Knowledge-based data enrichment for HBIM

Knowledge-based data enrichment for HBIM: exploring highquality models using the semantic-web. Ramona Quattrini*, Roberto Pierdicca, Christian Morbidoni [email protected], [email protected], [email protected] Polytechnic University of Marche Keywords: Data Standardization, Taxonomies, Interoperability, Ontologies, Shared parameters, Representation workflow.

Abstract In the last decade, the paradigm Historical Building Information Modelling (HBIM) was investigated to exploit the possibilities offered by the application of BIM to historical buildings. In the Cultural Heritage domain, the BIM-oriented approach can produce 3D models that are data collector populated by both geometrical and non-geometrical information related to various themes: historical documents, monitoring data, structural information, conservation or restoration state and so on. The realization of a 3D model fully interoperable and rich in its informative content could represent a very important change towards a more efficient management of the historical real estate. The work presented in these pages outlines a novel approach to solve this interoperability issue, by developing and testing a workflow that exploits the advantages of BIM platforms and semanticweb technologies, enabling the user to query a repository composed of semantically structured and rich HBIM data. The presented pipeline follows four main steps: (i) the first step consists on modeling an ontology with the main information needs for the domain of interest, providing a data structure that can be leveraged to inform the data-enrichment phase and, later, to meaningfully query the data. (ii) Afterwards, the data enrichment was performed, by creating a set of shared parameters reflecting the properties in our domain ontology. (iii) To structure data in a machine-readable format, a data conversion was needed to represent the domain (ontology) and analyze data of specific buildings respectively; this step is mandatory to reuse the analysis data together with the 3D model, providing the end-user with a querying tool. (iv) As a final step in our workflow, we developed a demonstrative data exploration web application based on the faceted browsing paradigm and allowing to exploit both structured metadata and 3D visualization. This research demonstrates how is possible to represent a huge amount of specialized information models with appropriate LOD and Grade in BIM environment and then guarantee a complete interoperability with IFC/RDF format. Relying on semantically structured data (ontologies) and on the Linked Data stack appears a valid approach for addressing existing information system issues in the CH domain and constitutes a step forward in the management of repositories and web libraries devoted to historical buildings.

1.

Introduction

Nowadays, Building Information Modelling (better known as BIM) has become a globally spread standard able to ease the project collaboration, the data integration and to support project activities. Even if it was initially conceived for the planning and design phases of a project, it is actually used in the construction for a wide range of applications such as 4D simulation, clash detection, and providing detailed spatial and material quantities [1]. The strength of its approach resides mostly in a shift of paradigm in the workflow of a project, focusing all the decisional process in the early stages, improving transparency and interoperability for all the actors involved in the decisional and managerial phases. However, despite BIM’s efficiency on new buildings is widely proven, there’s a lot of uncertainty when dealing with historical buildings. The reason of this mainly depends on the purpose that drove the development of BIM, conceived for architectural design. Consequently, parametric objects are not suitable for the modelling of existing historical architectural elements. Moreover, another bottleneck is represented by the lack of 1

