Application of a Structural Digital Twin on a Laboratory Model for Performance Monitoring of Aging and Degradation

Abstract

The application of digital twin technology offers the potential to significantly enhance the reliability and performance of structural health monitoring techniques by providing detailed information on, and measurements of, the as-built properties and performance of structures. In addition, a digital twin provides greater support for the application of probabilistic methods for prognostication and risk assessment following model updating and diagnostics provided by structural health monitoring. To date, digital twins have been predominantly applied in the advanced manufacturing space, although the increased use of modular and prefabricated construction has driven interest in the extension of digital twin technology to both civil and nuclear structures. This chapter details the implementation of a structural digital twin and investigates its influence on the performance of a Bayesian vibration-based structural health monitoring approach. In this chapter, a digital twin is developed throughout the construction of a model-scale steel-plate composite modular wall in a laboratory environment. Throughout the design, fabrication, and erection of the specimen, design calculations, numerical models, LiDAR point clouds, physical measurements of geometric features, and quality control inspection data are integrated into a digital twin framework to develop an as-built representation of the structural geometry, material properties, and condition throughout the stages of construction. The modular nature of the construction results in different effective section properties for wall modules as the early-age elastic properties of the structural concrete are developed. This time dependency of the section properties is leveraged to produce a series of datasets where structural parameters in the model evolve due to aging. Probabilistic model updating using modal parameters obtained from experimental modal analyses is used for structural identification of the as-built model. Static loading of the specimen with supplemental instrumentation is used to validate the parameter estimates. The performance of the vibration-based structural identification is evaluated both with and without the as-built information produced by the structural digital twin framework.