BIM-integrated semantic framework for construction waste quantification and optimisation

BIM-integrated semantic framework for construction waste quantification and optimisation Subarna Sivashanmugam, Sergio Rodriguez Trejo , Farzad Rahimian * School of Computing, Engineering and Digital Technologies, Teesside University, United Kingdom ARTICLE INFO Keywords: Construction and demolition waste Quantification and classification Information and communication technologies Conceptual framework Semantic Web Technologies BIM ABSTRACT Quantification and optimisation of Construction Waste (CW) in the design stages are vital to implementing preventive CW management measures. Previous ICT-integrated CW models are not efficiently upscaled to achieve an interoperable and automated workflow. Therefore, this paper presents a BIM-integrated semantic framework for CW quantification and optimisation from the early design stages. A CW data model using Semantic-Web-Technologies (SWT) was developed and integrated with BIM. The results proved that unified data structure, standardised and granular information, established semantic relationships between building material and CW data, and diverse measurement units proposed in the framework facilitate seamless and dynamic information flows between BIM and CW platforms. The research outcomes are critical to improving interoperability and automation across the CW assessment process, enhancing the accuracy and reliability of results, supporting timely and integrated decision-making, and easing communication and collaboration among the supply-chain members. A test-case building demonstrates the application of the framework. 1. Introduction Due to urbanisation and increasing population, raw material consumption and solid waste generation have risen alarmingly. In particular, the Architecture, Engineering, Construction, and Operation (AECO) industry consumes half of the total raw materials and generates one-third of global solid waste. For instance, the UK’s AECO industry consumes 25 % of raw materials and contributes 62 % of total solid waste generation [1]. The consequences of uncontrolled waste generation result in excessive raw material extraction, carbon emissions, energy usage, pollution, and landfilling. Besides, it is reported that material waste from new constructions produces 3.5 million tonnes of carbon dioxide equivalent (CO2e) emissions [2]. One-third of the UK landfill waste is contributed by waste generated from the built environment activities [3]. In addition to environmental aspects, construction waste (CW) significantly impacts project time, cost, and health and safety factors [4]. The last ten decades of the industrial revolution were dominated by a traditional linear economy model, in which the materials are purchased, used, and discarded at the end of their life, leading to increased waste generation. However, with the increase in global population coupled with changes in consumption patterns, considerable changes to the environment, health and well-being are becoming integral parts of our system. As an alternative to the linear economy, the Circular Economy principles are on the frontline of decision-making to enhance the productive use of CW and maintain their values for an extended period [5]. Besides, the shift from a traditional fragmented system to a collaborative digital environment within the AECO industry, underpinned by Building Information Modelling (BIM), is also becoming a significant player in managing the design and construction factors [6] that influence CW generation. The research conducted by [7] identified that BIM has the potential to overcome the process, policy and technology-related factors influencing the CW generation and management across the project life-cycle. However, on top of all these approaches, accurate quantification of CW is recognised as a prerequisite for developing solid waste management (WM) plans across the life-cycle stages [8]. The WM steps proposed in the ISO 14001:2015 also prioritise the measurement process to effectively store, treat and dispose of waste materials [9]. This led to the development of various quantification models that support CW estimation at the building, regional, and national levels. Advanced Information and Communication Technologies (ICTs) like BIM, Artificial Intelligence (AI), Geographical Information Systems (GIS), and autonomous robots were integrated to enhance the quantification, classification, and optimisation process [10]. The life cycle * Corresponding author. E-mail address: [email protected] (F. Rahimian). Contents lists available at ScienceDirect Automation in Construction journal homepage: www.elsevier.com/locate/autcon https://doi.org/10.1016/j.autcon.2024.105842 Received 26 March 2024; Received in revised form 7 October 2024; Accepted 19 October 2024 Automation in Construction 168 (2024) 105842 Available online 30 October 2024 0926-5805/Crown Copyright © 2024 Published by Elsevier B.V. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). assessment (LCA) models also integrate CW data into the environmental impact assessment process. A review of existing models reveals the critical problems associated with their application in the collaborative digital AECO environment. Existing research studies neglect to close the gaps related to heterogeneity in design, material, and CW information, which causes interoperability and automation issues. Other critical limitations include the lack of standardised waste measurement units, inadequate granularity in the waste results, narrow focus towards a specific life cycle stage, and unstandardised and ununified data structures. Besides, the stages after product manufacturing in LCA are considered scenario-specific, so the data from the Environmental Product Declaration (EPD) certificates and LCA databases do not reflect the actual site scenario of a building during its construction/installation, operation and maintenance and demolition/deconstruction. Thus, most LCA models make generic assumptions when reporting the waste generated by building products, which limits their application in developing comprehensive WM plans. These limitations compromise the reusability and wide-scale applications of existing models. Further, as a result of this, multiple noncompatible and non-reliable CW databases are acting in silos rather than a single collaborative CW data model. Thus, maintaining a unified data structure, standardising the information flows, integrating sitespecific factors, and ensuring consistency in reporting that could achieve dynamic data flows between the CW and BIM platforms is vital to overcoming existing limitations. This research aims to bridge the data gaps by developing a semantic framework that ensures seamless information exchange, data consistency and accessibility, and automation of the overall CW assessment process from the early design stages. The research aim is achieved through the following objectives: Objective 1 - Explore the applications of CW quantification and optimisation models. Objective 2 - Explore the role of BIM–Semantic Web Technology (SWT) integration in unifying and standardising the CW information. Objective 3 - Propose a BIM-integrated semantic framework to enhance interoperability and automation across the CW quantification and optimisation process. Objective 4 - Demonstrate the application of the proposed framework through ontology and user interface (UI) development. Objective 5 - Validate the proposed framework using real-life case study-building data. The rest of the paper is structured as follows. Section 2 briefly explains CW, CW quantification and optimisation, the benefits and limitations of existing CW quantification and optimisation models, the role of SWT in supporting interoperability across various digital platforms, and SWT-BIM integration within the AECO context. This section ends with highlighting the research gap. Section 3 explains the methodology adopted for the research. The research framework and its associated modules are discussed in Section 4. The application of the framework through the development of an ontology and BIM-based UI is discussed in Section 5. This section also presents the outputs of the framework validated using a test-case building. Section 6 discusses the results obtained from the research. The implications of the research outcomes are discussed in Section 7. Finally, Section 8 concludes the paper by highlighting the contributions the research has made and limitations that need to be accounted for in future works. 2. Literature review This section briefly overviews the definition of CW, CW quantification and optimisation, existing CW quantification and classification models, their technological features, applications across the project life cycle, and limitations (Objective 1). In addition, the role of SWT and studies focusing on SWT-BIM integration to maximise sustainability in the built environment are explored (Objective 2). These two objectives were underpinned using a literature review. Finally, this section highlights the research gap that justifies the significance of SWT-BIM integration in enhancing the CW quantification and optimisation workflow. 2.1. Construction waste (CW) Waste Framework Directive (WFD) defines waste as ‘any substance or object which the holder discards or intends or is required to discard’ [11]. Accounting for the whole life-cycle of the project, waste from the built environment comprises materials generated from a wide range of activities, including manufacturing, procurement, construction/installation, operation and maintenance, and demolition/deconstruction, as highlighted by [12]. However, the material waste from the built environment is generally termed Construction Waste (CW) or Construction and Demolition Waste (C&DW). So, the term Construction Waste (CW) used in the research represents the material waste generated across the whole life cycle of the project. 2.2. CW quantification and optimisation Due to increased material diversity across the project life cycle stages, a detailed evaluation of CW quantity, type, source, time, and impacts becomes essential to select the most suitable waste treatment routes and allow safe disposal [13]. This evaluation is beneficial in establishing appropriate guidelines, strategies, and policies supporting sustainable CW management [14]. Thus, evaluating the generation rates, treatment routes, and associated sustainable impacts of CW is termed ‘construction waste quantification (CWQ)’. The terms used to refer to CWQ would differ based on the scope covered in the evaluation process. If the quantification process is performed for a specific project, it is termed project-level quantification. Meanwhile, measuring the waste generation across a specific region encompassing multiple projects is termed regional-level quantification [8]. Each of these quantification types has unique applications on WM. For instance, project-level quantification aims to support comprehensive WM planning across the life-cycle stages, determine the generation rates, and support benchmarking based on the project functionalities. The regional-level quantification supports the government and policymakers in setting legal waste targets and developing WM regulations based on geographical scope [8]. Among the two major types, the research supports the projectlevel quantification process. The research proposes a framework applicable in the design stages of a building project, thus characterising the project-level CWQ model. Further, utilising the acquired CWQ results, the material deemed less waste-generating and less impactful is optimised. The computational process facilitating the material selection using CWQ results is termed ‘CW optimisation’ within the scope of this research. 2.3. Overview of existing CW models Due to CW’s increased importance, several digital tools, processes, and regulatory frameworks have been developed in the past decade. The scope of the research explores only the technological dimensions of the CW models that support quantification and classification but not the actual process and regulatory frameworks. The widely used ICTs include BIM, AI-supported waste prediction and image recognition models, autonomous robots, and GIS. The ICTs are broadly applied across three levels of CW assessment: data collection, processing, integration, and presentation. The application of ICTs across these phases is analysed briefly below. 