complex components or by the inconsistency in the geometric representation of irregular shapes. As a matter of fact, often the modelling tools in the existing platforms perform very simple operations that are not always sufficient to geometrically describe the complexity of the real object, because of the high level of geometric abstraction. However, in the last years, research was oriented toward the “HBIM” paradigm, which is the acronym of Historical (or sometimes Heritage) Building Information Modelling and investigates the possibilities offered by the application of the BIM approach to historical buildings [2]. There are several important differences between HBIM and BIM, arising from inherent characteristics of historical buildings like the uniqueness of components and the absence of the concept of life-cycle [3]. In fact, an historical building is the product of a non-industrial process of construction and all the uses of a management tool are motivated by analysis, conservation and maintenance purposes. Due to these issues, it is worthwhile to investigate the possibilities offered by management tools to be exploited also in the Cultural Heritage (CH) domain. The obstacles preventing HBIM towards a complete conversion in practical use are twofold: the data enrichment and the 3D interoperability, described in the following. The data enrichment (and its related information management) represents the crucial part of the HBIM model, because it is intended more as a data collector rather than a geometrical representation. In a BIM-oriented approach we must imagine our model populated by both geometrical and nongeometrical data which, in the CH field, embodies a large set of information related to various themes: historical documents, monitoring data, structural information, conservation or restoration state and so on. The realization of a 3D model fully interoperable and rich in its informative content could represent a very important change towards a more efficient management of the historical real estate. Nonetheless, storage, portability and interoperability issues among architects and cultural heritage actors dealing with 3D technologies remain a huge challenge [4]. In fact, digital repositories seem unable to guarantee affordable features in the management of 3D models and their metadata; this is mainly because of the nature of most of the available data format for 3D encoding, not suitable for the necessary portability required by 3D information across different systems. However, many ontologies and schemas are available which produce standard sets of metadata, providing rich collections of classes and properties to capture every degree of granularity required for the description of 3D models and for their enrichment with annotations. Even if some attempts have been carried out to facilitate accessing these information to the majority of the users [5], the research community is focusing his efforts in the portability towards the web, since it is able to reach the most of the users, exploiting its advantages like accessibility, data visualization, interactive interfaces, inquiry facilities and enhanced access to remote repositories [6]. Also BIM systems designed for CH could benefit from this process. Thus, whether the semantic structuring of the 3D model allows organizing representations over time, taking into account the variety of hypothesis, the data enrichment process is mandatory to relate attributes describing time, the hypothetical value of the reconstruction and its reliability [7]. In this perspective, several points must be discussed and analysed: how to manage that information in a hierarchical way, how to connect them to a model and how to ensure the interoperability of data enrichment with non-proprietary tools. The previous points could be summarized in the following research challenge: how one can use semantic-web techniques for a more efficient survey of multiple cases data as a tool of comparison in support of any decisional process? From the above said, emerged the need to create a shared parametric semantic-aware repository, based on an ontology-oriented approach, allowing a full interoperability among the implemented data enrichment and the developed 3D model. The work presented in these pages outlines a novel approach to solve this interoperability issue, by developing and testing a workflow that exploits the advantages of BIM platforms and semanticweb technologies, enabling the user to query a repository composed of semantically structured HBIM data. As stated in the research aim section, a description of interoperability between BIM generated resources and semantic technologies for WEB applications is provided. Will be addressed the topic of how semantic-web technology can be exploited to provide semantic interoperability,

2

working on complex Romanic architectures and other set of BIM models. The remainder of the paper is organized as follow: Section 3 will be dedicated to the description of the whole workflow and to the discussion of the various steps towards the realization of the prototype Web application. The latter allows to query and visualize specific and complex architectures, as explained in the section 4. The results obtained with the workflow and the potential of semantic-web applications for the CH domain are discussed in Section 4 as well. Concluding remarks and new scenario offered by our approach are outlined in Section 5.

2.

Research aim

From the introduction section emerged the need of a real semantic interoperability solutions between the models developed into a BIM platform and their interpretation, management and querying in a more practitioner-centered environment. This is mainly because, enabling interoperability at a semantic level plays a pivotal role for bringing the benefits of BIM modelling and data enrichment. The choice of using semantic-web technology is mandatory to convey the meaning, which can be structured in a machine readable format and interpretable by the insiders. Taking advantage from the historical case study that will be shown in the paper, the aim of this work is to validate a workflow capable of grouping and organizing all the available information in a database managed in a hierarchical way by means of ontologies (intended in our case as taxonomies). By establishing a connection between the database and a proper HBIM model with sufficient level of development and grade, we were capable to deliver in one solution the activities of building management. This work also provides a study of the interoperability of the data connected, analysing the possible usage of any actor of the process, even if not in possess of any BIM-authoring platform.

3.

Materials and methods

3.1 State of art: HBIM, interoperability and data management. The first definition of HBIM appeared in 2009, due to Maurice Murhpy [8]. According to this definition, a HBIM is composed by a library of parametric objects and a system capable of mapping those objects to survey data, like point clouds or images. Whilst other definitions appeared in the literature [9], the research community came up with a common thread defining HBIM as the process of creation of a model starting from survey data (i.e. Terrestrial Laser Scanning (TLS), digital photogrammetry etc.), aimed at the creation of a library of parametrical objects, combined with the informative data in order to enrich the object besides the modeling phase [10]. The difference between BIM arises from the goals of defying the state of conservation and planning future operations, rather than the realization or the computing phase (which is a BIM prerogative). More in general, literature shows that the existing gaps originate from differences between processes for as-planned building and processes for as-built model [11]. The development of BIM approach in Architectural Heritage (AH) domain faces the following challenges and needs: semantic object vs. unsegmented geometries, standardization vs. irregularity and parametric intelligence vs. geometric accuracy. A good dissertation of such aspects can be found in [12]. Moreover, avoiding the confusion between modeling environment and BIM approach, a major challenge nowadays is to obtain an informed model, in specific and usable context. The final product of the HBIM process should be a model that, before being used as a professional tool, must be populated by non-geometrical data for making it suitable for engineering and conservation applications (this process can be defined as “data enrichment phase”). In HBIM field, the modeling phase and the semantic sub-division in various elements are sufficiently robust, although timeconsuming [13] [14] and based on proprietary tools. On the contrary, the data enrichment needs new effective tools, ready to use in complex context [15] for conservation and retrofitting [16]. The model, which at the end becomes a proper database, can be exported, edited, converted and continuously updated in accordance with different applications and represents the optimal solution in a field of study where unambiguity and durability of database is extremely important [17]. 3

Another central topic for researchers (and final users as well) is the interoperability of the information; in other words, how this database can be interrogated, edited and upgraded avoiding, or at least minimizing, the issues of continuous conversions process caused by the difference of software between all the actors involved, enhancing a collaborative and transparent approach [1]. A fully interoperable framework, capable of enabling a decentralized BIM and specifically designed for the AH domain have not been implemented so far. The use of semantic-web technology represents a turning point thanks to whom this process can be boosted. In fact, as several works in literature point out, the use of Semantic-web standards, namely RDF and OWL, as formal knowledge representation language have several advantages [18]; the main one is the possibility to flexibly mix BIM data and additional data created by operators via data enrichment, and to represent multiple 3D models in the same computational space, thus being able to perform cross-model queries. Semantic technologies’ advantage is exchanging information in a meaningful way, in a human readable format and with a minimum of human intervention. Due to the nature and large amount of data collected to implement a HBIM, semantic-web has demonstrated to fit with the existing standard format developed for such domain [19]. The main benefit of formatting data in a semantic way is that they can be queried according to their meaning and not to the structure itself. The methodology proposed in these pages is oriented at enriching the model and then to make semantics explicit. To do this, we translated the original bulk of information into RDF, the W3C standard data model to represent resources and relations among them [38]. Examples of RDF based approaches are present in literature in domains such as safety [20], construction [21] and management [22]. In the following, this process will be described, using the showcase of Santa Maria of Portonovo church. 3.2 The workflow for knowledge-based HBIM exploitation in semantic web As stated in the introduction section, the purpose of this research is to permit one to discover an HBIM enriched repository taking advantage from a semantic structure of the data. The workflow designed to accomplish this task is described in Errore. L'origine riferimento non è stata trovata., where the main steps of our procedure can be summarized as follows:

Figure 1 – The workflow is here summarized through a schema made up of four steps.

 Ontology modeling The first step consists in modeling the main information needs for the domain of interest, providing a data structure that can be leveraged to inform the data-enrichment phase and, later, to meaningfully query the data. The Ontology Web Language (OWL) was used as ontology formalization language.  HBIM and data enrichment The data enrichment process has been performed contextually with the 3D modeling. This has the advantage of enabling users already familiar with Revit to perform data enrichment without the need of learning to use a new tool. In this way we could rely on IFC (Industry Foundation Classes) properties and their implementation in Revit for the following step. 

Data conversion to RDF and mapping to domain ontology 4

Once data enrichment has been done, the standard Revit data export functionality can be straightforwardly used to produce IFC data compliant with the EXPRESS schema. We then use the IFC-to-RDF conversion tool [32] to obtain RDF (Resource Data Framework) data, including both BIM data and data enrichments. This makes the data suitable to be queried in a semantic way.  Data storage and online data exploration Once all the data of interest is represented in RDF, we need to expose it to applications that wants to make use of such data. However, as pointed out by Do-Yeop et al. [23], where data enrichment of BIM models using RDF is discussed, it is too difficult for a non-technician to write SPARQL queries and understand the domain ontology. To overcome this problem, as a final step in our workflow, we developed a demonstrative data exploration web application based on the faceted browsing paradigm and allowing to exploit both structured metadata and 3D visualization. 3.2.1

Case study: operative background of Santa Maria of Portonovo

The case study used in this research is the church of Santa Maria of Portonovo, an early example on the Adriatic coast of Romanesque style with Lombards influences (see Figure 2). The church, located at the edge of a cliff, has uncertain origins but can be reasonably dated in the period 1070-1080 a.c. and has been severely damaged by earthquakes and invasions and restored different times throughout the years. The choice of this case study is motivated by the availability of a large set of information covering different disciplines like, surveys, historical documents and, recently, about assessment of the state of conservation and risk [24]. Furthermore this case of study was treated before in other papers for what concerns the modelling phase and the creation of the parametric items library, with a study about the reliability and the accuracy of the model, both carried out in Autodesk Revit software [25].

Figure 2 – A picture and images from the point cloud of the Santa Maria of Portonovo, the church chosen for the case study.

