Enhancing Semantic Interoperability of Heritage BIM-Based Asset Preservation

Abstract

Preservation of Cultural Heritage (CH) demands precise and comprehensive information representation to document, analyse, and manage assets effectively. While Building Information Modelling (BIM) facilitates as-is state documentation, challenges in semantic interoperability of complex cultural data often limit its potential in heritage contexts. This study investigates the integration of BIM tools with the buildingSMART Data Dictionary (bSDD) platform to enhance semantic interoperability for heritage assets. Using a proof-of-concept approach, the research focuses on a historic tenement house in Tarnów, Poland, modelled with the IFC schema standard and enriched with the MIDAS heritage classification system. The methodology includes transforming the classification system into bSDD data dictionary, publishing thesauri for components, materials, and monument types, and semantic enrichment of the model using Bonsai (formerly BlenderBIM) plugin for Blender. Results demonstrate improved consistency, accuracy, and usability of BIM data for heritage preservation. The integration ensures detailed documentation and facilitates interoperability across platforms, addressing preservation challenges with enriched narratives of cultural significance. This method supports future predictive models for heritage asset conservation, emphasizing the importance of data quality and interoperability in safeguarding shared cultural heritage for future generations.

1. Introduction

The preservation of cultural heritage (CH) has increasingly come to rely on digital methods that ensure accuracy in documentation, transparency in information management, and sustainability in conservation processes. Among the range of digital approaches developed over the past two decades, Heritage Building Information Modelling (HBIM) has emerged as a particularly promising methodology. HBIM extends the principles of Building Information Modelling (BIM) to the domain of historic architecture, accommodating the irregular geometries, material heterogeneity, and layered historical narratives that are characteristic of cultural heritage assets [1]. Unlike conventional BIM, which was originally conceived for contemporary design and construction, HBIM must address issues such as incomplete archival sources, the need to capture intangible cultural significance, and the integration of multidisciplinary knowledge across architecture, engineering, archaeology, and conservation [2].

The increasing body of literature demonstrates that HBIM is no longer limited to representing physical form but is gradually evolving into a comprehensive digital environment for knowledge management in conservation [1]. Recent state-of-the-art reviews underline its role in providing a structured framework for preventive preservation, risk assessment, and the long-term monitoring of historic sites. Furthermore, studies on the digitisation of cultural heritage have highlighted the importance of standardization, particularly in linking capturing-reality techniques with interoperable information structures. Such efforts show that HBIM is not only a geometric tool but also a means of ensuring consistency and transparency in heritage documentation [2].

Central to these developments are open standards and shared vocabularies. The Industry Foundation Classes (IFC), approved as ISO 16739, provide the internationally recognized open schema for the description and exchange of building and infrastructure information [3,4]. Complementing this, the buildingSMART Data Dictionary (bSDD) establishes a semantic environment that enables the alignment of concepts, properties, and classifications across different datasets [5]. Together, IFC and bSDD represent the backbone for semantic interoperability in HBIM, although their adaptation to heritage-specific conditions remains an ongoing challenge due uncertainty in source data and the complexity of historical materiality [1,2]. Addressing these challenges requires both methodological advances and domain-specific extensions to existing standards.

In response, academics start to propose integrated HBIM frameworks that systematize the flow from survey data acquisition to semantically enriched model creation. Such frameworks emphasize the need to retain not only geometric precision but also the contextual information that makes heritage datasets meaningful [2,6]. They enable conservation offices to bridge the long-standing gap between traditional analog practices and digital innovation, ensuring that HBIM models become reliable repositories for interdisciplinary decision-making.

The practical application of HBIM can be observed in different case studies. The Galleria dell’Accademia di Firenze project demonstrated how HBIM supports the integration of structural assessments with ongoing maintenance planning, providing both technical and managerial benefits [7]. Similarly, the City Walls of Pisa project developed a pipeline that combined large-scale laser scanning with historical documentation, illustrating the scalability of HBIM to complex architectural ensembles and urban heritage [8]. These cases confirm the feasibility of HBIM not only as a documentation strategy but also as a conservation management tool.

Beyond such implementations, HBIM research has expanded into new technological frontiers. Artificial intelligence (AI) techniques are increasingly applied to automate the segmentation and classification of point clouds, thereby accelerating the labor-intensive process of converting survey data into structured HBIM models [9]. At the same time, the integration of HBIM with Extended Reality (XR) technologies—including virtual, augmented, and mixed reality—has opened new opportunities for cultural mediation, offering immersive environments that enhance both professional collaboration and public understanding of heritage assets [10].

A central challenge in modelling historic information lies in the uniqueness of heritage assets and the specialized terminologies used to describe them. While the IFC schema provides comprehensive support for common building elements, it does not natively incorporate the detailed technical vocabularies or heritage-specific descriptors required for cultural assets. Recent developments in openBIM and Semantic Web technologies provide new possibilities to overcome these limitations. In particular, the buildingSMART Data Dictionary (bSDD), developed in alignment with ISO 23386 and ISO 12006-3:2022 [11,12], establishes a standardized framework for classifying and describing building components and systems across languages and professional contexts. The bSDD acts as a hub of interconnected dictionaries published by independent organizations, enabling BIM environments to reference authoritative definitions through persistent Uniform Resource Identifiers (URIs). This approach enhances data consistency and interpretability by reusing shared definitions rather than generating bespoke terminologies. The service provides both a web interface and an Application Programming-Interface (API) for direct software integration, which facilitates semantically enriched HBIM models that can incorporate not only common attributes but also historical interpretations and conservation values [13]. In practical terms, users often access the bSDD not through its web portal but via BIM authoring tools, that more and more often integrate with the bSDD through plug-ins and extensions, as demonstrated later.

Despite this potential, the integration of bSDD into HBIM workflows remains limited. Research has begun to explore its applicability in heritage contexts, such as documenting states of preservation through openBIM systems [14], and in the development of scientific reference models for cultural heritage that seek to define rigorous standards for 3D documentation and semantic enrichment [13]. Recent contributions, including the work of Argasiński and Tomczak, further demonstrate how bSDD can be applied to heritage-specific classification systems, highlighting both opportunities and methodological challenges in adapting generic construction vocabularies to the cultural heritage domain [14,15,16].

This paper addresses these gaps by presenting a proof-of-concept study that combines BIM with the IFC schema [4], semantic enrichment, and bSDD mapping in the context of a historic tenement house in Tarnów, Poland. The project investigates how HBIM, extended with heritage classification systems accessible directly within modelling environments such as Blender [17] and its Bonsai extension [18], can serve as a flexible and interoperable platform for cultural heritage documentation. Through this approach, we aim to establish a methodological foundation for enriched digital representations that preserve the narrative depth of architectural heritage while also supporting conservation decision-making and long-term management.

2. Materials and Methods

2.1. Synthesis and Further Analysis

Standardised terminologies improve data interoperability, making sharing, integrating, and analysing information across systems and organisations easier. Enhanced search capabilities allow efficient records retrieval, while a common language supports research, analysis and pattern identification. In this context, the term common language refers to a shared, standardised vocabulary that ensures all actors use the same terms and definitions when describing heritage elements. Such a common language eliminates ambiguity, facilitates data exchange between systems, and supports interdisciplinary collaboration. In HBIM community, integrating dedicated heritage dictionaries such as Getty AAT and similar resources, as seen in research projects such as DURAARK, has proved invaluable [19]. These dictionaries help categorise and label heritage building elements, which is essential for effective data management and retrieval. In addition, they support conservation efforts by providing accurate documentation of conditions and interventions, ensuring accurate communication and recording of conservation work [20].

