CONSERVATION PROCESS MODEL ( CPM ) : A TWOFOLD SCIENTIFIC RESEARCH SCOPE IN THE INFORMATION MODELLING FOR CULTURAL HERITAGE

The aim of the present research is to develop an instrument able to adequately support the conservation process by means of a twofold approach, based on both BIM environment and ontology formalisation. Although BIM has been successfully experimented within AEC (Architecture Engineering Construction) field, it has showed many drawbacks for architectural heritage. To cope with unicity and more generally complexity of ancient buildings, applications so far developed have shown to poorly adapt BIM to conservation design with unsatisfactory results (Dore, Murphy 2013; Carrara 2014). In order to combine achievements reached within AEC through BIM environment (design control and management) with an appropriate, semantically enriched and flexible The presented model has at its core a knowledge base developed through information ontologies and oriented around the formalization and computability of all the knowledge necessary for the full comprehension of the object of architectural heritage an its conservation. Such a knowledge representation is worked out upon conceptual categories defined above all within architectural criticism and conservation scope. The present paper aims at further extending the scope of conceptual modelling within cultural heritage conservation already formalized by the model. A special focus is directed on decay analysis and surfaces conservation project.


INTRODUCTION
The research has focused on architectural modelling developed by means of ontologies; this model is not a simple data repository on existing architecture, but has to support heritage conservation process.Within this methodology, concepts are represented through entities and properties.Entities are set up into hierarchic taxonomies by means of an analytic process.Different classesdeclined into super-classes and sub-classes according to the relationship existing between themorganise different meaning in architecture.Every meaning is associated to single concepts, specified through definitions and integrated by properties describing relationship with the context.This methodology, defined as the "formalization of a conceptualization" (Gruber 2009), makes it possible to hook modelling to a logic framework that grants the reliability of representation.In addition, it permits to establish relationships between concepts, enabling the description of hermeneutic processes, such as historical critics, not static but flexible and always open to new developments.However, rigour required by the ontology formalisation (formal modelling), on the one hand makes it possible to sharpen the analysis and to articulate the research upon more appropriate definitions (Ciotti 2014, Gigliozzi et al. 2003, Mc Carty 2005, Orlandi 2010); on the other hand, to avoid to fall in contradiction, it may lead to simplified representation.A special focus on the participation of specialists within the formalization process can help to avoid that kind of risk.Starting by cultural heritage description already been developed by means of ontologies (Crofts et al. 2010; http://www.cidoccrm.org/sites/default/files/cidoc_crm_version_6.2.pdf [14.04.2017]), the research aims at encompassing other domains involved in historical architecture investigation process and conservation design.All information required for the exchange and integration of heterogeneous scientific documentation developed within cultural heritage process establishes the intended scope of the proposed ontology.Such a model should convey the requirement that the depth and the quality of descriptive information that can be handled should be sufficient either for academic research or for professional activity.This ambition arises from the fact that conservation project advocates a special knowledge process.This process developed upon the building has to work out a critical assessment that gives direction to conservation project.Therefore, an important focus of the model must be oriented to scientific reliability.

THE CONSERVATION PROCESS MODEL
The proposed model has a twofold aim, it intends to capture and represent the semantic contents of cultural heritage conservation process and to point at working up a model that may achieve integration, mediation and interchange of information in the midst of cultural heritage conservation discipline.Such a model may become an advisable instrument able to support historical architecture conservation research.This kind of modelling has tried to follow existing major ontologies such as Cidoc CRm and FRBRoo adapting and integrating their structure to the considered discipline.Beyond technical items, a particular attention has focused conceptual issues, particularly aroused within architectural project modelling, considering that the scopes addressed to by Cidoc and FRPRoo are library and museum communities.In this direction, interesting researches have been developed (Noardo 2015;Noardo 2016;Felicetti et al. 2013, Ronzino et al. 2013, Ronzino et al. 2015, Guillem) although not especially focused on cultural heritage conservation process.Within Conservation Process Model (CPM), some existing classes have been introduced from scratch; others have been adapted to the context.Some classes are perfectly fitting conservation issues, while some contents needed an ad hoc formalization, either for technical requirements or for theoretical implications.The prefixes 'E' (for classes) and 'P' (for properties), followed by the subscript CPM and a progressive number, encode new declared classes and properties.

