Sledenje in obnavljanje identifikatorjev IFC v medplatformnih delotokih BIM

The Industry Foundation Classes (IFC) standard provides an open and durable basis for information exchange in Building Information Modelling (BIM). Its reliability depends on the persistence of the GlobalId (GUID – Globally Unique Identifier), which ensures stable references to model elements across versions and platforms. In practice, however, identifiers are frequently regenerated during routine modelling, parametric scripting, or import–export roundtrips. When identifiers change, elements are interpreted as new, links to associated information are lost, and comparison results become incomplete. This thesis analyses the conditions under which identifiers are altered and evaluates how selected BIM management platforms respond to such discrepancies. Empirical tests confirm that unstable identifiers are a recurring outcome of standard workflows. To address this issue, a custom Python-based analysis script was developed. The script compares two IFC versions, detects changes in geometry, placement, and properties of elements, and restores original GUIDs through a rule-based mapping process. The outputs include Excel audit reports and enriched IFC files that visualise and re-establish identifier continuity. Validation on simplified test cases and large-scale project models demonstrates that identifier mapping reliably recovers connections otherwise broken by regenerated GUIDs. The results confirm that identifier instability is a critical barrier to robust IFC-based workflows, while also showing that it can be transformed into a manageable process. By restoring identifiers, the method enables reliable change detection, stable information linkage, and a durable foundation for lifecycle-spanning project data management.