Assessing effect of LOcal SubSoil VARiability and Uncertainty in SSI

The accurate modelling of soil properties and soil-structure interaction (SSI) are an important issue for the seismic margin and safety assessment of industrial plants and critical infrastructures.
 
Concurrently, it has been shown in the past that the spatial variability may have a major impact on extended and multi-supported structures.
 
The spatial variability may also play a major role in the accurate computation of Soil-Foundation-Structure Interaction problems with complex boundary conditions, such as it is generally the case for embedded structures (excavation, backfill) or when seismic isolation is considered. 

LOSSVAR implemented probabilistic numerical SSI models and evaluate the impact of random soil variability and epistemic uncertainty on the expected seismic margin and safety factors in probabilistic risk assessment. The numerical simulations are backed up by experimental tests. 

For this purpose, laboratory tests with small scale soil-structure models exhibiting soil variability were carried out. Benchmarking with different simulation codes and simplified SSI models, together with sensitivity and uncertainty analysis, were performed. 

The proper distinction of random spatial soil variability and epistemic uncertainty is one of the keys to reduce epistemic uncertainty in the performance based assessment of plant safety. Another strong point of this research proposal is the implementation of a V&V procedure including benchmarking and sensitivity analysis.