2.3.1. Data collection Access to and use of accurate data has been shown to pave the way for efficient quantification of CW across the project life cycle. Past studies incorporated ICTs like BIM and GIS to extract accurate data required for quantification. Cheng and Ma [15] was the first study that adopted BIM to extract the Material take-off (MTO) details from the 3D S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 2 BIM model to estimate demolition and renovation waste. Further, the BIM data supported the development of a CW prediction model that aids in measuring waste generation levels from new buildings [16]. Besides, BIM and multi-criteria decision-making algorithms were integrated to measure the recycling value, energy consumption, carbon and cost of demolition waste materials from building projects [17]. The BIM information is further used to determine the influence of design options and construction schemes on CW generation levels [18]. The use of BIM in CW assessment was further expanded to other BIM dimensions. For instance, a 4D-BIM framework was developed to estimate concrete and drywall waste quantities during construction stages. The scheduling data was incorporated with design information to estimate the reuse and recycling potential of waste materials [19]. Another study used 7D BIM data in new construction to measure waste volume and cost, schedule waste pick-up time, and its reuse and recycling capabilities [20]. Based on the review, it can be seen that BIM is primarily used to extract design and material information from new constructions during the design and construction stages. In addition to BIM, GIS played a vital role in acquiring the dimensional and material information of existing built assets. The GIS images were used to capture the spatial and temporal data of existing buildings. For instance, Wu et al. [21] used GIS images to estimate demolition waste generation at the regional level. The building usage type, type of structural frame, GFA, and service life of each building were collected using ArcGIS to measure the expected waste generation levels. The same application can be seen in measuring demolition waste during the urban renewal process in China [22]. It has also been proved that errors occurring from other building records can be mitigated by incorporating GIS images into CW assessment [23]. In addition to measuring building layouts, the site locations and spatial features of waste recycling plants and landfilling sites were also analysed using GIS maps to facilitate efficient waste treatment and disposal [24]. 2.3.2. Data processing Unlike conventional methods, data processing is managed by integrating AI techniques, including machine learning (ML) and deep learning (DL) models. The predictive analytic capabilities of AI models were widely utilised in previous studies to forecast CW generation levels at project and regional levels. Analysing AI-based waste reports allows researchers to demonstrate the relationship between various project factors and CW generation levels. For instance, the results from a hybrid AI model show that the type of construction and Gross Floor Area (GFA) of the building are two leading indicators influencing the CW generation levels in new building constructions in the UK [16]. In addition to predicting or forecasting CW generation levels, AI models help discover patterns from digital CW images that ease the on-site waste classification, sorting, and pick-up process. Other advanced ICTs, like autonomous robots, were also integrated with AI models to automate the onsite waste-handling process [25,26]. This automated process further ensures the health and safety of stakeholders. These studies prove that incorporating AI lessens the quantification time, enhances data accuracy and minimises human intervention. 2.3.3. Data integration and presentation Results from the review showed that communication between CW models and BIM is enabled through the Application Programming Interface (API) of BIM authoring tools. For instance, [16] developed a plugin within the BIM environment to allow users to run the CW analysis and develop a CW model that supports decision-making in the design phase. The API was also used to integrate the CW information into the BIM model, which supports detailed estimation [15]. The results from the study also emphasised that the API features of BIM authoring platforms enhance their integration with CW tools [16]. As a result of this integration, researchers witnessed many benefits ranging from enhanced visualisation of waste results, user interaction, flexibility in design-optimisation, data-sharing, ease of communication, and collaboration among the project stakeholders. A detailed description of each model, the level of technology adoption, its applications across the building life-cycle, and limitations are discussed in [10]. 2.4. Semantic web technologies (SWT) The semantic web enables a shift from the traditional web from being a ‘document medium for people’ to a ‘data and information medium that can be exchanged and integrated automatically’ [27]. It accesses information based on the ‘meaning’ rather than the ‘syntax’, allowing machines to recognise data and extract knowledge across various sources. The growing technologies generate piles of heterogeneous data that must be seamlessly integrated to discover patterns and acquire knowledge. These gaps are filled by the semantic web, which allows machines to analyse data independently and have seamless interaction with various systems, applications and disciplines [27]. Apart from the traditional web, the data retrieved from the semantic web are more accurate, interoperable, and consistent. The logical connections/relations between the interconnected terms enable interoperability between different systems. Also, the meanings are derived from a structured data hierarchy to develop knowledge domains. Structured information that can enhance reasoning and knowledge representation is considered to facilitate semantic web functions. Overall, the versatility and the ability to store and process metadata call the semantic web a robust decentralised system that enables efficient collaboration between machines and humans [27], which is critical to enhancing the productivity and functionality of the CW platform. The primary part of the semantic web is ontology. Ontologies are metadata schemas that provide consistent vocabularies, each with explicit specification and machine-readable semantics [28]. In the computational field, an ontology defines standard semantics to ease understanding of the information and knowledge sharing within a domain by conceptualising the entities and relations among them [29], thus called an ‘explicit specification of conceptualization’ [30]. An ontology must be encoded in predefined formal languages to enhance the reasoning mechanism throughout the ontology lifecycle. Several logic and web-based standard languages are used for ontology modelling based on the semantic architecture/stack. The languages include OIL, DAML, DAML+OIL, SHOE, SHACL, and OWL [31]. However, the World Wide Web Consortium (W3C) recommends the Resource Description Framework (RDF) and its extension language, Web Ontology Language (OWL), as standard lightweight languages for representing semantic data [32], which is used in this research. In the current scenario, incorporating diverse technologies leads to the generation of big data that needs to be seamlessly integrated across multiple applications. Big data represents the data’s large volume, variety, and velocity, called 3Vs [33], a typical scenario in the AECO industry. The heterogeneity (variety) being a predominant barrier to achieving interoperability, the RDF provides a common framework that allows fragmented information to be linked and exchanged with multiple systems, applications, and disciplines on the web [34]. The RDF data will be represented in an easy-to-understand labelled graphical format called triples. The triples include subject, object, and predicate, in which the ‘subject’ and ‘object’ are linked via a ‘predicate’. Hence, the RDF data model is called a triple store. It creates a chain of triples, where an object of one triple would form a subject of another connected through a predicate. The predicates that have literal values are called data properties [32]. The standard way to extract the knowledge in the ontology is through query languages. Several languages, including SPARQL Protocol and Query Language (SPARQL), Semantic Query-enhanced Web Rule Language, RDF Query Language, Sesame RDF Query Language, Triple Query Language, and OWL 2 Query Language are used to query ontologies [35,36] based on the ontology syntaxes. Among all existing languages, SPARQL was selected to query the ontology designed for this research. The critical reason for selecting SPARQL is that it is one of the S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 3 W3C-acknowledged languages to query semantic web ontologies presented in RDF syntax. SPARQL aims to build queries to extract, analyse and manipulate RDF data models stored on the web or in an RDF store. SPARQL, a ‘tell and ask’ system, presents queries as statements that are easy to understand. It has structured query language syntax, representing its query patterns as triples [37]. 2.5. Integration of BIM and SWT for sustainable built environment BIM is a process that makes a complete paradigm shift in the project workflow and delivery methods. The development of BIM applications is perceived as a platform to enhance the integrated project delivery methods, collaboration and communication across the supply chain [38]. The output produced by BIM is called the Building Information Model, a data-rich, object-oriented, parametric and intelligent digital representation of the built asset, which characterises the existence of geometrical and non-geometrical information that can be utilised across the project life-cycle [38]. Inclusive of all, potential benefits of BIM across the life-cycle stages are driven by: i) technology (3D-nD modelling and simulation, federated BIM model, parametric modelling, clash detection and visualisation), ii) information (functional and nonfunctional information, interoperable Industry Foundation Classes (IFC) data, life-cycle data, information management, and shared data environment), and iii) process management (real-time collaboration, stakeholder integration, seamless information exchange, increased automation) factors [6,39]. The interoperability with data and information exchanged among different systems and participants is vital to enhance communication and collaboration across various digital platforms. In the BIM context, the interoperability is achieved using IFC. The desired IFC can be encoded in various formats, including STEP Physical Format, Extensible Markup Language (XML), JavaScript Object Notation, Turtle, and RDF. The Turtle and RDF formats are developed to support the semantic web and linked data applications of IFC models using OWL, called ifcOWL [40]. The applications of SWT in the AECO context can be seen across the whole life cycle, aiming to enhance interoperability, digital collaboration, performance analysis and information management. Several studies investigated the applications of SWT-BIM integration to achieve a sustainable built environment. Sobhkhiz et al. [41] developed a material ontology to estimate the embodied carbon emissions of materials during the product manufacturing stages and the impact of material supplier location on overall carbon levels. The study integrated the LCA data with BIM to develop the semantic knowledge base and support carbon estimation. The main goal of the research was to showcase the importance of SWT in BIM-LCA integration. Another study by [42] utilised the SWT to integrate the carbon data from Environmental Product Declaration (EPD) certificates with BIM. This study provided an overview of estimating carbon emissions during the product manufacturing stages based on the EPD data stored in the ontology. However, the carbon data from the ontology was manually imported into the BIM model as a custom parameter. Besides, the SWT applications can be seen integrated with BIM to develop a decision-support system that facilitates the development of sustainability practices from early life cycle stages [43]. Besides, the BIM-SWT integration was utilised to develop a cloudbased environment that provides real-time data to designers to assess the green building rating scores. The quality and quantity of design information in the BIM model and the semantic definitions established by SWT facilitated the assessment and scoring process [44]. Taking advantage of BIM and SWT, Jiang et al. [45] developed a GBEOntology to overcome the gaps associated with fragmented building information and facilitate the green building evaluation process. Other studies with SWT-BIM integration support energy analysis [46], project cost estimation [47], and incorporation of product manufacturer data into BIM [48] to develop an interoperable knowledge base. 2.6. Research gap As a result of complex non-linear material flows in the AECO industry, an enormous amount of information needs to be collected, organised, processed, and mapped to make informed judgments. Similarly, the information required to make informed WM decisions comes from various distributed data sources; its direct application into the CW models is constrained due to heterogeneity, inconsistency, and noninteroperability. The integration of ICTs in previous studies offered various benefits across data collection, processing and integration in the CW assessment process. However, there is a paucity of literature to identify and standardise the critical type, quality, and quantity of waste information required for automating and improving the CW quantification and optimisation process. Although a few studies [49,50] adopted coding systems to