From Building Information Modeling to Building Energy Modeling: Optimization Study for Efficient Transformation

Abstract

The conversion from building information modeling (BIM) to building energy modeling (BEM) based on the industry foundation classes (IFC) data standard is a crucial step for efficient building energy design and energy performance analysis. The scope encompasses analyzing limitations in existing BIM-to-BEM workflows and proposing an optimized strategy that addresses data loss and modeling inconsistencies. The research question revolves around enhancing conversion efficiency and precision, with the hypothesis validated through literature review, development of a conversion tool, and case study verification. The data collection and evaluation methods involve streamlining the conversion process by incorporating BIM model optimization, automatic repair of damaged geometric information, and automatic thermal zone division. The main findings reveal that the optimized strategy and tool significantly reduce information duplication, improve the precision of energy simulations, and validate the hypothesis, thereby contributing to more efficient and accurate building energy design and analysis.

1. Introduction

1.1. BIM and BEM Conversion

In the early 21st century, Autodesk, an American company, first introduced the concept of building information modeling (BIM). BIM is currently a popular information technology in the construction industry and is widely used in all phases of building planning, design, construction and operation. However, the functionality of BIM extends beyond these applications, with its data offering expansive possibilities in various peripheral domains [1]. Reducing energy consumption and managing energy use and carbon emissions over the entire lifecycle of a building are critical goals in the green development of the construction sector. BIM has emerged as an effective solution for integrating and managing information throughout the building lifecycle. To achieve energy consumption reduction and low-carbon management, it is essential to collect and analyze energy consumption data from buildings. Therefore, BIM encompasses various information necessary for building energy analysis. Accurate performance analysis during the early design stages can significantly save costs, time, and labor [2].

Moreover, during the project design phase, designers often need to simulate projects to predict the green and low-carbon performance of the design scheme. The design process requires seamless interoperability between design software, simulation software, and performance analysis tools, enabling smooth data and model conversions.

On the other hand, leading building energy consumption calculation engines include DOE-2, Energy Plus, and TRNSYS from the US, ESP-r from the UK, HASP from Japan, and DeST from China [3]. Among them, Energy Plus is the most used tool for building energy simulation, which offers precise simulations, a modular structure, and excellent scalability [4]. However, its input and output are in ASCII text format, requiring various variables and professional operation. Based on the situation that no direct interface medium can import Energy Plus simulation, the research community has developed several graphical user interface simulation software using Energy Plus as the simulation engine, such as common user interfaces Open Studio and Design Builder. The typical modeling process in Energy Plus involves manual redrawing based on BIM models or CAD construction drawings, consuming considerable time and effort from simulation personnel. Additionally, the building energy model recognized by Energy Plus is a closed, single-unit space with thermal zones, supporting only rectangular surfaces and not curved or arcuate surfaces. This may lead to issues such as information loss during BIM data conversion [5]. Therefore, addressing the BIM-to-BEM conversion is crucial for efficient energy consumption simulation [6].

The transition from BIM to BEM mainly involves three components: BIM modeling, interactive documents, and BEM simulation software [7]. The core is to utilize interfaces or plug-ins between BIM and BEM software to automatically convert data formats [8]. Issues arising from the conversion process may involve any or all three components, not just energy consumption simulation applications [9]. Recently, several studies have focused on improving the BIM-to-BEM conversion process. For instance, Kamel reviewed the applications of BIM in energy simulation, highlighting issues and solutions related to data exchange [10]. Truong investigated the BIM-to-BEM transition for optimizing envelope design, emphasizing the importance of efficient and cost-effective building energy performance [6]. Abd-Elnaby further discussed the integration of BIM and BEM for efficient data exchange and energy performance assessment [11].

The quality of BIM modeling also affects the completeness and accuracy of information contained in the exported interaction files. Fewer modeling errors in the early BIM stage led to fewer geometric information errors in the exported interaction files. However, modeling standards and individual requirements vary among designers or engineers, making BIM modeling errors unavoidable. Building energy simulation tools requires simplified, regular models with specific information parameters filled in. However, BIM software often creates complex models, especially in the early design stages where multiple design changes occur, leading to inconsistencies between the design and the model. Most BIM models established during the design process are not suitable for energy analysis. Therefore, model adjustment and simplification are necessary before exporting the interaction medium.

