HBIM and Information Management for Knowledge and Conservation of Architectural Heritage: A Review

Abstract

This paper presents a comprehensive review of research on Historic Building Information Modeling (HBIM), focusing on its role as a tool for managing knowledge and supporting conservation practices of Architectural Heritage. While previous review articles and most research works have predominantly addressed geometric modeling—given its significant challenges in the context of historic buildings—this study places greater emphasis on the integration of non-geometric data within the BIM environment. A systematic search was conducted in the Scopus database to extract the 451 relevant publications analyzed in this review, covering the period from 2008 to mid-2024. A bibliometric analysis was first performed to identify trends in publication types, geographic distribution, research focuses, and software usage. The main body of the review then explores three core themes in the development of the information system: the definition of model entities, both semantic and geometric; the data enrichment phase, incorporating historical, diagnostic, monitoring and conservation-related information; and finally, data use and sharing, including on-site applications and interoperability. For each topic, the review highlights and discusses the principal approaches documented in the literature, critically evaluating the advantages and limitations of different information management methods with respect to the distinctive features of the building under analysis and the specific objectives of the information model.

1. Introduction

HBIM has emerged as a widely debated topic, with a growing body of literature and numerous review [1,2,3,4,5] studies dedicated to it. However, most of these works tend to focus on geometric modeling and spatial representation, often marginalizing the role of the information system that underpins the entire HBIM framework. This oversight is particularly critical, as the informational component is not merely an accessory but a foundational element in the digital management of historic buildings.

In fact, buildings embody the complexity of skilled craftsmanship and a variety of construction techniques, having evolved through successive architectural interventions, multiple building phases, and the natural aging of materials. These cumulative transformations are often difficult to measure, and the inherent complexity of masonry behavior further complicates their analysis. In recent years, significant advances in geometric surveying and structural analysis have led to the development of highly accurate methodologies and automated tools, which are now fundamental to the documentation and preservation of cultural heritage. These technologies play a central role in the broader knowledge-building processes that inform any heritage conservation effort. Nonetheless, despite the high levels of precision and automation achieved, the seamless transfer of data between different tools and platforms remains a persistent challenge. Determining reliable methods for translating and exchanging information without compromising data integrity continues to be an open and unresolved issue. Addressing this issue is essential for enabling HBIM to function as a truly interdisciplinary and information-driven environment for heritage conservation.

This represents a critical issue, as the information associated with historic buildings—and the methods for linking and transferring it across different disciplines—is essential when working with heritage architecture. The issue is particularly relevant and challenging in the case of historic masonry buildings, which, unlike modern structures, do not feature standardized construction characteristics or materials, nor are they easily modeled.

To this end, the present literature review has been conducted with the aim of providing an overview of the current state of the art regarding various methodologies employed by different researchers, with a particular focus on the development of the informational component of the HBIM system.

This information can be organized into recurring themes: ontologies and semantic classification; management of building knowledge (including historical evolution, surface and structural condition analysis, and the interpretation of test results); representation of the project and in situ application of the tool to support planned conservation practices; and interoperability with structural models.

In addition to a brief discussion on how modeling issues influence model manageability and its capacity to handle information, the review highlights the main recurring approaches encountered in the literature for each of these themes. These approaches are considered essential for representing the conservation process within a BIM environment. With this aim, in the following, Section 2 provides a methodological overview on how the systematic literature review was conducted, Section 3 presents the review results from a bibliometric point of view, while in Section 4 current trends, well established and innovative approaches developed in the main research areas are discussed in detail. Finally, Section 5 points out research gaps and future trends, trying to advance some conclusions.

2. Materials and Methods

The review has been performed according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) methodology [6,7] to ensure a systematic review and objective discovery of the articles. The first step involved defining the keywords for the article search, which was conducted in the Scopus database (Figure 1). The search focused on the title, keywords, and abstract to maximize discovery success, using the following query:

(hbim OR h-bim) OR ((historic* OR heritage) AND (“building information model*” OR bim ))

Wild characters were used to account for slight variations in terminology such as Model, modeling, Modeling or historic and historical.

The initial search yielded 2238 articles. These results were further refined by selecting only Conference papers, Articles, Book chapters, Reviews, which were at their final publication stage, written in English, and published between 2008 and July 2024. The starting year was selected to ensure the inclusion of the first publication defining the HBIM by Murphy et al. [8], dating back to 2009. This filtering process reduced the selection to 1955 articles. Next, an in-depth review of the titles and abstracts was conducted to exclude those irrelevant to the review scope (e.g., those with content entirely unrelated to Building Information Modeling, such as studies in other research fields like chemistry) and those for which the full text was not accessible. Moreover, since the aim of the review was to analyze the current state of using BIM methodology to manage the entire knowledge and intervention process for built heritage, the following exclusion criteria were adopted: (i) articles that discussed HBIM only as a future perspective; (ii) articles that mentioned HBIM merely as a processing tool without explaining its implementation; (iii) articles focusing strictly on BIM-GIS integration. Articles with abstracts that did not clearly indicate their relevance were retained for a full-text reading.

Following this process, a total of 451 articles were comprehensively reviewed and utilized for bibliometric and research trend analyses. Bibliometric analyses involved the following aspects:

• Integrated analysis of publication year and document type (conference paper, article, book chapter, review), to assess, quantitatively and qualitatively, the temporal evolution of the research (Section 3.1);
• Geographical distribution, to identify regions leading in research, based on the location of case studies or authors’ affiliations for methodological studies (Section 3.2);
• Software packages, distinguishing between commercial and open-source solutions, and highlighting where specific software or plug-ins were developed to increase functionalities of existing tools (Section 3.3).

The research trend analysis thoroughly examined the content of each article, identifying three main topics that represent key aspects of the typical BIM workflow. Notably, the three sections—though presented in reverse order—correspond to the components of the BIM acronym, underscoring that they collectively represent the fundamental components in the development of an information system.

• Model: Definition of entities: focuses on how BIM objects are identified and defined, both geometrically and semantically, including segmentation, classification, modeling, and ontology definition;
• Information: Data enrichment: involves information and data management, encompassing overall database design and the identification of information to be included in the model (such as decay, structural deformations and cracks, monitoring data, interventions, etc.);
• Return to Building: Data sharing and usage: highlights the purposes for which the HBIM model is developed and addresses data interoperability across different software and methods for querying and accessing data.

Figure 2 summarizes the main research items addressed in the review.

Each article was analyzed to identify the investigated topics. The same categorization of topics was subsequently used to structure the analysis and discuss the state of the art, as presented in Section 4.

3. Bibliometric and Statistical Analysis

This section presents the results of the statistical analyses conducted on all 451 articles reviewed, to provide an overview of the general research trends.

3.1. Publication Year and Document Type

Figure 3 shows the results of the temporal analysis of the articles selected in the review. Already in 2008, Arayici [9] proposed the application of BIM to existing structures, but it was in 2009 that Murphy et al. [8] introduced the concept and acronym HBIM, understood as Historic Building Information Modeling. In the early years, contributions specifically dedicated to HBIM were limited, but they progressively increased starting in 2013, with a significant surge beginning in 2019. Currently, interest in the topic remains high, as evidenced by the large number of publications in 2022 and 2023. However, 2024 data is incomplete since the review only included articles published up to July, so the lower number of publications cannot be considered statistically significant.

The noticeable decrease in 2020, instead, can be attributed to the COVID-19 pandemic, which significantly curtailed scientific activity worldwide. Another noteworthy observation is the shift in the ratio of Journal articles to Conference papers. Until 2019, conference papers predominated, constituting between 66% and 100% of the annual publications. However, from 2020 onwards, there has been a reversal, with the majority of publications being Journal articles. This shift suggests an increased maturity and quality in the research being produced.

3.2. Geographical Distribution

A comprehensive geographical analysis of the publications was carried out, independently assessing both the location of the case studies and the institutional affiliations of the authors. This dual approach aims to highlight, on one hand, the countries where cultural heritage assets or historically significant sites are more frequently studied—suggesting both the presence of valuable heritage and a propensity to invest in HBIM methodologies—and, on the other hand, the countries most active in scientific research on this topic.

The spatial distribution of case studies was analyzed by counting the number of studies conducted in each country. Figure 4 presents a false-color map illustrating the number of publications focused on case studies located within national boundaries. Italy emerges as the most represented country, accounting for nearly 54% of all case studies, followed by Spain (9%), China (4%), Portugal (2.7%), and Saudi Arabia (2.4%).

With regard to author affiliations, international co-authorship was also taken into account by proportionally distributing each publication’s contribution among the countries of the affiliated authors, in order to provide a more accurate representation of collaborative research efforts. This analysis reveals a similar pattern: Italy accounts for approximately 55% of total contributions, followed by Spain (8.3%), China (3.5%), the United Kingdom (3%), France (2.6%), and Ireland (2.4%). The resulting distribution is visualized in Figure 5 using proportional circles, where the diameter of each circle reflects the relative concentration of author affiliations in each country.

Italy’s leading position can be partly justified by the extensive presence of architectural heritage in the country, which is only partially highlighted by the UNESCO census [10], but, more importantly, by the long-standing tradition of studies in the field of conservation and restoration [11].

3.3. Software Packages

Another aspect of the analysis examined the types of software used (Figure 6). While it was not always possible to definitively identify the software or tools utilized for creating HBIM in every contribution, it was considered valuable to present the findings to offer readers insights from this perspective as well. Specifically focusing on BIM authoring software, the analysis highlights a clear dominance of commercial solutions, with Autodesk Revit leading the field.

It is also worth noting that, although to a limited extent (2% of the articles), some research proposes the development and use of custom-built platforms designed to either replace or complement commercial solutions, aiming to overcome their limitations.

4. HBIM Main Research Topics

Figure 7 illustrates the most prominent topics addressed by the 451 publications analyzed in this review. The key findings related to these themes are discussed and elaborated upon in the subsequent paragraphs of this section, following the structural framework established in Section 2 (Figure 2). The data depicted in Figure 7 indicate the proportional attention allocated by researchers to each topic. It should be noted that many publications encompass multiple aspects; consequently, some publications are counted multiple times, leading to a total percentage exceeding 100%. To improve the clarity and comprehensibility of this densely detailed section, a summary table has been provided at the end of each thematic subsection, summarizing its main points.

In relation to the 451 publications analyzed, the most representative studies for each thematic area are explicitly cited and discussed in the subsequent sections, with particular emphasis on those that have introduced or advanced innovative approaches.

4.1. Model: Definition of Entities

Defining the entities that constitute the model is the initial step in developing an information system. Entities need to be defined both geometrically and semantically. Associating semantic meaning to entities involves classifying them, preferably according to recognized and shared standards. Giving geometric definition means three-dimensionally modeling the entities choosing a modeling strategy, according to the desired accuracy and the model’s purpose.

