Results

To reach the proposed scope, DIGITMAN has been organized comprising three main research steps.

The first step comprises the development of the conceptual model for the common digital framework and its related database, based on existing data, to be applied to pilots and considering the pillar-specific issues (WP1).

The second step comprises the development of predictive methods considering occupant-based evaluations, for maintenance (WP2), operation (WP3) and safety (WP4) tasks, in a separate manner, to hence support automation of facility manager’s daily activities and prediction of impact of future conditions depending on new occupancy scenarios, thanks to pillar-specific analytics

The third step comprises the development of a multicriterial approach to link the predictive-based analytics of the three pillars, and of a prototype tool to merge the different evaluations, considering the needs of building managers (daily management and probabilistic-based future predictions based on different probabilistic occupancy scenarios-WP5)

Research activities have been organized in WPs comprising one or more Tasks.

Overall WPs and Tasks structure in DIGITMAN

WP1-CONCEPTUAL MODELLING OF THE DIGITAL FRAMEWORK AND RELATED DATABASE DEFINITION

D1.1 – definition of an occupant-centric conceptual framework and correlation methodology

type: report; involved tasks: T1.1
Abstract: This deliverable outlines the foundational groundwork of the DIGITMAN project and establishes a strategic framework for an occupant-centric approach within digital building management. The document begins by delineating the scope of the research project, focusing on occupancy-based management strategies and identifying pertinent areas of interest. It then outlines a framework that details the logic and functionalities to be developed as project outcomes, followed by defining the ontology and semantics of relevant information pertaining to building management. Finally, the methodology for modeling building information is illustrated, and the pilot buildings chosen to demonstrate the project’s methodology are presented.
Download the deliverable:

D1.2 – integrated structuring of a common database

type: report; involved tasks: T1.2
Abstract: This deliverable presents the integrated approach for structuring the DIGITMAN project’s common database. It establishes the foundational data structures supporting the project’s objectives. The document first outlines the system architecture and the methodologies chosen for creating and managing the database. Then, it provides detailed reports to document the classification systems adopted for organically mapping maintenance activities, functional types of spaces, and equipment devices within the database. These classification systems align coherently with the different datasets from the partner institutions. Finally, the document reports the property sets and quantities used to assign data to the DIGITMAN database elements. They are based on a shared notation aligned with the partner’s datasets and international standards and schemas.
Download the deliverable:   

D1.3 – Data acquisition methodology

type: report + database; involved tasks: T1.3
Abstract: This deliverable reports on the data acquired and structured to build the knowledge base of the DIGITMAN project, encompassing an array of databases and models useful for the project’s purposes. The document details the creation of tables that synthesize building data from partner institutions’ buildings, structured in integrated relational and graph databases for consistency and ease of querying. In particular, the static elements and related semantic data are stored within a graph database, maintenance textual logs are deposited within a MySQL database, and sensors’ observation are store into JSON documents. Finally, the deliverable provides the building information models developed for three significant buildings within the sample analyzed. These models are stored in formats compatible with industry-standard software such as Autodesk Revit, Industry Foundation Classes (IFC), and Topologic JSON files. 
Download the deliverable:

