COGNIBUILD: Cognitive Digital Twin Framework for Advanced Building Management and Predictive Maintenance

Abstract

AbstractAccording to contemporary challenges of digital evolution in management and maintenance of construction processes, the present study aims at defining valuable strategies for building management optimization. As buildings’ and infrastructures’ Digital Twins (DT) are directly connected to physical environment through the Internet of Things (IoT), asset management and control processes can be radically transformed. The proposed DT framework connects building information model (BIM) three-dimensional objects to information about the planned maintenance of components, supplying system’s self-learning capabilities through input data coming from Building Management Systems (BMSs), ticketing, as well as maintenance activities’ data flow both as-needed or unexpected. The concept of real-time acquisition and data processing set the basis for the proposed system architecture, allowing to perform analysis and evaluate alternative scenarios promptly responding to unexpected events with a higher accuracy over time. Moreover, the integration of artificial intelligence (AI) allows the development of maintenance predictive capabilities, optimizing decision-making processes and implementing strategies based on the performed analysis, configuring a scalable approach useful for different scenarios. The proposed approach is related to the evolution from reactive to proactive strategies based on Cognitive Digital Twins (CDTs) for Building and Facility Management, providing actionable solutions through operational, monitoring and maintenance data. Through the integration of BIM data with information systems, BMS, IoT and machine learning, the optimization and real-time automation of maintenance activities are performed, radically reducing failures and systems’ breakdowns. Therefore, integrating different technologies in a virtual environment allows to define data-driven predictive models supporting Building Managers in decision-making processes improving efficiency over time and moving from reactive to proactive approaches.