Different knowledge-organisation instruments play complementary roles in this ecosystem. Taxonomies provide a hierarchical structure for classifying data and sub-categories. They are relatively static and focus on simple classification systems, ensuring consistency and common terminology across datasets. Thesauri play a crucial role in classifying heritage sites, providing a structured and standardised vocabulary for describing different aspects of these buildings [21]. They provide the consistency of terminology necessary for accurate recording and retrieval, improving communication between heritage professionals. The thesauri’s detailed descriptions and hierarchical structures help users understand the exact meaning and relationships between terms, facilitating comprehensive and nuanced classifications. Overall, thesauri support adequate documentation, research, conservation and management of heritage resources [22,23].

Ontologies extend these principles by offering formal, explicit specifications of entities, properties, and relations within a domain. Unlike taxonomies and thesauri, ontologies support constraints, logical axioms, and inference, which makes them well-suited to BIM/HBIM, where object–property–process relationships are central to modelling and query [24,25,26]. In practice, the same features that make ontologies expressive can complicate cross-system integration: aligning heterogeneous models and reconciling overlapping conceptualisations require rigorous methods to avoid semantic drift and fragmentation [25,26] and ([27], p. 3).

The concepts of Linked Data and the Semantic Web provide the foundation for representing, connecting, and exchanging structured information independent of the construction domain. Proposed by Berners-Lee et al. [28], the Semantic Web extends the current web by enabling machines to interpret and reason over data, not only display it. Central to this vision is Linked Data, which relies on four principles: using URIs as names for things, making these URIs dereferenceable via HTTP, providing useful structured data in standard formats, and including links to other URIs to enable discovery. The Resource Description Framework (RDF) is the core data model of the Semantic Web [29,30]. It represents information as subject–predicate–object triples, enabling the integration of heterogeneous datasets and the creation of machine-readable relationships [29,30]. RDF thus underpins knowledge graphs, allowing diverse data sources to be queried, linked, and reused consistently.

Together, Linked Data, RDF, and ontologies provide the infrastructure to move beyond simple classification schemes toward interconnected knowledge systems [31]. When applied in BIM and heritage contexts, these technologies enable not only consistent documentation but also cross-dataset integration, advanced querying, and reasoning—critical for conservation, management, and the long-term reuse of digital heritage information [25,26,32,33]. Integrating taxonomies with ontologies in BIM allows for a robust knowledge-organisation system. Taxonomies provide the necessary hierarchical structure, while ontologies add depth with properties and relationships. This combination effectively manages and visualises historical data, making it easier to link different datasets and derive meaningful insights. Creating knowledge graphs based on ontologies can further enhance the management of historical BIM data [34]. These graphs help visualise the relationships between historical objects and their attributes, supporting advanced applications such as predictive maintenance and restoration planning.

Pauwels et al. [25,27] highlight the potential of integrating BIM and Semantic Web technologies, specifically focusing on the semantic enrichment of design and construction information. Extending this to the heritage domain, Simeone et al. [35] demonstrated how semantic enrichment of BIM enhances the representation of built heritage, ensuring richer and more meaningful digital documentation. Iandanza et al. [36] further emphasised the role of Semantic Web BIM in connecting different users and supporting the interpretation of cultural heritage models. In this context, the WissKI 3D Repository studies by Bajena et al. [37] provide concrete implementation of semantic repositories for 3D heritage, facilitating exploration, preservation and metadata management. Furthermore, OntPreHer3D by Bajena [38] introduces an ontology for preservation of 3D heritage models that systematically captures not only component structure but also provenance, uncertainty, and interpretative choices underlying reconstructions. Lastly, Fonnet et al. [39] explore integrating mixed reality with HBIM to support preventive maintenance, illustrating how immersive technologies can leverage semantic infrastructures for better outcomes in heritage asset conservation.

MIDAS Heritage [40,41,42] is a British standard for recording cultural heritage information on various assets such as buildings, monuments, archaeological sites and shipwrecks. It outlines the minimum information required and procedures for understanding, protecting, and managing heritage assets. Used by government organisations, local authorities, heritage sector organisations, and researchers, the first edition was published in 1998 by the Royal Commission on the Historical Monuments of England, and the second edition in 2007 by English Heritage (renamed to Historic England) [40]. The standard was developed for the Forum on Information Standards in Heritage (FISH) to address standards and recording issues in the heritage sector.

