In geotechnical and structural engineering, predictive numerical models are key tools used to evaluate risk, optimize design, and anticipate ground-structure interaction behavior. However, the reliability of these models strongly depends on the quality and accuracy of the input parameters. Given the inherent uncertainty in soil conditions, performing parameter variation and sensitivity analysis becomes essential to support design and construction decisions.
This article explores the strategic implementation of sensitivity analysis using high-performance tools like DAARWIN, SAALG Geomechanics’ cloud-based platform that integrates finite element models with real-time field monitoring data.
What Is Sensitivity Analysis and Why Does It Matter?
Sensitivity analysis quantifies how uncertainties in input parameters (e.g., stiffness, cohesion, or permeability) influence model outputs (e.g., displacements, stresses, or pore pressures). In the context of finite element analysis (FEA), this approach addresses key questions such as: Which parameters actually drive the system's response? How does soil variability impact the stability of our solution?
Main approaches:
Local sensitivity analysis: Evaluates the effect of small changes in a single parameter while holding others constant. Useful for linear systems or well-characterized conditions.
Global sensitivity analysis: Simultaneously varies multiple parameters over their full range of values, ideal for nonlinear systems and those with strong parameter interactions.
Parameter Variation in Finite Element Models
Geotechnical parameters are often derived from limited site investigations and display high spatial variability. Systematically varying these parameters allows engineers to:
Identify dominant parameters that control structural performance.
Evaluate model robustness under worst-case scenarios.
Optimize field and laboratory testing by focusing efforts on the most influential variables.
Avoid overdesign, cutting unnecessary costs while maintaining safety margins.
DAARWIN: Scalable Sensitivity Analysis for FEM
DAARWIN is a cloud-based platform developed by SAALG Geomechanics that automates and scales sensitivity and parameter variation workflows in FEM software such as PLAXIS or MIDAS.
Key features:
Full compatibility with PLAXIS, including advanced soil constitutive models.
Flexible parameter selection, supporting both independent and correlated variable sets.
Cloud-based parallel computing, enabling hundreds of simulations to run simultaneously.
Advanced visualization, providing sensitivity indices and result plots that are intuitive and technically robust.
Practical Application in Complex Projects
In urban tunneling projects with mechanized excavation under heterogeneous ground conditions, DAARWIN has been used to reveal that over 70% of predicted surface settlement was attributable to the stiffness variability of a single clay stratum. This insight informed targeted field testing and reduced uncertainty in predicting TBM face behavior, leading to better excavation control and cost-effective support strategies.
Sensitivity analysis and parameter variation are no longer optional in advanced geotechnical engineering—they are fundamental tools for managing uncertainty. Their effective implementation, supported by platforms like DAARWIN, enables engineers to shift from static assumptions to data-informed, adaptive design, ultimately leading to safer, more efficient, and technically sound infrastructure projects.
