Advancements and Applications of Industry Foundation Classes Standards in Engineering: A Comprehensive Review

Abstract

The Industry Foundation Classes (IFC) standard has been widely implemented as an open data standard in the architecture, engineering, and construction (AEC) industry. IFC enables robust information representation and facilitates cross-disciplinary collaboration, serving as a critical data foundation for future intelligent development in the engineering field. However, current IFC research topics remain relatively fragmented, and there are still several challenges in the practical implementation of IFC. Therefore, this paper provides a comprehensive review of IFC research over the past two decades. The research progress is systematically summarized in three key areas: IFC applications, interoperability, and data processing. Through this review, the limitations in IFC development have been discussed, and future research directions are proposed. This paper aims to provide a comprehensive perspective on addressing data-related challenges in the AEC industry and contributes to facilitating the deep integration of emerging technologies such as artificial intelligence within the AEC domain.

1. Introduction

Building Information Modeling (BIM) has undergone nearly five decades of development since its initial conception [1]. During this period, emerging digital technologies such as the Internet of Things (IoT), blockchain, and Digital Twin have been widely implemented across various aspects of the architecture industry. Throughout the evolution of these digital technologies, data, as the fundamental carrier of information, have become a key factor for intelligent development in the architectural field.

With the advent of the artificial intelligence era, the demand for high-quality data in the architecture industry has increased significantly. Emerging artificial intelligence (AI) applications, including generative design and architectural large language models, all require large quantities of building information data as their foundation. This trend has led to growing attention on research into standardized expression of fundamental data in the architectural domain. Industry Foundation Classes (IFC) [2], published by buildingSMART (formerly the International Alliance for Interoperability, IAI), serves as a universal data standard for the building industry and plays an irreplaceable role in the semantic expression of building information.

Industry Foundation Classes (IFC) is an open data standard designed for the entire lifecycle of building projects. This standard is specified using the EXPRESS data modeling language based on ISO 10303-11 [3], enabling a formalized expression of building information models through strict schema definition language [4]. Over nearly three decades of continuous development and iteration, the IFC standard has evolved through multiple versions from IFC1.0 to IFC4.3, with each version expanding its capabilities [5]; the evolution of IFC standards and the number of entities is shown in Figure 1. The core concept of IFC standards is to establish a platform-independent data exchange system, enabling data sharing and collaboration among various stakeholders throughout the building lifecycle through standardized information models. Over the past two decades, IFC has been extensively implemented across all stages of the building lifecycle. During the design phase, IFC has facilitated design collaboration and information sharing [6]; in the construction phase, it has provided data support for progress management, quality control, and cost accounting [7]; in the operation and maintenance phase, it has enabled standardized information models for facility management and space optimization [8]. IFC has also been adopted in emerging fields such as prefabricated construction and intelligent construction. Researchers have extensively studied IFC as a fundamental data model, and its application has been reviewed across various domains. For instance, the use of IFC in the development of BIM technologies has been discussed in the fields of transportation [9], geographic information systems [10], structural engineering [11], and indoor navigation [12]. These studies highlight that IFC not only enables information representation across different domains but also serves as a bridge for information sharing. However, challenges related to information completeness and software interoperability persist in the practical application of IFC. In addition, there are reviews focusing on IFC extensions [13], summarizing methods for extending IFC and the application of IFC ontologies [14]. These approaches are considered as potential solutions to address the issues of interoperability and information representation in IFC. Most of the existing reviews primarily focus on the application of IFC within a single engineering domain, often mentioning IFC as the underlying data foundation for these applications. However, a comprehensive review of the development of the IFC standard across the entire engineering industry is still lacking. There remains a need to systematically summarize and discuss the developments of IFC within the industry, as well as to identify and analyze the common challenges and underlying causes that hinder its adoption in various fields.

To accommodate the intelligent development trend in the architecture industry and mitigate the constraints of data standard deficiencies on industry development, this study systematically examines the development route and application status of IFC over the past two decades. Through an in-depth analysis of relevant research findings, this study aims to reveal potential challenges and future development directions of IFC while providing references and research perspectives in building data standardization for researchers, thereby promoting the development of the construction industry.

2. Methodology

This study selected the Web of Science (WoS) Core Collection database as the data source. As an internationally recognized literature database, WoS is widely used for literature reviews in the AEC field and includes a large number of important journals. The initial time span of this study covers a period of 20 years, from January 2004 to December 2024. In order to provide a more comprehensive overview of the development of the IFC, the time frame will be extended to March 2025, which coincides with the commencement of this research. The specific process of literature retrieval and selection is detailed as follows:

Firstly, using the advanced search function of the WoS database, a search was conducted spanning the period from January 2004 to March 2025. Four search expressions were designed, where TS represents the topic, KP refers to the keywords automatically generated by WoS, TI denotes the title, and AK represents the author keywords. To minimize omissions and avoid irrelevant studies being included in the search results, the searches for keywords and titles were restricted to ensure that the full name of the IFC standard appeared in all fields. The search expressions and corresponding results are summarized in

Table 1. Search expressions and results.

Query Expression	Number of Publication
TS = IFC AND TS = Industry Foundation Classes	433
(KP = IFC OR KP = Industry Foundation Classes) AND ALL = Industry Foundation Classes	19
(TI = IFC OR TI = Industry Foundation Classes) AND ALL = Industry Foundation Classes	157
(AK = IFC OR AK = Industry Foundation Classes) AND ALL = Industry Foundation Classes	322

. After merging and deduplicating all retrieved records, a total of 455 valid articles were obtained as the research sample.

3. Literature Analysis

Through statistical analysis of 20 years of IFC-related research literature, Figure 2 illustrates the annual publication trends, reflecting the temporal evolution characteristics of IFC-related studies. This evolutionary trend demonstrates the academic community’s deepening understanding of IFC: progressing from initial exploratory cognition, through widespread application research, to the current phase of domain-specific implementation.

In the early stages, the increasing demand for building informatization drove substantial growth in IFC-related research. Subsequently, IFC gained widespread acceptance in the engineering field, with research focus shifting towards practical applications, leading to peak research output during this phase. However, this process also led researchers to recognize that IFC is fundamentally a data representation standard, with its core functionality lying in structured information expression and exchange. Currently, engineering research has become more specialized and refined, while IFC’s role as a fundamental data standard has become more clearly defined. Consequently, related research has evolved into developing specific solutions for engineering problems, and studies on IFC have entered a phase of steady development.

Based on the statistical analysis of publications source, Figure 3 presents the names of journals with more than three publications over the past 20 years. Among them, Automation in Construction accounts for 27% of the total literature, making a significant contribution to IFC-related research. Most journals are from the architectural engineering field, which accommodates interdisciplinary research directions. Additionally, journals from other fields have also included IFC-related studies, particularly in the geographic information domain, such as ISPRS International Journal of Geo-Information, Remote Sensing, and Sensors, as well as in informatics and computer science, such as Advanced Engineering Informatics, IEEE Access, and Computers in Industry. This indicates that although IFC serves as a data standard in AEC field, it is closely intertwined with information technology.

Based on a systematic analysis of 455 publications, the IFC-related research was categorized into three parts according to their keywords and topic classifications. The first category is application-oriented research, which utilizes IFC as the foundational data format in engineering applications. In these studies, the primary focus is on developing solutions for engineering problems. The second category concerns interoperability, which represents the core characteristic of the IFC standard. The third category focuses on IFC data processing research, which addresses issues related to IFC data utilization, including studies on data operations such as data comparison, compression. These three parts are discussed in detail in the following sections, along with a review of the literature closely related to IFC.