4.1.1. Semantic Definition

In BIM workflows, each building component must be semantically classified, meaning it should have a defined semantic significance. This classification ensures uniformity and standardization of processes while providing a shared foundation for cost estimation. Over time, international classification standards for building components, such as those outlined in ISO 12006-2:2015 [12], have been progressively introduced. Among the most prominent are the American OmniClass and the British UniClass which are also implemented in the main BIM authoring software. In addition to these classification systems, the international standard IFC (Industry Foundation Classes), approved as ISO 16739-1 [13] standard, plays a dual role: it is not only a digital description of the built environment (as other classification standards) but also an open format for data exchange, facilitating interoperability across diverse software platforms, including BIM authoring tools, structural analysis programs, and energy simulation software. However, all these standards were originally conceived for new buildings, making their application to historic structures not straightforward.

For heritage buildings, existing classifications lack dedicated classes for mapping typical elements of the historic built environment. Although the IFC schema is gradually being expanded to include new classes, much of the focus has been on the infrastructural and urban domains. As a result, as stated in [14], despite detailed modeling efforts, from a semantic point of view historic buildings are often reduced to the same logic as new constructions, failing to capture specific architectural components like arches and vaults, or to manage properties such as material decay or structural cracks.

Therefore, one of the main established research strands on HBIM is related to the semantic classification of historic buildings’ components and, more generally, to the search for a common vocabulary that, also through the definition of specific ontologies [15,16,17,18,19], provides a conceptual representation not only of physical elements, but also of properties, relationships, actions part of the conservation process.

Ontologies can be defined as formal and hierarchical representations of entities, properties and relations between the concepts [20]. A widely used ontology in cultural heritage is CIDOC-CRM. However, it was originally developed for movable assets and is more closely aligned with semantic web logic than with information models. Consequently, significant research has been directed toward developing new ontologies or introducing new classes tailored to the needs of architectural heritage.

To this aim, two main approaches for semantic definition in HBIM can be identified (Table 1): Case-study specific (CSS) and General-purpose (GP).

The former—CSS—focuses on creating tailored classifications and ontologies that are highly consistent with a specific case study [16] or building component. For example, an ontology for vaulted structure is developed in [15]. It describes subcomponents, masonry arrangement and texture, while aiming to relating these specifics to a more general ‘foundation’ ontology. CSS allows for precise alignment with the reality of a project, making it particularly suitable for detailed investigations. However its limited scope often results in poor interoperability, which undermines the primary goal of ontologies.

The majority of studies adopt the GP approach, which aims to extend existing ontologies, creating shared vocabularies and representations that can serve as references across the field. Examples include CIDOC-CRM extensions, such as the Conservation Process Model (CPM) [17,21,22] or the HBO ontology [23]. CPM introduces new classes to represent artifacts, investigation processes, actors, and both past and future lifecycles. Additionally, surface decay, investigation and conservation activities are defined and related to the artifact. Further studies are built on CPM framework.

In [18] the authors attempt to align the hierarchical structure of the ontology (class, sub-class, entity) with the structure used in modeling software (family, type, instance). Similarly, ref. [24] explores creating a correspondence between HBIM elements and ontology semantics. The HBO ontology conceptualizes various types of interventions on historic buildings (including maintenance, restoration and retrofitting) and integrates them with the IFC standard. Other expansions of CIDOC target the conceptualization of traditional heritage buildings of a specific region [25,26], or focus on distinct building typologies, such as elements of religious [27] or fortified architecture [28], the latter combining existing CIDOC, IFC and Getty classifications. In [29] the Getty Vocabulary serves as foundation to define a shared nomenclature system or glossary.

Some researchers have proposed expanding the IFC schema. In [30], IFC-OWL (Ontology Web Language) is used to map IFC extensible properties and a custom ontology is proposed to model additional domain data. In [19,31,32], the IFC schema is expanded with a particular focus on archeology, adopting the open-source tool FreeCAD, to overcome the limitations of commercial software. The open-source library IfcOpenShell is used to develop objects conceived as a container for customized properties. New IFC roles, such as ifcVault, ifcArch, and ifcUSM are introduced modifying the Python script of each entity.

Table 1. Main approaches for ontology development.

Approach	Reference Scheme	Related Papers	Pros	Cons
CSS Case study specific (development from scratch)	none	[15,16]	Major adherence to the real building and to the project’s needs	Poor interoperability
GP General purpose (extension of existing ontologies and standards)	CIDOC-CRM	[17,18,21,22,24,25]	Better interoperability. Possibility to be further developed by others	Starting point distant from the needs of historic buildings. Fixed scheme. Top-down standards
	IFC	[19,30,31,32]
	mixed	[23,28,29]

Table 1. Main approaches for ontology development.

Approach	Reference Scheme	Related Papers	Pros	Cons
CSS Case study specific (development from scratch)	none	[15,16]	Major adherence to the real building and to the project’s needs	Poor interoperability
GP General purpose (extension of existing ontologies and standards)	CIDOC-CRM	[17,18,21,22,24,25]	Better interoperability. Possibility to be further developed by others	Starting point distant from the needs of historic buildings. Fixed scheme. Top-down standards
	IFC	[19,30,31,32]
	mixed	[23,28,29]

4.1.2. Geometric Definition

3D modeling of built heritage is still a challenging task, primarily due to the lack of standardization in historic building components and their inherent irregularity. These factors make parametric object-oriented modeling (that is native BIM approach) not readily applicable and require careful consideration of both the most appropriate modeling techniques and the desired level of accuracy in representing the building irregularities. Numerous studies addressed this topic (see Figure 7) and outlined the main approaches for geometric modeling along with the technical possibilities offered by various software tools [33]. Additionally, previous review and articles [1,2,3,34] examined the state of the art regarding geometric data collection, 3D modeling (with particular attention to the development of standardized object libraries) and model validation in terms of metric accuracy.

Although extensively covered in the literature, this paper revisits the topic to provide a comprehensive overview. Additionally, a discussion about the current viewpoints concerning the appropriate level of accuracy and the strategies employed to prevent the loss of geometric information is presented. Existing modeling approaches are introduced highlighting the different levels of geometric coherence, automation, reproducibility, semantic recognizability of objects and data enrichment (Table 2).

The first modeling approach is Reverse modeling (RM), which is the quickest and most accurate method for generating surfaces from point clouds, themselves obtained from laser scanning or photogrammetric surveys [2]. RM is based on software like Geomagic or MeshLab, which employ automatic tools and algorithms to convert the point cloud into mesh surfaces. Meshes can represent complex shapes that cannot be described by simple geometric rules, and approximate curved surfaces with a quality depending on the size of the faces. RM is generally used for complex and unique elements such as decorative features [35], where high fidelity to the real shape is crucial, but replicability is not. This approach excels in automation and realism [36,37,38,39], but achieving semantic recognition of objects for subsequent information enrichment requires prior segmentation of the point cloud. Segmentation, if performed manually, is time-consuming, so the research is focusing on automating this operation [40]

The second approach is Direct modeling (DM). The model is created starting from geometric primitives (usually profiles obtained from point cloud slices or 2D drawings from previous surveys) and working with modeling operators. The most known modeling software are Rhinoceros, 3D Studio Max, Maya, but some tools for DM are also implemented in the main BIM authoring software. The use of pure modeling software is often encouraged by the existence of tools that ensure effective interoperability with BIM authoring platforms, such as Rhino.Inside.Revit or the Grasshopper–Archicad Live Connection. DM is used, for example, in [41,42,43,44], and, in general, when only 2D drawings are available as a starting point. It is also used when simplified models are needed for interoperability issues, or when it is necessary to model unique building components. While DM allows for manageable models, it is time-consuming, challenging to automate and results in unique elements that cannot be reused for similar parts.

The third approach is Parametric modeling (PM). This approach uses a finite number of parameters (e.g., length, height, thickness, radius) to define object geometry, enabling adaptability and reuse across similar objects. PM creates ‘intelligent’ models that streamline the modeling phase, by governing rules for geometry, topological relationships and constraints. This technique, which closely resembles current practices for new buildings, requires segmenting data into semantically defined objects and meticulously designing the modeling process to accurately understand and reproduce the objects’ geometric behavior. It is particularly effective for recurring construction elements, such as windows, moldings, frames and decorative features.

PM can be divided into two sub-approaches: ‘Generative PM’ and ‘Object-oriented PM’. The former approach can use NURBS (Non-Uniform Rational B-Splines), mathematical representations of geometry that accurately define shapes ranging from simple lines and curves to complex organic 3D forms. NURBS excel at describing curved surfaces, with flexibility for editing through knots and control points. Compared to the meshes, NURBS reduces also the memory usage and offer flexible and accurate design capabilities. Generative modeling leverages visual programming tools like Grasshopper [45] or Dynamo to flexibly create complex shapes, which can also be implemented in object libraries [40,46].

Object-oriented PM, instead, relies on the use or development of object libraries (as discussed in more detail in the following section). These objects are semantically defined, related to other objects through specific rules and topological constraints, and are associated with a database containing all the information and parameters necessary for their definition [47,48,49]. This approach is commonly used in BIM authoring software as it leverages the potential and the logic of BIM.

Literature frequently highlights the use of different modeling approaches within the same BIM model as the most effective solution, choosing the most appropriate method depending on the complexity of the object and whether it is unique or part of a series [50,51,52,53]. In [54], the advantages and limitations of such integration, including potential information loss and interoperability challenges, are explored. The architectural style or the age of a building also influences the choice of the modeling approach: modern or industrial heritage [55], with standardized elements, aligns better with BIM logic and encounters fewer modeling and interoperability issues; architectural styles adhering to treatises are often more adaptable to PM; while vernacular architecture, ruins, or archeological sites, with their high geometric irregularity, pose significant challenges for creating accurate yet manageable models.

In addition to identifying the most suitable modeling methodology, another widely debated aspect in the literature is the required level of detail and accuracy of the model with respect to the real object. On the one hand, some studies aim to achieve a high level of correspondence between the 3D model and the survey data, with the goal of faithfully documenting the object and enabling the extraction of all necessary outputs from the 3D model. The focus is extensively on modeling methodologies, integrating various tools and developing new ones to address specific needs in terms of geometric accuracy [56,57].

