Structural health monitoring of masonry structures, relying on strain-based strategies, requires capable techniques for distinguishing damage-induced strain variations from those caused by environmental fluctuations. This study presents a novel SHM approach developed within the framework of nonlinear cointegration, specifically designed to process strain measurements and eliminate the influence of external environmental factors such as temperature and relative humidity. The strategy leverages neural networks to model the complex nonlinear relationship between environmental variables and the acquired strain time series, enabling the extraction of damage-sensitive features that remain unaffected by environmental conditions. By ensuring that strain measurements are corrected for environmental influences, the proposed method enhances the reliability of damage detection, allowing for a more accurate assessment of structural integrity. The effectiveness of this approach is evaluated through its application to a full-scale masonry building mock-up subjected to controlled damage scenarios under naturally varying environmental conditions and equipped with innovative brick-like strain sensors. Results demonstrate that the cointegration-based strategy effectively isolates damage-induced strain changes while suppressing misleading variations due to environmental fluctuations. The findings confirm that nonlinear cointegration, in combination with neural networks and smart bricks, provides a robust framework for continuous strain monitoring in masonry structures. This methodology offers significant advancements in structural health monitoring by improving the sensitivity and robustness of strain-based damage detection strategies, paving the way for more reliable monitoring solutions in real-world applications. The study highlights the potential of data-driven techniques to enhance structural assessment and long-term maintenance strategies for heritage as well as modern masonry structures.
Strain-based damage detection using nonlinear cointegration theory: application to a masonry building model using smart bricks
Michele, Mattiacci;Meoni, Andrea;D'Alessandro, Antonella;Glisic, Branko;Ubertini, Filippo
2025
Abstract
Structural health monitoring of masonry structures, relying on strain-based strategies, requires capable techniques for distinguishing damage-induced strain variations from those caused by environmental fluctuations. This study presents a novel SHM approach developed within the framework of nonlinear cointegration, specifically designed to process strain measurements and eliminate the influence of external environmental factors such as temperature and relative humidity. The strategy leverages neural networks to model the complex nonlinear relationship between environmental variables and the acquired strain time series, enabling the extraction of damage-sensitive features that remain unaffected by environmental conditions. By ensuring that strain measurements are corrected for environmental influences, the proposed method enhances the reliability of damage detection, allowing for a more accurate assessment of structural integrity. The effectiveness of this approach is evaluated through its application to a full-scale masonry building mock-up subjected to controlled damage scenarios under naturally varying environmental conditions and equipped with innovative brick-like strain sensors. Results demonstrate that the cointegration-based strategy effectively isolates damage-induced strain changes while suppressing misleading variations due to environmental fluctuations. The findings confirm that nonlinear cointegration, in combination with neural networks and smart bricks, provides a robust framework for continuous strain monitoring in masonry structures. This methodology offers significant advancements in structural health monitoring by improving the sensitivity and robustness of strain-based damage detection strategies, paving the way for more reliable monitoring solutions in real-world applications. The study highlights the potential of data-driven techniques to enhance structural assessment and long-term maintenance strategies for heritage as well as modern masonry structures.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