3.1. Research on IFC Applications in Engineering Fields

Research on IFC applications in the engineering domain has encompassed various aspects of construction engineering. To systematically organize these research findings, this paper analyzes IFC applications across three major phases of the project lifecycle: design, construction, and operations and maintenance (O&M). Through systematic classification and statistical analysis of the literature keywords and topics, this study thoroughly investigates the application characteristics, research hotspots, and existing challenges of IFC across different phases.

3.1.1. IFC Applications in the Design Phase

IFC applications in the design phase refers to the utilization of IFC in domain applications during the design phase. Through analysis of the literature themes and keywords, the main areas of focus include rule checking, energy analysis, cost estimation, and collaborative design. The annual publication counts for these topics are shown in Figure 4.

The distribution shown in the figure reveals that rule checking stands as the most active research domain. From a temporal perspective, research on rule checking has been continuously conducted over the past two decades, with a notable growth trend in the last five years. In rule-checking studies, the mapping between IFC and regulatory content [15] represents a crucial step towards achieving automated checking. Early research adopted hard-coding approaches, directly converting IFC into program objects [16] for checking purposes. Subsequently, to enhance automation levels, researchers introduced machine learning methods to achieve semantic matching between IFC concepts and regulatory requirements [17]. An alternative implementation pathway involves transforming IFC information into more checking-friendly formats, such as semantic web expressions [18,19].

In the field of building energy analysis, research primarily focuses on IFC data interoperability issues [20]. Although IFC can provide material and property information for energy consumption analysis [21], additional geometric information is still required to generate spatial boundaries [22]. Researchers have also investigated methods for converting IFC data into energy analysis formats [23,24]. However, existing studies generally conclude that IFC has not yet fully met the requirements for energy analysis and lacks specific energy analysis concepts [25].

In lifecycle assessment and cost estimation studies, IFC primarily serves as a fundamental data carrier. However, implementation challenges have been highlighted, including difficulties in information exchange between various BIM software platforms [26] and issues of data incompleteness [27]. These challenges necessitate concept extension [28] and property set supplementation [29] to support domain-specific applications.

Research on collaborative design is concentrated in the literature from a decade ago, corresponding to the early stages of digital transformation in the engineering industry. These studies mainly focused on the development of collaborative system [30] and their interoperability with IFC data [6]. Most research indicates that IFC requires further long-term development for collaborative design applications. This is attributed to two main factors: poor compatibility between IFC and current design software [31], and IFC’s inability to fully support the complete scope of an engineering domain [32].

A small portion of research has extended the IFC data application to the design phases of other engineering domains, including tunnel design [33], bridge design [34], and sound insulation design [35], among others.

3.1.2. IFC Applications in the Construction Phase

IFC applications in the construction phase refers to the utilization of IFC in processes and activities related to the construction domain. Through an analysis of the themes and keywords in the literature, the primary focus areas include construction scheduling, construction risks, and quality control, as well as the applications involved in the construction process. The relevant studies were primarily published within the past decade. The annual publication counts are illustrated in Figure 5.

Among them, research related to construction scheduling has been particularly active, focusing on two aspects: schedule generation and progress monitoring. In construction schedule generation research, IFC provides building component dimensional information and representing construction resource information through entities such as IfcLaborResource and IfcConstructionEquipmentResource [36] for construction planning. This can be further integrated with machine learning methods to generate construction schedules [37]. Progress monitoring research primarily investigates methods to compare IFC model information with actual construction data, including approaches such as photo data integration [38] and point cloud data comparison [39]. Additional studies focus on the integrated representation of progress information and model data, utilizing Ifcworkschedule entities to represent work plans [40] and enable construction progress updates [41]. However, IFC still has limitations in progress management, such as the inability to effectively distinguish between as-built and as-designed data, and the lack of specialized software supporting IFC-based progress management [40].

Construction risk research primarily focuses on risk assessment and prediction methods, where IFC plays a crucial role in information support. These studies encompass various applications, including integrating IFC with geological information to assess building settlement risks caused by tunnel construction [42]; extending IFC to represent deep excavation information [43] for pit risk identification; extending IFC monitoring-related property sets to meet safety risk assessment requirements in tunnel construction processes [44]; and incorporating IFC structures into fall protection system ontologies [45] for fall activity assessment. Research related to construction quality mainly addresses quality inspection during the construction process [46]. Although IfcProcess can represent project construction processes, IFC still lacks entities specifically related to quality management [7].

Applications during the construction process include construction space planning research, which utilizes geometric information of building elements provided by IFC [47]. In completion acceptance studies, IFC data are employed for comparison and updates between actual buildings and design models [48]. However, discrepancies exist between the IFC geometric system and measurement forms. IfcMapConversion represents the mapping of the map coordinate system; its use only allows the placement of objects at a single point (specified by easting, northing, and orthogonal height) and orients the object towards the initial direction (north). In practical applications, when the project scale is large, it is necessary to divide the project into several parts using local coordinate systems or to apply coordinate corrections [49].

IFC has also been implemented in other domains during the construction phase, including the development of information models for virtual construction [50] and the representation of tunnel excavation and support documentation [51], among others.

3.1.3. IFC Applications in the Operations and Maintenance Phase

IFC applications in the operations and maintenance phase refers to the utilization of IFC in relevant applications within the operations and maintenance domain. Through an analysis of the themes and keywords in the literature, the primary focus areas include facility management, health monitoring, and defect diagnosis. Additionally, another significant topic, path planning and fire evacuation, was also categorized into this phase. The annual distribution of publications is illustrated in Figure 6.

As the core component of the operation and maintenance phase, facility management research focuses on resolving information representation and integration challenges. On one hand, IFC needs to effectively represent information in facility management processes through the extension of entities and properties [52]. On the other hand, it is essential to establish links between IFC and facility management data. This integration can be achieved through database approaches [53], where IFC and facility management information are consolidated within a single database, enabling information management through database queries. Semantic data models have been developed [8] to create semantic links between building information and facility management data.

In health monitoring research, although IFC schema includes monitoring device entities such as IfcSensor, manual or programmatic integration of monitoring information with building in-formation is still required. Most studies extend IFC by incorporating monitoring data as properties, such as cyber-physical systems for monitoring device environments [54] and bridge monitoring sensor types [55].

Defect diagnosis research primarily investigates methods for representing defect information. This includes utilizing IfcSurfaceFeature to represent surface damage in bridge components [56] and describing defects through entities such as IfcElementAssembly [57], IfcProxy, and IfcVoidingFeature [58]. Additionally, some studies have developed dedicated ontology models to represent tunnel defect semantics [59].

IFC has also been extended to the operation and maintenance phase of other domains, including building automation control [60] and asset monitoring [61], among others.

Path planning and fire evacuation research has primarily emerged in the past 10 years, with path planning using building information as its core focus. In path planning applications, IFC provides geometric and semantic information of building components, including spatial elements such as walls, columns, doors, and windows [62], as well as spatial connectivity relationships through entities like IfcRelSpaceBoundary and IfcConnectionGeometry [63]. However, spatial connectivity relationships may be lost during software export [64], making geometric analysis a more reliable method for determining spatial relationships [65]. Additionally, researchers developed a crowd simulation ontology that integrates IFC structure for building evacuation scenarios [66].