reduce heterogeneity between design and waste data, the mapping patterns are mechanical, study-specific and lack wide-scale applications. So, the absence of standardised and interoperable information to automate the quantification and optimisation workflows is one of the significant limitations of current studies. Another significant gap is the lack of dynamic interaction between the CW and BIM platforms. Most existing CW models have a static relationship of extracting design data from BIM and conducting CW assessment externally. This approach is highly iterative, and communication between the platforms is time-consuming. Also, the current studies lean towards weight and volume-based estimations, so the arguments made from those estimations tend to focus more on inert waste because of its heavy weight and ignore hazardous waste because of its negligible weight and volume. For instance, the study by [51] emphasised that hazardous materials like gypsum and paint contribute to merely 2 % of the total waste volume, but their environmental impacts are significant. Similarly, recycling aluminium waste contributed to 45 % of total carbon emissions despite its weight being 0.66 % of the total waste quantity [52]. So, as a result of volumetric estimations, some lightweight materials with high environmental impacts often become overlooked. Thus, accounting for these limitations, shifting from a weight-based to an impact-based waste metric system and setting targets based on CW impact levels are encouraged [53]. It is also emphasised that this shift would ensure that the waste domain supports Net Zero Carbon targets. Besides, the study conducted by [54] also emphasised that Global Warming Potential (GWP) is one of the critical determinants supporting the assessment of demolition waste performances. Although various studies have explored the applications of SWT for linking environmental data with BIM, there is still a gap in the CW dimension. This creates a scenario where CW-related information is not yet integral to the AECO domain, causing interoperability and communication gaps between these domains. Besides, the lack of homogeneity in representing CW information offers limited time for the stakeholders to plan for circular material flows [55]. So, based on these identified knowledge gaps within the current system and prospects of SWT-BIM integration, this research seeks to develop a BIM-integrated semantic framework enabling a unified data structure and standardised information for impact-based CW quantification and optimisation from the early design stages. The semantic CW data model and UI developed using the framework ensures interoperability and automation across the overall quantification, classification, and optimisation workflow. 3. Methodology The research follows the six-step flexible approach proposed in the Design Science Research (DSR) framework. Being a problem-solving paradigm, DSR aims to develop an innovative solution to real-world problems by generating innovation artefacts and design knowledge. Such generated design knowledge may be constructs, models, methods, and instantiations [56]. The research outcome aligns with the ‘method’ knowledge, where the guidelines and workflows are proposed to solve the interoperability and automation gaps within the CW domain. Fig. 1 S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 4 depicts the six steps of the DSR approach. It includes the following: i) problem identification, ii) define research objectives, iii) design and development of the solution, iv) demonstration or application of the solution, v) evaluation, and vi) dissemination and communication of the knowledge. The problem identified, and the objectives defined in the research are discussed in sections 1 and 2. The development of the solution is presented through a semantic framework (Section 4) that aims to develop an interoperable and automated system supporting CW quantification, classification, and optimisation from the early design stages of building projects. The existing research gaps and the perspectives of AECO stakeholders are integrated to develop a collaborative solution. The application of the proposed framework is demonstrated via a semantic web application (Section 5.1) and a BIM-integrated application (Section 5.2). These applications led to the development of an ontology (also called semantic CW data model or RDF model) and a BIM-enabled UI, respectively. The semantic data model (ontology) and the supporting UI clustered within the proposed framework are tested using the test case building data. The case study design helps researchers investigate ‘how’ and ‘why’ questions used to develop the overall context precisely [57]. Similarly, the final objective of the research is to identify how the BIMintegrated semantic framework is enhanced from the conventional CW measurement, reporting, and optimisation process in terms of interoperability and automation. In addition, this application-based evaluation is also considered appropriate for validating an ontology’s vocabularies, concepts, data, hierarchies, and semantic relations [58]. It helps to decide the quality of the ontology based on the accuracy, consistency, and completeness of the results to the domain specification. So, real-life test-case-building data was used to evaluate the application of the semantic data model and the UI. 4. Research framework The BIM-based semantic CW quantification and optimisation framework and its associated modules are discussed in this section (Objective 3). The framework comprises three main modules: data collection or extraction, data mapping and standardisation, and BIMbased waste assessment and decision-making (Fig. 2). Below is a detailed description of each module with desired inputs, processes, and outputs. 4.1. Data collection/extraction module The data collection/extraction module illustrates the data types required for CW quantification, classification, and optimisation from the early design stages. The three main data clusters include building and Fig. 1. DSR process (adapted from [56]). Fig. 2. BIM-based semantic CW quantification and optimisation framework. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 5 design details, waste factors, and technical, physical and GWP of materials. The type of data influencing the overall process was clustered based on the outcomes of the literature review and stakeholder perspectives. Since BIM is considered a rich data source, the first cluster, building information, including the building’s geometric and functional properties, is extracted from the BIM model. The Level of Model Definition (LOMD) decides the quality and quantity of design and material information. The LOMD represents the level of information existing within the developed BIM model to maintain data quality and ease information exchange. The two components of LOMD defined by BS ISO 19650 (replaced version of PAS 1192–2) include the Level of Model Detail (LOD) and Level of Model Information (LOI). The LOD describes the level of graphical data, and the LOI relates to the level of nongraphical information within the BIM model. Although there is no linear relationship between the LOMD and project life-cycle stages, the proposed framework meets the LOMD 3 requirements to integrate the CW analysis with early-stage design workflows. This level equals LOD 200 within the LOD framework defined by the American Institute of Architects. There are multiple CW data sources, including existing site waste records, statistical reports, waste disposal records, and secondary data sources to extract the waste factor of a building material [8]. Thus far, no single waste data source has been found to have broader implications across various applications. So, the accessibility, availability, and completeness of datasets should be accounted for before making a selection [10]. Two different conditions are presented in the model to obtain the waste factor. Firstly, if the extracted data is an analysed and cleaned report, it can be directly used for assessment. This includes data from CW benchmarking and data collection platforms. In contrast, the extracted data will be raw if extracted from site waste records or direct site visits. In such cases, statistical or AI-based analysis should be conducted before integrating the extracted waste data with other data clusters. The choice of analysis will depend on the volume, variety and velocity of the data (3Vs), as suggested in [33]. The third cluster, technical, physical, and GWP of materials, is vital to provide designers with the best-optimised materials. The primary data sources for collecting these data are EPD data repositories. These databases provide detailed information that can be used to measure the environmental impacts of building materials across their whole life cycle. The granularity of the data within the EPDs makes it appropriate to measure environmental impacts precisely. Despite the benefits, inconsistency and incompleteness of LCA impact values hinder the options to compare multiple EPDs of the same material type [59]. Therefore, a detailed investigation should be conducted to select EPDs that can be compared with other EPDs of the same material without any constraints. In case of missing EPDs, generic LCA data sources of building materials are also preferred. The type of data source should be selected based on the LOMD of the BIM model [59]. In general, it is preferred that if the LOMD≥3, product-specific EPDs are appropriate, while generic LCA reports are best fit for lower LOMDs [60]. Besides, the data extraction methods and data types differ depending on the databases selected. An increase in digital EPDs eases the data extraction process; however, it is also important to note if digital EPDs contain adequate technical and physical properties of a material in addition to emission values to facilitate a comprehensive CW assessment process. 4.2. Data standardisation and mapping module Integrating the CW model with the BIM platform is essential to demonstrate seamless communication between design, construction, and waste platforms. It is essential to verify that material, waste and EPD data used for computational analyses between various digital platforms are standardised before mapping. The difference in terminologies, data hierarchy and variable structures limit automation and dynamics within the CW quantification and optimisation process [10]. Besides, the Zero Avoidable Waste report highlights that differences in material and waste codes significantly impact interoperability [12], thus hindering the easy communication of building and CW platforms. Although the properties of these materials are similar, standard terminologies and classification systems are vital to achieving communication and automation within the process. Therefore, the Unified Classification System (Uniclass) 2015 [61] and European Waste Catalogue (EWC) codes [62] are proposed to ensure consistency with the design, material, and CW information. In addition, API interaction within the framework modelling emphasises using unique IDs of materials within the LCA databases. Thus, considering the limitations like varied data structures and formats, unstandardised information and lack of dynamics within the optimisation process, an ontology aided by the SWT is proposed. The diverse data collected from CW and LCA databases is standardised to an RDF format, thus leading to the development of an ontology. The ontology modelled in the RDF format is integrated with BIM via an API or an interface supported by the selected BIM modelling platform. This integration enables automation across the data extraction, quantification and optimisation, and presentation of the results. It also aims to structure the information flow and ease waste data exchange across the building life cycle. The standardised and consistent information in the ontology will also lead to the development of reliant and complaint CW databases. The ontology becomes valuable only if the knowledge stored in it is extracted. The semantic SPARQL queries are used to retrieve explicitly or implicitly stored information in the ontology. The outputs retrieved from the queries are the primary results incorporated into the BIM model to measure the CW and its environmental impacts. 