Modelling knowledge for conservation process
So far, five principal domains have been defined: artefact, investigation process, actors, lifecycle 1 and lifecycle 2 (Acierno et al. 2017).The artefact domain focuses on the architectural organism description.Actors domain refers to people concerned with the building existence or studies; cultural heritage process is addressed to the description of all the analysis worked out on the building and the information gathered; lifecycle 1 focuses on the transformation process description and lifecycle 2 is conceived for design and managing processes (Fig. 1).
The structure itself follows the CIDOC scheme.Moreover, other two existing models have been important benchmarks for the formalization, actually CIDOC extension: FRBRoo (Functional Requirements for Bibliographic Records, http://www.cidoc-crm.org/frbroo/ModelVersion/version-2.4) and AR model developed by Guillelm et al. 2016.Inside the model, classes and properties hierarchies have been extended within the domains, while entities and properties are defined through default specifications.In particular, class identification (Superclass of/ subclass of); scope note; order logic and properties identified different entities.Furthermore, the description of entity properties needs to specify: domain, range, scope note and quantification.Finally, the model, developed by means of ontology, will integrate BIM modelling.Highlighted by the existing literature (Pauwels et al. 2013), the work focusses on the possibility of achieving a satisfactory representation by enhancing interoperability between these two different applications.The main attention concerns the consistency between the two reference structures, in order to facilitate mapping activity.Each BIM element will correspond to an ontology entity that will provide a complete definition of its underlying semantics.A specific instrument, a 'BIM Semantic Bridge', ensures the connection between BIM database and the knowledge base (Acierno et al. 2017).

