Robust Solution Verification Experiments on Nonuniform Meshes

Conference · Tue Jul 01 00:00:00 EDT 2025

DOI:https://doi.org/10.1115/FEDSM2025-158275· OSTI ID:3012512

Weinmeister, Justin ^[1]; Sanjaya, Devina ^[2]

---

1. ORNL
2. University of Tennessee Knoxville

The activities of verification, validation, and uncertainty quantification (VVUQ) provide a comprehensive means to assess the credibility of computational models. Within VVUQ, solution verification assesses numerical errors and evaluates whether the simulation is sufficiently accurate for its intended applications. As computational modeling gains traction in the development of complex, high-consequence systems, the need for robust solution verification intensifies, particularly because experimental data for these systems are often limited. This work examines improvements in the robustness of Richardson extrapolation (RE), a method commonly used in solution verification to study the discretization error of computational models using a power law. Nonuniform mesh refinement is discussed alongside other pollutants that affect the robustness of the power law model. Maximum likelihood estimation (MLE) is proposed as a robust strategy to address the uncertainty generated by nonuniform mesh refinement. An exploratory computational fluid dynamics (CFD) study of a 2D planar Poiseuille flow is conducted to determine if nonuniform mesh noise can be modeled with this MLE approach for more robust RE.

• Research Organization: Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC05-00OR22725
• OSTI ID: 3012512
• Resource Type: Conference paper/presentation
• Conference Information: ASME 2025 Fluids Engineering Division Summer Meeting - Philadelphia, Pennsylvania, United States of America - 7/27/2025-7/30/2025
• Country of Publication: United States
• Language: English