A Data Requirement Framework for BIM–IoT Integration in Sustainable Building Rainwater Drainage Systems

2.1 Building Information Modeling for building system management

BIM has been widely recognized as a core digital platform for supporting building design, construction, and lifecycle management [10, 16-22]. By integrating geometric, material, and relational information within a unified digital environment, BIM enables systematic representation of building components and infrastructure systems, thereby improving coordination, reducing design conflicts, and supporting long-term asset management [8, 11].

Despite these advantages, numerous studies have highlighted that conventional BIM implementations are primarily designed to manage static information [16, 18]. Such data structures are well-suited to representing physical characteristics and design intent, but are limited in their ability to capture dynamic operational behaviour during the building use phase [18, 19]. This limitation is particularly significant for systems whose performance varies over time, such as rainwater drainage systems, which are highly sensitive to rainfall variability and transient flow conditions [3].

In the context of building rainwater drainage, BIM is commonly used to model pipe geometry, network topology, and material properties [10]. While this information provides an essential structural baseline, it does not offer insight into real-time system performance or operational risks [16]. As a result, BIM alone is insufficient to support performance monitoring, adaptive management, or sustainability-oriented planning of drainage systems [19].

2.2 Internet of Things and dynamic data for drainage systems

The rapid development of IoT technologies has enabled real-time monitoring of infrastructure and building systems through distributed sensor networks [2, 11]. IoT devices generate continuous time-series data related to water level, flow rate, pressure variation, rainfall intensity, and system status, thereby providing valuable information on actual system behaviour under varying operational conditions [19].

In water and drainage management, IoT-based monitoring has been applied to urban drainage networks, flood warning systems, and smart water infrastructure [12, 19]. These applications demonstrate the potential of dynamic data to enhance system responsiveness and risk mitigation. However, much of the existing research emphasises sensor technology, data acquisition, and communication protocols, with comparatively limited attention given to how dynamic data should be structurally integrated with existing building information models [17].

Within building-scale rainwater drainage systems, IoT data are often collected and stored independently from BIM models [15]. This separation results in fragmented data environments in which real-time monitoring data cannot be directly linked to the physical components represented in BIM. Consequently, the analytical and decision-support potential of IoT data remains underutilised in sustainable building system management [6].

2.3 Data gap between Building Information Modeling and Internet of Things integration

Although BIM and IoT each provide significant benefits when considered independently, their integration presents substantial challenges due to fundamental differences in data characteristics [6, 16]. BIM primarily manages structured, object-based, and static data, whereas IoT systems generate unstructured or semi-structured, time-dependent data streams [2, 11]. This mismatch creates a data gap that hinders seamless integration and limits the effectiveness of combined BIM–IoT systems.

Previous studies have identified the lack of semantic mapping and data alignment mechanisms as a major obstacle to BIM–IoT integration [16, 17]. Without clear definitions of data requirements and relationships, dynamic sensor data cannot be consistently associated with BIM elements, reducing the reliability of system assessment and analysis [6]. This issue is particularly critical for rainwater drainage systems, where operational behaviour is governed by rapid temporal variations in rainfall and flow conditions [3, 19].

Moreover, the absence of a structured data integration framework increases the risk of ad hoc solutions that are difficult to scale, maintain, or generalise across projects [18, 22]. These limitations highlight the need for a systematic approach to defining and managing data requirements for BIM–IoT integration.

2.4 Property Sets and limitations of Industry Foundation Classes-based data structures

The Industry Foundation Classes (IFC) standard has been developed to support interoperability in BIM by providing a common data schema for building information exchange [7, 10]. Within IFC, P-Sets enable flexible attachment of attribute information to BIM elements, supporting extensible representation of component properties [7].

However, existing standard P-Sets primarily focus on physical and static attributes, such as dimensions, materials, and classification codes [10, 22]. They provide limited support for representing dynamic, time-series data required to describe operational system behaviour, including water level fluctuations, flow variability, and abnormal system states [16, 18]. As a result, standard IFC-based data structures are insufficient to support BIM–IoT integration for drainage systems without structural extension [6].

Several studies have explored the use of customised P-Sets to accommodate sensor data and digital twin applications [16, 17]. While these approaches demonstrate technical feasibility, they often lack a systematic framework for defining data requirements, identifying data gaps, and structuring extensions in a manner that supports sustainable planning and long-term system resilience [6].