WP2-MAINTENANCE PREDICTIVE MODULE DEVELOPMENT

D2.1 – KPIs and predictive methods for maintenance tasks

type: report ; involved tasks: T2.1, 2.2, 2.3
Abstract: This deliverable describes the action performed to develop the module addressed to predict the priority to assign to end-users’ maintenance requests collected the two ticketing support systems in use at UNIVPM and POLIMI. A unified database has been created to align  UNIVPM and POLIMI tickets and to integrate structured and unstructured data to support decision-making processes. Following a rigorous text pre-processing phase, NLP methods, including normalisation, stop-word removal, stemming, lemmatization, and regex data cleaning, have been applied to optimize the dataset. Key Performance Indicators (KPIs) were defined to assess building health status, maintenance efficiency, and request prioritization. The most relevant KPIs are also selected for future multi-criteria evaluations and improve comparisons between different areas and activities. A Bidirectional Long Short-Term Memory (Bi-LSTM) model was implemented to automatically predict request priorities, demonstrating high accuracy (0.87) and reliable classification metrics. This predictive approach enables efficient allocation of maintenance resources and can support decision-makers in feature prediction, allowing real-time automated classification with human supervision. The integration of KPIs and ML-based classification methods within a digital dashboard facilitates a comprehensive assessment of maintenance needs, improving response times and resource optimization. Future developments within WP5 will focus on integrating the selected KPIs into assessment metrics, weighting and merging them with additional “how-to” microservices. The integration of these methodologies within a BIM-based and spatially structured framework will further enhance facility management by linking current and predictive maintenance insights directly to specific building stock elements.
Download the deliverable:

WP3-OPERATION PREDICTIVE MODULE DEVELOPMENT

D3.1 – KPIs and predictive methods for operation tasks

type: report ; involved tasks: T3.1, 3.2, 3.3
Abstract: This deliverable address operational challenges in building management within public administrations, focusing on university buildings.  As local public administrations, universities must strategically manage facilities, balancing effectiveness, efficiency, and stakeholder needs. This includes daily maintenance to prevent degradation and ensure sustained performance. Therefore, a performance-based approach to built asset management is essential. This research introduces a Digital Decision Support System (DDSS) for strategic and operational performance-based management of existing buildings. The DDSS utilizes digital technologies and data techniques to establish a data environment providing insights into energy performance relative to planned occupancy. Supporting both “how-to” (operational) and “what-if” (strategic) analyses, the system aims to increase building managers’ understanding of energy behavior. This is achieved by analyzing energy needs and connecting them to occupancy parameters using Key Performance Indicators (KPIs). This knowledge can inform decisions such as prioritizing energy-efficient spaces and identifying areas for energy efficiency improvements. The resulting knowledge informs decisions, prioritizing energy-efficient spaces and identifying improvement areas, with the ultimate goal of enhancing Indoor Environmental Quality (IEQ) while minimizing energy consumption.
Download the deliverable:

WP4-SAFETY PREDICTIVE MODULE DEVELOPMENT

D4.1 – KPIs and predictive methods for safety tasks

type: report ; involved tasks: T4.1, 4.2, 4.3
Abstract: Fire Safety is one of the most relevant topics in building safety. Well-established regulatory framework implies requirements on “deemed-to-satisfy” solutions to be easily and quickly checked by decision makers. In particular, evacuation-related topic surely represents a fundamental issue, being widely interconnected with building layout and occupancy alternatives, and being thus relevant for “how-to” (current scenario) and “what-if (design alternatives) tasks. Due to the complexity of buildings and their occupancy, modeling tools are needed to collect and manage these fire safety features and input data, and then assess regulation compliance using rapid Key Performance Indicators (KPIs). This report aims at defining simple, regulation-based KPIs for building safety assessment, mainly oriented towards “what-if” tasks. KPIs are implemented in BIM to Building Performance Simulation tools (i.e. Building Safety Model – BSM), and then applied to relevant DigitMan case study. KPIs trace geometrical issues under alternative scenarios, quickly supporting decision makers since they are associated with specific building levels, from micro (space, building component in the mans of egress) to macro (whole building/compartment) scale.
Download the deliverable:

WP5-SYSTEMATIZATION, TOOL DEVELOPMENT AND DEMONSTRATION

D5.1 – Multicriteria analysis methods description

type: report ; involved tasks: T5.1
Abstract: ongoing work
Download the deliverable:

D5.2 – Development of a comprehensive prototpe tool and user guide for building managers

type: tools + manual (report); involved tasks: T5.2
Abstract: ongoing work
Download the deliverable:

D5.3 – Application of the prototype tool on an exemplary case study

type: report; involved tasks: T5.3
Abstract: ongoing work
Download the deliverable: