An integration framework for heritage maintenance coupling HBIM towards cognitive digital twin (CDT): concept and ongoing implementation at Villa d'Este, Tivoli

The conservation and management of historic buildings require approaches that preserve cultural heritage while improving functionality, environmental performance, and occupant well-being. The Architecture, Engineering, Construction, and Operations (AECO) sector remains constrained by fragmented workflows, limited technological integration, and siloed data repositories that hinder informed decision-making across the building lifecycle. This paper proposes a methodological framework that couples Heritage Building Information Modelling (HBIM) with an Internet of Things (IoT) monitoring layer, and outlines its planned evolution toward a Cognitive Digital Twin (CDT) to support pre intervention decision-making in heritage maintenance and refurbishment. The framework defines a four-phase early-stage workflow with the scope to reconcile quantitative performance objectives with occupant needs and conservation constraints. The cognitive layer (semantic enrichment, knowledge graphs, and a Graph Neural Network, GNN) and the associated multi-criteria and multi-objective decision support are presented as a development roadmap rather than as already-implemented capabilities. The framework is being instantiated through an ongoing implementation at the Grotta di Diana, within the Villa d'Este complex (Tivoli, Italy), a UNESCO World Heritage Site, where a non-invasive IoT sensor network is being deployed in a constrained underground environment. As the installation and the acquisition of monitoring data are still in progress, this paper focuses on the definition of the framework and, through the implementation, on demonstrating the operational feasibility of non-invasive instrumentation in such settings; the analysis of monitoring data and the activation of the predictive components are identified as ongoing and future works.