The present work enables new models of data collection in HBIM, through the validation of 3D model and its enrichment, with all kind of useful analysis that have been separated so far. 3.2.2

HBIM

After a standard modelling phase in Revit, the quality assessment, in comparison with TLS point cloud, shows sufficient results: the whole model presents gaps under 3 cm in 63% of points. For single elements and shapes, outcomes are good also. The bi-dimensional drawings, automatically 5

extracted from the Revit model, were also evaluated in the case study. They show distances between 2-3 cm from the compared slices, therefore the model is able to represent the building in a 1:100 scale. This representation scale is not sufficient for restoration and detailed knowledge of historical buildings: an analysis of the model for assessing it’s LOD according to the INNOVance research group specifics [26] was carried out. Our model results a reliable “as-built” model which corresponds to a LOD 500 (following the INNOVance specs). A model capable of being used as a tool for the management activities must be a LOD 550. By comparing those two definitions the unique difference emerges in terms of richness of non-graphical information. In addition, although our HBIM is based on non - simplified parametric models, further drawings should be collected in order to manage all kind of documentation in broader heritage conservation practices. For example, in order to develop a set of documents reliable for knowledge of the church and its transformations during the centuries, it is mandatory to manage detailed drawings, such as Figure 3. This step enables the requested LOD and a Grade 3 for the data-set.

Figure 3 – A detailed drawing for the south wall of the church. The collection of this kind of document and his data enrichment to the 3D implements a Grade 3 HBIM model.

The second stage of this procedure is the ontology for managing all information dividing both building and metadata in different layers with different depth. Following a structural-oriented logica, a building is the central object in our model and is conceptually structured in three layers: ● Global: the whole building in its integrity. Global attributes can be historical information or documentation, photo surveys date, etc.. 6

●

Macro-areas: An architectural group of elements that are structurally connected (e.g. a nave of a church) and/or shared features, eg., elements having the same damage mechanism. ● Element: single architectural elements (like columns, walls, doors) that can be grouped in macro-areas. The model, as represented in the previous scheme, in this way is ideally segmented in various minimal units (Elements), each one identified by a code (e.g.. Column_001) that is necessary to organize and segment the related metadata. A group of those units with the same structural behavior is a macro-area (for example “Apse”) and at this level the information regards all the singular element code composing the element and all the metadata regarding the damage mechanism. In the case presented, the demonstrative domain is management and intervention planning for historical buildings (see Figure 4).

Figure 4 - At each level corresponds different kind of information. “Global” contains all the web references, historical documentation, photo and surveys data of the building. “Macro-areas” level contains information about the composition of the macro-area in terms of elements and all the analysis about possible damage mechanism that can be activated. “Elements” contain all the information about the state of conservation, the damage occurred and the monitoring data.

3.2.3

Modelling the domain ontology

As our final goal is to reuse the analysis data together with the spatial and 3D model, to provide querying and end-user browsing, we need to structure data in a machine-readable format. RDF and OWL was chosen in this research to represent the domain (ontology) and analysis data of specific buildings respectively. This is in line with recent research trends related to BIM data in construction industry ([27], [28]), where an increasing number of information management applications is relying on semantic-web technologies or tools from the Linked Open Data domain to support data interoperability [23]. In the domain of HBIM, however there are few research works that experiment this approach, even if the need for mixing 3D model features with additional, domain specific, information emerged [18]. Using RDF as a representation model not only allows to model such domain information in a machine understandable way, but also allows to include all this information in a unique knowledge graph where multiple buildings and associated 3D models can co-exist. The ability to query and browse information related to multiple models as the same time goes beyond the possibilities currently provided by the IFC representation model. 7

Given the conceptual domain modeling briefly discussed in the previous section, to demonstrate the proposed workflow, we modeled an OWL ontology capable of representing the resources that we want to describe and the all needed information. To model the concept of macro-area we reuse classes form an existing ontology (the Architectural Design ontology [32]). Specifically, we use the BuildingComponent class to represent macro-areas, which are composed of single elements, represented by the BuildingElement class. Global informations are modeled by using the Building class. As shown in Figure 5, a BuildingComponent can have sub-classes, indicating common macro-areas available in churches (e.g. Nave, Apse, Vestibule, etc.) and could be further extended as needed to capture other kinds of macro-area. A BuldingComponent is part of a Building and is composed of instances of the BuildingElement class, that can be of specific type, e,g, Doors, Walls, Columns, etc. Each BuildingElement can then be described by a set of properties hierarchically arranged in a taxonomy. For example, all the properties describing the architectural history, e.g., historical sources (“Fonti_Storiche”), or describing the graphical information such as LOD (“Grade”) and graphical sources ("Elaborati_2d_di_Dettaglio”) that are sub-properties of the property (“Grafica”).

Figure 5 - A snapshot of the OWL ontology developed to drive our experiments in semantic data representation, querying and visualization. Nodes represent classes (in purple) and properties (in green).

