Population based structural health monitoring framework using optimal sensor placement and pseudo-transfer learning
Loading...
Date
2026-06-20
Authors
Dahel, Ihab
Journal Title
Journal ISSN
Volume Title
Publisher
ENSTP-FJ
Abstract
Algeria's mass-housing programs use identical reinforced concrete buildings, making traditional Structural Health Monitoring (SHM) economically unfeasible. This thesis develops a Population-Based SHM (PBSHM) framework. A Ground+9-story (G+9) reference building (RPA 2024, BAEL 91/99) is analyzed. A Genetic Algorithm determines an Optimal Sensor Placement (OSP), identifying 20 sensors using the Modal Assurance Criterion (MAC) matrix. Inter-structure variations are modeled as Gaussian noise. A dual-branch Denoising Autoencoder (DAE) reconstructs sensor fields and predicts damage, achieving a coefficient of determination (R²) >0.93 and Mean Squared Error (MSE) <0.02, validating scalable deployment.
Description
90 p. : ill. ; 30 cm. + (1cd-rom).
Bibliogr.
Keywords
Population-based SHM , Denoising autoencoder , Optimal Sensor Placement (OSP) , Modal Assurance Criterion (MAC) , Flexibility-based damage indicator , Coefficient of determination (R²) , RPA 2024 (code) , Seismic monitoring , Surveillance basée sur la population (PBSHM) , Auto-encodeur débruiteur (DAE) , Placement optimal de capteurs (OSP) , Critère d'Assurance Modale (MAC) , Indicateur de dommage par flexibilité , Coefficient de détermination (R²) , RPA 2024 (règlement) , Surveillance sismique
Citation
Dahel, Ihab. Population based structural health monitoring framework using optimal sensor placement and pseudo-transfer learning. Mém. Ing., Alger, ENSTP-FJ, 2026.