Abstract
The DIGITMAN project (“Occupant-based DIGITal predictive MANagement to improve the built environment”) is intended to develop an occupant-based digital framework to support the predictive management of buildings during their operational phase. The research focused on integrating heterogeneous data sources, including BIM models, monitoring systems, IoT sensors, maintenance records, energy simulations, and occupancy-related information, into a unified decision-support system. The objective is to provide facility managers with practical tools to improve maintenance planning, building operation, energy performance, indoor environmental quality and safety assessment. The proposed system is implemented as a web-based prototype and validated on university buildings in Italy, demonstrating how occupant-centred data and predictive analytics can support more informed, proactive and resource-efficient decisions in complex public building stocks.
Results achieved
The project developed and demonstrated an occupant-based Digital Decision Support System (DDSS) for the predictive management of complex public building stocks. It addressed a key challenge for public administrations and universities: the need to manage buildings with integrated, reliable, and actionable information, overcoming fragmentation in data related to maintenance, operations, and safety. A major result of the project is the definition and implementation of a unified digital framework that connects heterogeneous data sources, including Building Information Models (BIM), Building Performance Simulation (BPS), IoT sensors, Building Automation Systems (BAS), maintenance records and occupancy-related information. These data are organized through a Knowledge Graph (KG), enabling different types of information to be linked to specific spaces, assets and operational conditions. This approach makes it possible to move beyond static digital models and to represent buildings as dynamic systems whose performance depends on actual use, occupancy patterns and management choices. The project produced a working prototype of the DDSS (TRL5), based on a modular microservice architecture. The prototype integrates four main analytical modules, each addressing a specific domain of facility management: Maintenance How-To, Operation How-To, Energy What-If and Safety What-If. The “How-To” logic supports short-term and daily operational decisions, helping facility managers understand how to respond to current or near-real-time conditions. The “What-If” logic supports medium- and long-term strategic planning, allowing decision-makers to compare alternative scenarios before implementing changes to building use, occupancy density, or technical operations. Within the Maintenance How-To module, the project developed Machine Learning (ML) procedures to analyse and classify maintenance requests. Historical maintenance data from a university building portfolio were processed using Natural Language Processing (NLP) and predictive classification methods. The system demonstrated the ability to automatically classify maintenance requests with good accuracy and to support their prioritisation according to technical category, urgency and programmability. The analysis showed that a large share of requests could be treated as low-priority and programmable interventions, supporting a shift from reactive maintenance to more planned and resource-efficient management. The Operation How-To module was developed to support daily monitoring of building operation and indoor environmental conditions. By integrating data from sensors and building automation systems, the prototype can monitor parameters such as temperature, relative humidity, CO₂ concentration, air quality indicators, particulate matter and illumination. These values are combined into a composite comfort and performance index, which provides an accessible indicator of indoor environmental quality and helps facility managers detect anomalies or comfort issues. The system also compares scheduled occupancy with estimated actual occupancy, helping identify mismatches between planned and actual use of spaces and adjust operations accordingly. The Energy What-If module enabled the evaluation of alternative occupancy and operational scenarios through simulation-based analysis. The project developed a BIM-to-BEM workflow to transform building information into energy simulation models and to compare different scenarios, such as changes in classroom density, space use or ventilation strategies. The simulations showed that building energy performance is strongly influenced by occupancy conditions and HVAC control strategies. The Safety What-If module addressed the evaluation of evacuation safety under different occupancy scenarios. The project developed a Building Safety Model (BSM) derived from the topological BIM representation, enabling assessment of evacuation routes, travel distances, and effective widths in relation to regulatory requirements. This module enables verification of whether a scenario that appears advantageous from the perspective of energy or space efficiency is also compatible with safety constraints. In this sense, the project demonstrated the value of cross-domain analysis, illustrating that a management option can be considered effective only if it performs adequately across multiple dimensions. The integrated DDSS was validated on two university case studies in Italy. The validation confirmed the applicability of the proposed approach in real facility management contexts and demonstrated the usefulness of the prototype for visualising key performance indicators, comparing scenarios and supporting informed decisions. The web-based interface allows users to explore buildings, select spaces and timeframes, monitor indicators and interact with predictive analytics through dashboards, charts and 3D visualisations. The project achieved a significant advancement in the digitalisation of building management by demonstrating how occupant-related data can serve as a central variable in predictive facility management. The results demonstrate that integrating occupancy, maintenance, energy, comfort, and safety information into a single decision-support environment can improve the allocation of technical and economic resources, support proactive planning, and reduce the risk of decisions based on isolated or incomplete evaluations. The project was funded by the European Union – Next Generation EU, Mission 4 Component 1, CUP: D53D23003630006.Dettagli del progetto
Responsabile scientifico: Riccardo Gulli
Strutture Unibo coinvolte:
Dipartimento di Architettura
Coordinatore:
Università Politecnica delle Marche(Italy)
Contributo totale Unibo: Euro (EUR) 59.804,00
Durata del progetto in mesi: 24
Data di inizio
28/09/2023
Data di fine:
27/09/2025