3.2.4

Data enrichment

The data enrichment phase is intended to achieve the affordable Level of Information: indeed, according to [29], the Level of Definition (LOD) is the sum of Level of detail and information. We already demonstrated the correct Level of Detail (see the paragraph 3.2.2), based on state-of-art methods, while the Level of Information needs improvements in the domain, both for the case study both for current trends. After the labelling and the organization of the database the proper data enrichment phase was carried out directly in Revit by using shared parameters (Figure 6). Those parameters are Revit’s internal parameters which can be created, grouped in different thematic areas and fully customized, accepting a wide range of different data types (text, number, external links references and so on), as suggested in [30]. For our experiments we created a set of shared parameters reflecting the properties in our domain ontology. The choice of this methodology was driven by the exportability of those parameters via IFC extension format. After the semantic organization of the database, the next step is the proper “data enrichment” phase. The first attempts to achieve this goal were tried using a third party plugin (Keynote Manager) and a Revit native plugin (Db link) [31] both with unsatisfactory results. In the first case the weak point was the nature of keynotes themselves (which are not capable of recognizing 8

different in terms of instances). On the other hand, the second attempt was aborted due to the difficulties of database visualization because we preferred to work in a 3d-based visualization instead of a database visualization tool, for a better and faster instances identification. The final expedient was carried out directly in Revit by using shared parameters [32]. The data-enrichment process at first starts with the attribution of a label code that is necessary to uniquely identify instances. Starting from an already modelled building this phase was carried out by using an appropriate shared parameter. This expedient is necessary also to perform queries on specific elements. Later, in the process all elements must be singularly labelled and all the relative datasheets must be linked to them. This stage of the process is the most time-consuming operation and requires a lot of efforts from professional figures and technicians. A key element for a correct handling of any building workflow is to take into account the management of the process in terms of timing [33]. Currently there are no plans for a future automation process of this phase. Nevertheless, the time effort for those operations can be deferred during all the process of data collecting and the modelling pipeline.

Figure 6 – The data enrichment phase in Revit environment. The information system paradigm is performed through the shared parameters. The figure shows, at the level of instance, the macro-element definition, some links about static analysis or detailed graphical attached data and a vocabulary for same labels.

3.3 Towards an interoperable framework The interest in RDF/OWL for representing BIM data is witnessed by the recent work on the IFCOWL ontology, which translates in OWL the IFC data model, usually represented as an EXPRESS schema, e.g., in construction industry domain [34]. After exporting the 3D model from Revit in IFC

9

format, which includes information added via shared parameters, we based on the tool provided at Aalto University and Ghent University [33] to convert data into IFC-OWL [19].

Figure 7: The transformation from the IFC representation pattern to the domain ontology pattern

The resulting data contains all the information added during the data enrichment phase, and can be queried via the standard Semantic-web query language (SPARQL). However, the complexity of the IFC model, and specifically the pattern used to represent custom properties, makes writing such queries cumbersome. In Figure 7 we show the IFC model structure where an architectural element (ifc:Element) is associated, by means of a ifc:RelDefinesByProperty definition, to a ifc:PropertySet, which includes all the properties associated to a given element or to a group of elements. The ifc:PropertySet is then connected to one or more properties that has a name and a nominal value. For example the query for retrieving all the element that as a damage type equal to “# GRADE” is the following: Example 1: select distinct ?object ?type ?name where { ?object rdf:type ?type ?object ifc:tag_IfcElement ?label_id. ?label_id expr:hasString ?id. ?object ifc:name_IfcRoot ?label. ?label expr:hasString ?name. ?definition ifc:relatedObjects_IfcRelDefines ?object. ?definition ifc:relatingPropertyDefinition_IfcRelDefinesByProperties ?set. ?set ifc:hasProperties_IfcPropertySet ?value. ?value ifc:nominalValue_IfcPropertySingleValue ?text. ?text expr:hasString "2". ?value ifc:name_IfcProperty ?property. ?property expr:hasString "Grade".