4.3. BIM-based CW assessment and decision-making module A BIM-based CW assessment and decision-making module is introduced to enhance the dynamics between the BIM platform and the semantic CW data model. This component will provide a UI within the BIM environment for selecting an optimal design or re-selecting another design option should the initial design be deemed highly wasteful. The decision-making layer will allow users to CW waste generation levels for each material layer of an element individually, which will then be extended to other elements within an element family. Hence, the waste results of multiple element families that make up the building structure will provide the total waste generation value of the building (Fig. 3). The GWP impacts and treatment methods of each waste material are also presented to allow users to select an optimised design that meets the project requirements. The technical, waste and carbon information of materials extracted from the ontology are presented as custom parameters within the BIM model. The assessment can be performed with one or multiple cycles until the user approves the specific material appropriate for the design. A new element family will be created as a final step, and the waste results will be presented in a tabular and graphical format. The quantified waste results incorporated into the BIM model allow the user to make waste-informed decisions from the initial stages of a project, thus overcoming the limitations of noninteroperable and static information flows between BIM and CW platforms. This decision layer further allows users to benchmark and compare multiple design options in the virtual environment before construction begins. So, the proposed semantic framework aims to enhance standardisation, consistency, and interoperability between BIM and CW platforms, further promising automation within the CW assessment workflows. 5. Demonstration and application of the framework This section demonstrates the application of the proposed framework through the development of a semantic CW data model and a BIM-based UI (Objective 4). This section also discusses the results of the proposed framework evaluated against real-life building data (Objective 5). S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 6 5.1. Ontology (semantic CW data model or RDF model) Various methods are used to model ontologies, including Cyc, Uschold and King’s, Toronto Virtual Enterprise (TOVE) modelling, METHONTOLOGY, Simple Knowledge Engineering Methodology (SKEM), and NeOn. Of all methods, the ontology modelling in this research is guided by SKEM. The critical reasons for selecting SKEM, among other methods, include the ease of understanding and use, accounting for reusing existing ontological resources, high level of specifications, flexibility, iterative and agile nature, quality documentation about the methodological process, and its integration with Prot´ eg´ e ontology editor (the selected tool for ontology modelling). Besides, this methodology is widely adapted across various domains [64,65] and has been highly preferred to date [66]. Seven steps are associated with the conceptualisation, formalisation, and implementation of the ontology using SKEM. The following sections will discuss each step and its application in the research. However, it is vital to note that ontology development is not linear and may undergo iterations across the modelling stages until it meets the requirements. 5.1.1. Determining the domain and scope of the ontology The primary step of ontology modelling is determining the domain and scope the ontology will cover. Some basic questions that must be answered to facilitate ontology designers determine the scope of the ontology include the following: •Purpose of the ontology - The ontology aims to develop an interoperable BIM-integrated semantic model with a unified and consistent data structure to enhance CW quantification and optimisation from the early design stages. •End-users of the ontology - The leading users are the design team members, including designers, architects, and sustainability professionals involved in making CW-related decisions in the design phase. •Intended uses of the ontology information - Develop a knowledgebased information system to provide waste generation levels and optimisation choices for building materials used in BIM designs. The information aids in making informed waste decisions and optimising materials and designs, thereby aiding the AECO industry in minimising CW generation, maximising building circularity, and reducing whole-life carbon levels. •Domains covered - The ontology covers the CW domain of building projects in the UK. It includes technical, environmental, and generic elements of CW that significantly impact its management. As ontology development progresses, designers have the flexibility to change or update the answers to these questions. So, any changes made from the initial answers will be highlighted in the following sections. 5.1.2. Considering the reuse of existing ontologies After the domain specification, the fundamental step of conceptualisation is to check for any existing general or domain ontologies that can be reused for modelling [29]. The ontology reuse is a process in which the existing ontologies (as a whole or individual terms/statements) serve as an input to model new ontologies. Reusing existing knowledge enhances the interoperability of the ontological model across the selected domain of interest and reduces information redundancies. However, different taxonomies, varying levels of formality, versions and updates increase the manual work associated with integrating ontologies and complicate the reuse process [67]. Considering these factors, ontologies that are most relevant for the research are reused in the research. Several ontologies within the AEC domain focusing on materials have been developed. The two ontologies, Building Materials Ontology (BMO) [68] and Materials Properties Ontology (MAT) [69], provide the explicit relationship between elements, materials and their properties aligning with BIM. Although multiple material properties are included in the BMO, they all lean towards technical aspects while holding limited information to aid in accurate and granular CW assessment. So, only some of the concepts and attributes of BMO and MAT ontologies supporting the building material classification are included. Further, the ‘QUDT:Unit’ concept within the Quantities, Units, Dimensions, and Types (QUDT) ontology, which provides a unified model for measurement, is incorporated into the research. 5.1.3. Enumerating important terms in the ontology The terms include taxonomies of concepts, attributes, and individuals required to fulfil the ontology specification requirements. In this research, the factors affecting CW generation, including building type, material type, materials’ dimensional features, density, service life, waste generation rate (WGR), embodied carbon, the quantity of raw and secondary materials used for construction, waste treatment scenarios, applications of materials, Uniclass codes, EWCs, and manufacturer details are some of the essential terms gathered to facilitate ontology modelling. The above-listed factors depend on the information collected from the non-ontological resources: a literature review, stakeholder interviews and domain-related knowledge sources. The results of the literature review and stakeholder interviews are presented in [10,70], respectively. The domain-related knowledge sources selected for modelling include BRESmartWaste data, EPDs, Uniclass, EWC, and whole-life cycle assessment guidance provided by [71]. Since the gathered data and knowledge are heterogeneous, they are organised and structured to develop a conceptual knowledge model that facilitates ontology modelling. This process is called conceptualisation [72]. As [29] suggested, these terms provide an outline to the developer to model ontologies, so any overlapping and redundancies existing with these terms are updated during formalisation. 5.1.4. Defining the class and class hierarchy of the ontology The conceptual model with an identified hierarchy of concepts and its associated properties are formalised and implemented into the ontology. Formalisation is a process of transforming informally defined knowledge in a conceptual model to a formalised ontology language. This transformation aims to enhance the ontologies’ computation factors [72]. The ‘concepts’ are also called ‘classes’ in the semantic web, so these two terms are used interchangeably. There are several approaches to class hierarchy development. This research borrows the three approaches suggested by [73] to develop a class hierarchy. It includes a top-down approach, a bottom-up approach, and a middle-out/ combination approach. Although no single method inherently Fig. 3. Building levels for CW quantification and optimisation – Cross-ref [63]. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 7 performs better than others, the choice depends on the ontology developers [29]. So, the middle-out combined approach is selected for the research. It is because of its flexibility in implementation while taking the edges of top-down and bottom-up approaches. Classes and their hierarchies: In the semantic web, each item is a resource identified by a Unique Resource Identifier (URI)/ Internationalized Resource Identifier (IRI). So, the names of classes, properties and instances used here are IRIs. They could also be written in a short form using their prefixed and local names. E.g. for the entity ‘http://tees.ac.uk/CW_Data/Lifecycle_CW_Impact_Assessment_Results#BuildingType’, the IRI http:// tees.ac.uk/CW_Data/Lifecycle_CW_Impact_Assessment_Results# is prefixed as PRODCIRO, and ‘BuildingType’ is the local name. So, the local names of the IRI entities are used in the following sections to ease communication. The main classes of the Product Circularity Ontology (PRODCIRO) include ‘BuildingType’, ‘BuildingElements’, ‘EPDProductType’, ‘EPDProductData’, ‘BuildingWasteMaterialType’, ‘EuropeanWasteCatalogue_EWC’, ‘UniclassPrCodes’, and ‘Unit’ (Fig. 4). Most of these top-level classes have middle-level and low-level classes. The type of information each class contains, its purpose in the research context, and its relationship with other classes are tabulated (Table 1). Each class within the ontology holds one or many relationships with other classes and attribute data, called object and data properties. These relationships are discussed in the following two sections. 5.1.5. Determining the properties of classes – Slots Defining classes is not sufficient to meet the purpose of ontology. A description of the internal structures and relationship of classes is also required, which can be achieved using properties. There are two main properties: object and data property. The object properties define the relationship between the classes of ontology. The list of object properties and their relationship with the ontology classes is presented in Fig. 5. The main object properties included in the ontology are ‘applicableForBuildingElements’, ‘applicableForEPDs’, ‘takesWGRFrom’, ‘hasEWC’, ‘hasUniclassPrCode’, ‘hasUnit’, ‘isEWCOf’, ‘isUniclassCodePrOf’, ‘isManufacturedBy’, ‘manufactures’, ‘isLocatedIn’, ‘isProductManufacturingLocationOf’. The object properties are selected based on their level of application across the CW assessment process. The properties will be the predicate that connects the subject with an object, called triples, in the RDF representation. The subject and object are individuals of two different classes. The object of one triple could be the subject of another triple. For instance, Fig. 5 shows that the ‘EPDProductType’ and ‘ProductManufacturer’ classes form a triple via the object property ‘isManufacturedBy’. Also, the ‘ProductManufacturer’ and ‘ProductManufacturingLocation’ classes form a triple via the object property ‘isLocatedIn’. In this case, the class ‘ProductManufacturer’ is an object in the first triple and a subject in the second triple. Establishing the relationship of multiple classes forms multiple triples, representing the RDF data as statements. The subject and object of the property are also called domain and range, respectively. Each of the object properties and its relationship with the classes are presented in Appendix A.1. Data properties are predicates that help to create a relationship between an individual and an attribute data. Fig. 6 presents the complete list of data properties added to PRODCIRO. These data properties are established based on the primary and secondary information collected from the listed knowledge sources (Section 5.1.3). The data property axioms also have additional constructs: domain and range. Like object Fig. 4. List of classes added to PRODCIRO. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 8 property, the domain asserts that the subject of the property belongs to a specific class expression. In contrast, the range asserts that the property’s value must belong to a data value in a specific data range. The data range in RDF is borrowed from XML schema data types, and it would be a primitive or enumerated data type used to restrict the range of properties [74]. Appendix A.2 presents the domain and range constructs added to the data properties. Each data property significantly impacts the CW assessment process. The estimations supporting the assessment are expressed using the SPARQL endpoint integrated with the BIM environment (discussed in Table 1 PRODCIRO Classes. Classes Type of information the class holds Justification for its inclusion in the ontology ‘BuildingType’Type of building usages: residential and non-residential (commercial, industrial, educational, healthcare, leisure and public) buildings. According to BRESmartWaste reports, building typology is one of the main factors influencing waste generation during the construction/ installation phase of the building. So, this class is incorporated into the CW assessment. ‘BuildingElements’This is a primary class with a collection of building elements. The building elements are classified according to the Uniclass Element and functions (EF) classification. This class serves two purposes: 1) Each product has a unique application across various building elements. So, this class is used to select and match the products (within the ‘EPDProductType’ class) applicable to a specific element (within the ‘BuildingElements’ class). 2) The information extracted from the BIM model is categorised at the element level to support designers in assessing CW. So, this class serves as a filter to support the mapping of materials (based on its application) extracted from the BIM model with a list of material choices in the ontology. ‘EPDProductType’The EPDs of each building product come under this class. It is one of the main classes added to the ontology to classify building products based on their material composition. It links with the generic, technical, and environmental data added to the ‘EPDProductData’ class. ‘EPDProductData’A primary class that holds generic, technical, carbon, secondary material usage, and expected waste treatment scenario details of building products extracted from the EPDs. The data extracted from this class is integrated with waste data from the ‘BuildingWasteMaterialType’ class to quantify, classify and optimise materials based on CW impact results. ‘BuildingWasteMaterialType’Information on the material composition of CW recorded in the waste reports is added to this class. One of the main classes used to classify CW. The classification is based on their material composition. The data extracted from this class is integrated with waste data from the ‘EPDProductData’ class to quantify, classify and optimise materials based on CW impact results. ‘EuropeanWasteCatalogue_EWC’Contain EWC codes of each waste material reported in ‘BuildingWasteMaterialType’. These codes are used to select an effective treatment method and allow safe disposal of waste materials, aligning with the legal waste duty of care. ‘UniclassPrCodes’The primary data within the class is the Uniclass Product (Pr) codes and their corresponding names adopted from the National Building Specification (NBS) platform. The Pr codes in the class serve as a primary variable to filter materials by mapping the data from the BIM model with ontology data. ‘Unit’Declaration units of all technical and environmental information. It includes the unit of density, dimensional (length, breadth, and thickness) unit, service life (calendrical), GWP metrics based on the EPD declaration unit, and unit of waste generation rate based on GFA. It is one of the core classes added to the ontology to ensure consistency of the waste results. Fig. 5. Object properties of classes and their relationships. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 9 a compound structure comprising seven different material layers. The material layers include ‘GSS_Brick Red Multi Stock’, ‘GSS_Cavity’, ‘Fiberglass Batt’, ‘Cementitious Board’, Metal - Stud Layer’, and ‘Plasterboard’. Fig. 15 shows the material layers of an element that come under the selected wall family. 5.3.2.1. CW parameters of optimised materials. Each material layer is analysed separately after the element selection and material extraction. The materials within seven layers of the selected element of the BIM model are mapped with alternate material choices in the ontology via the Uniclass Product (Pr) classification. This stage is followed by quantifying and optimising the materials based on their ranks. The optimised materials are created as new materials in the project library, and the CW information of the new materials is imported as custom parameters. Fig. 16 presents the custom waste parameters for 1 kg of ‘AggregateBricks_2’. 5.3.2.2. Optimised material layers of the selected element. The BWT runs through each material layer individually, quantifies and ranks materials based on CW impact values, and replaces the existing materials with the optimised ones. The optimised material layers are applied to the duplicate family type ‘EW-01 Brickwork with SFS_1’ created before optimisation. Fig. 17 presents the duplicate family type and the optimised material layers. Analysing a single element within a family Fig. 14. Elements in ‘EW01 Brickwork with SFS’ wall family. Fig. 15. Material layers of ‘EW01 Brickwork with SFS’. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 16 simultaneously changes other elements in that family. So, ten elements of the same family are updated (Fig. 18). The test-case operational results for a multi-layered external wall within the selected BIM model ensure BIM-integrated automated quantification and optimisation with one click. Besides, the optimised materials were precisely mapped to the same material layer without disturbing the original element structure, thus obtaining the expected outcomes. This validates the first check that integrating the ontology and BIM achieves the core requirements expected from the semantic framework: interoperability, standardisation, consistency, and automation. 5.3.3. Test case optimisation results This section discusses the outcomes of the second check: evaluating the accuracy of the CW data. Since walls are the type of building elements incorporated into PRODCIRO, the material details within the walls are quantified and optimised. The case study building has 488 wall elements grouped into 23 wall families. The LOMD of the BIM model is <5, so only the dimensional and material details are included in the model. The supply-chain details, including the product-specific names, manufacturer, and supplier details, are limited. Thus, the optimal material is selected from the list of available alternatives. Fig. 19 shows a screenshot of a sample MTO of the wall category in the selected case study building. Fig. 16. Custom waste parameters of ‘AggregateBricks_2’. Fig. 17. Duplicated element of ‘EW01 Brickwork with SFS’ family with optimised materials. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 17 The aggregated list of MTO showed that 19 different types of materials were used across the 488 elements. This includes ‘90 mm Timber Slat’, ‘Air Openings’, ‘Blockwork’, ‘Cementitious Board’, ‘Default Mesh Wall’, ‘Default Wall’, ‘Fiberglass Batt’, ‘GSS_Brick Red Multi Stock’, ‘GSS_Brick Black’, ‘GSS_Cavity’, ‘Metal - Stud Layer’, ‘Plasterboard’, ‘Render Material 255-255-255’, ‘Roofing - EPDM Membrane’, ‘Soffit Cladding’, ‘Wall Cladding – Panels’, ‘Wall Cladding –Vertical 300mm Panels’, and ‘Window Louvers’. Out of all, materials that had no specific Uniclass Pr codes were removed by the filter. For instance, the material named ‘Default Wall’ lacks a specific material composition to allow the tool to quantify CW. So, these materials were removed by the tool. Further, materials like cementitious boards and cladding panels had either no or no single EPDs that met the geographical scope criteria. Hence, the lack of data limits the inclusion of these materials in the CW assessment process. Notes 1 The term ‘blockwork’ in the Uniclass is classified under the ‘Systems’ category, which does not align with the classification supported by PRODCIRO. So, the tool accounts for the primary material concrete block used to construct blockwork. 2 The bricks are currently classified based on their material composition in the Uniclass Pr classification. Due to the lack of colour factors in this classification, the optimisation approach adopted for GSS_Brick Black is similar to GSS_Brick Red Multi Stock. The results may slightly differ if the GWP values for black bricks are different from red-coloured bricks. Finally, the tool filters the six material types that met the prerequisite conditions. These are ‘Blockwork’, ‘Fiberglass Batt’, ‘GSS_Brick Red Multi Stock’, ‘GSS_Brick Black’, ‘Metal - Stud Layer’, and ‘Plasterboard’. The geographical location plays a vital role in deciding the CW generation levels, so the alternatives are the products with the UK as its geographical scope. The geographical scope of products is verified using EPD certificates. The design details, technical factors influencing waste generation, and waste generation levels of optimised materials are tabulated in Table 2. The obtained values are the impacts of various CW indicators like mass, density, service life, reused and recycled materials percentage, and embodied carbon associated with the material manufacturing process. Table 3 presents the usage of secondary materials, predicted waste treatment scenarios, and EWC of each material. Appendix B provides the complete list of available alternatives for each material type. The obtained values are validated for their accuracy. Although the research assesses the CW generation and its impacts across the project life cycle, no databases provide a sufficient waste range for these stages. Based on the thorough analysis of existing CW reporting tools in the UK and the research studies, it was evident that only a few survey reports and databases provide a wastage range for building materials. In addition, these databases only focus on the construction/installation stages. Besides, the extraction of actual waste data generated on-site is at the material level for the whole building and lacks classification based on elements. So, utilising this recorded data will be inappropriate as the PRODCIRO entities currently focus on wall elements. So, accounting for the data limitations, the wastage levels of materials during the construction and installation stages are validated against survey reports and on-site waste data collection tools. In addition, the product loss expected by the manufacturer during the construction/installation of the product is also included in the validation process. Table 4 shows the estimated wastage values of materials during their construction/installation stage against the existing waste data reports. Out of six materials, only two have sufficient data from the manufacturers to compare the accuracy of estimated CW values. The estimated wastage level of ‘SteelFrames_3’ is 8 % of the total material quantity, which is 3 % higher than the expected product waste by the manufacturer. Secondly, the waste percentage of ‘Plasterboards_7’ is 6 %, 4 % lower than the expected product loss. Although there is no waste data for ‘ConcreteBlocks_2’ from the manufacturer, the survey conducted by Reusefully Ltd. states that the wastage rate of concrete blocks ranges from 3 to 5 % [76]. Considering this range, the estimated CW for ‘ConcreteBlocks_2’ is 1 % higher than the prescribed range. This increased CW in ‘SteelFrames_3’ and ‘ConcreteBlocks_2’ might be a result of unstandardised designs leading to increased offcuts [70], improper handling of materials, and not utilising the supplier’s takeback schemes [76]. Thus, the project should incorporate these factors to reduce the CW generation levels. In many cases, the products are designed to be in custom sizes as required for specific designs, which, when used, will prevent the waste generated during installation on-site. However, the wastage levels of all materials are approximately close to the ‘good’ wastage rate proposed by the Waste Resource and Action Fig. 18. Optimised elements in ‘EW01 Brickwork with SFS’ wall family. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 18 Programme [77]. Overall, the CW quantification results obtained from the research will provide the accurate and consistent data required for developing comprehensive WM from the early design stages. At present, the ontology primarily focuses on integrating the indicators at the material and product level to quantify and optimise CW generation at the early stages of the project. Further, in addition to product-level factors, the design and construction factors that influence the CW generation levels could be added to the ontology to facilitate a comprehensive waste assessment process at the building level. This integration further allows the user to optimise processes and activities carried out across the project life cycle to maximise the CW reduction potentials of buildings. 6. Discussion This study sought to investigate what type of information and how semantic web ontologies, in combination with BIM-based visual programming language, can support CW data structuring, automation, and interoperability across the quantification and optimisation process. This section will discuss the key findings obtained from semantic CW framework development results and its validation against a test case. 6.1. Granular and standardised CW data ensure quantification accuracy and automation In the current scenario, only the waste factors of materials are Fig. 19. Screenshot of MTO of Wall Category in the case study building. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 19 primarily employed for quantifying and optimising CW. Even the advanced ICT-integrated models use only the material volume and waste factors to quantify the CW generation [10] without considering other influencing factors. However, the findings from the research demonstrated the influence of technical factors, including density, service life, dimensions, and building usage, on CW generation levels. In addition to the type (‘what’) of CW data received, the following prominent challenge identified was ‘how’ the data should be standardised and structured to incorporate into the analysis. The existing CW platforms require high human intervention and significantly lack communication with the design and construction platforms. This escalates the time and cost of quantification, compromises automation, and reduces the chances of comparing multi-design options before site work begins. Hence, utilising data standardisation, seamless data exchanges, and linked data attributes of SWT, the proposed BIM-integrated semantic framework is aimed to ensure automation and interoperability throughout the CW quantification and optimisation process. Besides, the automated quantity estimation, life-cycle information management, parametric modelling, and API features of the BIM platform increase the information standards, accuracy, reliability, and dynamics of the developed semantic CW data model. Thus, the standardised and interoperable CW data achieved by integrating BIM and SWT allow the AECO industry to utilise the full potential of ICTs in the CW domain and develop waste-efficient designs with less human effort and minimal time. 6.2. Semantic relationships unify the difference between material and CW characterisation codes Even though the coding system enhances accuracy in the CWQ results, there was a lack of analysis exploring viable options to overcome the differences between the material and CW characterisation codes. For instance, the research by [49,50] linked the Andalusian Construction Costs Database and EWC to enable linear communication between the BIM material library and CW data. However, these integrations are database-specific and lack defined semantic relationships. Besides, the coding approach used for material descriptions and waste classification does not align with each other. For instance, this research adopted Uniclass and EWC to enable seamless information flows between BIM and CW data models. In this regard, it was identified that EWC classifies waste materials based on their reactive nature, while Uniclass Pr codes classification in the AECO follows a granular approach based on their raw material composition, manufacturing process and applications. Due to significant differences in the classification approaches between Pr and EWC codes, the complexity associated with tracking the input and output material flows increases. These inconsistencies are also highlighted as significant limitations impacting the AECO industry in achieving robust and timely CW data [12]. Thus, utilising the SWT functionalities, the developed CW data model provides a unified structure by establishing comprehensive semantic relationships between building materials and CW data. The relationship between the material and waste codes is also established to ease material tracking. This unification aims to enhance integration and interaction across the different CW platforms, thereby improving the tracking and management of material flow information across its life cycle. Further, the standardised output from the model would act as a catalyst to overcome one of the critical gaps identified by the AECO stakeholders: collaboration and communication within the supply chain [55]. 6.3. Diverse CW measurement units support integrated decision-making The other influencing theme that required attention is adapting to a standardised unit of waste measurement. The current CW system employs various measurement units, including weight, volume, generation rate estimation, and percentage analysis. However, the design team members stress that having a standardised unit aligning with other sustainability indicators could increase the chances for CW optimisation Table 2 Optimal material choices and their CW values. Existing materials Optimised materials Uniclass Pr classification Volume (m3) Product Density (Kg/m3) CW_MS (tonnes) CW__MS (tCO2e) CW_CIS (tonnes) CW__CIS (tCO2e) Service Life (Years) Replacements CW_OMS (tonnes) CW_OMS (tCO2e) Metal-Stud Layer SteelFrames_3 Pr_20_85_32_84 318.26 7750 0.00 0.00E+0 200.281 3.40E+01 60 0 0.00 0.00E+0 Plasterboard Plasterboards_7 Pr_25_71_35_65 126.68 660 1.166 1.05E-02 4.013 3.62E-02 60 0 0.00 0.00E+0 Fiberglass Batt StrawInsulation_1 Pr_25_57_06 116.49 100 0.126 4.37E-05 0.228 7.92E-05 75 0 0.00 0.00E+0 Blockwork 1 ConcreteBlocks_2 Pr_20_93_52_05 80.51 650 10.73 2.96E+01 3.061 8.45E+00 100 0 0.00 0.00E+0 GSS_Brick Red Multi Stock AggregateBricks_2 Pr_20_93_52_02 71.61 1950 3.944 4.94E-02 8.169 1.02E+02 100 0 0.00 0.00E+0 GSS_ Brick Black 2 AggregateBricks_2 Pr_20_93_52_02 11.94 1950 0.658 8.24E-03 1.362 1.71E-02 100 0 0.00 0.00E+0 Abbreviations: CW_MS – Construction Waste generated during the product Manufacturing Stage, CW_CIS – Construction Waste generation during Construction/Installation Stages, CW_OMS – Construction Waste generation during Operation and Maintenance Stages. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 20 in the early design stages [55]. This study emphasises that, in addition to the quantity of waste generated from a specific design, designers are leaning towards understanding the environmental impacts of CW generation and its significance in meeting the sustainable targets the built environment sets. Further, the impact-based waste analysis over volume and weight-based estimations are promoted to meet the Net Zero Carbon targets and capital budgets [53]. Although it is vital to introduce an impact-based analysis, the weight-based results are crucial for the construction and WM team to provide adequate resources for waste storage, handling, treatment, disposal, and reporting. Thus, considering these multiple dimensions of design, construction, and WM teams, the proposed system integrated the impact-based assessment into the CWQ to ease optimisation without compromising the benefits of weight-based results for the on-site team. Besides, integrating the CW information with diverse measurement units within BIM allows users to opt for an optimisation method based on their requirements and at what stage of project life the decisions are made. This BIM-integrated approach, therefore, eases the whole assessment process and maximises the chances of collaborative decision-making within the CW domain. 7. Research implications Designers and Architects are the primary beneficiaries of the research outputs. With increased focus on design stages to reduce CW generation, designers and architects are encouraged to develop wasteefficient designs. Although various design-out waste principles and CWM methods exist, reliable waste data is considered vital to enhance decision-making. Thus, the semantic model proposed in the research provides an accurate and consistent CW data model that is interoperable with the BIM platform to make informed decisions. This, therefore, facilitates designers and architects to quantify waste, optimise wasteefficient materials, benchmark waste levels, and compare multiple design options in a virtual environment in no time. Sustainability Professionals - Playing an active role in the design and construction phases, sustainability professionals aim to set targets for each building to minimise material waste and maximise circularity. So, one of the critical barriers confronted in the current system is the lack of quality data to set standardised waste targets. So, the waste and its impact results from the research would support them in setting clear targets and bridging linear communication between the design and construction teams. Policymakers – Similar to sustainable professionals, policymakers are actively involved in setting building and national-level targets to reduce the impacts of CW. Despite the existing CW targets, reports [53] suggest a need for impact-based targets to reduce carbon and cost associated with waste generation. Thus, the impact-based information extracted from the semantic model implies a significant improvement in putting enforceable targets and policies in place to prevent CW generation, maximise circularity and reduce whole-life carbon emissions. 8. Conclusions, limitations, and future work Accurate quantification and classification of CW in the design stages is a prerequisite to implementing the WM measures across the project life cycle. However, the lack of standardised and unified data across the current CW platforms leads to inconsistencies and inaccuracies when reusing the information. This data gap impedes the AECO industry from utilising the full potential of advanced ICTs within the CW quantification and optimisation context. The research proposed a framework underpinned by BIM and SWT to quantity, classify and optimise CW generation from the early design stages. The application of the framework is demonstrated through the development of a unified and standardised CW data model and a BIM-based UI named PRODCIRO and BWT, respectively. The framework aims to ensure four main CW dimensions: ‘interoperable and standardised data’, ‘unified data structure’, ‘automated quantification and optimisation’, and ‘consistent classification and reporting’ across the project life-cycle stages. The methodology incorporated in this research allows users to access granular, consistent, and timely CW data to make informed design decisions. Further, accessing CW data within the BIM environment through the SPARQL endpoint allows designers to reduce the complexities oriented with unstandardised CW classification and diverse units of CW measurement resulting from external data sources. Integrating CW data into BIM further encourages seamless communication and collaboration across the supply chain actors involved in CW-related decisions. Overall, access to CW data inside BIM would increase the chances of making early decisions, thereby allowing the built environment to shift from implementing reactive WM measures to proactive and preventive measures. This shift further maximises the value of waste materials and reduces whole-life carbon emissions. Overall, the main contributions from the research include the following: Table 3 Optimal material choices, resource use and waste treatment routes. Optimised materials Secondary material (%) Reusability (%) Recyclability (%) Energy Recovery (%) Landfilling (%) EWC SteelFrames_3 83 0 92 0 8 17–04-05 Plasterboards_7 4 75 15 0 10 17–08-02 StrawInsulation_1 0 0 0 95 5 17–06-04 ConcreteBlocks_2 NA NA NA NA NA 17–01-01 AggregateBricks_2 32 0 90 0 10 17–01-02 AggregateBricks_2 32 0 90 0 10 17–01-02 Table 4 Comparison of estimated wastage with existing CW data reports. Optimised materials Estimated Wastage_CIS Existing CW data sources Expected product loss assumed by the manufacturer Reusefully Ltd’s Survey Report WRAP wastage rate (Good)* WRAP wastage rate (Baseline)* SteelFrames_3 8 % 5 % NA 5 % 15 % Plasterboards_7 6 % 10 % NA 15 % 22.5 % StrawInsulation_1 3 % Assumed as 0 % NA NA NA ConcreteBlocks_2 6 % NA 3 (Good rate) – 5 (Baseline rate) 5 % 20 % AggregateBricks_2 6 % NA NA 5 % 20 % AggregateBricks_2 6 % NA NA 5 % 20 % * The latest report was published in 2008, so this should be accounted for while making decisions. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 21 i) The research is the first attempt to integrate BIM, CW, and lifecycle impact data semantically. Previous studies focused on the CW quantification and classification process with limited focus on the data. This research focused on enhancing the ‘CW data structure and standards’ that support the entire ‘process’. ii) A bidirectionally connected framework is proposed to enable seamless information flows between the BIM and semantic CW data models. iii) A semantic data model is developed with a granular set of technical and environmental information to enhance the integration of the CW-AECO domain. iv) The UI ensures that data extraction, material mapping, quantification, optimisation, and visualisation of waste results are automated and integral to BIM. v) Integrating the dimensions of environmental sustainability and innovational approaches within the AECO-CW domain, the research contributes to some of the Sustainable Development Goals proposed by the United Nations. It includes Goal 12 - Resource Consumption and Production, Goal 9 - Industry, Innovation, and Infrastructure, and Goal 13 - Climate Action. The research has some limitations that must be accounted for in future work: i) Although the research integrated a diverse range of site-specific scenarios to present the differences in CW, the data collected primarily represents the UK. So, the influencing waste indicators and generation levels might differ from other geographical locations. So, future research should account for a granular set of indicators to expand the semantic framework applications across various regions. ii) At present, the ontology primarily focuses on integrating the indicators at the product level to quantify and optimise CW generation at the early stages of the project. Further, the design, construction/installation factors that influence the CW generation levels could be added to the ontology to facilitate a comprehensive waste assessment process at the building level. This integration further allows the user to optimise processes and activities carried out across the project life cycle to