The conversion between BIM and BEM relies on data exchange files for information transfer. Currently, the commonly used interactive file formats, industry foundation classes (IFC) and gbXML, are supported by most energy simulation software [12]. IFC, first proposed in 1995, is a BIM data standard issued by the International Organization for Standardization, which can comprehensively describe information related to various disciplines of architectural engineering [13]. A large number of software tools support the IFC file format [14], which has become the de facto standard for international construction data exchange. In 2002, Hitchcock from the Lawrence Berkeley National Laboratory (LBNL) proposed using IFC as a standard format for interacting with energy simulation calculations, enhancing interoperability between design software and energy simulation software. IFC2 × 3 and IFC4 can be used for building energy consumption simulation. IFC4 is the latest version of IFC, released in 2013, which improves the efficiency and cross-platform interoperability of IFC. IFC2X4 contains the most comprehensive BIM data standards. This version of IFC has about 800 entities, 358 attribute sets, and 121 data types. Although these numbers indicate the complexity of IFC, they also reflect the richness of semantic information construction. Although IFC has very good building performance analysis data support capabilities, most building energy consumption analysis software still has problems importing or using IFC [15], while the gbXML data standard, funded and developed by organizations such as the US Department of Energy, NREL, Autodesk, ASHRAE, and Bentley Systems in 2000, is an open BIM standard with good building information-sharing capabilities. gbXML can describe all the attributes of a building, containing 346 elements and 167 simple types. The gbXML standard has become a recognized data standard in the field of building green performance analysis [16]. However, gbXML was proposed relatively late. According to official statistics from the IFC standard and gbXML standard organizing body, there are currently 45 types of design and modeling software that support IFC data export, while only 15 support gbXML data export. In addition, Lin analyzed the description mechanism of IFC and gbXML for green performance information, sorted out the information description standards of the two, and established the entity mapping relationship between the IFC model and the gbXML model, providing a research foundation for data conversion (

Table 1. Mapping relationship between IFC and gbXML.

Mapping Relationship between IFC and gbXML
Sort	Item	IFC Entity/Attribute	gbXML. Element/Attribute
Basic information	Unit of measurement	IfcUnit	lengthUnit, areaUnit, volumeUnit, temperatureUnit
	Due north	IfcGeometricRepresentationContext, True-North	Azimuth
	Document history	IfcOwnerHistory	DocumentHistory
	Geographical position	IfcSite, RefLatitude, IfcSite, RefLongitude	Location
Spatial organization structure information	Site	IfcSite	Campus
	Building	IfcBuilding	Building
	Floor	IfcBuildingStorey	BuildingStorey
	Space	IfcSpace, IfcZone	Space, Zone
	Space boundary	IfcRelSpaceBoundary	SpaceBoundary, Surface, AdjacentSpaceld
	Space hierarchy	IfcRel Aggrega tes	Campus, Building
	Relation	IfcRelContainedInSpatialStructure	Building, BuildingStorey, Building. Space
	Component	IfcBuildingElement and subclasses	Surface, surface TypeEnum
	Equipment	IfcDistributionElement and subclasses	LightingControl, AirLoopEquipment, cLoopEquipment, Hyfroni-cLoopEquipment
	Greening and roads	IfeGeographicElement	Vegetation, Transportation
Material information		IfcMaterialDefinition, IfcMaterial, IfcMar-terialLayer, IfcMaterialLayerSet	Material, Layer, Construction
Other information	Planned cost	IfeTimeSeriesSchedule	Schedule, YearSchedule, WeekSchedule, DaySchedule
		IfcCostItem, 1fcCostValue	Cost

) [17].

Through research, it is found that most energy consumption simulation software (Open Studio, Design Builder and GBS) use gbXML as the import file for information exchange, and a few use IFC files. There are also various problems and deficiencies in the interaction process. For instance, exporting a gbXML file from Revit and importing it into GBS only contains building geometry information without thermal information for enclosing structures. When exporting a gbXML file from Revit with the “Rooms/Spaces” option and importing it into Open Studio, the conversion process only includes relatively complete geometry and Zone information, lacking HVAC-related data. Although exporting with the “Energy Settings” option provides more comprehensive information, the model often has large broken surfaces that are difficult to repair during visualization in Open Studio (Figure 1). eQUEST requires indirect import of gbXML through third-party software [18]; Energy Plus also needs third-party software to indirectly import gbXML and IFC [19,20].

Some scholars have also studied the conversion from BIM to BEM. For instance, Yang reviewed existing BIM-to-BEM conversion methods and proposed a method that converts BIM models to 2D graphics and then automatically reconstructs 3D models using BIM data information [21]. Nevertheless, this method performs all geometric calculations in a 2D plane, making it unsuitable for handling special enclosing structures like tilted walls and roofs in engineering buildings. Liang suggested that gbXML is the best choice in the field of green building energy consumption simulation [22], whereas gbXML often encounters component missing issues during data export. Thus, improving gbXML information data by leveraging the advantages of IFC data was proposed. Through the evaluation of the information import process and the parameter setting function of software energy consumption simulation process of six BIM-based energy consumption simulation software, including Green Building Studio, PKPM-Energy, HY-EP, DesignBuilder, and OpenStudio, Wang found that due to the inconsistency of exported information, the incomplete compatibility of software to data format, and the differences in parameter setting standards, the current BIM-to-BEM conversion process is still imperfect [23].