In engineering applications, the primary function of IFC is information representation. IFC facilitates the description of concepts within the engineering domain through various forms of expression, thereby supporting the flow of information throughout the project lifecycle. By enabling different stakeholders to share information using a unified model, IFC significantly reduces the complexity of collaboration [30]. However, there remain numerous concepts that IFC cannot adequately represent, particularly those related to dynamic processes during the construction phase. In practice, most studies seek to overcome these limitations by extending IFC or customizing additional specifications to describe such concepts [33]. Nevertheless, these approaches often increase the complexity of information sharing.

3.2. Interoperability

In research concerning IFC in engineering domains, interoperability is also a crucial direction. The interoperability of IFC not only encompasses the need to satisfy diverse domain-specific expression but also demands seamless data exchange with other fields. The aforementioned analysis of IFC applications clearly indicates that interoperability remains the primary challenge in domain-specific IFC implementations. Accordingly, this study reviewed the literature related to IFC interoperability to examine its current research status. Through an analysis of the themes and keywords in the literature, interoperability research can be categorized into two aspects: information exchange and information integration. Additionally, a portion of the literature focused on evaluating IFC interoperability, which was also included in the statistical analysis. The annual distribution of publications is illustrated in Figure 7. As indicated in the figure, the trend in interoperability research corresponds to the overall trend of IFC studies.

In existing research, several studies have evaluated the current state of IFC interoperability, particularly focusing on the interoperability between IFC data and mainstream architectural engineering software. These assessments reflect IFC’s adaptability level within the contemporary built environment. The relevant studies encompass evaluations of IFC data compatibility with structural design software (including ArchiCAD, Tekla Structures, MagiCAD, and Autodesk Revit) [67,68], interoperability assessments with structural analysis software [69,70], compatibility evaluations with construction management software (such as ACCA Edificius, Autodesk Navisworks, and Synchro Pro) [71], and interoperability assessments in the LCA domain [26,72]. The evaluation results indicate varying levels of IFC content support across different software platforms: while geometric aspects such as component shape information are generally supported, the controllability of geometric representation remains insufficient. Moreover, regarding domain-specific concept support, the level of support in the structural engineering domain remains inadequate.

3.2.1. Information Integration

Research on information integration methods primarily focuses on the integration of IFC with information from other domains to achieve unified data representation during application processes. These studies emphasize semantic-level integration, aiming to represent data from different domains in a consistent format with IFC data. Semantic consistency in data representation is primarily achieved through semantic mapping [73] or transformation [74]. Among these studies, the integration with CityGML data standards from the Geographic Information Systems (GIS) domain has been extensively investigated [75]. Cross-domain data sharing is accomplished by establishing correspondence relationships between IFC and CityGML concepts [76]. However, due to the differences in application scenarios between these two standards, several key challenges need to be addressed during the transformation process, including coordinate system registration [77], geometric representation conversion [10], and Level of Detail (LOD) transformation [76]. Significant progress has also been made in data integration research across other domains, including 3D tiles standards [78], Transmodel in the transportation domain [79], and Building RoboAvatar in the robot domain [80]. During the integration process, researchers have supplemented and extended IFC specifications to address domains that were inadequately described in the original schema [81].

Beyond direct data format integration, researchers have proposed various indirect integration methods. These approaches include establishing associations through new model objects [82], achieving semantic integration through ontology models [83], and utilizing relational or graph databases [84,85] for data storage and integration.

The primary objective of various integration approaches is to achieve semantic consistency between IFC data and information from other domains. Most conceptual integration issues can be resolved through mapping techniques. However, for content that is not described within IFC data, mapping alone is insufficient. Such gaps must be filled using extend methods before mapping can proceed. Indirect integration methods can address these issues directly by incorporating IFC and domain-specific information into a different model. Nevertheless, this approach may result in the loss of original IFC semantics and requires the definition of additional rules to adequately describe certain concepts.

3.2.2. Information Exchange

Information exchange serves as a fundamental component of interoperability, enabling the transfer and sharing of IFC data across various domains. The process of information exchange involves two key steps: first, extracting the required information from IFC; second, converting IFC into domain-specific data formats to achieve cross-domain information exchange.

(1) IFC extraction

In the context of IFC extraction, Information Delivery Manual (IDM) and Model View Definition (MVD) are important technical pathways. While IDM defines information exchange requirements, MVD specifies the mapping relationships between required information and IFC schema, essentially functioning as a filter for project information. Related studies have primarily focused on developing domain-specific MVDs for various applications, including building performance simulation [86], building automation systems [87], steel fabrication [88], and seismic licensing [89].

The development methodologies for MVDs have exhibited diverse characteristics, encompassing the integration of existing MVD concepts for reuse [90], implementation of Extended Process to Product Modeling (xPPM) methodology [91], and ontology-based development of IDM and MVD [92]. Furthermore, some researchers have adopted alternative approaches by directly leveraging IFC ontology to extract exchange information through rule-based methods [93].

Although Model View Definition (MVD) serves as a template for defining IFC information exchange, its definition process is overly complex. Consequently, some researchers have developed approaches to directly extract required information based on IFC entity and property definitions. Several methodologies are list in

Table 2. IFC extraction method.

IFC Extraction Method	IFC-Based Solution	Ref.
No-Scheme IFC data extraction	All data instances referenced by IFC relationship entities are recursively extracted, together with the instances corresponding to the building elements selected by the user.	[94]
property-based IFC model extraction method	A property-based partial model view of IFC is defined, in which attributes specified in the partial model view definition are retrieved via IfcObject entities, while relationship attributes are accessed through relationship entities.	[95]
structural design information delivery method	The defined exchange requirements are mapped onto the IFC model, and an exchange model generation algorithm is developed to extract IFC data based on these requirements.	[96]
Selection Set-based IFC extraction method	The selection set is defined as a combination of IFC instances and rules and a general extraction language based on XML schema is proposed, which supports extraction requirements for object types, properties, and relationships.	[97]
instance-based IFC model extraction method	A hierarchical tree structure of IFC is constructed, and the tree nodes are iteratively traversed from the bottom up. Based on specific requirements, the physical IFC instances and attributes associated with IFC relationship entities are extracted during this process.	[98]

.

Extraction methods can be categorized into two types. The first type utilizes the MVD approach recommended by IFC, which enables comprehensive extraction of the entire IFC model. However, the process of defining MVDs is quite complex and requires project practitioners to have a deep understanding of IFC. The second type involves direct extraction methods, where information is retrieved from IFC based on specific requirements. This approach only necessitates the establishment of access rules for IFC, allowing targeted information to be extracted efficiently. However, the extracted data are limited to certain attribute information rather than the complete IFC model.

(2) IFC data conversion

In practical application scenarios, IFC data require effective conversion mechanisms with other data formats for information interoperability. The IFC data conversion also serves as foundational work for achieving information integration. Research on IFC transformation can be categorized according to various target data formats.

In the engineering domain, the primary focus is on data conversion for energy analysis and GIS applications. In the field of energy consumption analysis, the purpose of data conversion is to transform the material, geometric information, and other relevant content from IFC data into a format compatible with energy analysis engines. Current research primarily focuses on DOE-2 and EnergyPlus, which are the most widely used energy analysis engines. Some studies include the development of an interactive interface for IFC to Input Data File (IDF) conversion that enables users to verify and modify required energy analysis information based on identified IFC properties, ultimately generating IDF files compatible with EnergyPlus [99]; the generation of INP files compatible with the DOE-2 energy analysis engine, based on geometric and material data extracted from IFC data [21]; a conversion method based on object structure mapping between IFC and OpenStudio’s OSM files [23]; and a methodology for converting MEP equipment models from IFC to the IDF format [24], which first converts IFC to the IDF format containing only building geometry and material information, then analyzes the IFC to obtain HVAC system-related information, and finally creates HVAC objects within the IDF to generate a complete energy model.

Data models in the GIS focus on the expression and description of spatial relationships and geographic reference information. As one of the most influential data standards in this field, CityGML has played an important role in geographic information modeling and exchange since its adoption as an official standard by Open Geospatial Consortium (OGC) members in 2008. Currently, data conversion research in the GIS field mainly concentrates on IFC and CityGML. Researchers have developed various approaches, including the bidirectional mapping method between IFC and CityGML [100], where the core approach involves developing a semantic city model reference ontology as an intermediate ontology containing all entities and properties from both IFC and CityGML models; the conversion method based on Triple Graph Grammars (TGG) [101], which achieves conversion by establishing semantic and geometric relationships between IFC object graphs and CityGML object graphs; and the Feature Manipulation Engine (FME) conversion method [76], which serves as an Extract, Transform, and Load (ETL) tool to establish semantic mapping relationships between the two schemas and completes CityGML data construction through geometric conversion and link settings. Additionally, There is a conversion method for transforming IFC EXPRESS schemas into UML models consistent with geographic information standards developed by ISO Technical Committee 211 (ISO/TC 211) [102], which converts EXPRESS Schema and Entity to UML Package and classes with “FeatureType”, respectively, achieving standard conversion through data type mapping and semantic constraints.

In fundamental data representation, geometric data serves as critical information for the visualization of BIM. However, the current IFC geometric data format faces technical challenges such as rendering performance issues when applied in various 3D rendering engines. To meet the rendering requirements of geometric data in different 3D engines, several studies have investigated the conversion between IFC and 3D graphic representation formats. These include the IFC to 3D tiles conversion method [103], which decomposes IFC models into independent components through BIMServer API, converts them to OBJ format using IfcConvert tool, transforms them to glTF format via obj2gltf tool, and finally integrates them into b3dm format and constructs hierarchical tilesets through the improved 3D-tiles-tool, achieving complete preservation of geometric information and semantic properties; the open-source conversion method from IFC to Shapefile [104], which performs three-level coordinate transformation based on component geometric representation and position information, and proposes a polyhedron generation algorithm for sweep representations; its enhanced algorithm [105] further extends support for boundary representation, clipping, and mapping representation conversion, ensuring the generation of shapefile-compliant geometric shapes through different processing strategies; the conversion method for IFC clipping representation [106], which primarily addresses the geometric representation problem of multiple clipping with unbounded half-spaces by instantiating them into B-Rep representation through virtual boundary setting; and the conversion method based on the open-source computer graphics library OCCT (Open CASCADE Technology) [107], which converts IFC objects into temporary shape objects containing basic geometric elements, reorganizes them into OCCT B-Rep, and then converts them to Shapefile format while preserving key semantic information in attribute tables.

There are also studies on data conversion in other building-related fields, including the conversion between IFC and the Robot Operating System (ROS) simulation description format SDF [108], which extracts parameters such as placement position, dimensions, orientation, and height from its geometric representation to convert into Box elements in SDF; the conversion between IFC and the Universal Robot Description Format (URDF) [109], which parses IFC data to generate static building elements; and the conversion between IFC and the steel processing BVBS format [88], which extends entities and properties in the IFC standard and establishes data mapping relationships between IFC and BVBS.

IFC data conversion generally adopts two approaches. The first approach involves converting IFC files into a target format. This method is primarily used for energy performance data and fundamental data transformation. In this method, the required information is extracted from the IFC file and reformatted as needed. The main advantage of this approach is that only the necessary information is parsed, thus eliminating the need to process the entire IFC file. However, due to the lack of a comprehensive standardized property set in IFC, these methods can easily access geometric information, but it is challenging to reliably extract material and other attribute data. This limitation poses significant challenges for the development of IFC conversion methods [23]. The second approach focuses on bidirectional conversion between IFC and the target format, typically by establishing mapping relationships. This method is mainly applied in GIS data. It enables the creation of robust conversion mechanisms for concepts that are supported by both IFC and other data models. However, when certain concepts are not supported by one of the models, semantic information loss is inevitable [10]. Additionally, the absence of a LOD framework in the IFC model further complicates GIS data conversion [110].

(3) IFC Classification

During the information exchange process, semantic errors between IFC entity categories and actual component categories due to IFC data originating from different design software, consequently affecting interoperability. Therefore, some research efforts focus on addressing the IFC classification issues. Currently, almost all studies concerning IFC classification have adopted machine learning-related methods to improve classification accuracy and efficiency.

In terms of traditional machine learning approaches, several methods have been developed, including a Support Vector Machine (SVM)-based approach for IFC component matching [111] and an automatic classification system using random forest algorithms for prefabricated components [112]. The random forest-based method extracts feature information from IFC files, such as position, material properties, and component names, to implement automatic classification and coding for components. Additionally, a multi-algorithm classification approach [113] has been proposed that extracts six key features: Is_External, Is_Load_Bearing, length, cross-sectional area, volume, and number of Cartesian points. This approach integrates multiple algorithms, including SVMs, Artificial Neural Networks (ANNs), decision trees, and random forest algorithms for classification tasks.

In terms of deep learning approaches, Multi-View Convolutional Neural Networks (MVCNN), PointNet, and SVM have been proposed for IFC component classification [114]. Each component is captured from 12 different angles (at 30-degree intervals) as input for MVCNN, converted into point cloud data with 2048 points for PointNet processing, while SVM utilizes eight geometric parameters extracted from Boundary Representation (B-Rep) as feature input. Experimental results demonstrate that MVCNN performs exceptionally well in classifying door and wall components. Another approach involves the lightweight neural network model SpaRSE-BIM [115], which converts IFC models into point cloud data (2048 and 4096 points per object), enabling efficient component feature processing and accurate classification.

The current classification methods primarily focus on extracting features from the attribute information of IFC elements, as well as from images or point cloud data associated with these elements. By integrating image data with the semantic information provided by IFC, these approaches have achieved efficient IFC classification. Although image processing techniques are already well established and can be effectively leveraged, feature extraction from IFC semantic information remains limited to fixed attribute data. Significant implicit features in IFC, such as spatial relationships and inheritance relationships between objects, have not yet been fully considered. Addressing these aspects is crucial for further advancing classification methods.

3.3. Research on IFC Data Processing

After an in-depth analysis of Industry Foundation Classes (IFC) applications in the construction domain and interoperability, its core value in facilitating interoperability and data exchange between heterogeneous building information systems has been widely acknowledged. Although significant progress has been made in the implementation of IFC standards, its practical effectiveness as a data standard remains largely constrained by the operability of its data model. Based on these considerations, this section systematically reviews and discusses the research on IFC data processing from a data perspective. Through the analysis of literature topics and keywords, research on IFC data processing was categorized into aspects such as IFC extension, IFC storage and representation, and IFC generation. The annual distribution of publications is illustrated in Figure 8.

3.3.1. IFC Extension

IFC extensions primarily encompass three approaches: entity extension, property set extension, and type definition extension.