maximise the CW reduction potentials of buildings. iii) Although the research acknowledges the impact of accurate and standardised CW data from the early design stages on making informed decisions, the increase in the quantity of data might increase the complexity of BIM models. This overloading of data may present some challenges to quality and timely decisionmaking. Therefore, developing a UI module with traceable layers that allows users to prioritise and filter decision-making criteria per their project requirements could be a potential future research area. In addition, future research works could explore the possibilities of enabling interaction between PRODICRO and other ontologies to initiate the development of an integrated decision system. This, therefore, eases the integration of ontology data and BIM and reduces the data overloading in BIM. iv) The current approach is based on a single environmental criterion, i.e. CW impacts of materials. Although the impacts of embodied carbon are widely acknowledged, other dimensions, including economic, social, health and safety, may influence the optimisation. Thus, integrating multiple criteria that influence CW decision-making is significant. Another potential research area could be integrating AI elements with the developed BIMbased semantic framework to enable multi-criteria decisionmaking. Since the existing semantic waste model provides structured and consistent waste data, incorporating the AI models could support the researchers to reduce inconsistencies and gain valuable insights, thereby utilising the full potential of ICTs across the CW assessment and management process. Data is one of the critical pillars of the AI process, so the developed ontology with standard semantics could support future research to obtain quality CW data for modelling. CRediT authorship contribution statement Subarna Sivashanmugam: Writing – original draft, Visualization, Validation, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Sergio Rodriguez Trejo: Writing – review & editing, Validation, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Farzad Pour Rahimian: Writing – review & editing, Validation, Supervision, Methodology, Funding acquisition, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability The SPARQL queries and the Dynamo Revit scripts supporting the complete CW assessment are added to the GitHub page - https://github. com/SubaSiva95/CW-assessment-via-RDF-BIM. It can be accessed upon request. Acknowledgements This paper reports a research project that was part of the lead author’s PhD study, partially funded by Teesside University. This study also received a grant from UKRI Innovate UK-Horizon Europe Guarantee Extension (ref number. 10078543) within the Horizon Europe Project “A Territorial Construction System for a Circular Low-Carbon Built Environment-RECONSTRUCT” (GA-101082265). Appendix A. Facets associated with the attributes of the ontology A.1. Facets (instance type slots) assigned to the object properties Object Property Description Facets (Instance type slots) Domain Range ‘applicableForBuildingElements’Defines the relationship between the classes of ‘EPDProductType’ and ‘BuildingElements’. ‘EPDProductType’‘BuildingElements’ (continued on next page) S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 22 (continued) Object Property Description Facets (Instance type slots) Domain Range ‘applicableForEPDs’Defines the relationship between the classes of ‘BuildingWasteMaterialType’ and ‘EPDProductType’. It is also an inverse property of ‘takesWGRFrom’. ‘BuildingWasteMaterialType’‘EPDProductType’ ‘hasEWC’Defines the relationship between the class ‘BuildingWasteMaterialType’ and ‘EuroepeanWasteCatalogue_EWC’. It is also an inverse property of ‘isEWCOf’. ‘BuildingWasteMaterialType’‘EuroepeanWasteCatalogue_EWC’ ‘hasUniclassPrCode’Defines the relationship between the classes of ‘EPDProductType’ and ‘UniclassPrCodes’. It is also an inverse property of ‘isUniclassPrCodeOf’. ‘EPDProductType’‘UniclassPrCodes’ ‘hasWGRFor’Defines the relationship between the classes of ‘BuildingType’ and ‘BuildingWasteMaterialType’. It is also an inverse property of ‘isWGROf’. ‘BuildingType’‘BuildingWasteMaterialType’ ‘isEWCOf’Defines the relationship between the class ‘EuroepeanWasteCatalogue_EWC’ and ‘BuildingWasteMaterialType’. It is also an inverse property of ‘hasEWC’. ‘EuroepeanWasteCatalogue_EWC’‘BuildingWasteMaterialType’ ‘isLocatedIn’Defines the relationship between the class ‘ProductManufacturer’ and ‘ProductManufacturingLocation’. It is also an inverse property of ‘hasEWC’. It is also an inverse property of ‘isProductManufacturingLocationOf’. ‘ProductManufacturer’‘ProductManufacturingLocation’ ‘isManufacturedBy’Defines the relationship between the classes of ‘EPDProductType’ and ‘ProductManufacturer’. It is also an inverse property of ‘manufactures’. ‘EPDProductType’‘ProductManufacturer’ ‘isProductManufacturingLocationOf’Defines the relationship between the class ‘ProductManufacturingLocation’ and ‘ProductManufacturer’. It is also an inverse property of ‘isLocatedIn’. ‘ProductManufacturingLocation’‘ProductManufacturer’ ‘isUniclassCodeOf’Defines the relationship between the classes of ‘UniclassPrCodes’ and ‘EPDProductType’. It is also an inverse property of ‘hasUniclassCode’. ‘UniclassPrCodes’‘EPDProductType’ ‘isWGROf’Defines the relationship between the classes of ‘BuildingWasteMaterialType’ and ‘BuildingType’. It is also an inverse property of ‘hasWGRFor’. ‘BuildingWasteMaterialType’‘BuildingType’ ‘manufactures’Defines the relationship between the classes of ‘ProductManufacturer’ and ‘EPDProductType’. It is also an inverse property of ‘isManufacturedBy’. ‘ProductManufacturer’‘EPDProductType’ ‘takesWGRFrom’Defines the relationship between the classes of ‘EPDProductType’ and ‘BuildingWasteMaterialType’. It is also an inverse property of ‘applicableForEPDs’. ‘EPDProductType’‘BuildingWasteMaterialType’ ‘hasUnit’Defines the relationship between the class 1) ‘EPDProductType’ and ‘Unit’ 2) ‘BuildingWasteMaterialType’ and ‘Unit’ ‘EPDProductType’, ‘BuildingWasteMaterialType’ ‘Unit’ A.2. Facets (value type slots) assigned to the data properties Data property SubPropertyOf Facets (value-type slots) Domain Range ‘hasWGRValue’‘hasBREWasteResults’‘BuildingWasteMaterialType’xsd:float ‘hasDeclarationQuantityValue’‘hasProductGeneralData’‘EPDProductType’xsd:integer ‘hasEPDID’‘hasProductGeneralData’‘EPDProductType’xsd:string ‘hasProductName’‘hasProductGeneralData’‘EPDProductType’xsd:string ‘hasLength’‘hasProductDimension’‘EPDProductType’xsd:float ‘hasBreadth’‘hasProductDimension’‘EPDProductType’xsd:float ‘hasThickness’‘hasProductDimension’‘EPDProductType’xsd:float ‘hasDensityValue’‘hasProductTechnicalData’‘EPDProductType’xsd:float ‘hasAreaDensityValue’‘hasProductTechnicalData’‘EPDProductType’xsd:float ‘hasServiceLifeValue’‘hasProductTechnicalData’‘EPDProductType’xsd:integer ‘hasDryBulkDensityValue’‘hasProductTechnicalData’‘EPDProductType’xsd:float ‘hasGWPValue_A1-A3’‘hasGWP’‘EPDProductType’xsd:float ‘hasWasteValue_MS’‘hasProductWasteResults’‘EPDProductType’xsd:float ‘hasCRU’‘hasProductWasteResults’‘EPDProductType’xsd:float ‘hasMFR’‘hasEoLResults’‘EPDProductType’xsd:float ‘hasMER’‘hasEoLResults’‘EPDProductType’xsd:float ‘hasSM’‘hasResourceUseResults’‘EPDProductType’xsd:float Appendix B. Material choices and their associated CW indicators Below is the knowledge extracted from the ontology using queries in Apache Jena Fuseki, a web-based SPARQL server. The given data is for 1 kg of S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 23 material. The materials listed are the ones filtered after the stage 1 filter, i.e., based on its ‘applicableForBuildingElements’ property. Material choices for ‘Metal – Stud Layer’. Material choices for ‘Fiberglass Batt’. Material choices for ‘Plasterboard’. Material choices for ‘GSS_ Brick Red Multi Stock’. Material choices for ‘GSS_Brick Black’. Material choices for ‘Blockwork’. Notes 1) Acronyms - CW_MS – Construction Waste generated during product Manufacturing Stage (Column 6), CW_CIS –Waste generation during Construction/Installation Stages(Column 8), CW_OMS – Construction Waste generation during Operation and Maintenance Stages (Column 12), CW_MS_GWP – GWP of Construction Waste generated during product Manufacturing Stage (Column 7), CW_CIS_GWP – GWP of Construction Waste generated during Construction/Installation Stages (Column 9), CW_OMS_GWP – GWP of Construction Waste generated during Operation and Maintenance Stages (Column 13), SM – Secondary Material, MFRe – Material for Reuse, MFR - Material for Recycling, MFER - Material for Energy Recovery, MFL – Material for Landfilling, EWC – European Waste Catalogue, ‘NA’ denotes that the specific data is ‘not available.’ 2) The units of measurement are not displayed due to space constraints. So, below is the list of units for each indicator. o Density (based on volume) – kg/m3 o Density (based on area) – kg/m2 o Service life – years o Replacements – numbers o CW – Kg per mass of the product o GWP – kgCO2e per mass of the product o SM, MFRe, MFR, MFER, MFL – Percentage (%) Appendix C. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.autcon.2024.105842. References [1] DEFRA, UK Statistics on Waste, Retrieved from: https://www.gov.uk/govern ment/statistics/uk-waste-data/uk-statistics-on-waste, 2022. Last Access: 04 January 2023. [2] Construction Leadership Council, The Routemap for Zero Avoidable Waste in Construction, Retrieved from: https://www.constructionleadershipcouncil.co. uk/wp-content/uploads/2021/07/ZAW-Interactive-Routemap-FINAL.pdf, 2021. Last Access: 05 January 2023. [3] Z. Liu, M. Osmani, P. Demian, A. Baldwin, A BIM-aided construction waste minimisation framework, Autom. Constr. 59 (2015) 1–23, https://doi.org/ 10.1016/j.autcon.2015.07.020. [4] S. Nagapan, I.A. Rahman, A. Asmi, A.H. Memon, I. Latif, Issues on construction waste: The need for sustainable waste management, 2012 IEEE colloquium on Humanities, Science and Engineering (CHUSER), IEEE (2012) 325–330, https:// doi.org/10.1109/CHUSER.2012.6504333. [5] K.T. Adams, M. Osmani, T. Thorpe, J. Thornback, Circular Economy in Construction: Current Awareness, Challenges and Enablers, Proceedings of the Institution of Civil Engineers-Waste and Resource Management vol. 170, Thomas Telford Ltd, 2017, pp. 15–24, https://doi.org/10.1680/jwarm.16.00011. [6] S. Azhar, Building information modeling (BIM): trends, benefits, risks, and challenges for the AEC industry, Leadersh. Manag. Eng. 11 (3) (2011) 241–252, https://doi.org/10.1061/(ASCE)LM.1943-5630.0000127. [7] J. Won, J.C. Cheng, Identifying potential opportunities of building information modeling for construction and demolition waste management and minimization, Autom. Constr. 79 (2017) 3–18, https://doi.org/10.1016/j.autcon.2017.02.002. [8] Z. Wu, T. Ann, L. Shen, G. Liu, Quantifying construction and demolition waste: an analytical review, Waste Manag. 34 (9) (2014) 1683–1692, https://doi.org/ 10.1016/j.wasman.2015.12.026. [9] ISO, Environmental management systems — Requirements with guidance for use, 2015. Retrieved from: https://www.iso.org/standard/60857.html. Last Access: 28 June 2023. [10] S. Sivashanmugam, S. Rodriguez, F.P. Rahimian, F. Elghaish, N. Dawood, Enhancing information standards for automated construction waste quantification and classification, Autom. Constr. 152 (2023) 104898, https://doi.org/10.1016/j. autcon.2023.104898. [11] EC Directive, Directive 2008/98/EC of the European Parliament and of the Council of 19 November 2008 on Waste and Repealing Certain Directives, Retrieved from: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32008L0098, 2008. Last Access: 28 June 2023. [12] Construction Leadership Council, Zero Avoidable Waste in Construction, Retrieved from: https://www.constructionleadershipcouncil.co.uk/wp-content/uploads/20 16/05/ZAW-Report-Final-Draft-25-February-2020.pdf, 2020. Last Access: 15 May 2023. [13] P.A. Yost, J.M. Halstead, A methodology for quantifying the volume of construction waste, Waste Manag. Res. 14 (5) (1996) 453–461, https://doi.org/ 10.1006/wmre.1996.0044. [14] Y. Li, X. Zhang, Web-based construction waste estimation system for building construction projects, Autom. Constr. 35 (2013) 142–156, https://doi.org/ 10.1016/j.autcon.2013.05.002. [15] J.C. Cheng, L.Y. Ma, A BIM-based system for demolition and renovation waste estimation and planning, Waste Manag. 33 (6) (2013) 1539–1551, https://doi.org/ 10.1016/j.wasman.2013.01.001. [16] O.O. Akinade, L.O. Oyedele, Integrating construction supply chains within a circular economy: an ANFIS-based waste analytics system (A-WAS), J. Clean. Prod. 229 (2019) 863–873, https://doi.org/10.1016/j.jclepro.2019.04.232. [17] D. Han, M. Kalantari, A. Rajabifard, The development of an integrated BIM-based visual demolition waste management planning system for sustainability-oriented decision-making, J. Environ. Manag. 351 (2024) 119856, https://doi.org/ 10.1016/j.jenvman.2023.119856. [18] W. Lu, C. Webster, K. Chen, X. Zhang, X. Chen, Computational building information modelling for construction waste management: moving from rhetoric to reality, Renew. Sust. Energ. Rev. 68 (2017) 587–595, https://doi.org/10.1016/j. rser.2016.10.029. [19] B.C. Guerra, F. Leite, K.M. Faust, 4D-BIM to enhance construction waste reuse and recycle planning: case studies on concrete and drywall waste streams, Waste Manag. 