DIGITAL TWIN ANALYTICS; LIFE-CYCLE MODELING OF STRUCTURES FOR PRESENT AND FUTURE CONDITION ASSESSMENT
Date
2021
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The evaluation of structural systems is a necessary task in order to maintain the integrity of structures over time. These assessments are designed to detect damages of structures and ideally help inspectors to estimate the remaining life of structures. Current methodologies for monitoring structural systems, while providing useful information about the current state of the structures, are limited in the monitoring of structural defects over time and linking them to predictive simulation. A digital twin (DT), as defined here, integrates monitoring observations and geometric survey data with numerical simulations in order to provide depictions of life-cycle performance. The objective of this research is to propose an integrated framework that supports digital twin modeling of structures. Two main aspects of DT model are studied in this dissertation. First, tracking the evolution of remotely sensed defects, along with linking them to numerical simulation is studied in order to provide structural performance characteristics over time. Second, integrating survey information from various data sources is investigated. In the first section, remotely sensed defects are parametrized using feature extraction techniques, and a stochastic dynamic model is then adapted to features to model their evolution over time. Then the future state of defects is predicted through the dynamic model. Later, the Finite Element Model of structural components is linked to the future state of defect for predictive simulation. In the second section, results from multiple non-destructive evaluation (NDE) techniques are integrated and used as input in a machine learning classifier to provide a feature-level data fusion of NDE measurements. This integrated framework supporting the life-cycle modeling of structural defects provides more reliable forecasting capabilities and a more comprehensive understanding of structural performance, which improves decision-making processes for asset management. The accuracy, effectiveness, adaptability, and feasibility of the presented framework was evaluated with sets of synthetic and laboratory scale data.