2.5 Digital twin concepts and the need for conceptual data frameworks

Digital twin concepts have gained increasing attention as a means of linking digital models with real-time operational data to support system monitoring, analysis, and decision-making [6, 13]. In the context of buildings and infrastructure, digital twins are expected to enhance predictive capabilities and improve system resilience under changing environmental conditions [17, 18]. Similar concepts have also been discussed in sponge city initiatives aimed at improving urban stormwater resilience and adaptive water management under climate variability [20].

Recent studies emphasise that successful digital twin development depends not only on sensor deployment or software platforms but also on the availability of well-defined data structures and integration frameworks at the conceptual level [6, 12]. Without clear data requirements and integration logic, digital twin implementations risk becoming fragmented, project-specific, and difficult to sustain [21].

For building rainwater drainage systems, the absence of a dedicated conceptual data framework limits the ability to translate BIM–IoT integration into practical decision-support tools for sustainable planning and risk management [16, 17].

2.6 Research gap and study positioning

Based on the reviewed literature, three key research gaps can be identified. First, there is a lack of dedicated DRF specifically designed for integrating BIM and IoT in building rainwater drainage systems [6, 17]. Second, existing BIM standards and data structures do not adequately support the representation of dynamic, time-dependent operational data required for drainage system monitoring and assessment [16, 18]. Third, the development of digital twins for drainage systems lacks a robust conceptual data foundation to support sustainable planning and long-term resilience [13, 21].

To address these gaps, this study proposes a DRF that systematically defines static and dynamic data types, identifies data gaps, and establishes structured integration mechanisms between BIM and IoT. By focusing on data requirements and architectural design rather than system implementation, the proposed framework provides a scalable and sustainability-oriented foundation for future digital twin development and decision-support applications in building rainwater drainage management.

2.7 Comparative positioning of Building Information Modeling-Internet of Things integration frameworks

To position the proposed DRF within the broader context of BIM–IoT integration research, a structured comparison of representative prior frameworks is presented in Table 1. The comparison focuses on drainage specificity, time-series integration capability, IFC extension strategy, and architectural layering. This analysis clarifies the distinctive scope and technical positioning of the proposed DRF.

Table 1. Comparative analysis of Building Information Modeling-Internet of Things (BIM–IoT) integration frameworks

As shown in Table 1, existing BIM–IoT integration studies generally focus on middleware connectivity, digital twin platforms, or general building monitoring applications. However, few frameworks explicitly address drainage-specific data requirements, structured time-series linkage at the schema level, and IFC-aligned extension mechanisms within a layered architectural model. The proposed DRF distinguishes itself by systematically categorising drainage-related data gaps and formalising their integration through customised P-Sets within a four-layer architecture.

Following the methodological framework presented in Section 3, this section reports the results derived from the identification of BIM static data, the definition of IoT dynamic data requirements, the data gap analysis, and the development of data integration mechanisms under the proposed DRF. The results are discussed to demonstrate how the proposed framework addresses structural limitations of conventional BIM and supports sustainability-oriented planning and management of building rainwater drainage systems.

4.1 Identification of Building Information Modeling static data

Based on the methodology described in Section 3.3, BIM static data representing the physical configuration of the building's rainwater drainage system were systematically extracted from the BIM model of the case study building. These data constitute the structural baseline of the system and correspond directly to the static data domain (B) within the proposed DRF.

To support structured utilisation and subsequent integration with dynamic data, the extracted BIM static data were classified into three main categories: geometric properties, structural/topological properties, and material properties, as summarised in Table 2. Geometric properties include pipe diameter, length, slope, and elevation, which define the physical capacity and spatial characteristics of the drainage system. Structural and topological properties describe connectivity relationships, flow direction, and network hierarchy, enabling consistent mapping of operational data to specific BIM elements. Material properties provide supporting information related to durability and maintenance considerations.

The variables listed in Table 2 were derived using the structured selection strategy defined in Section 3.2.

Table 2. Categories of Building Information Modeling (BIM) static data for building rainwater drainage systems

Table 2 summarises the categories of static data extracted from the BIM model, which constitute the structural baseline (B) of the building rainwater drainage system within the proposed DRF. Geometric and topological properties provide reference attributes for linking time-series operational data to physical BIM elements, while material properties support interpretation related to durability and long-term planning.