Entities form the foundation of the IFC structure, describing various physical objects and their properties through type definitions and inheritance mechanisms. Extending new entities requires adherence to IFC structural specifications [116], careful selection of base classes for inheritance, and the definition of domain-specific subclasses. For example, in the field of structural monitoring, sensor node subclasses are extended based on IfcDistributionControlElement [117], while in land management, spatial classes are extended under IfcSpatialElement [118]. buildingSMART, in collaboration with industry partners, has released extension standards for domains such as railways and roads [119], and new versions like IFC4.2 continue to incorporate content from emerging fields such as geology, railways, highways, and ports.

Property set extension enhances the semantic expressiveness of entities by defining new property sets [120]. This method does not require changes to the entity structure, offering high flexibility. The definition of property sets must comply with IFC naming and syntax conventions and support various data types, thereby meeting specialized information requirements.

Type definition extension supports more complex data descriptions by adding new value domains to existing attributes [13]. For instance, the association between IfcDoor and IfcDoorType allows the PredefinedType attribute to be extended to cover additional types. This can be achieved through proxy entities such as IfcProxy (removed in IFC4.3) and IfcBuildingElementProxy [121], which are used to represent concepts not yet included in IFC.

Each of these three extension methods has distinct characteristics: entity extension facilitates inheritance but must strictly follow EXPRESS syntax, which may impact standard compatibility; property set extension offers flexibility but has limited semantic expressiveness and a lower degree of standardization; type definition extension balances framework stability and flexibility, though excessive customization may compromise the uniformity of the standard.

3.3.2. IFC Data Storage and Representation

The IFC data model is characterized by its complex structure and extensive data quantities, which poses significant challenges for efficient data querying and storage. To address this issue, researchers have conducted extensive studies and identified the following storage and representation approaches:

(1) Relational database

Relational databases, with their mature technological ecosystem and widespread implementation foundation, emerged as the primary focus of early research in IFC data storage and querying. Initial studies employed a direct conversion method from EXPRESS entities to database tables, where each entity was mapped to a separate table with its attributes stored as table elements [122]. However, this one-to-one mapping approach imposed significant performance overhead on the relational model due to IFC’s inherent complex structure and extensive inheritance relationships, driving researchers to explore more optimized storage solutions. A simplified relational database schema [123] was proposed that categorized IFC types for storage and established table associations using ElementID as the primary key. In this schema, entities such as IfcProduct and their subclasses were consolidated into a single table, retaining only common attributes (e.g., GUID, name, and description), while other attribute information and material properties were stored separately in two additional tables. IfcRelationship entities were distributed across different tables based on their categories. Geometric data were managed independently through the Oracle Spatial database to facilitate geometric and spatial operations.

Object-oriented relational databases demonstrated distinctive advantages in handling inheritance and aggregation relationships within IFC. Research proposed conversion rules from IFC to object-based relational databases [124], where each specific EXPRESS type was mapped to a corresponding database type, and entities were converted into classes in standard SQL; IFC entities were reconstructed as IfcEntity objects [125], with instance files represented as IfcEntityInstance, enabling IFC file contents to be stored in object databases following this parsing approach.

In practical applications, most researchers adopted an approach of parsing and storing IFC data based on specific project requirements. For instance, in a construction quality assessment database [126], relevant data from IFC models and construction quality evaluation were extracted and stored in a relational database. This approach essentially restructured IFC data by establishing a database schema aligned with project-specific information requirements. The database design retained GUID as the primary key. Other examples include: a domain-specific relational database for building performance maintenance based on IFC structure with customized entity-relationship models [53], and a relational database storing physical properties of construction site objects parsed from IFC [127].

(2) Graph databases

As a data management system based on graph structures, graph databases are fundamentally composed of three basic elements: nodes, edges, and properties. Due to the complex object relationships and inheritance hierarchies in IFC data, which naturally align with graph structures, several researchers have proposed that graph databases are more appropriate for IFC data storage.

Various graph storage approaches have been implemented in IFC-related research. A 4D construction information model for prefabricated buildings [40] employed a simplified graph storage schema. In this approach, each entity from the instance file was represented as nodes with associated attributes values stored as key-value pairs, while relationship properties between entities functioned as relationship indicators. In a study integrating monitoring data with building information [84], graph structures were established through entity attributes traversal: entities were represented as nodes, reference-type attributes were transformed into edges connecting entities, and data-type attributes were represented through new nodes and relationships. In a BIM version control methodology [128], all IFC entities were represented as nodes, with numerical attribute values attached to nodes as key-value pairs, entity associations were represented by edges, with properties defined the relationships between nodes. Another approach involved converting IFC to LPG (Labeled Property Graphs) [129], where three types of entities were transformed into nodes: relationship entities, entities without GlobalID containing only value-type attributes, and entities whose attribute values were reference-type entities. Attribute relationships between entities were converted to edges, text-type attributes were directly assigned as node properties, and entity inheritance structures were represented as node labels.

In general, the pattern for storing IFC data in graph databases is to represent entities as nodes and entity attributes as edges. The main differences among various approaches primarily lie in the handling of numerical attribute values, where some studies selected to store these as node properties to avoid creating additional nodes.

(3) RDF and ontology

As one of the significant approaches for Building Information Model (BIM) data storage and representation, Resource Description Framework (RDF) and ontology technologies have received significant research focus in the academic community due to their advantages in semantic expression. RDF, recommended by the World Wide Web Consortium (W3C) as a standard metadata framework, employs a triple model consisting of Subject, Predicate, and Object to precisely describe semantic relationships between resources, thereby forming a structured semantic network. Ontology technology establishes a knowledge system with explicit logical specifications by formally defining domain elements such as concepts, relations, properties, and constraints, thus providing technical support for cross-domain knowledge integration and semantic reasoning.

Based on these characteristics, researchers have actively explored the feasibility of transforming IFC standards into RDF and ontological representations. The officially released ifcOWL ontology was developed through the conversion methodology proposed by Pauwels [130]. This methodology transforms the IFC schema, defined in EXPRESS language, into Web Ontology Language (OWL) format, enabling ontological expression while preserving the original semantic structure. However, this work left several issues unresolved. Subsequent research has conducted multifaceted optimization studies on ifcOWL, including converting IFC rule constraints into class expressions to enhance semantic reliability [131], adopting Well-Known Text (WKT) to optimize geometric data representation [132], and implementing standardized rule validation through Shape Constraints Language (SHACL) [133].

In research on IFC storage using RDF and ontology approaches, various methodologies have been developed: the conversion of IFC data to ifcOWL and Building Topology Ontology (BOT) via the IFCtoLBD converter [134]; there are studies focusing on IFC data representation based on ifcOWL. In the semantic alignment between IFC concepts and regulatory concepts [17], a novel IFC knowledge graph was proposed. While the overall structure of the knowledge graph is based on ifcOWL, it differs in that all predefined types within IFC entities are connected as child nodes to their corresponding entity nodes, enabling a more granular classification of entities. In the application of natural language to information retrieval [135], the IFC Natural Language Expression (INLE) ontology was created to complement ifcOWL, mapping natural language queries to the IFC ontology and enabling the transformation of natural language queries into SPARQL queries on RDF-based IFC models.

RDF, which is inherently a graph structure, demonstrates superior capabilities in data sharing and linking compared to other graph-based IFC representation methods, offering significant advantages in supporting Semantic Web technologies and reasoning capabilities. However, due to its non-native graph database characteristics, its graph traversal efficiency is relatively low. In contrast, while Labeled Property Graphs (LPG) can fully preserve IFC data structures and support efficient graph traversal algorithms, they have limitations in cross-domain data sharing [136].