In order to, on one hand, make queries easier and, on the other hand, obtain data that conform to our domain ontology, we performed additional transformations on the RDF graph. To do so we used SPARQL construct queries as transformation rules to map from the IFC pattern to our domain ontology. Detailed discussion of such a transformation technique are out of the scope 10

of this paper, interested readers can rely on the SPARQL specifications to gain detailed technical understanding [34]. In Figure 7 we compare the representation pattern used in IFC, specifying that a particular element (:Comlumn_12) has a grade equal to “2”, with the representation of the same information in our domain ontology. After the transformation is performed the example query in example 1 become the following: Example 2 select distinct ?object ?type ?name where { ?object rdf:type ?type. ?object :grade “2”. }

4

RESULTS AND DISCUSSION

To demonstrate the step in our workflow, we built a sample web application targeted to operators in historical building management. The purpose of such a demonstrator is that of showing how, relying on a semantically structured representation of metadata and on IFC, information can be browsed in a human friendly way via the faceted browser paradigm. The results are divided in two connected parts: a) a configurable faceted browser, b) the 3D visualization handling. The user navigates data by means of facets, which allows to filter architectural elements by their attributes. Facets can be configured by associating to each one a SPARQL query. This provides the maximum flexibility about what facets to display depending on the domain ontology used and on the target users of the systems. In our case, the selected facets reflect relevant properties of our domain ontology, but can also show IFC based information, as the “Element IFC type”, as shown in Figure 8.

Figure 8. The faceted browser. Users can select a specific building or navigate across the architectural elements placed in different buildings, based on their IFC type or other domain specific features. An example of query performed in the web application; in this case the user immediately sees selected walls part of a macro-element, with the required grade.

Once the user has selected a subset of all the elements, e.g., those belonging to the same historical building, or having the same damage type, etc., these can be contextually visualized in the originating 3D model, that constitutes the part b) of the achieved results. In the example reported in Figure 8, a user selected the “Santa Maria di Portonovo” building and filtered element of belonging to a specific macro-element and with a specific grade. Such elements are shown in yellow in the 3D model. Users can then further refine the model visualization by selecting and deselecting elements. Once a set of filters if set, available values in the facets (left menu) show the distribution of domain 11

features across the selection and suggest possible further navigation axes. For example, among the elements selected in the example (Figure 8), 16 are columns, 5 are walls, and so on. The tool is completely based on standard uniform description languages (RDF, OWL and IFC) and on open source software, making it completely decouple from specific proprietary 3D modeling tools, e.g., Revit. It is important to notice that, thanks to the schema flexibility of the RDF model, the application can be easily customized for different domains, i.e. building facets on top of different domain ontologies. Hence, architectural elements can be iteratively filtered wrt domain specific information, and visualized in their 3D context within the browser. This can be very helpful especially in the case of monitoring the status of conservation of a historical building. The proposed solution in fact enables the practitioners to investigate over each single architectural component, getting precise and reliable information related to the CIM. Is it so to be noted that the BIMExplorer also contains all the information extracted from the enriched HBIM models, such as additional links. It is possible therefore to combine in a single query features like dimension of elements and placements and, in general, 3D model features, along with domain features, e.g. querying the system for all the damaged walls with last inspection before 2014, with no openings and below a given height. 5

CONCLUSION

The work, here presented, supports a complex management of data relating to the facility management and maintenance of monumental historical buildings. More in general, it enables the spreading of suitable HBIM starting from survey tools and professional analysis about conservation: the effectiveness of the work lies in adopting standard data management and allowing LOD and Grade model compliant with the state of art knowledge for historical heritage, until now not granted by BIM modeling. This research indicates that it is possible (1) to represent a huge amount of specialized information models with appropriate LOD and Grade in BIM environment and then guarantee a complete interoperability with IFC/RDF format and (2) to interrelate these RDF graphs into an aggregate or enriched RDF graph. Consequently, the linked data strategy appears a valid approach for addressing existing information system issues in the CH domain. The performed workflow, although time consuming and computationally heavy, shows the advantage of a unique environment of modelling/survey and data enrichment: this means that expert users in representation and conservation will work in a useful and daily environment. An achieved goal is the demonstration of a complete interoperability of the enrichment in nonproprietary sw. Another achieved goal concerns the validation of the data migration from HBIM to the semantic-web tool, in which users can browse the model filtering properties and viewing the 3D. The data displayed could be of different typologies and they can refer: a) 3dimensional models or parts of them b) digital work sheet or details drawing c) multimedia contents such as pdf, videos or pictures d) web links. Weak point of the research is the lack of linking between the ontology modeling and the data enrichment phase. Indeed, at the present stage, the ontology implementation is not readable in the Revit model and in the IFC format because it is not allowed by the proprietary sw, but it is able to help the translation of elements hierarchy. In this light, it could be useful also to automate the link between IFC/RDF format and OWL taxonomy. As future work, we will provide other enriched HBIM models to populate the web database, according to different domain and use cases. Regarding the domain ontology already defined, that is an ontology about structural analysis storage, it is planned to add other informed models enabling several users to consult and browse a model dataset in a wider field. This step will consist in querying the database and to stress the system, with the purpose to enable the validation of the tool robustness and, if necessary, to improve it. Another challenge to be faceted by our research will be the automation or facilitation of the data enrichment phase through a dedicated tool. This system will enable complex query for the 3D dataset and adding information in various times, without operating in BIM proprietary sw. Similar kind of tool will optimize the proposed workflow (Figure 1), improving the link and the data transfer between “data enrichment” and “faceted browser” (dashed red line). 12