The results demonstrate that BIM static data can comprehensively represent the physical layout and connectivity of the rainwater drainage network. However, these data are inherently time-invariant and do not reflect variations in system behaviour during rainfall events or operational disturbances. As a result, BIM static data alone are insufficient to support performance monitoring or risk assessment, underscoring the necessity of integrating dynamic, time-series data within the DRF.

4.2 Internet of Things dynamic data requirements

Following the limitations identified in Section 4.1, where BIM static data were shown to be insufficient for representing time-dependent system behaviour, this section presents the results of defining dynamic data requirements derived from IoT-based monitoring. These dynamic data represent the operational domain (I) within the proposed DRF and are essential for capturing the actual performance of building rainwater drainage systems under varying rainfall conditions.

Based on the functional characteristics of rainwater drainage systems and potential operational risks, a set of key dynamic variables was identified. These variables include water level, flow rate, pipe pressure, rainfall intensity, and system status indicators such as overflow or blockage conditions. All identified variables are characterised as time-series data that reflect temporal variations in system behaviour and external forcing, as summarised in Table 3.

The dynamic variables summarised in Table 3 follow the selection logic described in Section 3.2.

Table 3. Internet of Things (IoT) dynamic data requirements for building rainwater drainage systems

Table 3 summarises the key time-series data required to represent the operational behaviour of building rainwater drainage systems. These dynamic variables constitute the IoT data domain (I) within the proposed DRF and provide essential information for monitoring system performance and identifying operational risks that cannot be captured by static BIM data alone.

The identified IoT dynamic data requirements are summarised in Table 3, which presents the required variables, their functional relevance, and their roles in system monitoring and assessment. Although this study does not involve physical sensor installation or real-time data acquisition, defining these data requirements provides a structured foundation for subsequent BIM–IoT integration and prepares the data architecture for future implementation and digital twin development.

The results highlight that dynamic data are indispensable for interpreting system performance during rainfall events and identifying abnormal operating conditions. However, when considered independently, IoT-derived data lack inherent linkage to physical system components represented in BIM models. This limitation reinforces the need for a structured integration mechanism, which is addressed through data gap analysis in Section 4.3.

4.3 Data gap analysis

Building upon the identified BIM static data (Section 4.1) and IoT dynamic data requirements (Section 4.2), a data gap analysis was conducted to examine discrepancies between the two data domains. The analysis aims to identify information required for monitoring and managing building rainwater drainage systems that cannot be represented within conventional BIM data structures.

The data gap was formally defined using the conceptual relationship G = I−B, where G represents missing or unsupported information, I denotes required dynamic, time-series data, and B represents static data extracted from BIM models. By systematically comparing these two data domains, critical gaps related to operational system behaviour were identified.

Table 4. Linkage between BIM–IoT data gaps and developed P-Sets

The results of the data gap analysis are summarised in Table 4, which links BIM-supported information, required IoT dynamic data, identified data gaps, and corresponding data extension mechanisms. As shown in the table, while BIM static data comprehensively describe the physical configuration and connectivity of the drainage system, they do not support key operational variables such as water level variation, flow performance, rainfall influence, and abnormal system states.

Table 4 presents the results of the data gap analysis by mapping BIM static data against required IoT dynamic data for building rainwater drainage systems. The table highlights operational variables that cannot be represented within conventional BIM data structures and identifies corresponding gaps that necessitate structural data extensions.

The results presented in Table 4 demonstrate that the identified data gaps arise primarily from the inability of BIM to accommodate time-dependent and event-driven information rather than from deficiencies in model completeness or geometric detail. These gaps are therefore structural in nature and cannot be resolved by adding further static attributes to the BIM model.

Consequently, addressing the identified data gaps requires a structured data integration mechanism capable of extending BIM attributes to support dynamic information. This requirement provides the rationale for the development of customised P-Sets, which are presented and discussed in Section 4.4 as a core component of the proposed DRF.

4.4 Development and role of Property Sets

Based on the data gaps identified in Section 4.3, customised P-Sets were developed as a structural mechanism for extending BIM data models to support systematic integration of IoT-derived time-series data. The development of P-Sets focuses on enabling dynamic operational information to be associated with relevant BIM elements without modifying the core BIM schema.