On the other hand, some studies accept a higher degree of geometric approximation, seeking an appropriate compromise between the level of detail and accuracy required for the model’s purpose, the simplicity of the modeling process, and effective information management. In [58], this concept is achieved by using ‘federated models’. Specifically, the study is based on the premise that a geometrically simplified model can sometimes enhance information management and model manageability. Consequently, two ‘federated’ models are developed. The first, derived from the point cloud and highly accurate from a geometric point of view, aimed to incorporate all the complexity of the geometry acquired through instrumental surveys. The second model, more schematic, is conceptualized as a three-dimensional repository of non-geometric data, meticulously designed with maximum ‘granularity’ (modeling individual and regularized stone ashlars) to associate information with the smallest identified geometric element. Alternatively, as presented in [59], simplified modeling can be supplemented by associating a false color map representing geometric accuracy.

In [60], two different types of models are discussed. The first one, referred to as a ‘digital surrogate’ (or black box), metrically accurate, is an uncritical representation. The other is an ‘interpretative model’ (white box), simplified and ideal, providing an insightful interpretation of the asset. However, while the reliability of the first model lies in its metric accuracy (the trust is in the model itself), in the second case reliability is contingent upon the correctness of the process (the trust is in the creator).

Table 2. Main approaches for modeling historic buildings, ordered according to an increasing level of ‘interpretation’ of the object and a decreasing level of automatization of the process.

Approach	Sub-Approach	Related Papers	Good in Case of	Pros	Cons
RM Reverse modeling	Mesh	[35,36,37,38,39]	Very complex shapes and need of accuracy	Extreme accuracy. Automatization of the process	Heavy model Poor interoperability and semantic significance
DM Direct modeling	From point cloud	[42,44]	Non repetitive building	Good accuracy	Time consuming. Need of good hardware. No possibility to adapt to similar objects
	From 2D drawings	[41]	Necessity of a simplified modeling for interoperability issues	Fast process. Light and manageable model	High simplification
PM Parametric modeling	Generative	[40,45,46]	Repetitive objects with complex shape	Fast adaptability to similar objects Makes the best of BIM logic	Time consuming for the first object
	Object oriented (in BIM software)	[47,48,49]	Repetitive objects
Mixed modeling	-	[50,51,52,53]	Buildings with parts having different features	Accuracy, time spent, complexity that fit the needs of the project	Necessity to use various software

Table 2. Main approaches for modeling historic buildings, ordered according to an increasing level of ‘interpretation’ of the object and a decreasing level of automatization of the process.

Approach	Sub-Approach	Related Papers	Good in Case of	Pros	Cons
RM Reverse modeling	Mesh	[35,36,37,38,39]	Very complex shapes and need of accuracy	Extreme accuracy. Automatization of the process	Heavy model Poor interoperability and semantic significance
DM Direct modeling	From point cloud	[42,44]	Non repetitive building	Good accuracy	Time consuming. Need of good hardware. No possibility to adapt to similar objects
	From 2D drawings	[41]	Necessity of a simplified modeling for interoperability issues	Fast process. Light and manageable model	High simplification
PM Parametric modeling	Generative	[40,45,46]	Repetitive objects with complex shape	Fast adaptability to similar objects Makes the best of BIM logic	Time consuming for the first object
	Object oriented (in BIM software)	[47,48,49]	Repetitive objects
Mixed modeling	-	[50,51,52,53]	Buildings with parts having different features	Accuracy, time spent, complexity that fit the needs of the project	Necessity to use various software

4.1.3. Development of Libraries of Parametric Objects for Historic Buildings

With regard to the definition of the entities of the model, the development of libraries of parametric components for historic buildings aims to create objects with both semantic and geometric significance, similar to those used for standardized components in new constructions. The acronym HBIM originated from the potential to parametrize construction and decorative elements of historic buildings [8,61]. This approach is sometimes supported by information from historical treatises [62] or manuals and is particularly applicable to classical or neoclassical buildings, where decorative elements are more ‘standardized’ [63]. For example, treatises by Leon Battista Alberti, Francesco Di Giorgio Martini or Guarino Guarini have guided the parametric modeling of vaults [16], and geometric rules for masonry domes have been used to build ideal parametric objects of real domes [64]. Other applications involve parametrizing classical architecture orders [49,65,66,67,68] and moldings profiles [40], mainly using algorithmic modeling and VPL (Visual Programming Language) to define a complete parameterization of the entire range of possible variables identified.

The developments of parametric libraries is also addressed to make available components typical of specific architectural styles or construction typologies, such as Arab-Norman [50], Catalan [48], Arabic [69] or Chinese [70] architecture. Some initiatives aim to create open-source and shared object libraries to be used in similar buildings [57]. Additionally, studies like [66,71] compare the capabilities offered by different software in supporting library development.

Research groups like [57,71] are evaluating the possibility of including library objects for some elements of the historic built environment (particularly vaults and wooden-beam floors) so as to convey, through BIM software, knowledge about the construction features of historic buildings, thus facilitating both surveying and maintenance practices. This approach requires a preliminary reflection about the semantic breakdown of the elements in search of a common vocabulary, as in [72], where the Getty Thesaurus is used as reference.

Some software houses are starting to develop specific tools and objects for historic building components. For example, the tools available in Edificius software [73] allow a geometrically simplified modeling of wooden floors and vaults with different geometries. From an informative point of view, no dedicated properties have been defined to assist users during data entry; however, such properties can be introduced by users in order to meet specific needs.

4.2. Information: Data Enrichment

In this section, the information typically associated with elements in the HBIM model are examined, along with the strategies used to manage information contents proper of historical buildings and conservation processes. The presentation is organized as follows: initially, all the information used to characterize the element itself is discussed (Section 4.2.1), including historical and descriptive data, materials and mechanical characteristics, and information derived from diagnostic investigations. Next, attention is given to the state of conservation (Section 4.2.2), encompassing both surface degradation and damage that may affect the element’s overall integrity and structural behavior. Finally, the management of the time dimension is explored (Section 4.2.3), considering various aspects such as the documentation of construction phases, the integration of monitoring data from sensors, and the recording of conservation activities undertaken on the structure.

4.2.1. Characterization of the Element

Historical and descriptive information can be categorized into two main types: textual data and document-based data. Textual data includes specific details, such as dates related to construction, transformation, or prior restoration works, as well as information about owners, authors, or notable events connected to the building. These data are generally managed as attributes or properties associated with individual elements in the model, allowing for easy querying and retrieval.

Document-based data, on the other hand, consists of materials such as drawings, photographs, reference books, or archival records that provide additional context or historical evidence. Unlike textual data, these documents are managed through links to external files, which must be systematically organized for accessibility. Since such files cannot be directly queried, their content must often be translated into textual or numerical attributes if it is to be fully integrated into the model’s database. Some research has focused on how to catalog documents (bibliographic or archival, graphic and textual, etc.) in order to structure a database that can be easily queried and used for designing conservation interventions [74]. Others are focusing on designing the most appropriate environment, also thanks to VR platforms [75], for the visualization of historical data, with the aim of making BIM a tool for sharing knowledge about architectural assets.

When structured effectively, the HBIM model functions as a spatially referenced digital archive [41,76], in which documents and textual information are easily accessible because directly related to specific portions of the building or construction element to which they refer. This facilitates the comprehension of the history of the building, particularly in case of complex architectures.

Materials, mechanical and thermal characteristics are assigned to model’s elements using properties, similar to the approach for historical information. However, the level of geometric detail and even the modeling strategy can determine the granularity at which these properties are applied. For models with low geometric detail or those created through direct or reverse modeling, material information, along with related mechanical and thermal properties, is typically assigned descriptively to the element as a whole.

In contrast, higher levels of detail, particularly when parametric modeling strategies are employed, enable the stratigraphy of elements—such as walls or floors—to be explicitly modeled using specialized tools in BIM software [77]. Each layer can then be assigned specific material, mechanical, and thermal properties, including transmittance values, which are particularly useful in research focusing on building energy performance [78,79]. Similarly, masonry texture information can be added, either as custom properties [15,80], or by modeling each stone or brick unit [81].

Mechanical properties are often already preset in BIM authoring software through material libraries. However, ongoing research aims to refine these properties and address the interactions with elements’ construction features. For example, in [82] mechanical characteristics and other parameters for structural assessment are added using custom properties tailored to specific construction entities (walls, slabs, etc.). Some properties’ values are automatically calculated using mathematical expressions in Archicad, establishing correlations between different parameters. In [41,83] the mechanical properties of wooden structures are assigned on the basis of visual inspections, following the Italian UNI 11119:2004 [84] standard, which classifies timber based on defects and damage observed. In [85,86] structural properties are derived from diagnostic analysis performed on construction elements. Further research aims to reach the automatic calculation of indices like the Masonry Quality Index (MQI), which evaluates the structural properties of masonry based on specific attributes—such as quality of mortar, presence of diatons, alignment of vertical joints, etc. In [87] the MQI classification is integrated into BIM using Visual Programming Language (VPL), enabling the automatic calculation of the MQI categories (A,B,C) based on the values assigned to both construction parameters. These classifications are visually represented, with masonry elements adopting distinct colors corresponding to their MQI category.

To characterize stratigraphy and mechanical properties of elements, diagnostic investigations are fundamental and integrating the results of these analyses into HBIM models is a common research topic. By spatially linking diagnostic data directly to the tested building element, HBIM enhances accessibility compared to traditional reports, which typically reference tests to structural elements through codes.

Different approaches can be identified to manage data from investigations (Table 3). As will be discussed later, similar approaches are also employed to manage other types of information, such as surface decay and damage. In the former approach (hereinafter referred to as Properties of Elements—PoE), the results of the tests can be incorporated as properties of the investigated construction elements. This approach is straightforward for numeric data (e.g., mechanical properties or quantitative test results) and allows correlation with other parameters, such as material characteristics. However, it does not support precise localization of tests, as properties are applied to entire entities rather than specific parts.

Following the PoE approach, in [41,85,86,88] quantitative test results are recorded as information, while images and reports are linked to the properties of walls and pillars as external documents. In [82] the mechanical properties assigned to masonry walls are derived from in situ double flat jack tests, whose reports are linked to the wall objects. In [83], graphs deriving from resistance drilling tests are linked to the wooden beam objects and used to evaluate the effectively resistant section. In [89], both numeric and raster data from different diagnostic tests are included among the structural entities’ parameters using the IFC format, though not all data fit the schema.

The second method (Symbolic Modeled Objects—SMO) uses custom markers to represent the investigations symbolically. These objects include properties such as the type of test, execution date, instrumentation, quantitative results, interpretation summaries, and links to images or reports. This method enables precise spatial location of investigations and facilitates comparisons between tests on the same element. Nevertheless, it complicates the semantic connection between the investigation and the construction element since they are separate objects. Concerning the SMO approach, examples can be found in [90,91,92], where symbolic shapes (e.g., cubes or spheres) are used to punctually represent the different types of tests regardless its actual spatial extent.

