Rockfall impact conditions vary widely in energy, trajectory, and location, requiring their probabilistic rather than deterministic characterisation. The behaviour of the articulated modular concrete-block rockfall protection structure presented in this project involves multi-mechanism energy dissipation and is highly sensitive to both impact conditions and wall geometry. Building on the existing Non-Smooth Contact Dynamics (NSCD) model implemented in SICONOS, this work systematically investigates the influence of wall geometry on the structural response of the system across ten configurations and a realistic range of impact conditions. Given the computational cost of full NSCD simulations, Polynomial Chaos Expansion-based surrogate models are calibrated in UQLab from a Sobol-sampled dataset of Impact Condition Parameters, enabling rapid response evaluation. Leave-One-Out and Mean Absolute Error assessments confirm that the metamodels are sufficiently accurate for comparative purposes. Sobol sensitivity analysis identifies the impacting kinetic energy as the dominant driver of response, with impact position and inclination playing secondary, geometry-dependent roles. The calibrated surrogates are then used to rank the geometries against realistic impact scenarios and user-defined, multi-criteria design requirements. Results show that no single geometry performs best across all conditions: the optimal configuration depends jointly on the scenario and the criteria considered, especially when spatial, maintenance, and budget constraints are combined. This constitutes a first step toward a simulation-based, site-specific design framework for this passive rockfall protection technology.
Rockfall impact conditions vary widely in energy, trajectory, and location, requiring their probabilistic rather than deterministic characterisation. The behaviour of the articulated modular concrete-block rockfall protection structure presented in this project involves multi-mechanism energy dissipation and is highly sensitive to both impact conditions and wall geometry. Building on the existing Non-Smooth Contact Dynamics (NSCD) model implemented in SICONOS, this work systematically investigates the influence of wall geometry on the structural response of the system across ten configurations and a realistic range of impact conditions. Given the computational cost of full NSCD simulations, Polynomial Chaos Expansion-based surrogate models are calibrated in UQLab from a Sobol-sampled dataset of Impact Condition Parameters, enabling rapid response evaluation. Leave-One-Out and Mean Absolute Error assessments confirm that the metamodels are sufficiently accurate for comparative purposes. Sobol sensitivity analysis identifies the impacting kinetic energy as the dominant driver of response, with impact position and inclination playing secondary, geometry-dependent roles. The calibrated surrogates are then used to rank the geometries against realistic impact scenarios and user-defined, multi-criteria design requirements. Results show that no single geometry performs best across all conditions: the optimal configuration depends jointly on the scenario and the criteria considered, especially when spatial, maintenance, and budget constraints are combined. This constitutes a first step toward a simulation-based, site-specific design framework for this passive rockfall protection technology.
NSCD simulation-based design optimization of modular rockfall barriers
GUIDOLIN, CLAUDIA
2025/2026
Abstract
Rockfall impact conditions vary widely in energy, trajectory, and location, requiring their probabilistic rather than deterministic characterisation. The behaviour of the articulated modular concrete-block rockfall protection structure presented in this project involves multi-mechanism energy dissipation and is highly sensitive to both impact conditions and wall geometry. Building on the existing Non-Smooth Contact Dynamics (NSCD) model implemented in SICONOS, this work systematically investigates the influence of wall geometry on the structural response of the system across ten configurations and a realistic range of impact conditions. Given the computational cost of full NSCD simulations, Polynomial Chaos Expansion-based surrogate models are calibrated in UQLab from a Sobol-sampled dataset of Impact Condition Parameters, enabling rapid response evaluation. Leave-One-Out and Mean Absolute Error assessments confirm that the metamodels are sufficiently accurate for comparative purposes. Sobol sensitivity analysis identifies the impacting kinetic energy as the dominant driver of response, with impact position and inclination playing secondary, geometry-dependent roles. The calibrated surrogates are then used to rank the geometries against realistic impact scenarios and user-defined, multi-criteria design requirements. Results show that no single geometry performs best across all conditions: the optimal configuration depends jointly on the scenario and the criteria considered, especially when spatial, maintenance, and budget constraints are combined. This constitutes a first step toward a simulation-based, site-specific design framework for this passive rockfall protection technology.| File | Dimensione | Formato | |
|---|---|---|---|
|
Guidolin_Claudia.pdf
embargo fino al 15/07/2027
Dimensione
21.91 MB
Formato
Adobe PDF
|
21.91 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/110643