The developed P-Sets address key operational variables that cannot be represented by conventional BIM attributes, including water level variation, flow performance, rainfall influence, and abnormal system states. These dynamic attributes are defined as time-series properties and are linked to specific BIM components such as pipes, inspection chambers, and drainage terminals. Through this approach, P-Sets function as an intermediary data structure that bridges static BIM data and dynamic IoT data within the proposed DRF.

The developed P-Sets and their associated data variables are summarised in Table 5, which presents the relationship between dynamic operational parameters and corresponding BIM elements. The table demonstrates how each identified data gap is addressed through a structured extension of BIM attributes, thereby enabling consistent association of time-series data with physical system components.

Table 5 provides the technical specification of the developed P-Sets, including attribute data types, unit definitions, temporal referencing formats, and sensor linkage attributes. Time-series data are structured using TimestampID and are associated with BIM elements through IFC-based element identifiers and SensorID fields. This specification ensures consistent data interoperability without modification of the core BIM schema.

Table 5. Developed Property Sets (P-Sets) for building rainwater drainage systems

The results presented in Table 4 indicate that customised P-Sets can effectively resolve the structural data gaps identified in Section 4.3 by enabling dynamic operational data to be coherently linked to BIM elements. This structured linkage allows BIM models to conceptually reflect system behaviour under varying rainfall and operational conditions while preserving their original design-oriented structure.

The role of P-Sets within the overall DRF architecture is conceptually illustrated in Figure 4, which shows how static BIM data and dynamic IoT data are integrated through the Integration Layer, temporal indexing, and BIM element identification to ensure structured transformation of time-series attributes within the DRF. As shown in the figure, P-Sets serve as the primary mechanism for associating time-series data with BIM elements, enabling integrated data flow toward application-level use cases such as monitoring, risk assessment, and preparation for digital twin development.

Figure 4. Building Information Modeling–Internet of Things (BIM–IoT) data integration through Property Sets (P-Sets) under the Data Requirement Framework (DRF)

Figure 4 illustrates the conceptual integration of BIM static data and IoT dynamic data through customised P-Sets under the proposed DRF. The figure highlights the role of the Integration Layer in linking time-series operational data to BIM elements to support monitoring, risk analysis, and future digital twin applications.

By establishing a structured mechanism for BIM–IoT data linkage, the developed P-Sets provide the foundation for systematic data transformation and utilisation within the DRF. Building upon this integration mechanism, the following section presents data flow modeling results to demonstrate how integrated data are transformed into decision-ready information under the proposed framework.

4.5 Data flow modeling under the Data Requirement Framework

Building upon the integration mechanism established through customised P-Sets in Section 4.4, data flow modeling was conducted to demonstrate how integrated BIM–IoT data are transformed into decision-ready information under the proposed DRF. The data flow modeling focuses on logical data movement and transformation across the DRF layers rather than on software-specific implementation. The modeling explicitly incorporates data typing, temporal indexing, and IFC-based element identification to ensure structured transformation of time-series attributes into BIM-linked dynamic properties.

The data flow begins at the IoT Dynamic Layer, where time-series data representing system operation—such as water level, flow rate, and rainfall intensity—are generated. These data are subsequently transferred to the Integration Layer, where validation, consistency checking, and semantic alignment with BIM elements are performed. Through the use of customised P-Sets, validated time-series data are then associated with corresponding BIM components, enabling structured linkage between dynamic operational data and static system representations.

The sequence of data flow processes under the DRF is summarised in Table 6, which outlines the transformation of data from acquisition to application-level utilisation. As shown in the table, raw IoT-derived data are progressively refined through mapping, aggregation, and contextualisation steps to support monitoring, analysis, and planning-oriented decision support.

Table 6. Data flow under the Data Requirement Framework (DRF)

Table 6 summarises the data flow processes under the proposed DRF, illustrating how IoT-derived time-series data are validated, mapped to BIM elements through P-Sets, and transformed into decision-ready information for application-level use.

The results presented in Table 6 indicate that the DRF provides a structured and coherent pathway for managing heterogeneous data types without modifying the core BIM data structure. By explicitly defining data flow stages and responsibilities across architectural layers, the framework reduces data fragmentation and enhances the consistency of integrated information.