According to Pezeshki, both BIM and BEM currently rely heavily on IFC and gbXML exchange formats and can store geometric graphic information with attribute data. However, current BIM tools often cannot accurately export this information, and BEM tools cannot parse and apply it [24]. Kamel conducted an in-depth study on the challenges, problems, and deficiencies in the BIM-to-BEM interoperability process. They believe that future research in this field should focus on generating accurate data exchange files and developing energy simulation tools that can read complete energy consumption information files and implement simulations. Porsani studied the interoperability between BIM and BEM using IFC and gbXML formats, and identified several shortcomings, primarily related to errors when reading BIM data into gbXML and IFC file formats [25]. Although some commercial software, such as Designbuilder, has some automatic error repair capabilities, the software is only suitable for simple building models. Moreover, Porsani pointed out issues such as missing spaces, broken roofs, missing exterior walls and floors, and unrecognized external shading in the energy model export module of Revit, a leading BIM platform renowned for its parametric modeling capabilities, intelligent design tools, and seamless data management [7]. Similar to Revit, Archicad, a BIM solution known for its user-friendly interface and support for international building codes and standards, enables the creation of detailed BIM models and supports data exchange in multiple formats, including IFC [26].

1.2. Research Gaps

Literature research highlights several research gaps in the BIM-to-BEM conversion process, including issues of inaccurate data and low efficiency. These gaps form the basis of the current research, which aims to address these issues by proposing an optimized conversion strategy and tool [11]. This work investigated the interaction methods between BIM and BEM, as well as the interaction files IFC and gbXML. Additionally, various energy consumption simulation software was studied and application experiments were conducted to verify the causes of information errors in the BIM-to-BEM conversion process.

Based on the literature review and experiments, two main issues were identified in the current BIM-to-BEM conversion. The first issue related to the quality of BIM data is that differences in BIM modeling standards and varying requirements among modeling personnel lead to conversion failures[10]. The second issue pertained to BIM software, where discrepancies in data standards adopted by different BIM software resulted in difficulties in interaction and conversion between these software.

To address the issues at the BIM data standard level, simplifying the large amount of repetitive or irrelevant information in the IFC standard, standardizing BIM modeling requirements, and improving the ability to parse IFC entities can be effective. As for the BIM tool level, developing a tool with a unified data conversion standard that can automatically repair lost or damaged information during the BIM-to-BEM conversion process would be beneficial.

1.3. Contribution of the Work and Structure of the Paper

The primary objective of this paper is to contribute to the optimization of the BIM-to-BEM conversion workflow, specifically addressing the challenges of information loss and inefficiency during the conversion process. The research aims to propose and validate an efficient conversion strategy and tool that enhances the efficiency and accuracy of building energy simulations. The primary contribution of this work lies in proposing and validating an efficient BIM-to-BEM conversion strategy based on the IFC data standard. This strategy not only addresses the limitations of current conversion methods but also introduces a tool that automates and improves the information transfer process.

The paper is structured as shown in Figure 2. Section 1 provides an overview of BIM and BEM, introduces the current state of BIM-to-BEM conversion, and highlights the research gaps identified through a comprehensive literature review. Section 2 presents the proposed optimization method, including the workflow for BIM model optimization before conversion, the conversion method from IFC to gbXML, and the automatic repair and thermal zone division functions. Section 3 details the development of the BIM-to-BEM conversion tool based on IFC files, including its functional analysis and reliability verification through a case study. Section 4 discusses the findings and limitations of the study, and Section 5 concludes the paper by summarizing the contributions and outlining future research directions.

2. Research Procedure

The primary research objective of this study is to address some issues arising from the BIM-to-BEM information conversion process. These issues include the accuracy and adaptability of information conversion, as well as considering how to automatically fix errors generated during the conversion process and how to automatically optimize the converted BEM model, in order to reduce related setup work before simulation. Based on the original BIM modeling standards, a set of modeling requirements are proposed to correct and add relevant information, optimize the BIM model, and prepare for exporting an interactive file suitable for energy consumption simulation, ensuring accuracy and adaptability before information interaction. By using the C# programming language and developing a programming tool with VS.net 2022, a conversion tool has been developed. This tool can automatically convert file formats, fix information errors, and optimize information content, thereby solving some of the problems arising from the BIM-to-BEM information conversion process.

2.1. Overview of Research Procedure

This subsection provides an overview of the research procedure employed in this study to optimize the conversion process from BIM to BEM (Figure 3). The research procedure is structured in a systematic and methodical manner, encompassing BIM model optimization, conversion tool development, and validation through a case study.

The first phase involves BIM model optimization before conversion. This step is crucial to ensure the quality and integrity of the BIM models used for the conversion process. It encompasses model simplification, where non-essential components are removed to reduce complexity and errors, model error checking and manual correction to identify and rectify any modeling issues, and setting and naming rooms and spaces to facilitate accurate transmission of information during conversion.

The second phase focuses on the development of a BIM-to-BEM conversion tool based on IFC files. The tool is designed to automate the conversion process, addressing challenges such as information loss and inefficiency inherent in manual or semi-automated conversions. Key functionalities include parsing IFC files, extracting relevant information, automatic repair of damaged geometric information, and automatic thermal zone division.

The third and final phase involves validation of the proposed optimization strategy and conversion tool through a case study. A real-world building project is selected for this purpose, allowing for a practical evaluation of the feasibility and effectiveness of the tool in real-life scenarios. The BIM model of the building is optimized, converted using the developed tool, and then imported into an energy simulation software for performance analysis. The results are analyzed to assess the accuracy and efficiency gains achieved through the optimized conversion process.

Overall, this research procedure adopts a holistic approach to BIM-to-BEM conversion, combining model optimization, automated tool development, and empirical validation to enhance the efficiency and accuracy of the conversion process. The objective is to contribute to more sustainable and cost-effective building energy design and analysis.

2.2. BIM Model Optimization before Conversion

The BIM model optimization before conversion involves three main steps: model simplification, model error checking and manual correction, and setting and naming rooms and spaces.

2.2.1. Model Simplification

The aim of BIM model simplification using Revit is to improve the speed of exporting files and reduce errors during the conversion process while ensuring that the information used for energy consumption simulation remains intact. Simplification includes removing model components that do not significantly impact energy consumption analysis, such as furniture and columns, and retaining only key building components related to energy analysis, such as floors, exterior walls, interior partitions, roofs, doors, and windows. As shown in Figure 4, if complex geometric shapes, such as structural columns, are not removed during the preliminary stage, errors may occur in the thermal zone space information of the exported building energy consumption information file, leading to inaccuracies in subsequent simulation work. On the other hand, complex geometric shapes can be simplified into basic shapes that accurately represent the main thermal load and thermal comfort, such as simplifying curved or bending walls into straight walls.

2.2.2. Model Error Checking and Manual Correction

After simplification, error checking and model correction are necessary to ensure there are no overlapping or missing geometric components, unsealed spaces, or other issues. It is essential to confirm that all walls and partitions used to define rooms are correct, continuous, and properly placed.

The main focus is to check for gaps or overlaps between walls, and between walls and floors, which are related to the modeling precision of the BIM modeling engineer. Higher precision leads to fewer such errors and smaller discrepancies in the exported model. Common modeling errors in BIM models are illustrated in Figure 5. These errors usually cause inaccuracies in the spatial information of the exported building energy model. Therefore, it is crucial to check and fix these conventional errors before exporting and converting to ensure the integrity of the model.

Apart from avoiding or correcting basic modeling errors, it is also important to ensure that each building component (such as walls, roofs, floors, windows, etc.) in the BIM model has correct physical and thermal conduction properties. This process ensures the accuracy of the converted building energy model information, while mistakenly using floor components to represent roof attributes leads to inaccuracies in the energy model. For instance, if a roof component is directly replaced with a floor component for modeling convenience, the exported and converted building energy model will automatically identify the roof as a floor, affecting the accuracy of subsequent energy consumption simulations (Figure 6A,B). On the contrary, using the correct roof component in the BIM model ensures accurate roof attributes in the converted energy model (Figure 6C,D).

2.2.3. Setting and Naming the Rooms and Spaces

In BIM, the accurate establishment of room details, along with their labeling and nomenclature, aims to facilitate the referencing and extension of BIM spaces to the converted BEM. This process guarantees the precise transmission of information. Through labeling and naming conventions, one can ascertain the spatial correspondence between the converted BEM and the original BIM model. Verifying the consistency of room data is crucial to ensure the accuracy of the converted information, thereby preventing unwarranted errors during subsequent simulations using the converted BEM.

Rooms and spaces are two distinct components that serve different purposes. Rooms are architectural elements that are utilized to maintain information about occupied areas. The space contains information about the parameters related to the heat and cooling loads in the area where it is located. It is mainly used for system analysis. In the context of calculating the volume of an enclosed space, the boundaries of the space are defined by the surfaces of the room elements (e.g., walls, floors, ceilings, roofs, and space dividers). The volume of the space extends in both the horizontal and vertical directions to the faces of the room boundary elements. Once the entire building model has been constructed and the space allocated, a comprehensive system analysis can be conducted at a later stage.

Room settings are configured in the Revit model, and each room should be marked and named. To prevent misalignment, it is essential that the height offset, top offset, and base offset of the room are in accordance with the top elevation, floor height, and base elevation of the model space. As shown in Figure 7, the positions of rooms can be verified and modified in a sectional view. During this stage, the spaces and thermal zones information in the exported building energy model is automatically default to match the set room information, verifying the correspondence between the information set in BIM and BEM.

Space settings and naming are applied to the model, with spaces automatically placed and named based on the previous room settings. The height offset, top offset, and base offset of the spaces should also align with the elevations of the model spaces. As illustrated in Figure 8, incorrect positioning during the conversion process can lead to the automatic misrecognition of the original unspaced enclosure structure as a spaced relationship. Consequently, the characteristics of the enclosure structure undergo corresponding changes. Initially, both sides of the wall denoted a spatial relationship; nevertheless, erroneous settings have altered this, now representing a connection between the space and the external environment. Precise environmental data are paramount for energy consumption simulations; hence, accurate location settings have a direct influence on the exactness of subsequent simulation results. Spaces with similar usage, temperature settings, and load characteristics can be merged through zones, ensuring that the thermal zone information in the exported building energy model matches the zone settings in Revit. This helps reduce the complexity of energy consumption analysis after conversion without affecting overall accuracy.

Once the aforementioned settings have been completed, it is then possible to export IFC files and subsequently parse and convert them. In light of the comprehensive nature of the document description model, the most recent iteration of IFC4 has been selected for analysis in this paper.

2.3. BIM-to-BEM Conversion Method Based on IFC

This study focuses on the conversion method from IFC to gbXML. It involves parsing the information content and structure of both IFC and gbXML, comparing the information relationships between these formats, and finding a suitable conversion path. In this study, the IFC file is parsed and extracted using the IFCviewer development library, while the gbXML file employs a recursive algorithm to traverse all nodes and obtain all model data to form an analytical data structure.

Through parsing, it is found that the IFC format file consists of four levels: resource level, core level, interactive level, and control level. Information is interrelated through references, inheritance, etc. in the IFC file [27]. The gbXML file includes detailed data describing single or group buildings, mainly covering building geometry, building environment, space division, systems and equipment, personnel and equipment operation information, and other data nodes.

Parsing of the IFC file with the IFCviewer development library revealed that the various information contained within the file is composed of multiple statements. Each statement is expressed in the form of a string. The structure of each statement is an independent number plus a set of letters and numbers composed of the code. It should be noted that the code of each statement in the same file can only be used once, and there is no limit to the order. As illustrated in Figure 9, the number “#2926” denotes the statement number, “IFCAXIS2PLACEMENT3D” represents the name of the entity instance described by the statement, and the contents in parentheses indicate the attributes of the entity instance described by the statement. The attributes are separated by the symbol “, ” and the values of the attributes can be of any data type in IFC. Alternatively, they can be referenced or inherited from entities described by other statements already defined in the document. By understanding the specific meanings attributed to each component of the statement, it becomes possible to identify the definitions and attributes associated with the various elements of the information contained within the IFC file.

The gbXML information structure differs from that of the IFC. It is a tree-structured, hierarchical form of description, as illustrated by the specific information contained in Figure 10.

By comparing the parsed files of IFC and gbXML, relationships between the contained information are identified, and basic rules are established for disassembly, classification, and corresponding conversion. Classifications mainly include one-to-one, one-to-many, and many-to-one relationships. For information with a one-to-one correspondence between IFC and gbXML, direct conversion can be performed. When one type of information in IFC corresponds to multiple types of information in gbXML, further conversion rules are developed based on IFC entity attributes to achieve model conversion. If multiple types of information in IFC correspond to one type of information in gbXML, the conversion is determined based on the completeness and conversion difficulty of the IFC entities, considering their types and attribute information.

Table 2. Examples of IFC and gbXML information classification conversion.

No.	Category	Description of Conversion
1	One-to-one conversion	(IFC) Ifc Owner History can be directly converted to (gbXML) Document History.
2	Many-to-one conversion	(IFC) Ifc Slab, Ifc Wall, Ifc Window, etc., can be converted to (gbXML) Surface, distinguished by corresponding Types and boundary positions in the gbXML file.
3	One-to-many conversion	(IFC) Ifc Geographic Element can be converted to either (gbXML) Vegetation or Transportation, depending on the corresponding Type in the IFC file to determine the specific content in gbXML.

presents examples of different types of conversions.

2.4. Comparison with Other Tools

To position this work within the broader context of BIM-to-BEM conversion, it is essential to compare our proposed optimization strategy and developed tool with existing methods and tools. Several commercial and academic tools have been developed to facilitate the conversion process, each with their strengths and limitations. Commercial tools such as GBS from Autodesk and DesignBuilder often rely on gbXML as the interaction file format. These tools automate the conversion process to a certain extent but suffer from information loss, particularly for complex models [7]. GBS, for instance, while efficient in geometry conversion, lacks comprehensive thermal information. Similarly, DesignBuilder, while offering a relatively smooth conversion process, requires manual adjustments for models with curved surfaces or complex enclosing structures [5]. On the academic front, researchers have proposed various methods to improve the interoperability between BIM and BEM. A method that converts BIM models to 2D graphics and then reconstructs 3D models has been proposed, but this approach is unsuitable for handling complex architectural features [21].