(4) Other representation methods

There are several other approaches for IFC storage and representation. The MongoDB document-based database approach [137] decomposes the IFC model into five components, which serve as the basis for serializing IFC entities into documents. For IFC-to- Hierarchical Data Format 5 (HDF5) storage representation [138], data types in EXPRESS structures are represented through HDF5 Compound datatype, while HDF5 Groups and datasets are utilized to encapsulate the serialized IFC data.

Among the various methods for storing and representing IFC data, the relational database approach can decompose the IFC structure and store it in the form of tables. This method allows for the comprehensive storage of all IFC information. However, due to the inherently complex structure of IFC and its numerous interrelationships, relational databases often encounter challenges in terms of query and storage efficiency. Graph databases, on the other hand, are well suited to the structure of IFC, enabling the complete storage of inheritance relationships between IFC entities and providing a clear reflection of the overall IFC architecture. Nevertheless, the substantial amount of redundant geometric information contained in IFC files, when stored in graph databases, can lead to significant resource consumption, especially when processing large-scale IFC files. In the context of RDF and ontology representations, the official release of ifcOWL has offered a viable alternative for applying IFC within semantic web and knowledge graph environments, and continuous improvements are being made to this standard. Other storage methods primarily focus on addressing specific shortcomings of IFC storage, such as enhancing parsing efficiency [138]. However, these approaches are often difficult to scale and implement in real-world project applications.

3.3.3. IFC Validation

In engineering practice, IFC frequently encounters issues such as inconsistent data quality, incomplete model integrity, and imprecise syntactic expressions. To ensure data reliability and application effectiveness, several researchers have conducted studies on IFC validation.

IFC data validation primarily encompasses three levels: syntactic validation, semantic validation, and design requirement validation [139]. Specifically, syntactic validation, based on EXPRESS language specifications, covers compliance checking of data structures, entity relationships, property types, and constraint rules including GLOBAL, UNIQUE, and WHERE clauses. Semantic validation focuses on the semantic and syntactic conformance to Model View Definitions (MVD). Design requirement validation extends to regulatory compliance at the engineering specification level. Regarding validation standard refinement, a four-dimensional quality assessment framework has been established [140]. This framework consists of data structure validation, which aims to maintain syntactic correctness of IFC files and the integrity of geometric-topological structures; semantic validation, which emphasizes the rationality of data behavior; MVD conformance validation, which ensures models meet specific exchange requirements; and import/export integrity validation, which focuses on data fidelity during transfers between BIM tools.

Regarding specific validation methods, the MVD-based IFC validation approach [141] evaluates whether IFC instance files comply with the syntax and semantics established in MVD modular concepts according to different types of MVD rule sets. This validation method encompasses six types of verification. Similarly, the MVDlite algorithm [142], based on MVD validation for IFC, introduces a “rule chain” structure to integrate and reorganize templates and rule statements from mvdXML rule sets. However, due to IFC’s inherent nature as a data exchange standard rather than a modeling standard, MVD as its architectural subset exhibits limitations in aspects such as validation scope definition, rule logic formalization, and validation process standardization [143].

In terms of geometric data validation, a method for verifying spatial boundary geometric errors in IFC models [144] employs a Monte Carlo-based boundary detection algorithm to examine geometric errors, including gaps between spatial boundaries, overhangs, overlaps, and incorrect surface normal orientations.

IFC validation primarily aims to ensure the correctness and completeness of IFC data. Most studies focus on the validation of Model View Definitions (MVDs), which can assess whether an IFC file meets specific requirements. However, the lack of clear and quantifiable definitions within IFC leads to inconsistencies in both the creation and interpretation of IFC data [140], making it challenging to standardize validation rules.

3.3.4. IFC Comparison

IFC comparison involves analysis of IFC instance files to identify differences between them. Several approaches have been proposed in this work.

A metric framework for quantitatively evaluating IFC file differences was developed [145], which achieved automated model comparison through an independently developed tool (CompareP21). This research provided a theoretical foundation for the quantitative expression of IFC model differences, although the detailed implementation of its comparison algorithms has not been fully disclosed. An IFC model comparison method based on RDF graphs [146] calculates differences by applying three algorithms to generate signatures for blank nodes. An automatic content-based IFC file comparison method [147] constructs a hierarchical structure of IFC, eliminates redundant instances, and compares terminal nodes. The comparison proceeds iteratively from bottom to top until reaching the root node. An IFC submodel comparison method [148] removes redundant information from IFC models and generates signatures. Changes are then extracted by comparing signatures through breadth-first search based on the directed acyclic graph representation of IFC.

The advantage of these methods lies in the fact that they rely almost entirely on the content of the IFC files for comparison, rather than depending on GUIDs. However, due to the presence of numerous redundant nodes and the inherent complexity of IFC structures, even after processing and optimizing the IFC data, the hierarchical structure remains unchanged. As a result, the improvement in comparison efficiency is still limited.

3.3.5. IFC Compression

IFC compression aims to reduce IFC file storage size while maintaining model information integrity. Several relevant research approaches have been developed.

The content-based compression algorithm [149] analyzes and eliminates identical entities based on IFC’s hierarchical structure, followed by reassigning all reference numbers. In this approach, frequently referenced entities are assigned smaller numbers. The reference-based IFC compression algorithm [150] constructs subgraph structures for each IFC instance and performs depth-first traversal to remove redundant instances while updating their reference numbers. Unlike the content-based compression method [149], this approach builds subgraphs for individual entity instances rather than a complete IFC model graph. The IFCXML file compression method [151] follows a similar approach to reference-based IFC compression algorithm [150] but focuses on identifying duplicates among independent entities (entities that do not reference others) before removing redundancies in their parent instances based on reference relationships.

IFC compression methods typically involve removing redundant IFC information, reorganizing IFC entities, and reducing the overall file size. However, these approaches often overlook the potential impact of changes in entity reference order when determining redundancy [150].

3.3.6. IFC Generation

As a widely recognized standardized data schema in the architecture and engineering domain, IFC provides a normalized data foundation for information exchange and sharing throughout the building lifecycle. Therefore, creating IFC models has become one of the essential requirements in building informatization. Literature analysis indicates that current IFC generation research primarily focuses on two directions: one is IFC model reconstruction based on point cloud data, and the other is IFC model generation based on CAD drawings.

In the transportation infrastructure domain, research on Mobile Laser Scanning (MLS) point cloud conversion into standardized road IFC models has been conducted [152]. This approach identifies key elements such as road centerlines, traffic signs, and guardrails through geometric feature recognition and achieves standardized IFC entity construction using the xBIM toolkit. For bridge infrastructure, researchers have developed methods for generating IFC models of truss bridges from point cloud data [153]. This method processes truss components in point clouds using Bounding Box technology to generate information models compliant with IFC 4.1 standard. Studies on railway infrastructure have explored IFC modeling methods based on Airborne Laser Scanning (ALS) point clouds [154]. The approach identifies rails and track bed components through point cloud segmentation, determines geometric forms through track model fitting, and generates standardized IFC models using IfcOpenShell. The methodological approach for IFC generation from CAD drawings primarily consists of two steps: component recognition through deep learning techniques and subsequent IFC model generation based on the recognized information. This approach has been applied to recognize components such as walls, doors, and windows in architectural drawings [155] and to identify components like pipelines in MEP drawings [156].