The choice of the most appropriate method depends on the objective of the model, but, above all, on the type of test and its spatial extent. Some tests, such as endoscopies, flat jacks, penetrometry, hygrometric or some sonic tests, are punctual and their precise localization is essential. In contrast, surface-wide investigations, such as thermography or radar scans, provide raster outputs (e.g., thermographic images) that can be linked to extended surfaces, whether that of the construction element or an attached one.

Instead of being symbolic markers, the objects representing the tests can faithfully reproduce the investigated area: in this case, the approach is defined as Realistic Modeled Objects (RMO). In [47], radar tests are modeled as surfaces with the same spatial extension of the investigation and are placed in the exact point where the it took place. The same is performed in [93] where surfaces representing radargrams are positioned at different depths, producing a kind of 3D tomography. In [37], thanks to a custom platform implemented by the research group, the representation method varies according to the type of test: punctual tests use symbolic objects, while large-scale investigations are mapped as additional layer of the surface. In [94], thermographic images and GPR (Ground Penetrating Radar) results are projected onto the wall objects, not as added surface but as a new ‘material’ whose properties correspond to the diagnostic data. This visualization aims at having an immediate correlation in the three-dimensional model between thermographic investigations and surface degradation.

Finally, researchers experimented ontologies to better conceptualize diagnostic activities. For instance, the ontology developed in [17] defines the ‘Investigation process’ domain. Customized entities and relationships are created to relate the test to the surface (here the focus is architectural surfaces conservation) on which the investigation was carried out.

Table 3. Main approaches for representation of investigation results.

Approach	Related Papers	Good in Case of	Pros	Cons
PoE Property of Elements	[41,82,85,86,88,89]	Investigations conducted on large surfaces	Correlations between test’s result and mechanical properties of structural element.	No precise location of the test (unless written indication among the properties).
SMO Symbolic Modeled Objects	[90,91,92]	Punctual investigations	Precise location of the test.	Poor interoperability and semantic connection with the structural element.
RMO Realistic Modeled Objects (geometry that reproduces the tested area)	[47,93]	Tests on large areas	Possibility to map the precise surface on which the test is conducted. Reconstruction of 3D tomography in case of GPR tests.	Poor interoperability and semantic connection with the structural element.
Mixed approach	[37]	Representation strategy depending on the type of investigation.	Precise location when needed.	Poor semantic significance since different strategies are adopted for same topic.
Projection on the surface	[94]	Tests that provide images of the surface (thermography, GRP, etc.)	Possibility to compare with decay mapping and orthophotos.	Difficulties in projecting on curved surfaces. Need to use also other strategies for information enrichment.

Table 3. Main approaches for representation of investigation results.

Approach	Related Papers	Good in Case of	Pros	Cons
PoE Property of Elements	[41,82,85,86,88,89]	Investigations conducted on large surfaces	Correlations between test’s result and mechanical properties of structural element.	No precise location of the test (unless written indication among the properties).
SMO Symbolic Modeled Objects	[90,91,92]	Punctual investigations	Precise location of the test.	Poor interoperability and semantic connection with the structural element.
RMO Realistic Modeled Objects (geometry that reproduces the tested area)	[47,93]	Tests on large areas	Possibility to map the precise surface on which the test is conducted. Reconstruction of 3D tomography in case of GPR tests.	Poor interoperability and semantic connection with the structural element.
Mixed approach	[37]	Representation strategy depending on the type of investigation.	Precise location when needed.	Poor semantic significance since different strategies are adopted for same topic.
Projection on the surface	[94]	Tests that provide images of the surface (thermography, GRP, etc.)	Possibility to compare with decay mapping and orthophotos.	Difficulties in projecting on curved surfaces. Need to use also other strategies for information enrichment.

4.2.2. State of Conservation

Mapping surface decay is a crucial step in any restoration project. It serves the following purposes: identifying the areas affected by decay, measuring the extent of the deterioration on the surfaces, and determining potential interventions that can be quantified, including their economic feasibility. Beyond these fundamental and traditional functions, it may also be beneficial to assess how surface degradation impacts the mechanical properties of the element in question, as well as the surrounding elements.

Surface decay is traditionally depicted on 2D drawings, such as elevations and sections, by overlaying hatches on the deteriorated areas, carefully outlined to match the exact contours of the decayed regions. A legend is used to correlate different types of degradation with specific hatch patterns. In 2D drawings the surface affected by degradation is automatically calculated.

Currently, various approaches have been developed in HBIM for mapping decay (Table 4).

The first method depicts degradations as a specific Property of the construction Element (PoE), as for diagnostic tests results. This method enables a direct correlation between degradation and the element, which facilitates query and thematic mapping of the model [88,95] and is essential for types of decay that impact the material’s properties. In [96] this method is further developed, by utilizing the relational database features connected to the 3D model to add additional information to the degradation, such as severity, extension of the phenomenon or approximate localization. Decay information can be effectively used to evaluate intervention priorities [97]: for this reason, PoE is an appropriate approach in the case of planned conservation processes. This approach inevitably involves extending the degradation across the entire element [50], preventing the precise identification of the area affected by the decay. Indeed, PoE is more often used for degradations that affect small elements, such as stone ashlars [88,95] or wooden beams [41], maybe linking the spread of the phenomenon as a parameter. A similar approach is used in [98] for a synthetic reception of indices related to the conservation status, vulnerability and transformability of construction and spatial units, with the aim of defining a master plan for the building’s reuse.

A second approach, similarly to what highlighted for tests results, consists of attaching Realistic Modeled Objects (RMO) with a minimum thickness retracing precisely the decayed area, to the wall surface [42,43,99]. These objects can include hatches, ‘morphs’, ‘filled regions’ [100], and primarily ‘generic adaptive models’ (special objects with flexible geometry that can be adjusted to both flat and curved surfaces) depending on the software ‘lexicon’. This method enables the precise spatial localization of degradation on the object’s surface, facilitating thematic visualizations. However, adapting the overlaid object to complex and irregular shapes is not always simple. For this purpose, in [101] a specific grid object designed to fit irregular surfaces is developed. From a semantic perspective, objects representing degradation can be enhanced with various properties, such as a degradation severity or suggested interventions [102,103], and inspection dates [104]. However, the connection between decay object and construction element is challenging. Currently, this task is not easily performed in major BIM authoring software and requires either the use of external databases properly connected to the 3D model [105], or manual association of the element identifier with the degradation. An exception is the use of Revit’s ‘nested families’ to link decay objects with the structural entity [18]. In [90] a custom parameter is proposed to assess the impact of pathology on the structural stability of the building component.

A variation in this approach involves modifying the outermost, fictitious layer of the wall (using Revit’s ‘divide surface’ and ‘paint’ tools), to preserve the connection between the deterioration and the substrate [106]. A ‘mixed’ approach—using both RMO and PoE—is used in [36], where the exact shape of surface decay is traced on annotation planes and additional synthetic information regarding condition assessment, accessibility and intervention are inserted to support planned conservation process.

Another variation involves using again Symbolic Modeled Objects (SMO), simple markers like parallelepipeds or cylinders. These are rich in information and identify decays, without modeling the exact surface [107]. They can be enriched with various attributes such as the relevance of the damage, percentage of affected area, severity and urgency of intervention. Symbolic markers with color codes linked to the type of decay are also employed in [108]. However, the general condition of the elements is recorded through custom properties assigned to the construction entities from which intervention priorities are derived and displayed in thematic views.

Lastly, some degradations involve a significant material loss [106]—such as missing parts and gaps—for which it must be assessed whether, based on the specific purpose of the model, it is useful or not to model the loss or simply include it as information.

Customized objects and elements created using tools developed for other purposes often lack strong semantic significance. To address this, in [17], as part of the previously mentioned ontology, entities for decay analysis are formulated as ‘subclasses’ of the ‘Condition assessment’ entity derived from CIDOC ontology.

Alternative approaches focus on enriching the tools available in commercial software. Some research has led to the development of plug-ins that help overcome some software limitations, although these have the disadvantage of not being interoperable or widely adopted. In applications using RMO reproducing the decayed area, software implementations (primarily scripts in the Dynamo visual programming language) aim at representing causes, such as average moisture, on the wall surface [91], or calculating the area affected by degradation [76] (something that can be easily performed with 2D CADs if the surfaces are flat). In [54], a plug-in is developed to associate orthophotos with surfaces and create thematic maps of a 3D asset’s surface within a BIM environment, thus offering capabilities for calculating areas and running queries. This approach attempts to overcome BIM’s limitations in mapping, bringing some of the functionality of GIS tools into the BIM environment and making the work of restorers, who are more used to 2D mapping, easier.

Other research efforts involve collaboration with software companies to integrate specific tools for mapping degradation phenomena into commercial software. One example is the tool added to the Edificius software as part of a more comprehensive toolbar developed specifically for HBIM [109]. This tool enables mapping on curved surfaces and allows users to select the type of degradation according to the classification provided by the Italian standard UNI 11182:2006 [110], as well as to define the depth of the affected area. Furthermore, ongoing research is focused on providing semantic meaning to the object representing the decay. To that end, the ‘State of preservation’ knowledge domain has been formalized in the Building Smart Data Dictionary (bSDD) [111,112]. The newly created bSDD entities can be assigned to IFC entities, linking them with systematized information such as type of decay, spread of the phenomenon and affected material.

An additional advancement being pursued by some groups involves the development of machine learning algorithms for automatically recognizing superficial decays, sometimes starting from the point cloud data [90], and then visualizing the results within information models [113,114,115].

Table 4. Main approaches for decay representation.

Approach	Related Papers	Good in Case of	Pros	Cons
PoE Property of the construction Element	[41,50,88,97]	In case of a planned conservation process. Decay on small objects Decay that affects the structural behavior.	Connection between component and decay. Properties for suggested intervention. Possibility to link 2D drawings for further detail.	Decay extended to the entire object and not to the part really affected to the pathology
(with external database)	[96]		Synthetic risk assessment and graphic queries.
(with specifically developed tool)	[95]		Load past situations, enter maintenance activities.
RMO Realistic Modeled Objects (which reproduces the area affected by the decay)	[42,43,90,99,102,103,104]	Decay that does not affect the structure but only the surface. Large building components with different kinds of decays	Possibility to map the precise area and to insert specific properties. Possibility to compare decay information from different surveys.	Time consuming. Difficulty in linking decay and substrate. Difficult semantic significance and interoperability. Difficulty in mapping on curved and irregular surface
(with external database)	[116]
(with software implementation)	[76,88,91,100]		Depending on the experimentation, possibility to calculate area with decay, or to map on curved surface	Poor interoperability
(with specifically developed tool)	[109,111,112]		Appropriate information are preset.	Still under development solutions
SMO Symbolic Modeled Objects (with no real shape)	[107,108]	Need for fast and light model for interoperability issue	High Level of Information	Low Level of (geometric) Detail
As modification of the external layer	[106]	Decay on wall objects. Decays that do not affect the structure	Connection between decay and substrate	Difficult semantic significance. Difficulty in mapping on curved and irregular surface. Allowed only with some software
(with software implementation)	[54]		3D mapping retracing orthoimages
As subtraction of part	[106]	Decay resulting in a loss of material	High accuracy	Time consuming process, heavy model
Automatic mapping	[90,113,114,115]	-	No time consuming for modelers	Poor control

Table 4. Main approaches for decay representation.