In contrast, the optimization strategy and tool developed by this work aim to address the limitations of existing methods by offering the following: (1) comprehensive IFC parsing and lightweight information extraction; (2) automatic repair of damaged geometric information; and (3) efficient thermal zone division. The comparison highlights that this approach offers a more comprehensive and efficient conversion process, particularly for complex BIM models. This is achieved through the development of a tool that automates the repair and optimization steps, significantly reducing manual intervention and improving the overall conversion efficiency and accuracy.

3. Optimization Process of the BIM-to-BEM Conversion

3.1. File Conversion

During the BIM-to-BEM conversion process, the foremost priority lies in ensuring accurate format transformation, specifically converting IFC files to gbXML format. Additionally, it is imperative to guarantee the precise conversion of the physical and thermal conductivity properties of the components. This ensures that the building energy consumption model information remains accurate after the conversion, thereby yielding realistic and usable simulation data from the converted building energy consumption model in subsequent stages. The conversion software presented in this paper incorporates various automated functions, including automatic searching, selection, information merging, and information referencing. Furthermore, it features built-in gbXML file format templates and optimization programs, facilitating the generation and preservation of new gbXML format files for application purposes.

This paper selects the conversion of interior partition walls from a BIM model of a certain project as the research object to study this process. Based on the modeling requirements mentioned above, the model is simplified, corrected, and set up in Revit 2020 to complete the BIM model optimization work, ensuring that the exported IFC file contains accurate information suitable for energy consumption simulation. Using Revit 2020 software, the optimized BIM model is directly exported as an IFC model file. The IFC file is then imported into the conversion software for automatic retrieval, extraction, citation, and conversion, realizing the reorganization of all “Ifc Wall” in the IFC file into “Surface” with wall attributes in gbXML. The specific process is shown in Figure 11 and Figure 12. It is very important to associate IDs with attribute information in the process. If this reference is missing and only “Ifc Wall” is referenced, it will not be possible to distinguish the location and attributes of the boundary Wall. The wall will be defaulted as thermally insulated, preventing thermal exchange between spaces and affecting subsequent simulation results. Therefore, it is necessary to find the positional relationship between the partition and the space to enable thermal exchange and ensure accurate simulation effects.

(Surface) aim0194 describes an interior wall component in the project, serving as an example to demonstrate how an interior wall component can be associated with its location and attributes. As shown in Figure 13, the process involves finding (Surface) aim0194 Reversed with the same vertex coordinates as (Surface) aim0194 and combining them into a single surface.

As illustrated in Figure 14, the two spatial information pieces, (space) aim0191 and (space) aim0204, associated with (Surface) aim0194 and (Surface) aim0194 Reversed are identified and referenced.

As shown in Figure 15, returning to OpenStudio for verification, the original adiabatic internal partition surface is converted to a heat exchange internal partition. This example verifies the feasibility of the many-to-one conversion method for the information contained in the IFC. Therefore, it is useless to only reference associated components in the process. It is also necessary to extract and associate ID and attribute information to achieve complete information conversion of components.

3.2. Geometric Information Auto-Repair

During the BIM-to-BEM conversion process, varying degrees of information loss often occur, especially geometric information. Complete geometric information is fundamental for building energy consumption model simulation applications. Therefore, to address incomplete geometric information, it is necessary to repair the geometric data to form one or multiple complete thermal zone spaces. Given the complex file structure and large data volume of energy consumption models, manual repair is practically impossible. Thus, it is sensible to develop a tool with automatic repair capabilities.

This study designed a tool with an automatic repair function for damaged geometric information in energy consumption models, following the process outlined in Figure 16. First, the building energy consumption model is imported to obtain spatial geometric information, thermal zone information, energy consumption load, etc. Then, the spatial geometric information is organized into a room data structure. Next, each room is checked using a room integrity verification algorithm. If a damaged room is detected, a repair algorithm is applied to fix damaged walls, floors, and ceilings. The room integrity verification is performed again until it passes. Finally, the repaired spatial geometric information, thermal zone information, and energy consumption load are used to generate a new gbXML file for subsequent energy consumption calculation and analysis.

The repair algorithm employed in this paper screens and repairs each space individually. As illustrated in Figure 17, within each space, every surface is judged based on the boundary formed by other surfaces. If there is a gap between the surface and the boundary, the coordinates of the points on the surface are adjusted to match the boundary points, filling all gaps. The repair is complete when all gaps are filled. This screening and repair process is automated through a developed tool.

After automatic repair, the repaired information is updated in the corresponding nodes and stored in the repaired file. A gbXML format file is then exported. The three-dimensional view before and after repair (Figure 18) can be obtained in Open Studio.