Although both IFC generation and IFC conversion aim to obtain IFC models, IFC generation focuses on constructing IFC models from unstructured or heterogeneous data sources (such as point clouds, CAD drawings, etc.), while IFC conversion involves standardized translation to IFC format based on existing structured data, primarily addressing data interoperability issues across different domains. Literature analysis reveals that the challenges in current IFC model generation research primarily lie in raw data processing and component information extraction. Research on IFC data itself mainly focuses on its semantic and geometric information expression mechanisms; the final generation of IFC models relies on mature open-source toolkits such as IfcOpenShell and xBIM.

3.3.7. IFC Merging

IFC merging refers to the integration of separate IFC model files into a complete IFC model. One approach is the graph structure-based IFC data merging method [157], which represents IFC structures as graph structures and identifies identical data across different IFC files through maximum common subgraph mining, followed by merging the non-identical portions. This method fully represents the structure of IFC; however, it does not perform redundancy processing on the IFC graph, which may affect the efficiency of merging during the mining process.

4. Discussion

4.1. Development Path of IFC-Related Research

As the most influential data standard in the AEC domain, IFC has evolved for nearly 30 years since the release of IFC1.0 by the IAI in 1997 and has now been updated to the official version of IFC4.3. Its application scope has expanded from the construction domain to related fields such as autonomous driving and robotics, which fully demonstrates the significance of building data standards. From the temporal distribution of research publications, IFC-related studies showed relatively steady growth during 2004–2014, followed by a significant increase from 2015 to 2020. This phenomenon is closely associated with the widespread attention given to BIM technology and the digital transformation of the construction industry. It is noteworthy that after reaching a peak during 2020–2022, the number of publications has tended to stabilize. This trend may be attributed to researchers’ growing recognition that IFC only serves as a data standard in building intelligence processes, while addressing practical engineering challenges remains the core task. Additionally, the inherent limitations of the IFC data standard are unlikely to be fundamentally resolved in the near term.

From the distribution of research topics, IFC application-oriented studies dominate the field, accounting for approximately 48%. Studies related to interoperability and data processing account for 36% and 35%, respectively. It should be noted that the total exceeds 100%, as application-oriented studies often involve both data processing and interoperability topics. This distribution is evident within the context of the architecture, engineering, and construction (AEC) industry. IFC is a data schema specifically developed for building information; research on its applications is inherently rooted in the architectural domain. The primary focus of IFC application research is to address challenges within the engineering domain, rather than issues inherent to the IFC data itself. Nevertheless, this line of research remains highly significant for the advancement of IFC, as it often uncovers practical problems encountered during implementation. These findings provide valuable insights and references that can inform the further development of the IFC standard. Research on IFC processing tends to align more closely with data handling studies in the field of computer science. Such research primarily focuses on addressing issues related to the use and manipulation of IFC data. However, for researchers whose background is rooted in AEC industry, these topics are typically not the main focus of their investigations. From the perspective of construction project phases, design phase studies are the most abundant. This is attributable to the design phase being the primary stage of building information generation, where the accuracy and completeness of semantic representation directly influence downstream applications. The operation and maintenance phase ranks second, primarily focusing on the in-depth application of IFC data. Research in the construction phase is relatively limited, mainly because IFC is better suited for static building information description, while its expression mechanisms for dynamic construction data and process-related information remain inadequate.

Interoperability, as IFC’s core value proposition, has received continuous attention throughout the entire development process of IFC. From the perspective of research topics, whenever information exchange and sharing across different domains occur during application, issues of information exchange and integration inevitably arise. Most studies have used IFC data as the foundational data format to achieve interoperability with traditional domains such as GIS, energy consumption analysis, and geometric data through data conversion. Most of these studies were conducted before 2020. However, interoperability challenges in this context are not solely attributable to the IFC standard itself but are often influenced by factors such as the level of support provided by authoring software. This has led to the emergence of research topics focused on IFC classification. Ideally, the semantics defined by IFC should be clear and perfectly aligned with the models used in authoring software. In practice, however, the mapping between these two components remains incomplete, necessitating additional semantic interpretation of models within the software. Clearly, this gap poses a significant obstacle to achieving true interoperability. Since 2020, in addition to studies focused on IFC classification, the scope of interoperability research has expanded to encompass emerging fields such as robotics.

Regarding IFC data processing research, IFC extension studies have spanned various periods to meet the information representation requirements across different domains. This further demonstrates that the current IFC does not sufficiently support the information representation needs within the AEC industry. In terms of data representation and storage, the research focus has shifted from traditional relational databases and document-oriented databases (pre-2020) to graph databases (2020–2024). IFC representation research based on RDF and ontologies began over 10 years ago, with the ifcOWL version released in March 2017 marking a significant milestone in this direction. The methodologies and technologies adopted in these studies are consistent with the prevailing trends in technological development. IFC validation work was predominantly conducted before 2020, focusing on syntax and semantic correctness validation, as well as MVD related verification. This is mainly attributed to the limited compatibility between IFC and currently available authoring software. In the absence of robust application software, it is challenging to use IFC data independently in projects. Consequently, the practical impact of validation efforts is limited. Other operation-related research, including compression, comparison was predominantly carried out before 2020, as researchers had become familiar with IFC’s data organization patterns and gradually implemented the basic operations.

From a technical methodology perspective, IFC classification and generation research has widely adopting advanced technical approaches such as machine learning. The relevant literature has emerged since 2020. Through the analysis of keyword citation bursts as shown in Figure 9, it is evident that deep learning has emerged as the focal points in current IFC-related research. This trend reflects the active integration of contemporary advanced technologies into IFC research, indicating a transformation from data standardization to intelligent processing.

4.2. Limitations and Challenges in IFC Development

Although IFC has been proposed and applied for more than 30 years, it is still challenging to achieve widespread adoption across various domains. Moreover, research trends indicate a growing focus on domain-specific application issues, while studies dedicated to IFC data itself are gradually declining. Accordingly, this paper conducts a comprehensive analysis of the relevant literature to identify the factors limiting the development of IFC, thereby providing guidance for future research.

4.2.1. Inherent Limitations of IFC Data Format

Although IFC has been established as the universal data standard for BIM, its core data model remains primarily centered on traditional architecture engineering domains. One of the major challenges in the application of IFC is its insufficient support for building-related disciplines. This limitation is also evident in the literature, where most studies involving the application of IFC to other domains have focused on extending the standard to meet specific requirements. buildingSMART has also worked to advance domain extensions through collaboration with national organizations. For instance, the China BIM Alliance has developed IFC Rail SPEC, and Korea’s KICT has proposed IFC Road SPEC. However, these extension schemes have not yet been incorporated into the official standardization system. IFC 4.3 has introduced new entities such as IfcRailway and IfcTrackElement for the infrastructure domain; these additions still fall considerably short of supporting the representation requirements of actual project implementations. The insufficient support for domain-specific information has made it difficult for IFC to be widely adopted in the intelligent development of additional fields.

IFC provides a comprehensive descriptive framework for building information modeling, it still has certain limitations in terms of representation. IFC primarily focuses on static building objects, and its data model lacks native support for dynamic information such as construction progress variations and equipment operational parameter flows. IFC introduces the IfcTask entity for task scheduling, its original design primarily targets project progress management and may face performance and efficiency challenges when handling real-time monitoring data. Additionally, the definitions of certain entities (particularly relationship entities) are overly broad and lack contextual constraints for specific use cases, making it difficult for users to determine their applicability. The diversity of geometric representations offers flexibility; the absence of clear conversion rules between different forms leads to data redundancy and precision loss. This is an inevitable challenge in standard development, as it is difficult to achieve both flexibility and strict standardization simultaneously.