The bSDD platform plays important role in the context of applying semantic standards to BIM [5]. The platform is based on ISO 23386 standards for data dictionaries and ISO 12006-3, the framework for classifications. The bSDD helps streamline the standard and provides them to users in their authoring software. The open API of the bSDD allows accessing the content from any application as JSON or RDF data [27]. Integrating standard vocabulary, bSDD and BIM ontology-based heritage dictionary within BIM environments can significantly enhance the documentation, maintenance, and semantic interoperability of Cultural Heritage Assets.

2.2. Case Study: Historic Tenement on Chopina Street, Tarnów

The study focuses on a historic residential tenement located at F. Chopina 11 in Tarnów (formerly ul. Klikowska 13), which is listed in the Heritage Inventory (Ewidencja Zabytków) [43] under Inspire ID PL.1.9.ZIPOZ.NID_E_12_BK.388404 [44]. The building is officially recognized as a cultural heritage asset in the Małopolskie Voivodeship (Figure 1).

The Tenement was constructed in 1912–1913 for Dr. Zygmunt Doliwa Dzikowski (1848–1928), a municipal and county physician, and his wife Katarzyna Henisz-Dzikowska (1855–1933). The design was prepared by Franciszek Hackbeil Jr. (1879–1921), a prominent Tarnów architect known for residential and sacral projects. The house was later associated with Jan Szczepanik (1872–1926), a renowned Polish inventor often referred to as the “Polish Thomas Edison”. Szczepanik lived and worked in the building in his later years, and today the façade bears a commemorative plaque unveiled in 1972.

Architecturally, the building represents an eclectic style with neobaroque features and traces of secession. It is a brick masonry structure with sandstone detailing, erected on an irregular plan with a two-bay layout. The tenement consists of a cellar, ground floor, first floor, and a usable attic. Its distinctive features include a corner turret (baszta) topped by a dome and spire, richly articulated façades with projecting bay windows, decorative gables, and stucco ornamentation. The roofs are dual-pitched and mansard, while the interiors preserve vaulted cellars, a central staircase with metal balustrades and wooden handrails, parquet floors, and historic joinery.

The total cubic volume of the building is approximately 1680 m³, with a usable floor area of around 268 m² (Figure 2). The property is currently undergoing renovation and adaptation works under private ownership (Jasko Development Sp. z o.o.), with conservation authorities emphasizing the necessity of protective supervision.

It should be noted that the digital model developed for this case study was limited to the façades and the main entrance, since this was the explicit scope of the professional survey commission at that time [45]. In practice, the entrance portal and adjacent features were documented in high detail using terrestrial laser scanning, which enabled precise segmentation of richly ornamented architectural elements. The case study therefore reflects a real-world documentation scenario rather than a hypothetical academic exercise, grounding the research in practical constraints and actual heritage conservation workflows.

In terms of methodology, the case study builds upon previous work described in detail in [46], where semantic segmentation was applied to the tenement based on a laser-scanning survey. The authors proposed a method for dividing the object into semantic units corresponding to architectural, material, and functional features, following the IFC schema implemented in Graphisoft Archicad [47]. Special attention was given to distinguishing elements by stylistic taxonomy, construction techniques, and decorative detailing. This segmentation provided the foundation for subsequent enrichment of the model with domain-specific information, demonstrating that semantic structuring is a crucial step in operationalizing historic data within openBIM and bSDD environments.

In the present paper, this line of research is extended by integrating the resulting IFC model with the MIDAS classification system and the buildingSMART Data Dictionary platform, as well as by developing semantic model enrichment workflow by utilising open-source environment based on Blender and Bonsai. For this purpose, a combination of terrestrial laser scanning and archival documentation was used to capture the building’s geometry, which was then interpreted and modelled in accordance with the IFC schema.

3. Results

3.1. Semantic Structuring and Classification

To facilitate semantic enrichment, the classification system of choice is first structured in the form of a data dictionary. According to relevant standards [3,4], a data dictionary is a database that contains metadata or a centralised repository of information about data, such as meaning, relationships to other data, origin, usage, and format. In this case, the implementation of data dictionaries as a digital platform the bSDD is applied.