Approach	Related Papers	Good in Case of	Pros	Cons
PoE Property of the construction Element	[41,50,88,97]	In case of a planned conservation process. Decay on small objects Decay that affects the structural behavior.	Connection between component and decay. Properties for suggested intervention. Possibility to link 2D drawings for further detail.	Decay extended to the entire object and not to the part really affected to the pathology
(with external database)	[96]		Synthetic risk assessment and graphic queries.
(with specifically developed tool)	[95]		Load past situations, enter maintenance activities.
RMO Realistic Modeled Objects (which reproduces the area affected by the decay)	[42,43,90,99,102,103,104]	Decay that does not affect the structure but only the surface. Large building components with different kinds of decays	Possibility to map the precise area and to insert specific properties. Possibility to compare decay information from different surveys.	Time consuming. Difficulty in linking decay and substrate. Difficult semantic significance and interoperability. Difficulty in mapping on curved and irregular surface
(with external database)	[116]
(with software implementation)	[76,88,91,100]		Depending on the experimentation, possibility to calculate area with decay, or to map on curved surface	Poor interoperability
(with specifically developed tool)	[109,111,112]		Appropriate information are preset.	Still under development solutions
SMO Symbolic Modeled Objects (with no real shape)	[107,108]	Need for fast and light model for interoperability issue	High Level of Information	Low Level of (geometric) Detail
As modification of the external layer	[106]	Decay on wall objects. Decays that do not affect the structure	Connection between decay and substrate	Difficult semantic significance. Difficulty in mapping on curved and irregular surface. Allowed only with some software
(with software implementation)	[54]		3D mapping retracing orthoimages
As subtraction of part	[106]	Decay resulting in a loss of material	High accuracy	Time consuming process, heavy model
Automatic mapping	[90,113,114,115]	-	No time consuming for modelers	Poor control

Representing structural damage is a challenging task, even in the traditional approach to a restoration project. While the D.P.C.M. 2011 [117] (in the context of Italian regulations) attempted to provide guidance for its representation, there are no universally accepted standards for how damage should be depicted. In the context of ancient masonry buildings, ‘damage’ is primarily characterized by the formation of cracks and deformations in structural elements.

Cracks are typically depicted in 2D drawings with a line that accurately represents the damage when seen in projection (e.g., cracks on walls seen in elevation). The same cracks are represented in plans with a symbolic sign on the structural element where they are located, following some conventions that help in distinguishing between passing or not-passing cracks, as well as varying degrees of severity. In many instances, additional information is required to identify the collapse mechanism responsible for the damage, which is sometimes illustrated using 3D schemes.

Different approaches are employed for crack mapping in BIM environment, depending on the specific objectives of the model (Table 5). In some cases, a geometrically simplified representation is favored, compensated by a rich set of associated information. Other studies prioritize geometrically accurate modeling, that capture the actual shape and extent of the damage.

The first approach involves again modeling cracks as customized realistic objects (RMO), such as linear paths or surfaces in the case of passing cracks, that closely replicate the damage. These cracks are often modeled using tools originally developed for other purposes (primarily ‘generic adaptive models’) [92,99] or created in 3D modeling software (e.g., Rhinoceros) alongside the structural element, and then imported into BIM authoring software with customized classification [89]. In the case study illustrated in [105], a high level of detail is achieved by mapping cracks on individual stone ashlars, allowing for the evaluation of their possible replacement as part of a post-earthquake reconstruction project. While relevant properties can be assigned to these objects, challenges remain in correlating the cracks with their relative substrates (e.g., walls or vaults).

Much more time consuming is to represent cracks as fully detailed 3D solids or void, modeling variations in crack width along the length of the damage [118,119].

In [120], cracks are instead mapped symbolically using rectangular markers attached to the wall (SMO). These markers are enriched with properties that assist in interpreting the damage (e.g., identifying failure mechanisms and their potential causes) and the risk associated with it. Additionally, the informative apparatus is structured using IFC-format data types to enhance interoperability. Symbolic objects are similarly used for crack representation in [107], where parameters such as crack length and width are included to assess the severity of the damage. Another development pathway involves linking cracks with well-known collapse mechanisms, through attached data-sheets [30].

As an alternative to these methods, new tools for software are developed to address limitations in existing approaches. For example, the aforementioned implementation of Edificius software [109] led to the creation of a dedicated tool for crack representation as RMO. This allows users to draw cracks on elevations and in 3D, as demonstrated in [121]. The crack object is automatically associated with a symbolic representation in plan, adhering to conventional symbology used in 2D drawings. It also incorporates specific properties, such as whether the crack is passing, its width, and the hypothesized causes, meeting the typical requirements of structural conservation projects.

Structural damage, however, is not limited to cracks but also includes deformations, which are often identified through precise instrumental surveys. Detailed 2D drawings typically depict these deformations and may include symbolic markers (explained via legend) to aid in understanding the damage such as out of plumb or bulging of walls, beam deflection or distortions in vaults or pavings.

The modeling of structural deformations in BIM environment is quite debated. On one hand research focuses on achieving high metric accuracy by faithfully reproducing the irregularities of historic buildings. For example, in [122], out of plumb walls are recorded by directly modifying the geometry of modeled objects. On the other hand, it is common practice to model structural elements using their ideal and rectified geometry. In [123], deformations are represented as displacement values in Cartesian coordinates, following a common practice in FEM software visualization. This approach enhances the interoperability and manageability of the information model. However, in case where rectified modeling is applied, deformation is recorded within the properties of the construction element with rectified geometry [120]. This enables thematic queries while retaining the option to attach detailed 2D drawings for a more comprehensive representation. Moreover, the metric difference between the idealized BIM model and point cloud data [124,125] (or the mesh or NURBS model deriving from the point cloud [126,127]) can be evaluated. This highlights the distinct roles of point clouds, which are geometrically accurate representation of reality, and HBIM models, which simplify geometry while enriching the model with informative content. HBIM models not only provide representations but also facilitate the interpretation of the ongoing phenomena. For instance, interpretative modeling allows for the differentiation—and distinct representation—of deformations originating during the construction (which may not indicate current structural issues) and those directly connected to stability problems, which are more critical for planned conservation efforts [103].

Table 5. Main approaches for crack representation.

Approach	Related Papers	Pros	Cons
RMO Realistic modeled object (path that retrace the crack)	[89,92,105]	Detailed representation useful for crack interpretation. Possibility to add customized properties.	Time consuming. Poor interoperability. No semantic connection with the structural element.
(with specifically developed tool)	[109,121]	Realistic representation in elevation, symbolic in plan. Appropriate properties for critical analysis.	Still under development solutions
SMO Symbolic modeled object (no real shape)	[120]	High Level of Information. Properties for critical analysis of crack pattern.	Low Level of (geometric) Detail (but possibility to attach more detailed 2D surveys).
As solids/voids	[118,119]	High accuracy (modeling of crack width)	Time consuming process, heavy model, interoperability issues.
With linked datasheets	[30]	Possibility to have detailed representation and correlation with possible collapse mechanisms.	No real BIM representation

Table 5. Main approaches for crack representation.

Approach	Related Papers	Pros	Cons
RMO Realistic modeled object (path that retrace the crack)	[89,92,105]	Detailed representation useful for crack interpretation. Possibility to add customized properties.	Time consuming. Poor interoperability. No semantic connection with the structural element.
(with specifically developed tool)	[109,121]	Realistic representation in elevation, symbolic in plan. Appropriate properties for critical analysis.	Still under development solutions
SMO Symbolic modeled object (no real shape)	[120]	High Level of Information. Properties for critical analysis of crack pattern.	Low Level of (geometric) Detail (but possibility to attach more detailed 2D surveys).
As solids/voids	[118,119]	High accuracy (modeling of crack width)	Time consuming process, heavy model, interoperability issues.
With linked datasheets	[30]	Possibility to have detailed representation and correlation with possible collapse mechanisms.	No real BIM representation

4.2.3. Time Management

Modeling in a BIM environment enables the effective management of change in both information and geometry over time. The management of dynamic data is commonly referred to as 4D-BIM [1]. In the context of an HBIM model for a historic building, time management encompasses both the building’s historical evolution and its projected future developments. This distinction is highlighted in [17], where the developed ontology includes two domains for time management: one for past time, representing historical construction phases and previous interventions and the other for future time, focused on managing conservation projects. More broadly, the management of dynamic information related to historic buildings can involve the past (mapping physical transformations over time—Construction Phases), the present (integrating Monitoring data) and future (representing design interventions and Conservation Process). In the existing literature, various methods are employed to manage each of the highlighted time periods.

In relation to the past, Section 4.2.1 has already explained how to handle historical data that can be entered simply as ‘static’ information within the object properties.

Instead, in this section, the focus is on the methods for managing the physical transformations undergone by the building (Construction phases), whose reception in HBIM involves an evolution in the geometry or in the information associated with the model entities. Indeed, the Historical evolution can go beyond being ‘narrated’ as textual information within object properties; it, can also be modeled by linking objects to the different construction phases, including prior restoration interventions [128].

Software tools, such as the ‘phases’ filter in Revit [52], or ‘graphical overrides’ in Archicad [129], facilitate the representation of temporal ‘macro-phases’. Each object is assigned to a specific construction phase, and elements from different periods can be automatically displayed in distinct colors to simplify the visualization of the building evolution. This approach has been applied in numerous case studies, including some of the earliest HBIM experiments [130,131]. A similar method is used in [132] to model different construction hypotheses and later evaluate the structural behavior of the different configurations. In [123], construction phases, past interventions, and earthquake-related collapses and deformations are modeled, with particular attention to the structural consequences of these changes.

The management of stratigraphic units (SUs) present significant challenges. Originally introduced for the analysis of archeological structures, this methodology is now widely applied to historic buildings: each masonry section, identified as a portion with homogeneous texture, corresponds to a SU. Relationships of contemporaneity, posteriority, and anteriority are established among these units to reconstruct the building’s relative chronology. SUs are categorized as positive, when a new portion is added, or negative, when traces of demolition are present.

One primary challenge lies in classifying SUs as a constituent element of a larger wall structure. To address this semantic limitation, ontologies are developed to encode stratigraphic temporal relationships [21,22]. Additionally, workflows for the semi-automatic generation of the Harris Matrix [133], a widely used tool for describing the relationships between SUs, are proposed. Alternatively, the potential of open source software, which can be freely customized to include relationships between units, are explored [19,32]. A dedicated plug-in is developed to annotate SUs directly on the 3D model, with further research focusing on assigning semantic significance to SU objects. These objects, often modeled as an additional layer on the wall surface with minimal thickness, are proposed as new IFC entities (currently absent in the default IFC schema) with customized properties.

Moreover, 3D information models enables the mapping of SUs not only as flat surfaces, as traditionally performed in 2D drawings, but also as volumes [77]. This approach highlights discontinuities confined to specific masonry layers, such as replaced plaster or restored outer masonry faces. However, this method is applicable only to positive units. Negative units, such as traces of demolitions or other subtractive actions, remain challenging to represent within this framework [134].

With regard to the building’s present time, the management of Monitoring data is an interesting and still underexplored issue.

Monitoring can be broadly understood as the periodic assessment of certain building parameters. For instance, the model developed in [104] records the progression of surface pathology detected during subsequent inspections. At each inspection, the type of degradation and the severity of the phenomenon are documented and entered into custom properties associated with the objects. The system described in [95] is specifically designed to track the progression of surface decay. Previous inspection data can be loaded into the system, allowing users to navigate through the building’s history: both past and current situations can be viewed simultaneously for comparison.

A more complex approach involves integrating real-time monitoring data obtained from sensors installed on the physical asset. These data are typically used for Structural Health Monitoring (SHM) or Environmental Monitoring (EM).

When the model receives real-time data from sensors and updates itself accordingly, it becomes what is often referred as a Digital Twin (DT) [135]. This Digital and Dynamic Twin represents an evolution of HBIM—that is a static archive of data—by establishing a dynamic link to the real asset. The primary objectives of DTs are to simulate the actual behavior of the historic building, estimate the most probable damage scenarios [136], automatically detect potential hazards, and suggest possible solutions [135].

Although this approach is still in its early stages and numerous studies are working to define and establish methodologies that are continuously evolving [137,138,139], there is an increasing number of applications related to BIM-based Structural or Environmental monitoring. By modeling specific object corresponding to sensors and linking them to external databases that receive their data, the BIM model effectively becomes a repository for monitoring data [140]. This enables querying of raw real-time data as well as pre-processed graphs [141].

Case studies in the literature demonstrate progressively higher levels of complexity and experimentation in the integration of monitoring data. The simplest approach uses only existing functions within commercial software, obviating the necessity for additional implementation or plug-in development. This approach links data through basic URL connections to external platforms or Drive folders, which are then inserted as properties of the modeled objects corresponding to the sensors. This method has been tested for both microclimatic [137] and structural static [85] or dynamic monitoring data [47]. However, it does not allow real-time updating of the model (which is typically noted as the next step of the research [47]), but rather facilitates a simpler connection to the data. This is what commercial software currently permits. Consequently, subsequent applications focus on the creation of new plug-ins or customized platforms.

In [141], structural monitoring data stored in a Microsoft Access database is linked to the BIM model using the Revit DB-Link plug-in. Autodesk Platform Service (formerly Forge) is then used to enhance sensor data management. A customizable reference application is created using Autodesk Forge APIs to connect BIM models with sensor data. In [142], the authors developed a new Python-base application as a Revit plug-in to integrate structural and environmental monitoring. The data are stored in a cloud-based repository. The developed plug-in ensures synchronization between the repository and the properties of the HBIM objects representing the sensors. The add-in includes various modules for both SHM and EM, which enables pre-processing of row data (e.g., linear regression). The results of this analysis can then be plotted as graphs and visualized within the model.

The web-based platform developed in [143] serves as a portal for visualizing environmental monitoring data. By selecting the sensors, represented as symbols in the 3D model, a window opens displaying real-time data acquired over a 24 h period, presented on a graph form. In [144], different types of sensors are modeled using customized objects classified as IfcSensorType entities. However, the interoperability is only partially successful, as some properties are not successfully exported. Raw data are stored in .xls sheets and displayed as ‘key schedules’ through the programming of a Dynamo script, enabling the user to select specific dates and times for comparison of various parameters. In [145], the study focuses on environmental and energy consumption monitoring. A Dynamo script is developed to facilitate automatic data exchange between the sensors and the BIM model. This script acts as a ‘bridge’ between the IoT platform for data storage and the model, enabling direct viewing of raw data (queried according to the observation period) associated with Revit ‘rooms’ object, representing the monitored environments. In [107], sensors data are stored in a MySQL database, allowing for the plotting of graphs to analyze the evolution of parameters. Again, an in-house plugin is developed to display the data and the corresponding graphs within the model.

The third temporal dimension to be managed is the future, through the planning of a Conservation Process. This is particularly relevant in the context of what is commonly referred to as a ‘planned conservation’: a series of interventions planned according to their urgency, which are carried out at different stages over time, and typically do not result in significant alterations to the architecture’s geometry. Conservation plans are traditionally structured as documents or databases, in which key data such as the date of inspections, interventions, planned future activities and risk assessment are documented.

Conversely, restoration projects, which often entail substantial modifications to the building’s volume and structure, are usually represented through 2D drawings that compare the existing condition with the proposed design. In these drawings, areas designated for demolition are typically highlighted in yellow, while new construction are shown in red. The approach to representing intervention—in BIM as well—varies depending on whether the model supports a single, transformative event (with significant changes in volume and architecture), or a continuous conservation process that unfolds over time with minimal geometric changes (Table 6).

In the BIM environment, for single restoration projects two models can be employed: one for the current state of the building (‘as found’) and one for the proposed design (‘as designed’) [105]. The latter incorporates new structural and architectural elements, as Modeled Objects (MO) categorized as part of a new ‘phase’ of the building. This allows for multiple project proposals to be modeled and compared, facilitating the evaluation of different solutions [146]. Additionally, elements from past interventions can be classified [85]. For example, in [121], structural strengthening interventions (such as tie rods, new lintels and reinforced plaster) are represented as new objects, and the mechanical properties of the masonry are updated based on the interventions’ beneficial effects. Additionally, technological details can be integrated as 2D drawings [76]. Following the intervention, the ‘as designed’ model evolves into the ‘as built’ one, useful in turn as a reference for future periodic maintenance actions. Scheduled future inspections, maintenance works, and their recurrence can be recorded as properties of the respective building elements [147]. Moreover, in accordance with traditional representation methods, as outlined in [63], areas marked for demolition are shown in yellow and new construction elements are highlighted in red [147]. These functions are readily accessible through software ‘filters’ tools (like in Archicad). Graphic queries and visualizations allow users to quickly display elements affected by specific interventions, whether they are past or planned [129].

Conservation plans can be enhanced by information modeling, allowing data to be spatially referenced and directly linked to the building component to which they refer.

For intervention involving maintenance or activities with minimal impact on the geometry of the building elements (such as architectural surface conservation), these actions are typically recorded in HBIM as custom properties of construction elements (again, PoE approach). However, a limitation exists since properties can only be assigned to the entire object, not to specific areas undergoing intervention. This limitation is particularly noticeable in cases like surface conservation on large walls, where it may be more effective to assign intervention properties to the object representing degradation [90,102]. On the contrary, setting properties for intervention is particularly effective for small-sized modeled objects, where the conservation activities (including past ones, as in [88]) encompasses the entire object. For instance, in the well-known case of the main spire of the Milan Cathedral (developed in a self-implemented platform), each stone ashlar corresponds to a geometric object of the model. Conservation activities are recorded associating a color to the ashlar depending on the type of interventions [95,148,149]. This is also used in [105,150] for ‘scuci-cuci’ intervention, that means substitution of single stone block in the pillars. It is notable that in [105] different representation strategies are used according to the type of intervention. Color-coding is also used in [151], where a Dynamo Script suggest the intervention, according to the type of decay, and an urgency assessment.

From the perspective of planned conservation, intervention planning is based upon a comprehensive risk assessment. In [96] the ‘risk’ connected to a modeled object is stored in an external database. The database allows for graphical query and the model is texturized according to query results regarding vulnerabilities. In [41], risk is assessed according to the damage condition, the historical-architectural value of the construction element itself, and the time elapsed since the last inspection and recorded again as property of the objects. Intervention priority is then based on this risk assessment. Priority levels can be displayed in the model through color coding, as in [152]. The same approach is applied on a broader scale, to evaluate the vulnerability of spatial units useful for the definition of a master plan [98].

The interventions (whether planned or previously executed) can be conceptualized and related to the element on which they are carried out, as entities within the ontologies discussed in Section 4.1.1 [17,23]. Specifically, in [17] the focus is on the conceptualization of surface conservation activities (divided into the well-known classes of preconsolidation, cleaning, consolidation, and protection) and linked to the ‘artifact’ thanks to the property ‘is extended to’.

Table 6. Main approaches for representation of restoration projects and conservation activities.

Approach	Sub-Approach	Related Papers	Good in Case of	Pros	Cons
MO Adding Modeled objects	‘as found’ and ‘as design’ -models with object for new elements	[76,85,105,121,146]	Project with a strong impact on the geometry, also structural strengthening intervention and previous intervention	Use of ‘filters’ to compare different phases or project proposals. Precise location and accuracy of the representation.	Poor visibility of interventions with small impact on geometry
	with ‘yellow’ and ‘red’	[63,147]
PoE Property of Elements	of the construction element	[41,88,95,96,105,149]	Planned conservation process (almost no influence on geometry). For intervention on elements of small dimension.	Connection between decay, risk, priority index and intervention.	No precise location of the intervention, extended to the entire element
	of the decay object	[90,102]	Surface conservation activity	Precise location of the intervention.	Poor correlation between construction entity and intervention

Table 6. Main approaches for representation of restoration projects and conservation activities.

Approach	Sub-Approach	Related Papers	Good in Case of	Pros	Cons
MO Adding Modeled objects	‘as found’ and ‘as design’ -models with object for new elements	[76,85,105,121,146]	Project with a strong impact on the geometry, also structural strengthening intervention and previous intervention	Use of ‘filters’ to compare different phases or project proposals. Precise location and accuracy of the representation.	Poor visibility of interventions with small impact on geometry
	with ‘yellow’ and ‘red’	[63,147]
PoE Property of Elements	of the construction element	[41,88,95,96,105,149]	Planned conservation process (almost no influence on geometry). For intervention on elements of small dimension.	Connection between decay, risk, priority index and intervention.	No precise location of the intervention, extended to the entire element
	of the decay object	[90,102]	Surface conservation activity	Precise location of the intervention.	Poor correlation between construction entity and intervention

4.3. Return to Building: Data Sharing and Use

Following the model’s information flow, the next step is Data sharing, which can be seen as a ‘Return to Building’. The extensive domain of data sharing encompasses all the solutions developed for on-site use of the information system by those responsible for managing the architectural asset, as well as exportation for possible further developments. Interoperability issues are central to this phase. Indeed, a correct exchange of both geometric and non-geometric data is necessary for using the HBIM model as a base for other analysis or applications.