EXPRESS, serving as the modeling language of IFC, introduces complexity to IFC standard development. Any modifications to IFC must strictly adhere to the existing hierarchical inheritance mechanism. Furthermore, IFC files based on the STEP physical format employ explicit text encoding, leading to file size inflation due to their deeply nested structure, particularly in geometric representations.

4.2.2. External Implementation Limitations

The practical implementation of IFC is constrained by software vendors’ differentiated implementation strategies. A large proportion of IFC data originates from these software applications. However, major BIM software applications (e.g., Revit and Archicad) demonstrate varying levels of compatibility with the IFC standard, and semantic mapping errors frequently occur. As a result, the lack of an authoritative, certified IFC model library limits the availability of large-scale data required for future applications. On the other hand, the development of IFC parsers requires in-depth understanding of both EXPRESS language and domain semantics, making it challenging for software vendors with limited resources to achieve comprehensive support. Consequently, the persistent inadequate IFC support from software vendors negatively impacts the market acceptance of the IFC data format, leading to reduced adoption rates and creating a vicious cycle.

4.3. Future Research Directions and Recommendations

With the rapid development trend of building informatization, IFC, as a data standard for building information, is expected to attract increasing attention and undergo further development by researchers. Among its current challenges, the primary limitation of IFC lies in its insufficient capability to represent other domain information. However, from the perspective of theoretical and methodological research, these limitations do not prevent IFC from serving as the underlying data format for the implementation of more intelligent applications and methods. Therefore, considering the existing limitations and research gaps, future development of IFC can be focused on the following aspects:

Semantic Enhancement of IFC: A significant limitation of IFC is its inability to fully support information within various domains. Therefore, it is necessary to explore methods to extend IFC’s capability to represent knowledge from a wider range of domains. This can be achieved by leveraging Semantic Web and ontology technologies to construct a more refined IFC ontology. Although ifcOWL already exists, the conversion from IFC to ifcOWL still faces several challenges, such as the conversion of constraint rules. It is recommended to enrich the IFC ontology with additional domain-specific definitions and rules, and to further clarify the ontological descriptions of IFC objects, properties, and relationships, thereby enhancing machine readability and reasoning capabilities. Furthermore, integration with knowledge models from other domains, such as GIS and the Internet of Things (IoT), should be considered to achieve semantic-level data interoperability and knowledge sharing during practical applications.

Development of Standardized Datasets: The lack of standardized datasets has constrained the advancement of IFC applications in artificial intelligence. Based on current domain requirements, it is essential to investigate the establishment methods of standard IFC model libraries at different levels and their data quality assessment systems. The standardized IFC model libraries should include at least: (1) component-level datasets for characterizing the features of various component entities in IFC; (2) spatial-level datasets for describing topological relationships between spatial components; and (3) building-level datasets for providing standardized representations of complete building instances across various domains. At the technical level, these datasets can provide structured training samples for deep learning algorithms, addressing the insufficient generalization capability of AI models in the architectural domain. At the standardization level, they can enhance the semantic interoperability of IFC models and support model validation and consistency checking.

Optimization of IFC Feature Representation Methods: With the growing number of machine learning applications, optimizing the methods for representing features in IFC data has become increasingly important. Current studies predominantly focus on image processing of IFC geometric graphics, which overlooks the rich semantic information inherent in IFC data. Therefore, future research should investigate IFC feature extraction methods adapted to different application scenarios and IFC data vectorization representation methods. Additionally, it is important to develop representation learning models capable of comprehensively capturing geometric, semantic, and relational features, with the aim of enhancing the applicability of IFC data in machine learning tasks.

Development of Standardized IFC Processing Tools: To lower the barriers to IFC adoption, it is necessary to develop intelligent and standardized IFC processing tools. Although several IFC parsers are currently available, their functionality is mainly limited to the display and presentation of basic information. Standardized IFC processing tools should not only be capable of understanding and handling the semantic information within IFC files but also support more advanced model operations based on this semantic understanding. Furthermore, research should focus on methods for processing large-scale IFC data, including the exploration of distributed computing and cloud computing technologies to enhance the effectiveness of IFC data processing in practical applications.

Finally, from the perspective of domain-specific applications, the design phase can focus on IFC-based generative design methods, which integrate data from various disciplines to support the generation of design solutions tailored to different requirements. In the construction phase, research efforts focus on IFC-based descriptions of construction sequences and methodologies to enable comprehensive digital representation of construction processes. For the operation and maintenance phase, the seamless integration of dynamic monitoring data with IFC models is crucial, allowing the IFC model to represent the real-time operational status of buildings and supporting the exploration of more Digital Twin applications.

5. Conclusions

As a widely adopted data standard in AEC, IFC has played an important role in promoting engineering informatization and collaborative innovation. This study comprehensively reviews the research progress of IFC standards over the past 20 years, revealing its development patterns and limitations from three dimensions: engineering applications, interoperability, and data processing methods.

From the overall trend perspective, IFC-related literature reached its peak during 2020–2022, which highly correlates with the development process of BIM technology and civil engineering digitalization. Subsequently, research focus has gradually shifted towards specific engineering domain issues, with publication numbers showing a moderate declining trend. Analysis of engineering application research reveals a differentiated distribution pattern across project phases. The design phase, as the core stage of engineering information creation, accounts for the largest proportion of research; the operation and maintenance phase ranks second, while research in the construction phase is relatively weak, reflecting the current standard’s limitations in dynamic process description and real-time data support. The interoperability aspects of IFC have maintained continuous academic attention throughout the entire research cycle. In terms of data processing research, early studies primarily focused on data processing-related research, such as data validation and comparison. With the rapid advancement of artificial intelligence technology, IFC data generation and classification based on machine learning have emerged as new research hotspots, reflecting the development trend of integrating IFC standards with emerging technologies.

Subsequently, two limitations of IFC development are summarized: (1) inherent limitations of the IFC data format, primarily reflected in its insufficient support for information representation in the current AEC industry and issues related to its information representation schema; (2) external implementation factors, particularly the interoperability barriers presented by the software products currently available in the market.

Finally, the future development directions of IFC were discussed based on the analysis of IFC limitations and research gaps. The research directions of IFC focus on the following aspects: the semantic enrichment of IFC, development of standardized IFC datasets, optimization of IFC feature representation methods, development of IFC standardization tools, and application-oriented research on IFC across different project phases.

Although this paper reviews the core IFC-related journal articles published since 2004, it does not include the literature from other databases such as SCOPUS, nor does it cover conference papers, books, or other types of publications, which presents certain limitations. Furthermore, this study does not analyze citation author relationships and thus lacks an assessment of the contributions made by different research teams to the development of IFC. Future research should pay closer attention to the progress of various research groups.

As the most influential data standard in the architecture and engineering domain, IFC has gained widespread industry recognition. Under the trend of data-driven intelligent transformation, IFC data serve as a fundamental cornerstone for the advancement of the construction industry. This review systematically analyzes the current limitations of IFC from multiple perspectives, aiming to provide a comprehensive view for addressing existing data-related challenges. The purpose is to facilitate the deep integration of the construction engineering industry with emerging technologies such as artificial intelligence during future development, thereby enhancing the level of intelligence and automation in the engineering field.