancy asset-information review: reconcile the as-built model against a freshly surveyed point cloud, correct the discrepancies, and deliver a structured asset register the FM team could actually run the building on.

The first phase was geometric reconciliation. The point cloud was registered to the model’s shared coordinates and each building system was compared in Navisworks — structure, envelope, MEP, and fit-out run as separate analyses. Measured deviations were reported at the element level, with the majority of the building falling within ± 12 mm of the survey. Where the model drifted further — a relocated door in the baggage hall, a plant-room partition moved 200 mm during construction — the geometry was corrected to the survey rather than argued with it. Every change was logged so the FM team could see exactly what the as-built record had been, and what it now claimed.

Reconciling the asset data

The geometry was the visible half of the problem; the data was the larger one. The model carried 12,400 asset items, but their attributes had been typed by different contractors to different conventions over the two construction phases. A set of Dynamo scripts was written to reconcile the model’s parameters against the FM database: equipment tags were matched on a combination of location and type, warranties and commissioning dates were pulled in where they were missing, and anything that failed the match was exported to a review list for a human decision rather than a silent default.

IFC export and property-set mapping

The register was only worth anything if it could leave Revit in a form the FM team could use. The export was validated against the ISO 19650 Stage 6 handover requirements: property sets mapped to the facility-management schema, spaces exported with their room data, and every asset carrying the required classification. The IFC was checked in Solibri for empty property sets and orphaned elements — the kinds of errors that only appear when someone actually tries to use the file — until the export opened clean in the FM’s own software.

---

The delivered register gave the facilities team a single source of truth: an IFC-ready asset model, a matching database export, and a deviation report that told them exactly where the building differed from its drawings. The lessons fed back into the process for the next project — lock the as-built rules at the start of construction, not the end, and the digital twin stops being a review exercise and becomes a working asset.