ThermoFluidCalc Report

Grid Convergence Index (GCI) Report
2026-09-26 20:44:29

📊 Grid Convergence Index (GCI)

Quantify CFD discretization error numerical uncertainty across 3 grid levels using Roache GCI and ASME V&V 20 standard.

⚡ Fortran 90 Engine Double Precision (IEEE 754) ✓ ISO / ASME Validated
Grid Convergence Index (GCI) Cfd
📊 Solver Telemetry ● ACTIVE
👁️ Views 45
⚡ Solves 37
💾 Downloads 540 📦 Fortran Code 4.4 KB
📅 Released Jun 2026
⏱️ Latency < 1 ms
⚡ TOOLS & REPORTS:
💾 Download Fortran 90

📝 Three-Grid Verification Setup

Presets: ASME 2D Benchmark Vehicle Drag Cd Pipe Pressure Drop Heat Exchanger Nu
⚙️ Mesh Specification
🟢 Grid 1 — Fine Mesh ($h_1$ / $N_1$)
🟡 Grid 2 — Medium Mesh ($h_2$ / $N_2$)
🔴 Grid 3 — Coarse Mesh ($h_3$ / $N_3$)
🛡️ Safety Factor
ASME V&V 20 Formulations:

$$r_{21} = \frac{h_2}{h_1}, \quad r_{32} = \frac{h_3}{h_2}, \quad \epsilon_{21} = \phi_2 - \phi_1$$ $$p = \frac{1}{\ln(r_{21})} \left| \ln\left|\frac{\epsilon_{32}}{\epsilon_{21}}\right| + q(p) \right|$$ $$\phi_{ext}^{21} = \frac{r_{21}^p \phi_1 - \phi_2}{r_{21}^p - 1}$$ $$GCI_{fine}^{21} = \frac{F_s \cdot e_a^{21}}{r_{21}^p - 1}$$

📊 Verification Uncertainty Report

Verification Summary

Richardson Extrapolated Value ($\phi_{ext}^{21}$ at $h \to 0$)
0.276977
Fine Grid $GCI_{fine} = \mathbf{\pm 3.306\%}$  |  Apparent Order $p = \mathbf{1.7}$
Asymptotic Range: EXCELLENT: Grid within asymptotic range (0.95 <= AR <= 1.05). (Ratio $AR = \mathbf{0.9777}$)
Apparent Order ($p$)1.7
Fine $GCI_{21}$3.31 %
Medium $GCI_{32}$6.02 %
Rel Error ($e_a$)2.28 %
Ratio $r_{21}$1.44
Ratio $r_{32}$1.41

📈 Solution Convergence vs Normalized Grid Spacing $h/h_1$

📋 Grid Convergence Analysis Breakdown

ParameterValueDescription
Convergence TypeMonotonic ConvergenceMonotonic / Oscillatory behavior
Apparent Order ($p$)1.7ASME V&V 20 solution
Richardson Extrap ($\phi_{ext}$)0.276977Continuum solution at $h \to 0$
Fine Grid GCI3.306 %Numerical uncertainty on fine mesh
Medium Grid GCI6.024 %Numerical uncertainty on medium mesh
Asymptotic Ratio ($AR$)0.9777Target $AR \approx 1.0$

Fortran Solver Raw Output:

===============================================================
  THERMOFLUIDCALC — ASME V&V 20 GRID CONVERGENCE INDEX (GCI)
===============================================================
Dimension           =  3
Safety Factor Fs    =     1.2500
Fine Size h1        =   4.054801E-03
Fine Solution phi1  =   2.845000E-01
Medium Size h2      =   5.848035E-03
Medium Solution phi2=   2.910000E-01
Coarse Size h3      =   8.220707E-03
Coarse Solution phi3=   3.020000E-01
---------------------------------------------------------------
Ratio r21 (h2/h1)   =     1.4422
Ratio r32 (h3/h2)   =     1.4057
Convergence Type    = Monotonic Convergence
Apparent Order p    =     1.7004
Richardson Extrap   =   2.769765E-01
Approx Error ea21   =   2.284710E-02
Approx Error ea21 % =     2.2847%
Extrap Error e_ext  =   2.716296E-02
Extrap Error e_ext %=     2.7163%
GCI Fine (GCI_21)   =   3.305581E-02
GCI Fine (GCI_21) % =     3.3056%
GCI Medium (GCI_32) =   6.023842E-02
GCI Medium (GCI_32)%=     6.0238%
Asymptotic Ratio AR =     0.9777
---------------------------------------------------------------
Asymptotic Check    = EXCELLENT: Grid within asymptotic range (0.95 <= AR <= 1.05).
===============================================================
--- GCI CONVERGENCE PROFILE ---
Grid_Index   h_norm      phi_val       GCI_pct
   0    0.000000E+00    2.769765E-01    0.000000E+00
   1    1.000000E+00    2.845000E-01    3.305581E+00
   2    1.442250E+00    2.910000E-01    6.023842E+00
   3    2.027401E+00    3.020000E-01    6.023842E+00
--- END GCI PROFILE ---

📘 Calculation Methodology: Grid Convergence Index (GCI) & Richardson Extrapolation

Mathematical Model & Theory

The GCI method (ASME V&V 20 / Roache) calculates numerical discretization error and confidence bounds by comparing results across three geometrically scaled grids:

$$p = \frac{1}{\ln(r_{21})} \left| \ln\left|\frac{f_3 - f_2}{f_2 - f_1}\right| \right|, \quad GCI_{fine} = \frac{1.25 |(f_1 - f_2)/f_1|}{r_{21}^p - 1}$$
$$f_{extrapolated} = f_1 + \frac{f_1 - f_2}{r_{21}^p - 1}$$

Assumptions

  • Monotonic asymptotic grid convergence.
  • Uniform refinement ratio $r \ge 1.3$.

Academic References

  1. Roache, P. J. (1998): Verification and Validation in Computational Science and Engineering.
  2. ASME V&V 20-2009 Standard.

Worked Engineering Example

Problem Statement:
Three grids with $r = 1.414$ yield $C_D = [0.0345, 0.0328, 0.0322]$. Calculate apparent order $p$ and $GCI_{fine}$.

Step-by-step Solution:
1. $p = \ln((0.0345-0.0328)/(0.0328-0.0322)) / \ln(1.414) \approx 3.01$.
2. $GCI_{21} = 1.25 \times (0.0006/0.0322) / (1.414^{3.01} - 1) \approx 1.27\%$.
Final Result:
Fine grid uncertainty is $C_D = 0.0322 \pm \mathbf{1.27\%}$.