The workflow showed how it is possible to move from geometric HBIM representation towards web 3D objects management, supporting enhanced comprehension of the single elements within the model information of the overall building organism, and allowing the connection to descriptive thematic database (constructive technologies, elements abacus), in a logic of semantic content models. Our approach, if performed through a widespread standardization, will be a key element for the diffusion of HBIM in professional practices, but it could lead to interesting results in the field of representation and information technology research. In this framework, the present research answers to the following relevant topics: definition of the different categories of items from a reality-based survey and their semantic organization in HBIM; connection between model and information considering each single item like a large repository; management of items/repositories of information in web libraries devoted to historical buildings.

ACKNOWLEDGMENTS The authors would acknowledge the FAI (Italian Ambiental Found) and the Superintendence for Architectural Heritage and Landscape of the Marche Region for allowing the acquisition of the Church of S. Maria at Portonovo. The data enrichment phase was carried out by Andrea Carassai, in his degree thesis supervised by Ramona Quattrini, Eva S. Malinverni and Roberto Pierdicca. The development of the web tools was possible thanks to the contribution of Manolo Micozzi. The authors also want to thank Paolo Clini, Enrico Quagliarini and Mirco Ripanti for sharing drawings and analysis used in the data HBIM collection.

REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9]

[10] [11] [12] [13]

[14] [15]

E.P. Karan, J. Irizarry, Extending BIM interoperability to preconstruction operations using geospatial analyses and semantic web services, Autom. Constr. 53 (2015) 1–12. R. Quattrini, E. Baleani, Theoretical background and historical analysis for 3D reconstruction model. Villa Thiene at Cicogna, J. Cult. Herit. 16 (2015) 119–125. S. Logothetis, A. Delinasiou, E. Stylianidis, Building Information Modelling For Cultural Heritage: A Review, ISPRS Ann. Photogramm. Remote Sens. Spat. Inf. Sci. 2 (2015) 177. A. Felicetti, M. Lorenzini, Metadata and tools for integration and preservation of cultural heritage 3D information, Geoinformatics FCE CTU. 6 (2011) 118–124. S. Boeykens, E. Bogani, Metadata for 3D models. How to search in 3D model repositories?, in: ICERI 2008 Proc., IATED, 2008: p. 11p. É. Meyer, P. Grussenmeyer, J.-P. Perrin, A. Durand, P. Drap, A web information system for the management and the dissemination of Cultural Heritage data, J. Cult. Herit. 8 (2007) 396–411. L. De Luca, C. Busayarat, C. Stefani, P. Véron, M. Florenzano, A semantic-based platform for the digital analysis of architectural heritage, Comput. Graph. 35 (2011) 227–241. M. Murphy, E. McGovern, S. Pavia, Historic building information modelling (HBIM), Struct. Surv. 27 (2009) 311–327. J. Garcia-Fernandez, J. Anssi, Y. Ahn, J.J. Fernandez, Quantitative + qualitative information for heritage conservation: An open science research for paving “collaboratively” the way to historicalBIM, in: 2015 Digit. Herit. Int. Congr. Digit. Herit. 2015, 2015: pp. 207–208. doi:10.1109/DigitalHeritage.2015.7419495. F.I. Apollonio, M. Gaiani, Z. Sun, 3D modeling and data enrichment in digital reconstruction of architectural heritage, ISPRS Arch. 5 (2013) W2. R. Volk, J. Stengel, F. Schultmann, Building Information Modeling (BIM) for existing buildings— Literature review and future needs, Autom. Constr. 38 (2014) 109–127. F.I. Apollonio, M. Gaiani, Z. Sun, A Reality Integrated BIM for Architectural Heritage Conservation, Handb. Res. Emerg. Technol. Archit. Archaeol. Herit. (2016) 31. S. Garagnani, A.M. Manferdini, Parametric accuracy: Building Information Modeling process applied to the cultural heritage preservation, in: 3DArch2013, Conf. Proc. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. Trento, Italy, Citeseer, 2013: pp. 87–92. L. Barazzetti, F. Banfi, R. Brumana, M. Previtali, Creation of Parametric BIM Objects from Point Clouds Using Nurbs, Photogramm. Rec. 30 (2015) 339–362. M. Murphy, E. McGovern, S. Pavia, Historic Building Information Modelling–Adding intelligence to laser and image based surveys of European classical architecture, ISPRS J. Photogramm. Remote Sens. 76 (2013) 89–102.