4.3.1. On Site Use

Research efforts aimed at developing information systems to support planned conservation processes often focus on implementing desktop or web-based platforms, expected to facilitate on-site use of the model, even by non expert users. Ideally, these platforms allow real-time in situ updates with information regarding the current state of conservation, interventions, or inspections conducted. Web applications overcome interoperability issues by freeing the users from reliance on any specific software and by avoiding the need for data exchange [96]. In addition, usability and portability of the model are enhanced, as users can view it directly on mobile devices and input data simultaneously, thus reducing working time and transcription errors [96]. As experimented by [141], online viewing can also facilitate the management of data from sensors connected to the BIM model.

Major commercial BIM authoring software provide desktop or mobile platforms, such as Graphisoft BIMx [41] or Autodesk Drive (formerly A360 [134]). However, these platforms only allow visualization and not on-site modification of data, since they are primarily meant for designers to showcase their projects and not to on-site record of data.

For this reason, some research has experimented the development of new platforms. This is the case of BIM3DSG, created for managing Milan’s Cathedral conservation activities [95,149], and applied to other case studies [37,153]. The web-desktop system, designed for on-site use on mobile devices, enables the input of any maintenance activity that does not modify the geometry of the model (which can be updated later by 3D experts). Another web application is developed in [30] with the objective of enabling non-expert users to perform semantic inquiries over the models, organized through IFC format, and to visualize the results of their research directly on the browser. An open-source cloud-based platform is developed in [32] to meet the specific needs of archeological heritage management.

Other studies have further developed services starting from commercial software, for in situ management of information. In [154], Autodesk Forge is used to develop a platform which provides various operational tools to support professionals in managing an asset’s scheduled conservation. Structured data related to building components can be inserted, with multiscale visualization and BIM/GIS integration. The database developed in [96] is accessible in two ways: firstly, through a desktop application, which acts as a plug-in for the BIM software; and secondly, through a web interface. The latter has been designed with the specific aim of ensuring data sharing and usability by both skilled and unskilled users, who can edit, view information and query the database, visualizing the results with false color maps.

In addition to BIM-based solutions, research is concentrating on the development of reality or point cloud-based platforms for in situ annotation, such as [155], which enables direct annotation on the point cloud. Web and mobile applications for virtual reality (VR) and augmented reality (AR) are much widespread [53,131,156]. In most cases these are aimed at sharing information on architectural heritage for tourism purposes. However, some interesting applications are focused on using VR platforms for on-site decay mapping [157].

4.3.2. Interoperability

Interoperability means that multiple systems are able to exchange information with each other and then to use it for further applications. The most effective approach for achieving interoperability between BIM software and other types of software is the use of open formats. This approach is often defined as open-BIM [111,120,158]. Interoperability is prevented, instead, by the use of proprietary formats. The IFC standard, as introduced in Section 4.1.1, is the reference open format for BIM applications. IFC is organized with entities and properties and allows the exchange of both geometric and non-geometric data. However, as previously stated, not all components of historical building can be properly described by IFC entities, and this precludes correct export and thus interoperability. Semantic issue and interoperability are therefore strictly linked: semantic definition is required for classification in open format. Open-BIM does not necessary entail the adoption of open-source software, which is used instead, for example, in [32]. Indeed, also commercial BIM software allow to read and export in IFC format.

Interoperability between BIM and other software is pursued for different further developments. Originally conceived as a ‘container’ for interdisciplinary data accessible to all project stakeholders, BIM supports integrated workflows that often require exporting the model for use in specialized tools. This process should preserve both geometric and informational content. Among the possible applications, Structural and Energetical analysis play a fundamental role in conservation projects.

With respect to Structural analysis, different types of assessments can be carried out on historic buildings. The most common one is Finite Element (FEM). In this approach, masonry is represented as a continuous, homogenized deformable body, discretized through the generation of a mesh comprising finite elements of defined shape [159]. This analysis is conducted on models that can be theoretically derived from HBIM.

Regarding interoperability between BIM and FEM, it becomes evident that the current state of the art on the topic suffers from the IT and semantic limitations of software and exchange formats. Indeed, when IFC is used, the exchange process works effectively only for ‘particular’ historic buildings composed only of construction elements which are used also in modern architecture and so can be translated into existing IFC entities (e.g., walls, columns, beams, roof, etc.). This is demonstrated by [158] and [160], which tested interoperability with finite element and macro-element analysis software, respectively. For these cases, both the geometry and the mechanical characteristics of materials can be set in the BIM model using the IFC property scheme and exported. A continuous exchange of data is also achieved in [127] but only for beam objects. However, some applications show interoperability shortcomings even for elements properly mapped to IFC entities [121]. In [123], although BIM elements are generically classified as IfcBuildingElementProxy, structural information are inserted as IFC properties and not lost in the export process for FEM in a self-developed platform. An additional drawback when operating with historic masonry buildings is that FEM software, such as Abaqus or Ansys, do not support the open IFC format. Hence, only the geometry is imported, while the information apparatus is lost.

In the experiments presented in [83,130,150,161,162,163,164], the geometry is transferred from BIM, while the information is re-entered into the structural analysis software. In [130], the subdivision of wall elements modeled in BIM to report the different construction phases is maintained in the export for structural analysis, while the different mechanical properties attributed to masonry belonging to different periods are reinserted in the FEM software. Only structural elements are transferred from BIM; non structural elements are excluded and converted into dead loads applied in the structural model.

Most of case studies bypass the BIM software, likely to avoid interoperability issues as explained in [165], and transfer the geometry directly from 3D modeling platform (primarily Rhinoceros) to the structural analysis tool. This approach aligns with a well-established workflow better described as Scan (or Cloud)-to-FEM, for which some interesting developments are introduced in [136,166].

The workflows presented in [167,168] are even more simplified and far from achieving true BIM-to-FEM interoperability. In these cases, the BIM model is ‘sliced’, and only 2D plans or sections are imported into the structural analysis software. Consequently, the geometry of the asset is rebuilt in a highly simplified form.

Interesting, although only the geometry of the wooden beam is exported for structural analysis, is what experimented by [83]. Starting from the assumption that decay in wood leads to the destruction of the external layer of the structure, the authors model in BIM both the ‘apparent section’, whose geometry is that obtained from the survey, and the ‘resistant section’, which is reduced compared to the apparent one, according to diagnostic test, and exported for finite element analysis. Possible developments to mitigate the loss of information include using precompiled tables to convert BIM model properties into properties to be applied in the structural model [169].

Other authors develop their own platform for structural analysis to overcome interoperability limitations. This is tested by [82] with a ‘macro-element analysis’ tool for masonry structures. ‘Macro elements analysis’ is based on models in which the structure is idealized and simplified in panel components, typically piers and spandrels [159]. The self-developed platform has a real-time connection with the BIM model, where mechanical and seismic parameters (according to the Italian regulation) are already collected. This approach also enables to verify the effect of strengthening intervention, which can be inserted in the information model and received in the structural one that refreshes automatically as a result. However, the self developed algorithm can only analyze simple geometries, as stated by the authors.

Another open issue regards the level of metric accuracy and geometric detail that the information model should have if its main purpose is the interoperability with models for structural analysis. Indeed, an exaggerated level of detail not only makes the models more difficult to manage, but leads to errors in the import phase which, as highlighted in several case studies, require manual ‘corrections’ [130]. As underlined in [161], it was necessary to manually edit the results of the auto-meshing algorithm to fit the requirements of the software for structural analysis. However, in other cases all the morphological complexity of each element without any simplification is kept for exportation [116], and the solids generated in the 3D modeling software are discretized by default libraries of structural analysis software (i.e., beam, rod, etc.) [150]. Interesting, from this point of view, is the experiment carried out in [170], with the aim of modeling a 3D geometry with an intermediate level of detail, made of simplified mesh of segmented point cloud (to keep irregularities and out of plumb of walls), which assures a sufficient interoperability. Although manually reinserted into the structural model, the mechanical properties are the result of the process of idealization of the building according to the interpretation of structural behavior which, for example, lead to underestimate the average modulus of elasticity on the masonry to consider damage information.

In addition to the issue of interoperability between information and computational models, other research concentrated on incorporating data into the BIM model with the objective of conducting other types of structural analysis. Indeed, as it is well known, the structural assessment of historic buildings includes not only global analyses on FEM, but is often resolved into verifications of local collapse mechanisms on simple two-dimensional diagrams. Therefore, in [126], 2D sections are extrapolated from the BIM model, to verify the equilibrium conditions. In [171], the objective is to assess the possible activation of collapse mechanisms, according to the Italian D.P.C.M. of 9 February 2011 for Evaluation and Reduction in Seismic Risk of Cultural Heritage [117]. Indeed, customized parameters are proposed for traditional bell towers as a kind of check-list of the vulnerability indicators, due to the geometric configurations or to specific building techniques. Similarly, in [122] symbolic objects are placed on structural elements and collect data about local collapse mechanisms assessment; detailed analysis are linked in the properties. A similar approach is developed in [47]: the symbolic objects that represent the Safety Index, again according to the 2011 D.P.C.M., change their color depending on whether the analysis is satisfied or not and have customized properties for the parameters used for the assessment. In [92] the results of static analysis on slender masonry walls and of seismic analysis with evaluation of collapse mechanisms are visualized in the HBIM with a color scale ranging from green (analysis satisfied) to red, depending on the safety level of each element.

Finally, the field of Energy analysis encompasses a range of analytical techniques and approaches. In some cases, the workflow seems to be quite straightforward. This is because certain types of analysis useful for historical buildings can be directly performed by BIM authoring software and available plug-ins. For example, in [172], solar and heat gain analysis of the building envelope are performed with Revit Insights, thanks to the reception of thermal properties of construction elements and local weather data. The amount of daylight and solar irradiance in rooms is also calculated, by modeling rooms as voids.

