A transfer Bayesian learning methodology for structural health monitoring of monumental structures. (15th November 2021)

• Record Type: Journal Article
• Authors: Ierimonti, Laura; Cavalagli, Nicola; Venanzi, Ilaria; García-Macías, Enrique; Ubertini, Filippo
• Is Part Of: Engineering structures. Volume 247(2021)
• Publication Date: 2021-11-15
• Subjects: Structural health monitoring -- Bayesian model updating -- Damage detection -- Surrogate modeling -- Transfer learning; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2021.113089
• Languages: English
• ISSNs: 0141-0296

Abstract

Abstract: A critical aspect in structural health monitoring (SHM) applied to civil engineering structures is the lack of diagnostic labels able to assign a damage class to the measured data. In this context, a semi-supervised learning methodology, designated as transfer Bayesian learning (TBL), is proposed with the main objective of labeling post-processed data in a probabilistic way by selecting a limited number of informative elements. The proposed method allows to define multi-class labels by making use of a surrogate model (SM) of the structure considering specific damage-sensitive mechanical parameters. The methodology is applied in a monumental building, the Consoli Palace, located in Gubbio, central Italy. The structure is instrumented with several sensors in order to measure vibrations, temperature and possible variation of existing cracks' amplitudes. Several nonlinear pushover analyses are carried out on a calibrated finite element (FE) model to use them in conjunction with Engineering judgment for the definition of the damage-sensitive regions. The SM, consisting of a simplified model that continuously exchanges information with the physical reality observed through the measurement system, is then used as a class classifier by means of a sensitivity damage chart (SDC). Finally, a Bayesian model updating of the damage-dependent parameters allows the probabilistic damage identification. Graphical abstract: Highlights: A new Bayesian transfer learning method for structural health monitoring is proposed. Real-time long-term monitoring data are used to validate the proposed methodology. A surrogate-based sensitivity damage chart enables the transfer learning paradigm. Local probabilities of damage allow to take SHM-informed decisions.