3.3. Automatic Thermal Zone Division

A thermal zone is the basic unit for energy consumption simulation calculations, representing a division of the building from a thermal perspective [28]. A thermal zone includes a uniform air volume temperature and all heat transfer and storage surfaces within or on the air boundary. During simulation, each room in a building is designated as a separate thermal zone or multiple rooms can be grouped into a single thermal zone. The principle of thermal zone division is to classify rooms with unified heating methods and air conditioning systems into the same thermal zone [29].

Besides automatic geometric information repair, automatic thermal zone division is another core method in this work, aiming to optimize the BIM-to-BEM conversion process. Automatic thermal zone division is achieved through the following steps: (1) extract rooms or areas with unified naming in BIM, reading information such as geometric surfaces, vertex coordinates of surfaces, and material types; (2) distinguish between rooms with shared boundary surfaces and those without; (3) group rooms or areas with unified naming and shared boundary surfaces into one thermal zone, and those without shared boundary surfaces into another; and (4) repeat the above method to divide thermal zones for other room functions. This process is automated through the developed tools [30]. Using the “Dormitory” function as an example, the specific steps for automatic thermal zone division are illustrated in Figure 19.

4. Development and Reliability Verification of the BIM-to-BEM Conversion Tool Based on IFC File

4.1. Functional Analysis of the Conversion Tool

After conducting literature reviews and related practical studies, the functions of the conversion tool were determined, and the software was developed using the C++ programming language. The conversion tool integrates the following functionalities: (1) reading IFC files and exporting gbXML files; (2) information lightweighting, primarily simplifying the geometric information of solid models with thickness walls into a model expressed in a single-face form while preserving the thermal information of the building; (3) automatically repairing damaged geometric information to form a complete thermal zone space; (4) optimizing the automatic division of thermal zones; and (5) extracting, converting, or adding other information such as building materials.

The tool utilizes the IFC viewer development library to read and parse IFC files, extracting relevant information about geometry, thermal zones, and thermal properties related to the building energy consumption model. By reading the building instance model in the IFC, a tree structure is generated through deconstruction and reorganization. As illustrated in Figure 20, the editing interface, located in the middle of the main interface, allows for information browsing, searching, and editing. The data structure interface, on the left side of the main interface, provides a convenient and intuitive way to view the data of the generated tree structure model. The processing information interface, positioned at the bottom of the main interface, displays software processing information, including any exceptional information.

4.2. Reliability Application Verification of the Conversion Tool

Taking one building project in Shenzhen, China as an example, a BIM-to-BEM tool conversion test and verification were conducted. The project involves auxiliary office buildings for a grain depot, which are simple rectangular structures with a total floor area of 260.82 square meters and a footprint of 130 square meters. It consists of two floors, mainly including a lobby, offices, equipment rooms, a weighbridge room, and an initial inspection room. The project is located in a subtropical monsoon climate zone, with long summers and short winters. Summers are hot and rainy, while winters are cold and dry. The annual average temperature ranges from 22.8 °C to 28.8 °C. The average temperature in January is 10.7 °C, and the average temperature in July is around 25.9 to 29.1 °C. The accumulated temperature greater than or equal to 10 °C is over 5000–6000 h. The annual precipitation is about 1500 mm, and the relative humidity is 78–82%. Therefore, air conditioning is often required to adjust the indoor temperature, resulting in high energy consumption. The designer used BIM software to perform conventional modeling of the building, which mainly included building-related information but did not include information about the air conditioning system. Then, according to the optimization requirements for converting the BIM model, the BIM model was simplified and standardized and export settings were configured in Revit. After exporting the IFC file (Figure 21A), the development tool (optimizegb IFC) was used to optimize the model conversion, generating a gbXML format file (Figure 21B) through tool processing. The generated gbXML file is imported into Open Studio, where parameters such as air conditioning system and internal disturbances caused by occupants are set, and an IDF file is exported for compatibility with Energy Plus. Then, Energy Plus was called for simulation calculations. In this case, no errors occurred during the simulation process, and relevant simulation results were obtained smoothly.

Through the above verification, it can be demonstrated that the gbXML file generated by the developed conversion tool can be directly imported into Open Studio to invoke Energy Plus for building energy consumption simulation. It achieves efficient conversion from BIM-to-BEM and significantly improves the efficiency of energy simulation calculation.

