Validation and Verification

What problem does it solve?

This Skill addresses the critical need to ensure that scientific simulations are both correctly implemented (verification) and accurately represent the physical phenomena they model (validation), thereby increasing confidence in simulation results.

Core Features & Use Cases

• Code Verification: Implements unit testing for numerical code, method of manufactured solutions (MMS), and checks for conservation properties.
• Solution Verification: Utilizes grid convergence index (GCI) and Richardson extrapolation to assess numerical accuracy.
• Model Validation: Compares simulation outputs against experimental data using metrics like RMSE, bias, correlation, and skill scores, including statistical hypothesis testing.
• Uncertainty Quantification (UQ): Employs Monte Carlo methods and sensitivity analysis (Sobol indices) to quantify input and output uncertainties.
• Use Case: A research team developing a climate model can use this Skill to rigorously verify their code, validate its predictions against historical weather data, and quantify the uncertainty in future climate projections.

Quick Start

Use the validation-verification skill to perform a grid convergence study on the simulation results provided in 'simulation_data.csv' against experimental data in 'experimental_data.csv'.