13

[16]

[17]

[18] [19] [20] [21]

[22] [23]

[24]

[25]

[26]

[27] [28] [29] [30] [31]

[32]

[33]

[34]

L.M. Khodeir, D. Aly, S. Tarek, Integrating HBIM (Heritage Building Information Modeling) Tools in the Application of Sustainable Retrofitting of Heritage Buildings in Egypt, Procedia Environ. Sci. 34 (2016) 258–270. P. Pauwels, R. Verstraeten, R. De Meyer, J. Van Campenhout, Architectural information modelling for virtual heritage application, in: Int. Conf. Virtual Syst. Multimed., Archaeolingua, 2008: pp. 18– 23. G. Costa, L. Madrazo, Connecting building component catalogues with BIM models using semantic technologies: an application for precast concrete components, Autom. Constr. 57 (2015) 239–248. P. Pauwels, D. Van Deursen, IFC/RDF: adaptation, aggregation and enrichment, in: First Int. Work. Linked Data Archit. Constr. (LDAC 2012), 2012. L.Y. Ding, B.T. Zhong, S. Wu, H.B. Luo, Construction risk knowledge management in BIM using ontology and semantic web technology, Saf. Sci. 87 (2016) 202–213. H. Schevers, J. Mitchell, P. Akhurst, D. Marchant, S. Bull, K. McDonald, R. Drogemuller, C. Linning, Towards digital facility modelling for Sydney opera house using IFC and semantic web technology, ITcon. 12 (2007) 347–362. S. Törmä, J. Oraskari, N.V. Hoang, Distributed transactional building information management, LDAC 2012. (2012) 9. D.-Y. Lee, H. Chi, J. Wang, X. Wang, C.-S. Park, A linked data system framework for sharing construction defect information using ontologies and BIM environments, Autom. Constr. 68 (2016) 102–113. E. Quagliarini, P. Clini, M. Ripanti, Fast, low cost and safe methodology for the assessment of the state of conservation of historical buildings from 3D laser scanning: The case study of Santa Maria in Portonovo (Italy), J. Cult. Herit. (2016). R. Quattrini, E.S. Malinverni, P. Clini, R. Nespeca, E. Orlietti, From TLS to HBIM. High quality semantically-aware 3D modeling of complex architecture, Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 40 (2015) 367. V. Caffi, B. Daniotti, T.M. LO, D. Madeddu, M. Muscogiuri, G. Novello, P.M. PAVANA, Il processo edilizio supportato dal BIMM: l’approccio InnovANCE-BIMM enabled construction processes: the InnovANCE approach, ISTeA, Bari. (2014) 91–109. E. Curry, J. O’Donnell, E. Corry, S. Hasan, M. Keane, S. O’Riain, Linking building data in the cloud: Integrating cross-domain building data using linked data, Adv. Eng. Informatics. 27 (2013) 206–219. J. Hois, M. Bhatt, O. Kutz, Modular Ontologies for Architectural Design., in: FOMI, 2009: pp. 66–77. AEC (UK) BIM Technology Protocol. Practical implementation of BIM for the UK Architectural, Engineering and Construction (AEC) industry., (2015) 47. C. Santagati, M. Lo Turco, From structure from motion to historical building information modeling: populating a semantic-aware library of architectural elements, J. Electron. Imaging. 26 (2017) 11007. A. Scianna, M. Serlorenzi, S. Gristina, M. Filippi, S. Paliaga, Sperimentazione di tecniche BIM sull’archeologia romana: il caso delle strutture rinvenute all’interno della cripta della chiesa dei SS. Sergio e Bacco in Roma., Establ. by Mauro Cris. Riccardo Fr. (2015). D. Simeone, S. Cursi, I. Toldo, G. CARRARA, B (H) IM-Built Heritage Information Modelling, Extending BIM approach to historical and archaeological heritage representation, in: FusionProceedings 32nd eCAADe Conf., n.d. C. Bianchini, C. Inglese, A. Ippolito, D. Maiorino, L.J. Senatore, Building Information Modeling (BIM): Great Misunderstanding or Potential, Handb. Res. Emerg. Technol. Digit. Preserv. Inf. Model. (2016) 67. P. Pauwels, W. Terkaj, EXPRESS to OWL for construction industry: Towards a recommendable and usable ifcOWL ontology, Autom. Constr. 63 (2016) 100–133.

Web references [32] [33] [34]

https://github.com/mmlab/IFC-to-RDF-converter http://www.ontospace.uni-bremen.de/ontology/modSpace/BuildingArchitecture.owl https://www.w3.org/TR/rdf-sparql-query/

14