Digital Twins in Civil Engineering: Conceptual Framework and Real-World Implementations

Abstract

Civil engineering stands at a critical juncture in the XXI century, facing a confluence of challenges and opportunities that demand a paradigm shift in practices and methodologies. As most structures built at the beginning of the past century globally approach the culmination of their designed lifespans, the need for sustainable, resilient, and technologically advanced solutions becomes paramount. To confront the complexity, this doctoral thesis endeavours to develop a comprehensive digital twin conceptual framework tailored for civil engineering, aware of ageing infrastructure, digitalisation, environmental impact, and the imperative to minimize waste. With numerous constructions approaching the end of their expected life cycles, the challenge lies not only in preserving the integrity of these structures, but also in reimagining them for a sustainable future. This work aims to investigate strategies for operation and maintenance, retrofitting, and sustainable policies. Regarding digitalisation, while strides have been made in embracing digital tools for design, construction, and project management, a fragmented landscape of technologies still persists. Siloed efforts limit the transformative potential of integrated digital solutions. This thesis aims to provide the pathways toward a cohesive and collaborative implementation of technologies under the umbrella of the digital twin. In the pursuit of efficiency and quality, civil engineering is witnessing a wave of innovation driven by emerging technologies. Building Information Modelling (BIM), the Internet of Things (IoT), and Artificial Intelligence (AI) are reshaping the industry landscape. This work critically evaluates the adoption and impact of these technologies, assessing their potential to revolutionise the practice of civil engineering in the Architecture, Engineering, Construction, and Operations and maintenance (AECO) sector, as the backbone providing the technical expertise needed to design, build, manage, operate and maintain the physical assets that support our societies. This thesis has confronted several challenges, encompassing the thorough conceptualisation of the Digital Twin (DT) for civil engineering with application in structures. It has also involved the integration of data and models within a Bayesian statistical framework to address updating and uncertainty quantification, the management of digital twin workflows through a high-level Petri net, the limited availability of data for training, and the establishment of a pipeline of surrogate models to facilitate diagnosis and prognosis within the DT framework. To overcome these challenges, AI strategies have been introduced, relying on Neural Networks (NN) and Deep Learning (DL) models.