116 (2020) 79–90, https://doi.org/10.1016/j.wasman.2020.07.035. [20] A. Bakchan, K.M. Faust, F. Leite, Seven-dimensional automated construction waste quantification and management framework: integration with project and site planning, Resour. Conserv. Recycl. 146 (2019) 462–474, https://doi.org/10.1016/ j.resconrec.2019.02.020. [21] H. Wu, J. Wang, H. Duan, L. Ouyang, W. Huang, J. Zuo, An innovative approach to managing demolition waste via GIS (geographic information system): a case study in Shenzhen city, China, J. Clean. Prod. 112 (2016) 494–503, https://doi.org/ 10.1016/j.jclepro.2015.08.096. [22] B. Yu, J. Wang, J. Li, J. Zhang, Y. Lai, X. Xu, Prediction of large-scale demolition waste generation during urban renewal: a hybrid trilogy method, Waste Manag. 89 (2019) 1–9, https://doi.org/10.1016/j.wasman.2019.03.063. [23] F. Kleemann, H. Lehner, A. Szczypi´ nska, J. Lederer, J. Fellner, Using change detection data to assess amount and composition of demolition waste from buildings in Vienna, Resour. Conserv. Recycl. 123 (2017) 37–46, https://doi.org/ 10.1016/j.resconrec.2016.06.010. [24] S. Su, S. Li, J. Ju, Q. Wang, Z. Xu, A building information modeling-based tool for estimating building demolition waste and evaluating its environmental impacts, Waste Manag. 134 (2021) 159–169, https://doi.org/10.1016/j. wasman.2021.07.025. [25] X. Chen, H. Huang, Y. Liu, J. Li, M. Liu, Robot for automatic waste sorting on construction sites, Autom. Constr. 141 (2022) 104387, https://doi.org/10.1016/j. autcon.2022.104387. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 24 [26] Z. Wang, H. Li, X. Yang, Vision-based robotic system for on-site construction and demolition waste sorting and recycling, J. Build. Eng. 32 (2020) 101769, https:// doi.org/10.1016/j.jobe.2020.101769. [27] T. Berners-Lee, J. Hendler, O. Lassila, The semantic web, in: Scientific American vol. 284, 2001, pp. 34–43. Retrieved from: https://www-sop.inria.fr/acacia/cour s/essi2006/Scientific%20American_%20Feature%20Article_%20The%20Semantic %20Web_%20May%202001.pdf. Last Access: 02 March 2023. [28] A. Maedche, S. Staab, Ontology learning for the semantic web, IEEE Intell. Syst. 16 (2) (2001) 72–79, https://doi.org/10.1109/5254.920602. [29] N.F. Noy, D.L. McGuinness, Ontology development 101: A guide to creating your first ontology, Stanford knowledge systems laboratory, 2001. Retrieved from, http s://protege.stanford.edu/publications/ontology_development/ontology101.pdf. Last Access: 06 August 2024. [30] T.R. Gruber, A translation approach to portable ontology specifications, Knowl. Acquis. 5 (2) (1993) 199–220, https://doi.org/10.1006/knac.1993.1008. [31] D. Fensel, F. Van Harmelen, I. Horrocks, D.L. McGuinness, P.F. Patel-Schneider, OIL: an ontology infrastructure for the semantic web, IEEE Intell. Syst. 16 (2) (2001) 38–45, https://doi.org/10.1109/5254.920598. [32] D. Allemang, J. Hendler, Semantic Web for the Working Ontologist: Effective Modeling in RDFS and OWL, Elsevier, 2011. ISBN: 0123859662. [33] M. Bilal, L.O. Oyedele, J. Qadir, K. Munir, S.O. Ajayi, O.O. Akinade, H.A. Owolabi, H.A. Alaka, M. Pasha, Big data in the construction industry: a review of present status, opportunities, and future trends, Adv. Eng. Inform. 30 (3) (2016) 500–521, https://doi.org/10.1016/j.aei.2016.07.001. [34] F. Manola, E. Miller, B. McBride, RDF Primer, W3C Recommendation, vol. 10, 2004, p. 6. Retrieved from: https://www.w3.org/TR/rdf-primer/, Last Access: 02 December 2023. [35] M.J. O’Connor, A.K. Das, SQWRL: A Query Language for OWL vol. 529, OWLED, 2009, pp. 1–8. Retrieved from, https://ceur-ws.org/Vol-529/owled2009_submissi on_42.pdf. Last Access: 04 March 2023. [36] J. Bailey, F. Bry, T. Furche, S. Schaffert, Web and semantic web query languages: A survey, Reasoning Web: First International Summer School 2005, Msida, Malta, July 25-29, 2005, Revised Lectures (2005) 35–133, https://doi.org/10.1007/ 11526988_3. [37] B. DuCharme, Learning SPARQL: Querying and Updating with SPARQL 1.1, O’Reilly Media, Inc., 2013 (ISBN: 1449371485). [38] C.M. Eastman, P. Teicholz, R. Sacks, K. Liston, BIM Handbook: A Guide to Building Information Modeling for Owners, Managers, Designers, Engineers and Contractors, John Wiley & Sons, 2011. ISBN: 0470541377. [39] A. Costin, A. Adibfar, H. Hu, S.S. Chen, Building Information Modeling (BIM) for transportation infrastructure–literature review, applications, challenges, and recommendations, Autom. Constr. 94 (2018) 257–281, https://doi.org/10.1016/j. autcon.2018.07.001. [40] P. Pauwels, W. Terkaj, EXPRESS to OWL for construction industry: towards a recommendable and usable ifcOWL ontology, Autom. Constr. 63 (2016) 100–133, https://doi.org/10.1016/j.autcon.2015.12.003. [41] S. Sobhkhiz, H. Taghaddos, M. Rezvani, A.M. Ramezanianpour, Utilization of semantic web technologies to improve BIM-LCA applications, Autom. Constr. 130 (2021) 103842, https://doi.org/10.1016/j.autcon.2021.103842. [42] Y. Schwartz, S. Eleftheriadis, R. Raslan, D. Mumovic, Semantically enriched BIM life cycle assessment to enhance buildings’ environmental performance, Proceedings of the CIBSE Technical Symposium. Edinburgh, UK, 2016. Retrieved from: https://www.researchgate.net/profile/Stathis-Eleftheriadis/publication /296949281_Semantically_Enriched_BIM_Life_Cycle_Assessment_to_Enhance_ Buildings’_Environmental_Performance/links/571ddf2208ae7f552a4a7dd7/ Semantically-Enriched-BIM-Life-Cycle-Assessment-to-Enhance-BuildingsEnvironmental-Performance.pdf, Last Access: 08 December 2023. [43] E. Petrova, P. Pauwels, K. Svidt, R.L. Jensen, Towards data-driven sustainable design: decision support based on knowledge discovery in disparate building data, Archit. Eng. Des. Manag. 15 (5) (2019) 334–356, https://doi.org/10.1080/ 17452007.2018.1530092. [44] D. Zhang, J. Zhang, J. Guo, H. Xiong, A semantic and social approach for real-time green building rating in BIM-based design, Sustainability 11 (14) (2019) 3973, https://doi.org/10.3390/su11143973. [45] S. Jiang, N. Wang, J. Wu, Combining BIM and ontology to facilitate intelligent green building evaluation, J. Comput. Civ. Eng. 32 (5) (2018) 04018039, https:// doi.org/10.1061/(ASCE)CP.1943-5487.0000786. [46] M. Niknam, S. Karshenas, Sustainable design of buildings using semantic BIM and semantic web services, Proc. Eng. 118 (2015) 909–917, https://doi.org/10.1016/j. proeng.2015.08.530. [47] M. Niknam, S. Karshenas, Integrating distributed sources of information for construction cost estimating using semantic web and semantic web service technologies, Autom. Constr. 57 (2015) 222–238, https://doi.org/10.1016/j. autcon.2015.04.003. [48] M. Niknam, F. Jalaei, S. Karshenas, Integrating BIM and product manufacturer data using the semantic web technologies, J. Inf. Technol. Constr. 24 (2019) 424–439. doi: https://www.itcon.org/paper/2019/22. [49] R. Qui˜ nones, C. Llatas, M.V. Montes, I. Cort´ es, A multiplatform BIM-integrated construction waste quantification model during design phase. The case of the structural system in a Spanish building, Recycling 6 (3) (2021) 62, https://doi.org/ 10.3390/recycling6030062. [50] R. Qui˜ nones, C. Llatas, M.V. Montes, I. Cort´ es, Quantification of construction waste in early design stages using Bim-based tool, Recycling 7 (5) (2022) 63, https://doi. org/10.3390/recycling7050063. [51] P. Sun, N. Zhang, J. Zuo, R. Mao, X. Gao, H. Duan, Characterizing the generation and flows of building interior decoration and renovation waste: a case study in Shenzhen City, J. Clean. Prod. 260 (2020) 121077, https://doi.org/10.1016/j. jclepro.2020.121077. [52] J. Wang, H. Wu, H. Duan, G. Zillante, J. Zuo, H. Yuan, Combining life cycle assessment and building information modelling to account for carbon emission of building demolition waste: a case study, J. Clean. Prod. 172 (2018) 3154–3166, https://doi.org/10.1016/j.jclepro.2017.11.087. [53] Defra, Our Waste, Our Resources: A Strategy For England, Retrieved from: https ://assets.publishing.service.gov.uk/government/uploads/system/uploads/attach ment_data/file/765914/resources-waste-strategy-dec-2018.pdf, 2018. Last Access: 02 April 2023. [54] D. Han, M. Kalantari, A. Rajabifard, Identifying and prioritizing sustainability indicators for China’s assessing demolition waste management using modified Delphi–analytic hierarchy process method, Waste Manag. Res. 41 (11) (2023) 1649–1660, https://doi.org/10.1177/0734242X231166309. [55] S. Sivashanmugam, R. Sergio, R. Farzad, D. Nashwan, Maximising the construction waste reduction potential – barriers and catalysts, Proc. Inst. Civil Eng. Civil Eng. 176 (6) (2022) 1–9, https://doi.org/10.1680/jcien.22.00163. [56] J. Vom Brocke, A. Hevner, A. Maedche, Introduction to design science research, design science research, Cases (2020) 1–13, https://doi.org/10.1007/978-3-03046781-4_1. [57] R.K. Yin, Case Study Research: Design and Methods, Sage, 2009. ISBN: 1412960991. [58] J. Raad, C. Cruz, A survey on ontology evaluation methods, in: Proceedings of the International Conference on Knowledge Engineering and Ontology Development, part of the 7th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management, 2015, https://doi.org/10.5220/ 0005591001790186. [59] E. Palumbo, B. Soust-Verdaguer, C. Llatas, M. Traverso, How to obtain accurate environmental impacts at early design stages in BIM when using environmental product declaration. A method to support decision-making, Sustainability 12 (17) (2020) 6927, https://doi.org/10.3390/su12176927. [60] R. Santos, A.A. Costa, J.D. Silvestre, L. Pyl, Integration of LCA and LCC analysis within a BIM-based environment, Autom. Constr. 103 (2019) 127–149, https://doi. org/10.1016/j.autcon.2019.02.011. [61] NBS, Uniclass - Unified Construction Classification, Retrieved from: https://u niclass.thenbs.com/, 2022. Last Access: 12 March 2023. [62] Environment Agency, Guidance on the classification and assessment of waste (1st Edition v1.2.GB) - Technical Guidance WM3, 2021. Retrieved from: https://assets. publishing.service.gov.uk/government/uploads/system/uploads/attachment _data/file/1021051/Waste_classification_technical_guidance_WM3.pdf. Last Access: 02 June 2023. [63] M. R¨ ock, A. Hollberg, G. Habert, A. Passer, LCA and BIM: integrated assessment and visualization of building elements’ embodied impacts for design guidance in early stages, Procedia CIRP 69 (2018) 218–223, https://doi.org/10.1016/j. procir.2017.11.087. [64] T. Elsaleh, S. Enshaeifar, R. Rezvani, S.T. Acton, V. Janeiko, M. Bermudez-Edo, IoTStream: a lightweight ontology for internet of things data streams and its use with data analytics and event detection services, Sensors 20 (4) (2020) 953, https://doi. org/10.3390/s20040953. [65] A. Rhayem, M.B.A. Mhiri, F. Gargouri, Semantic web technologies for the internet of things: systematic literature review, Internet Things 11 (2020) 100206, https:// doi.org/10.1016/j.iot.2020.100206. [66] A. Stadnicki, F.F. Pietro´ n, P. Burek, Towards a modern ontology development environment, Proc. Comp. Sci. 176 (2020) 753–762, https://doi.org/10.1016/j. procs.2020.09.070. [67] E. Simperl, Reusing ontologies on the semantic web: a feasibility study, Data Knowl. Eng. 68 (10) (2009) 905–925, https://doi.org/10.1016/j. datak.2009.02.002. [68] P. Valluru, J. Karlapudi, K. Menzel, T. M¨ at¨ asniemi, J. Shemeikka, A semantic data model to represent building material data in AEC collaborative workflows, in: Boosting Collaborative Networks 4.0: 21st IFIP WG 5.5 Working Conference on Virtual Enterprises, PRO-VE 2020, Valencia, Spain, November 23–25, 2020, Proceedings 21, Springer, 2020, pp. 133–142, https://doi.org/10.1007/978-3-03062412-5_11. [69] S.C.-F. María Poveda-Villal´ on, Material Properties Ontology. http://www.oeg-up m.net/ (web archive link, 10 March 2023), 2021. Retrieved from: https://bimerr. iot.linkeddata.es/def/material-properties/. Last Access: 10 March 2023. [70] S. Sivashanmugam, S. Rodriguez, F. Rahimian, N. Dawood, Maximising the construction waste reduction potential–how to overcome the barriers, in: Proceedings of the Institution of Civil Engineers-Civil Engineering, Thomas Telford Ltd, 2022, pp. 1–9, https://doi.org/10.1680/jcien.22.00163. [71] RICS, Whole life carbon assessment for the built environment, 2023. Retrieved from: https://www.rics.org/profession-standards/rics-standards-and-guidance/se ctor-standards/construction-standards/whole-life-carbon-assessment. Last Access: 08 August 2024. [72] M. Fern´ andez-L´ opez, A. G´ omez-P´ erez, N. Juristo, Methontology: from ontological art towards ontological engineering, 1997. Retrieved from: https://cdn.aaai.org/ Symposia/Spring/1997/SS-97-06/SS97-06-005.pdf. Last Access: 10 September 2023. [73] M. Uschold, M. Gruninger, Ontologies: principles, methods and applications, Knowl. Eng. Rev. 11 (2) (1996) 93–136, https://doi.org/10.1017/ S0269888900007797. [74] W3C, OWL 2 Web Ontology Language - Document Overview (Second Edition), 2009. Retrieved from: https://www.w3.org/2012/pdf/REC-owl2-overview-20121 211.pdf, Last Access: 20 March 2023. S. Sivashanmugam et al. Automation in Construction 168 (2024) 105842 25