Modelling decay analysis
As far as ontologies are concerned for decay analysis, some studies have already proposed a dedicated model (Caciotti 2015, Noardo 2014) that deals specifically with the professional edge of the process.However, we focus here the challenge to develop a model that may play a twofold role coping with both research and operational facet.
Decay analysis has been modelled within the template already proposed for the other types of investigation described by the model (Acierno et al. 2016).In CPM each analysis is described through the resources needed to be developed -as investigation methods, tools and samples-, the actors who carry out the job, the inputs needed and the information provided (fig.2).
Figure 2. Heritage process: investigation model Investigation methods are manifold, spreading from the simple direct observation to more sophisticated techniques such as thermography or chemical analysis.Tools are also rather diversified, as decay may be surveyed through cameras, videos, thermal imagery etc.Finally, samples may be needed in case of more specialized analysis.The actors who run the survey may be architects, engineers, diagnosticians, all of them described within a specific domain.Differently, inputs needed to develop the analysis are described within the cultural heritage investigation domain.These inputs range from the geometric survey to the materials survey and the existing lexicons, provided by the Institutions concerned by cultural heritage protection (i.e.Illustrated glossary on stone deterioration patterns published by Icomos).Formalization of decay analysis results deserves however a special attention.Albeit its nature as activity that aims at identifying building state of conservation, hence pertaining to the Cultural heritage process domain, it necessarily deals with the changing nature of architecture, coping therefore with the transformation process domain called lifecycle 1.Although the content of the investigation refers quite clearly to the changing of physical conditions, 'decay' encoding is not a foregone job.Actually, two Cidoc classes may present the right content: modification and condition state.The former, previously adopted, may certainly well describe the fact that something arose to change the primary consistency; nevertheless this statement may prove to be quite inappropriate when referring to historical architecture decay.As a matter of facts, the main attention within the scope of the decay analysis must focus on the building physical consistency in relation to its current state of conservation and not to the presumed primary state or to any other previous condition, actually out of our knowledge.Assuming this stance, the best class to represent decay analysis proved to be condition assessment, in turn a subclass of condition state.Cidoc System has formalized the former (encoded as E14) as "the act of assessing the state of preservation of an object during a particular period."In addition, Cidoc definition refers to the operational context of condition assessment further clarifying the conceptual frame of the class: "The condition assessment may be carried out by inspection, measurement or through historical research.This class is used to document circumstances of the respective assessment that may be relevant to interpret its quality at a later stage, or to continue research on related documents."In conclusion, decay analysis is formalized as a sub class of 'condition assessment', and in a first order logic it is expressed as follows: Ecpm1(x) ⊃ E14 (x).
The results of decay analysis are described through the property identifies and by the classes condition state and type.Therefore the statement will be: decay analysis (Ecpm1) has identified (PP35) condition state (E4) that has type (P2): decay phenomenon (Ecpm 2).Decay phenomena will be described according to the existing lexicons as ICOMOS Illustrated glossary on stone deterioration patterns, Normal 1/88, UNI 11182 and will be articulated into two main classes, superficial and structural decay.Each phenomenon identified will be then related to the surfaces concerned through the property is extended to (Pcpm1) and finally to the building (artefact).Therefore, instances, representing the particular area or areas interested by decay will be created.The single areas are referred to the artefact through the well-known assembly properties such as 'part of'/whole of' and are modelled as information objects that will provide information for the surfaces conservation project (fig.3).In order to clarify the content of this class, it is important to observe that although architectural conservation project requires several inputs to be planned, provided by geometrical and material survey, historical investigation, diagnostics, there are no bi-univocal relations linking these inputs to the actions required to realize the project.Indeed the project is not worked out upon each single result, but springs from their overall critical assessment.Therefore, the model will not foresee a direct link between a decay phenomenon and the intervention to fix it, as the projectto be effective -has always to mediate this relation with other contents.For instance: if the decay phenomenon is related to an event that has an historical importanceas in the coats of arms on the palaces in Feltre (Bl) erased by Venetiansprobably filling the gaps will not have any sense.Again, cracks due to the simple juxtaposition of two different walls cannot implicate the destruction of the archaeological evidence and the reconstruction of the whole masonry, and so on and so forth.Undoubtedly, only a-critical approach may conceive a set-up procedure, actually highly undesirable within the scope of cultural heritage conservation.The special feature of unicity of historical architecture entails singularity of intervention.Moreover, some conservation interventions do not concern just decayed areas.As an example, limewater is generally applied to the whole surface of a building to harmonize lacking parts with integer ones.A model developed just on the correspondence between decay phenomenon and intervention would miss to represent that situation.Further modelling follows AR development adapting its structure to conservation process.As to translate the architect idea -developed within an in depth knowledge process and appraisal -to documents that give direction to operational activities, many planning steps are required, the model foresees a number of subclasses that follows the construction process.Moreover, the law also prescribes planning activity.In Italy, a project has developed in three main work phases: feasibility planning, definitive planning and executive planning1 .Then, the model has to follow the same scheme.Starting from the main class Architectural Conservation project, that will represent the whole project, several subclasses will further specify the intervention activity (fig.5).Thus far, the research has focussed on surface conservation design, but the full ontology in its future developments will consider all the other issues.Thereby Surfaces Conservation architectural project (encoded Epcm4) will present the subclass Surfaces Conservation executive planning (encoded Epcm5), that is equal to AR class, Construction project.Following the first order logic representation, this will be expressed as follows: Going further, the model reflects the workflow structure and each phase of surfaces conservation project, such as preconsolidation (Epcm6), cleaning (Epcm7), consolidation (Epcm8), protection (Epcm9), addition (Ecpm10) and removal (Ecpm 11) will correspond to a class.These classes are actually equal to AR class Construction plan [encoded AR 4].Once defined the specific intervention, the model will specify the materials employed and finally the link to the peculiar activity described in the Institutional prices lists (fig.6).The instances that bridge the ontology to the BIM modelling and are represented in a BIM environment are either the decayed areas or the interventions.Specific efforts been addressed to conceive a structure that allows an easy mapping between the two environments.Integrating ontology modelling with BIM environment makes it possible to verify the correspondence to geometrical representation, to facilitate visualization and to extend the possibility of interaction to a wider community of actors and traditionally involved in cultural heritage conservation and managing process.The combined use of the two instruments, BIM and ontology may provide an interesting enhancement of the managing potential of the whole process.

CONCLUSIONS
This work focuses on the overcoming of a specific drawback of BIM current use within cultural heritage conservation scopethe low level of in the representation-integrating BIM and semantic web methodologies.As the conservation design moves from a hermeneutical process enhanced from building knowledge, each ICT (Information Communication Technology) instrument addressed to cultural heritage protection may not provide adequate knowledge representation.The use of ontologies, as well as their integration with a Building Information Modelling environment, allows a homogeneous, accessible and computable structured formalization of both direct and indirect knowledge necessary for the full comprehension of an architectural artefact.As a matter of facts, the proposed model intends to merge potentialities of new technologies, concerning strictness and capability of managing complex data systems, with humanistic issues that characterize each activity addressed to cultural heritage.

Figure 4 .
Figure 4. Classes formalization and integration within existing models

Figure 5 .
Figure 5.Comparison between model classes and traditional representation of the surfaces conservation project .

Figure 6 .
Figure 6.Surfaces conservation project model