Other studies address the topic of interoperability with software for energy analysis. In [78], custom properties for thermal performance (mainly transmittance values) of walls and windows are added to the IFC script, which is subsequently imported into the energy analysis software. In [173], energy analysis is performed at both building and urban scale, with regard to a small historic center. At urban scale just a plan is extracted from HBIM model and then used as starting point for fluid dynamics simulations. At building scale graphical algorithmic modeling is used to have active parametric link between HBIM and dynamic building performance simulation environments. Then, the results of daylight analysis are projected on HBIM wall elements. BIM to BEM (Building Energy Model) is the focus of the research presented in [174] and aimed at developing a Digital Twin, intended as decision support system. The proposed approach entails the creation of a ‘Topological model’, as an intermediate step and a vehicle for data exchange between BIM and BEM. In fact, BIM volume-based geometries must be transformed into BEM surface-based geometries for building envelop components. In order to achieve this objective, visual programming algorithms in Grasshopper are employed. The authors argue that geometry complexity of models can lead to computational bottlenecks. Indeed, the case study analyzed belongs to contemporary architectural heritage with more simple geometries. In [79], gbXML data exchange scheme is used to achieve interoperability between BIM and BEM. The same approach is applied in [175]: a simplified geometry of the building is exported from HBIM model, while thermal properties of the envelope and HVAC systems and added in the energy analysis tool.

5. Discussion on Research Gaps and Future Trends

Following the literature review conducted, several issues have emerged, along with ongoing challenges within the various categories of problems addressed in the development of an HBIM. According to the proposed organizational structure of the article, observations regarding the individual strategies follow the same framework, aiming to identify, for each macro-topic—Model definition (M, Section 4.1), associated informational framework (I, Section 4.2), and the ideal return to the building (B, Section 4.3)—possible critical issues and potentialities. Nonetheless, this organization is not always linear and feasible, as issues tend to be overlapping and not rigidly separable when their implications are evaluated from an application perspective.

First of all, several overarching considerations emerge across all the sections analyzed.

As general remark, it is important to emphasize that the majority of the referenced regulations and case studies pertain to the Italian context, reflecting the country’s well-established and prominent tradition in the field of architectural conservation (see Section 3.2).

A key distinction should be drawn between software-oriented solutions and those approaches—frequently encountered in the reviewed literature—that involve the development of custom implementations or plug-ins [37,95,96,101,141,142,143]. While these may offer greater adaptability and effectiveness for specific case studies, they often risk to limit broader interoperability. This also applies to the numerous proposals for specific information framework tailored to historical construction elements or conservation processes. Although software platforms offer sufficient flexibility to add new attributes, such data often remains confined to use by the original developers. Such solutions may also prove redundant and resource-intensive in terms of both computational time and memory. In this context, the collaboration between academic researchers and software developers is particularly noteworthy and presents both significant opportunities and challenges. Where such partnerships have been established [109], they have yielded promising results.

In this regard, alternative methodologies such as Geographic Information Systems (GIS) can be considered to address the current limitations of the HBIM approach and commercial software. Indeed, the high demand for both human and financial resources typically required by standard HBIM workflows may justify the adoption or the integration with alternative solutions, depending on the specific needs and objectives of the project [4,126]. A further consideration is that most studies focus on historical masonry buildings. Generally, HBIM of contemporary architecture tends to be less challenging—from a geometric point of view, due to the reduced presence of irregularities, and in terms of information management, as materials and components are often more consistent with those for which BIM was originally conceived. Nevertheless, the application of HBIM to contemporary buildings has been explored in only a limited number of contributions, including [44,55,60,89]. A broader adoption of this approach is likely to give rise to new open questions.

Finally, the review highlights a growing interest in applications that leverage artificial intelligence. In particular, specific stages of the process have been identified where AI offers promising developments, such as the automatic segmentation of point clouds (for semi-automatic Scan-to-BIM methods) [40] and the automated detection of surface degradation [113,115]. Automating the more “mechanical” parts of the workflow is clearly advantageous, as it can reduce both processing time and the risk of human error. However, it remains crucial to preserve human expertise and oversight in those phases of the process that involve interpretation, critical judgment, or context-specific decision-making.

5.1. Modeling: Not Only a Geometric Issue

As highlighted at the beginning, most studies on HBIM focus heavily on the development of precise geometric modeling. Indeed, significant progress has been made in defining geometries that increasingly reflect reality, even in the case of complex geometries such as historical structures. A major limitation of this approach is its exclusive focus on geometric modeling, often neglecting the informational component.

Nevertheless, there is a risk of producing models that are esthetically satisfactory and highly detailed but do not accurately reflect reality; instead, they may represent an “idealized” version based on historical documentation. To assess the effectiveness of a model, several metrics have been established, including the Level of Development (LOD), Level of Accuracy (LOA), Level of Information (LOI), and the Level of Reliability (LOR) [105,176], which measures the geometric and semantic correspondence between reality and the model. In particular, the concept of the Level of Information Need (as defined in ISO 19650 [177]) emphasizes the importance of the specific objectives of the model. In this context, greater “effectiveness” can be achieved even with lower levels of geometric detail, provided that interoperability and manageability are prioritized.

However, geometric simplification involves more than just streamlining the modeling process and reducing complexity. It requires a critical interpretative process of the building, justifying the selection of specific simplifications to better represent the structure or a particular part of it [58,60]. This critical simplification process is especially challenging when aiming to interpret the structural behavior of historic buildings. It is essential for transforming potentially infinite data into a coherent “model”, which inherently involves an interpretative understanding of the original data. This process not only allows for the enrichment of the model with information acquired through building knowledge, but also facilitates complex structural analyses.

A well-known limitation is the tendency toward excessive standardization, which can be mitigated by incorporating parameters that make objects adaptable to different projects. While each structural element of a historic building may be unique, the associated information scheme—comprising a list of properties filled with data specific to each element—can be more readily standardized and integrated into software through the use of classification systems [41,96]. This approach enables the software to serve as a “guidance tool” for surveying and documenting the specific construction features of historic built heritage, thereby streamlining the data collection and analysis process.

5.2. Information Framework: An Open Issue

Firstly, it is interesting to note that the issue of element characterization is indeed the least explored topic in the existing literature, as already highlighted in Section 4 (Figure 7). This underscores a significant gap in the development of HBIM, despite its central importance in the study of historic buildings.

From the informative point of view, the inclusion of damage and deterioration data within HBIM environment has received particular attention and represents a central aspect in the exchange of information, offering insights that can be directly employed by restoration professionals. As seen, various strategies and approaches have been adopted to address this issue.

The mapping of crack patterns within a BIM environment should harness the advantages of three-dimensional visualization to enhance the understanding and correlation of damage distributions. Additionally, the use of customized attributes can support the interpretation of underlying failure mechanisms. Moreover, it is necessary to further investigate the appropriate level of detail for representing cracks in BIM models, in order to ensure effective interoperability with structural analysis software.

Despite its potential, the application of crack pattern mapping remains limited compared to other areas, likely due to the inherent challenges in standardizing structural damage within the framework of information modeling. This may also reflect the still limited—though steadily increasing—engagement of structurally focused professionals in HBIM research. Ironically, when the emphasis is placed primarily on structural analysis (and ideally on achieving interoperability with computational models, which are often based on undeformed geometries), a more simplified modeling approach tends to prevail, as already stressed [120]. Such simplification facilitates model management and promotes interoperability but they need a deep work of interpretation of mechanisms of damage which is not always simple. Moreover, custom classification systems—typically created manually—remain non-interoperable, highlighting an area in need of further development.

5.3. Possible Return to Building

BIM was originally conceived—and continues to function effectively in the context of new constructions—as a comprehensive repository of interdisciplinary information, accessible to various stakeholders and exportable for structural analysis. However, when applied to historic buildings, information modeling has not yet fully realized this potential. This shortcoming is largely attributable to the limited interoperability between BIM platforms and the structural analysis software commonly used for historic masonry buildings. Moreover, accurately representing the structural behavior of historic buildings is inherently complex and not easily reducible to the kinds of simplifications and automated processes typically handled by software tools. A reliable structural assessment requires consideration of construction details as well as the building’s condition and deterioration—factors that still lack adequate representation within BIM environments and therefore cannot be effectively transferred to computational models. In this context, a crucial role is represented by the managing of monitoring data into the HBIM environment, which thanks to the assessment of the evolution of the damage, is fundamental for evaluating whether strengthening intervention are really necessary and for planning them.

When BIM is intended as a tool to support an intervention on an architectural asset which consists not in a one-time restoration project but in a planned conservation process, it is necessary that the model is updated together with the real building. Therefore, the model becomes a Digital Twin, or more than one (structural, energy twin, etc., different models of the same reality serving different purposes) interoperable with each other.

Considering this issue, from the review it is clear that efforts have focused almost exclusively on the resolution of IT problems (to obtain the difficult real-time connection between sensors and model). However, this focus has somewhat overshadowed methodological considerations regarding the most valuable information to incorporate into the model, so that it becomes an effective tool for planning interventions. It should be emphasized that an accurate and detailed modeling of crack geometry—including direction, extent, and the variation in width along their development—is essential for the proper modeling and interpretation of structural behavior and damage. For example, the case studies analyzed focused mainly on receiving raw data and graphs that simply plot the data, which can only be understood and used by experts [141]. It could perhaps be even more interesting (and paradoxically easier from an IT point of view) to incorporate synthetic indicators of monitoring trends, from which even simpler alarms are derived in the event of exceeding the set thresholds. This simplified information can be understood even by non-experts, architects and practitioners involved in conservation activities, and therefore can be more easily used for the detection of risks and the planning of activities.

In general, some applications that attempt to relate diagnosis and intervention—widespread mainly for the conservation of architectural surfaces where there is a more evident connection between degradation and solution—perhaps bring out the risk of excessive automation of the process, which seems to minimize the role of the restorer [151]

As final remark, it has to be noted that recent advancements in geometric surveying and structural analysis have yielded highly accurate and automated tools that are essential for the study and preservation of cultural heritage. These technologies underpin the knowledge base required for effective conservation efforts. Nonetheless, challenges remain in achieving seamless data interoperability and developing reliable methods for information exchange without loss.

Current research, as seen and demonstrated by this review, primarily focuses on refining these tools to create detailed Digital Twins that accurately replicate historic buildings and simulate their structural behavior with quantitative rigor. Paradoxically, however, the greater challenge often lies in reducing data complexity to facilitate more reliable analyses and foster a qualitative understanding of structural performance. Although precise geometric data and advanced structural calculations are critical for restoration design, the growing volume of data demands careful simplification. This process must be informed by interdisciplinary insights—including historical context, structural intuition, and empirical knowledge—to generate conceptual models that are both manageable and faithful to the building’s characteristics. Balancing geometric accuracy, resource limitations, and data usability remains a central challenge in heritage surveying and modeling. As Della Torre highlights [14], research on Historic Building Information Modeling (BIM) should move beyond mere architectural representation and leverage interoperable digital tools to streamline conservation workflows. This requires a comprehensive understanding of historic building, achievable only through deep empirical knowledge, which remains the most vital means for effectively comprehending and preserving cultural heritage.