To further illustrate the overall data movement and interaction between DRF layers, the conceptual data flow under the proposed framework is presented in Figure 5. The figure highlights how dynamic data from the IoT layer is processed through the Integration Layer and prepared for application-level analysis while maintaining linkage to BIM elements.

Figure 5. Data flow modeling under the proposed Data Requirement Framework (DRF)

By demonstrating a structured pathway for data transformation and utilisation, the data flow modeling results confirm the functional coherence of the proposed DRF architecture. These results provide a basis for evaluating the architectural readiness of the framework, which is further examined through conceptual validation results in Section 4.6.

For conceptual evaluation, “Meets” indicates that all required architectural and data integration conditions were addressed within the corresponding DRF layer. “Partially meets” indicates that approximately 50–80% of the expected conditions were addressed within the current conceptual scope but may require further implementation-level refinement.

Table 7. Conceptual validation matrix of the DRF

As shown in Table 7, the Integration Layer achieves the strongest validation outcome because it satisfies structural consistency, data completeness, and integration feasibility. The BIM Static Layer and Application Layer partially meet the data completeness criterion because they depend on future operational data and implementation-level deployment. The IoT Dynamic Layer partially meets integration feasibility because dynamic monitoring data requires a structured linkage to BIM elements through the Integration Layer. These results indicate that the proposed DRF is conceptually coherent, while further empirical implementation is still required.

The conceptual validation examines three key aspects: structural consistency, data completeness, and feasibility of data integration across the DRF layers. Structural consistency is evaluated by examining the alignment between BIM static data, IoT dynamic data, and customised P-Sets within the Integration Layer. The results indicate that the proposed framework maintains consistent relationships between data entities, enabling dynamic operational data to be associated with corresponding BIM elements without modifying the core BIM data structure.

Data completeness is assessed by examining whether the defined data requirements sufficiently represent key operational conditions of building rainwater drainage systems. The results confirm that the combination of BIM static data and IoT dynamic data requirements captures both physical system characteristics and time-dependent operational behaviour, thereby addressing the critical data gaps identified in Section 4.3.

The feasibility of data integration is evaluated through the logical consistency of data flow and data linkage mechanisms under the DRF architecture. The validation results demonstrate that the proposed framework provides a coherent pathway for transforming heterogeneous data into integrated, decision-ready information suitable for monitoring, risk assessment, and planning-oriented analysis.

Overall, the conceptual validation results indicate that the proposed DRF is structurally consistent and architecturally ready for future implementation and extension toward digital twin development. While no physical sensor deployment or real-time system testing is conducted in this study, the validation demonstrates that the framework provides a robust and scalable foundation for subsequent empirical studies and sustainability-oriented planning applications.

4.6 Transferability of the Data Requirement Framework

Although the proposed DRF was developed and examined in the context of building-scale rainwater drainage systems, its architectural structure is not limited to drainage applications. The identification of dynamic variables such as water level, rainfall intensity, and overflow status is drainage-specific; however, the layered integration logic, structured data gap analysis approach, and IFC-aligned customised P-Set extension mechanisms are transferable to other building systems requiring time-series integration.

Specifically, the four-layer architecture (BIM Static Layer, IoT Dynamic Layer, Integration Layer, and Application Layer) can be applied to HVAC systems, water supply networks, energy monitoring systems, and other infrastructure domains where static design information must be linked with dynamic operational data. The drainage-specific variables may change, but the schema-level extension logic and temporal data structuring strategy remain reusable.

This transferability demonstrates that the proposed DRF functions as a scalable data architecture rather than a single-use case framework, while still maintaining domain-specific precision in the present study.

Although this study focuses on conceptual framework development and validation, several directions for future research can be identified. First, the proposed DRF can be applied and evaluated through real-world implementation using IoT sensor deployment in building rainwater drainage systems to assess its performance under actual operational conditions. Such studies would enable quantitative evaluation of data integration effectiveness and support refinement of data requirements.

Second, the DRF may be extended to support predictive analysis by integrating hydraulic or hydrological models with real-time monitoring data. This extension would enhance the capability of the framework to support early warning, risk forecasting, and adaptive drainage management.

Finally, future research may explore the application of the DRF within a full digital twin environment, including automated data updating, system simulation, and decision-support tools for long-term sustainability planning. These extensions would further strengthen the role of BIM–IoT integration in resilient building and infrastructure management.