The following results could be obtained from the tests and validations. First, in this validation project, it was essential to establish a suitable BIM model to provide the correct IFC file, which was also the key to the correct conversion of the conversion tool. Second, this study attempted to use complex models for conversion, but it could not achieve a perfect conversion like the auxiliary room mentioned above. Issues such as incorrect geometric information and thermal zone information arose during the conversion of complex models. The reason was that the tool cannot currently handle all types of building shapes. Further development of the tool can be targeted at this aspect in the future. Third, this study compared the process of manually establishing an energy consumption model with using a conversion tool. It was found that if the initial IFC file information is accurate, the information converted using the tool is more accurate and less error-prone than manual modeling. This is because in the manual modeling process, “Surface Matching” is very error-prone. When performing a one-time “Intersect” and “Match” on the entire model, various errors may occur, including missing surfaces, incorrect surface attributes, corresponding errors, etc. Typically, manual modeling can only perform one-to-one “Intersect” and “Match” operations between two spaces, which requires a huge workload. Therefore, it is very necessary to develop a suitable conversion tool for BIM-to-BEM conversion.

5. Discussions

This paper presents an optimization study of the conversion process from BIM to BEM based on the IFC data standard. The proposed method, including BIM model optimization and the developed tool, aims to address the challenges encountered during the conversion process, such as information loss and inefficiency. The key findings, performance data, and limitations of this work are discussed below, along with suggestions for future research.

5.1. Key Findings and Performance Data

The primary objective of this research was to improve the BIM-to-BEM conversion process through the development of an optimization strategy and a conversion tool based on IFC. The proposed methodology includes BIM model optimization before conversion, automatic repair of damaged geometric information, and automatic thermal zone division.

BIM model optimization: By simplifying BIM models and ensuring the removal of unnecessary components, the exported IFC files contained more relevant and accurate information for energy simulation. This model simplification significantly reduced errors during the conversion process, as demonstrated in Section 2.2.

Automatic repair function: The developed tool exhibited remarkable performance in automatically repairing damaged geometric information. In the case study, the tool successfully repaired broken surfaces and gaps, leading to a complete thermal zone space in the BEM model. This improvement facilitated smooth simulation and improved the overall accuracy of the results.

Automatic thermal zone division: The automatic thermal zone division feature enabled the grouping of rooms or areas with similar heating and cooling requirements into a single thermal zone. This not only simplified the energy simulation process but also reduced the computational complexity. The automatic division algorithm proved to be efficient and reliable in the case study.

Performance data: The case study using a building project in Shenzhen, China, verified the effectiveness of the proposed optimization strategy and conversion tool. The BIM model was successfully converted into a gbXML file, which was then imported into Open Studio and used to generate an IDF file for Energy Plus simulation. The simulation process ran without errors, demonstrating the feasibility and practicality of the proposed method.

5.2. Limitations

Despite the promising results, this research still has some limitations that require attention in future iterations.

Complex model support: The current conversion tool is capable of converting relatively simple BIM models but struggles with complex models involving curved surfaces or unusual architectural features. Research conducted by Feng also indicates that only simple model conversions can be achieved at present [3]. Parameter settings: The tool cannot convert all required parameters, necessitating manual parameter setting in energy consumption simulation software before simulation. Integration with other software: To expand the application of the tool, future research could explore its integration with other architectural software, enhancing interoperability and usability.

Based on the limitations of the above research and discoveries, it can actually be summarized into a conclusive issue: there is currently no standard and normalized interactive information available and there may be many reasons for this, such as the complexity of the model, the lack of early parameters, or the loss of information during the export process.

6. Conclusions

This paper presents an optimization study on the BIM-to-BEM conversion process, utilizing the IFC data standard, to address the challenges of information loss and inefficiency in existing conversion methods. The key findings of this study demonstrate that the proposed optimization strategy, comprising BIM model optimization, automatic repair of damaged geometric information, and automatic thermal zone division, significantly enhances the efficiency and accuracy of the BIM-to-BEM conversion process. The conversion tool developed in this research, based on IFC files, has been empirically validated through a case study, confirming its feasibility and practicality in real-world applications.

In the context of the “dual carbon” goal, the demand for the integration of BIM with energy consumption and carbon emission calculation, analysis, and optimization applications throughout the building’s lifecycle has become increasingly prominent. The interoperability between BIM and BEM has also become increasingly important. Innovating new tools or institutional standards to handle relevant information interoperability will better promote the high-quality development of BIM technology. Through research, it has been found that the conversion from BIM to BEM mainly involves several aspects that affect the workflow of the conversion. The first is standardized and normalized interactive information, including input and output information; the second is accurate conversion between BIM and BEM. In this study, we attempted to improve the efficiency of BIM-to-BEM conversion by establishing standard modeling and output requirements and developing conversion tools. During the process, we discovered that with idealized and fully standardized BIM modeling (correct input and output information), efficient BIM-to-BEM conversion can theoretically be achieved through simple software repairs and optimizations. However, due to differences in buildings, complexity, and human factors, it is a significant challenge to achieve idealized and fully standardized BIM modeling with correct output information for most buildings. In the future, how to automatically generate standardized and accurate interactive information for various buildings may become a key research focus in the workflow of BIM-to-BEM conversion. Promoting more accurate and efficient building energy simulations will contribute to sustainable building design and operation.