Because the MIDAS was not structured in a format directly compatible with the bSDD data model, a data transformation process was necessary. Since MIDAS was available as a spreadsheet, and the bSDD provides a spreadsheet template for uploading data dictionaries to the platform, the transformation was carried out in MS Excel. While most of the MIDAS content could be transferred directly into the corresponding bSDD columns, certain attributes required reformatting or the use of lookup functions to preserve all information. Additional input was also supplied, such as the mapping of selected classes to the IFC dictionary, which required expert consultation to ensure accurate correspondence between the two classification systems.

Furthermore, because MIDAS was not originally provided in a structure that could be utilised within BIM environments, unlike established systems such as UniClass [48], it has to be translated into a BIM-readable format. This transformation not only enhanced its compatibility with BIM workflows but also reduced the time required for importing and implementing historical classification data, following the example of established practices in the United Kingdom.

3.2. Tools and Technologies

The methodological framework of this research relies on a combination of BIM authoring tools, point cloud processing software, semantic enrichment environments, and reference ontologies to ensure accurate documentation and enhanced semantic interoperability of heritage assets.

Graphisoft Archicad was employed as the primary BIM authoring environment to produce the geometric model of the tenement house. Its openBIM support, particularly the capability to export models in the IFC standard, enabled interoperability with other software environments. The tool facilitated the creation of an as-is state model that formed the foundation for semantic enrichment.

TLS (terrestrial laser scanning) data were processed using FARO SCENE, which provided precise registration and filtering of point cloud datasets. Photogrammetric data, particularly useful for capturing ornamental and fine-scale details, were processed with RealityCapture. The combination of TLS and photogrammetry ensured both geometric accuracy and visual fidelity, delivering a high-resolution point cloud dataset of the heritage asset. PointCab Origins was used to structure, segment, and optimize the point cloud data for downstream BIM applications. By generating orthophotos, sections, and floor plans from the raw scans, the software provided a manageable dataset for import into BIM tools. This preprocessing stage ensured that the high-density point cloud could be effectively integrated into the modeling workflow without unnecessary computational overhead (Figure 3).

The bSDD platform was integrated into the workflow to facilitate referencing standardized semantic definitions. By transforming the MIDAS heritage classification system into bSDD-compliant dictionaries of components, materials, and monument types could be consistently defined and referenced.

Blender, extended with the Bonsai plugin, was used to enrich IFC-based models with semantic metadata. While capable of geometric modelling similar to Archicad, Blender was used as an environment for extending the semantic layer of the already created IFC model. BonsaiBIM, thanks to the bSDD API integration, allowed the assignment of property sets, classifications, and relationships that link the model to heritage-specific ontologies, ensuring compliance with semantic interoperability goals.

In the Italian context, several researchers have demonstrated how custom thesauri within the bSDD can address gaps in existing ontologies, particularly for heritage-specific attributes. Scandurra and Di Luggo [14] proposed a dedicated “State of Preservation” domain in bSDD to classify conditions of decay and alteration not covered by standard schemas. In subsequent work, Scandurra et al. [49] extended this approach to the decorative apparatus of historic buildings, structuring new classes and properties to represent ornamental and artistic elements absent from conventional BIM taxonomies. These examples highlight how bSDD enables the development of localized or project-specific vocabularies, ensuring that semantic enrichment can incorporate cultural features such as regional ornament types alongside international standards.

3.3. Semantic Enrichment of Architectural Components

We applied three MIDAS data (Figure 4) dictionaries relevant to the case study: Components, Materials, and Monument Types. Components include 1397 terms, such as Stoup or Voussoir, with their definitions, usually consisting of one sentence. Materials include 636 terms, from general to very specific, such as Baveno Granite or Bethersden Marble. The Monument Types classification describes a whole facility or a place and contains 7897 terms, like Abbey Gatehouse or Anti Submarine Searchlight Battery. Because MIDAS was not published by the owners of the classification in the bSDD before, after acquiring their permissions, we mapped and published the classification for public use in the bSDD, available at: [50].

MIDAS and IFC mapping were only partially done for the components dictionary, as for many specialist terms, it requires expert interpretation. We performed semi-automated mapping for common terms by finding keywords such as Beam, Column, etc., and matching them with their equivalents in the IFC schema. Some terms are semantically close but have different names in IFC; for example, MIDAS’ area and room should be mapped to IfcSpace. A sample of the mapping between the two systems is presented in Figure 5, which also shows the hierarchical structure of both. Not shown in the figure, subelements like Jib Door were also mapped to their closest generalisation in IFC (here: IfcDoor).

Once published the classification systems instantly become accessible in all integrated software through bSDD API. In our case study, the initial model already contains objects classified using the foundation classes (IFC).

For example, the element highlighted in Figure 6 and Figure 7 is not a generic object (IfcBuildingElementProxy) but the element of class IfcBeam of a predefined IFC type CORNICE. The IFC classification stops at a particular information depth, but the IFC schema allows extending data with further semantic meaning by assigning classification references. In the case of the element from Figure 6 and Figure 7, after launching the dedicated module in Bonsai and typing ‘cornice’ in the search bar or using the ‘Filter Active IFC Class’ option, the application provides a list of terms from MIDAS related to cornices. In the case study, it was decided to classify the object as an ‘Angle Modillion’ [51].

The HBIM model of the tenement house in Tarnów with its layered semantic structure was successfully integrated in the model. Architectural elements were categorized not only by geometric typology (e.g., pilaster, cornice, bay window) but also by stylistic period and decorative function. The use of the MIDAS classification system, aligned with the bSDD, allowed for the standardized naming and encoding of elements, which were then extended with custom vocabularies to reflect local stylistic nuances.

Each element in the model was enriched with:

• IFC entity and type when applicable
• Relevant MIDAS classification reference when applicable:
• Component classification (one of 1397 available classes in the MIDAS Component ontology)
• Construction material (one of 636 available in MIDAS Materials ontology)
• Monument class (one of 7897 available in MIDAS Monument Type ontology)
• Related Property Sets when applicable
• Decorative style (e.g., Neo-Baroque, Secessionist)
• Construction technique
• Condition and historical relevance
• Data source and confidence level
• Historical phase (where identifiable)
• Confidence level in the attribution

The Bonsai plugin for Blender enabled the dynamic assignment and editing of semantic properties directly within the design software, bridging the gap between modeling and classification referencing. All within an open-source, publicly available tools.

3.4. Interoperability and Export

The enriched model was successfully exported as an IFC file with semantic data intact and tested across various openBIM-compatible viewers and platforms. The inclusion of bSDD-linked properties facilitated cross-platform data exchange, with consistent display of metadata across other openBIM tools.

This interoperability validates the viability of using bSDD-enhanced IFC models for heritage documentation and potential future use in national heritage repositories or conservation planning databases.

4. Discussion

The proof-of-concept demonstrates that embedding a heritage thesaurus in the bSDD data dictionary service is more than a technical exercise: it reframes HBIM as an evidence-rich, queryable knowledge graph. First, modelling information uncertainty explicitly, through confidence scores, temporal ranges and provenance flags, transforms ambiguous archival facts into actionable metadata. Instead of hiding doubts in free-text notes, the workflow surfaces them so that conservation officers can filter those by risk or evidential strength, and future researchers can revisit earlier attributions as new data emerge. This aligns HBIM with ongoing work that demand audit trails for every modelling claim.

Second, by linking MIDAS classes to resolvable URIs in the bSDD, each BIM element becomes a doorway to external knowledge systems, national monument registers, museum object graphs, and potentially also Linked Data. Such connections enable semantic queries that cut across institutional silos.

Third, URI persistence guards against vocabulary obsolescence. Even if MIDAS were superseded, the RDF entity for terms like “Angle Modillion” would still reference to its definition, retaining meaning for humans and machines that encounter the IFC decades later. Of course, that is if the bSDD service continues to exist. The study therefore answers a long-standing preservation paradox: How can we future proof a digital twin when both software and classification systems evolve faster than heritage decay? Embedding canonical identifiers at authoring time provides a viable path.

Fourth, the case shows genuine scalability but also surfaces the bottlenecks. While façade-level enrichment was efficient, modelling an entire historic district would require policies, mapping tools and crowd-assisted validation. Automated NLP or computer-vision tools—not new in the context of BIM—could pre-classify common elements, reserving expert time for ambiguous features such as detailing. Developing and open sourcing such assistive tooling would benefit the international heritage community.

Fifth, the pilot exposes difficulty of ontology mapping. National thesauri (e.g., Polish “Krajowa Ewidencja Zabytków” lists) encode jurisdiction-specific legal terms that have no MIDAS equivalent, while MIDAS contains detailed typologies, most absent from IFC. Bridging these gaps will need formal alignment strategies, such as SKOS mapping tables, OWL equivalence axioms or at minimum crosswalk spreadsheets maintained by heritage agencies. The authors note that partial manual mapping already consumed significant effort; sustained funding and clear governance are essential if the approach is to mature beyond research prototypes.

Missing from the current discussion, however, are socio-economic and ethical dimensions. Semantic HBIM can democratise heritage by powering AR experiences for tourists or participatory planning consultations, but only if data licences remain genuinely open and interfaces are accessible to non-experts. Conversely, richer digital twins may expose sensitive information—structural weaknesses and security layouts—that could aid vandalism. A risk-benefit framework and clear data-sharing tiers should accompany technical rollouts.

Furthermore, the long-term sustainability of enriched HBIM models depends on institutional support, ongoing maintenance of vocabularies, and the adoption of FAIR (Findable, Accessible, Interoperable, Reusable) principles [52]. Long-term stewardship must consider not just file format longevity (addressed via IFC) but organisational continuity. Who maintains the classification alignments when project grants end? Who funds the data dictionary database? Embedding the workflow into statutory reporting requirements, similar to how Historic England mandates MIDAS reports, could guarantee resourcing while incentivising practitioners to contribute mappings upstream. Such policy integration will decide whether semantically enriched HBIM remains a research curiosity or becomes standard practice in heritage management.

5. Conclusions

This study demonstrates that the HBIM can and should be integrated with rich heritage classification systems such as MIDAS. Thanks to publicly available and interconnected and standardized data dictionaries, the data can be structured, retrieved and interpreted in a consistent and unambiguous way. Platforms such as the bSDD enable smooth integration of the native BIM tools and data dictionaries, facilitating easy data creation.

Application of MIDAS ontology meaningfully enhanced semantic interoperability and data reusability in the demonstrated example of architectural heritage documentation. Using the example of the Hackbeil-designed tenement in Tarnów, the research illustrates how HBIM can function as more than a geometric container, but as a semantic environment for recording, interpreting, and communicating cultural value.

Furthermore, the use of openBIM standard—IFC—ensures data will be retrievable and readable decades from the moment of creation [53]. Geometrical modelling is no more important than proper classification and structuring of the semantic information, as that is what is possible to query, retrieve and interpret by both humans and machines. The approach outlined here ensures that heritage assets are not only visually documented but also contextually and narratively encoded, enabling conservation efforts grounded in both precision and interpretive richness. It addresses a critical gap in heritage digitalization by merging openBIM standard schemas with domain-specific ontologies, opening a path toward standardized, data-driven heritage management practices for both existing buildings and scientific reconstructions [54].

Future work should aim at further integration with existing institutional databases and explore how predictive analytics might benefit from this semantic structuring to model degradation processes, maintenance cycles, or restoration scenarios. Additionally, studies should look into the process of automating the tedious information modelling process, by harnessing the AI and in particular computer vision to label geometrical data [9,55] keeping in mind that the creation should be assessed and supervised by a specialist, i.e., an architect, conservation officer or architecture historian.