HAS A SHAPE.

00THE QUESTION

A CFD picture is easy. A defensible pressure signature is harder.

How can student-scale computation compare low-boom-style aircraft conditions without pretending it has certified ground noise? The answer was not a more dramatic render. It was a workflow built to preserve, sample, and challenge the waveform itself.

ROLE: Computational researcher · X-59 team
FOCUS: CFD workflow, pressure signatures, validation
TOOLS: OpenFOAM · ParaView · Python · HPC
STATUS: Baseline complete · AoA campaign continuing

6,124,664MESH CELLS

0.650 sFINAL SIMULATION TIME

2,401 / 2,401VALID R/L = 3 SAMPLES

0.999809CONVERGENCE CORRELATION

C608 mesh architecture with nine ray-following refinement corridors — FIG. 01 / MESH ARCHITECTURE
Nine characteristic corridors preserve wave content from the airframe to R/L = 3.
NOT UNIFORM REFINEMENT

01THE METHOD

Build the test before trusting the result.

Define the reference

A C608 half-model, Mach 1.4, zero additional angle of attack, and the NASA Sonic Boom Prediction Workshop probe convention created a comparison that could be repeated rather than hand-picked.

C608 / M1.4 / SBPW3

Resolve the wave

Nine continuous, ray-following refinement corridors carried pressure information from the airframe toward R/L = 3. The mesh spent resolution along the path of the signal, not everywhere at once.

1,487—1,513 SEGMENTS / RAY

Extract the signature

Fixed off-body probes converted a dramatic field image into normalized pressure traces, feature locations, peak-to-peak amplitude, total variation, and radial attenuation.

ΔP / P∞ / 6 RADII

Make it survive scrutiny

A processor-boundary sampling artifact was identified, rejected, and replaced with isolated reconstructed-field sampling. The final result had to pass two consecutive time-stability checks before it counted.

0.600 → 0.625 → 0.650 s

Near-field pressure-gradient visualization around the C608 aircraft

The field is the spectacle. The line beneath it is the evidence.

R/L = 0.1 · 0.25 · 0.5 · 1 · 2 · 3

02THE RESULT

0.0115648PEAK-TO-PEAK ΔP / P∞ AT R/L = 3

The refined workflow carried eighteen detectable features to the official probe.

Pressure signatures sampled at six radial distances from the C608 aircraft — FIG. 02 / RADIAL EVOLUTION
The waveform attenuates with distance while retaining its overall structure.

03THE GATE

A frame at an arbitrary time is not a result.

The accepted trace had to remain stable across consecutive reconstructed-field samples. At R/L = 3, the final 0.625-to-0.650-second comparison reached 0.999809 correlation with 0.549% peak-to-peak drift.

NRMSE: 0.433%
MAX DIFFERENCE: 1.542%
BEST LAG: −1 sample

Nearly overlapping pressure signatures at 0.6, 0.625, and 0.65 seconds — FIG. 03 / TEMPORAL CONVERGENCE
Three nearly overlapping traces; two consecutive comparisons passed.

04THE HONEST COMPARISON

Moderate agreement. Visible limits.

Against the published, unshifted LAVA direct-CFD trace, the result captures the gross waveform but not every sharp feature. That is progress—not validation.

Unshifted comparison between the OpenFOAM result and published NASA LAVA direct CFD — FIG. 04 / OPENFOAM × LAVA
The main comparison remains unshifted; no alignment is used to inflate agreement.

0.81577CORRELATION

10.943%NORMALIZED RMSE

73.6%AMPLITUDE RATIO

WHAT THIS IS

A convergence-checked, airframe-only pressure-signature baseline.

Reproducible off-body sampling
Documented mesh and solver provenance
A serious anchor for comparative AoA work

WHAT THIS IS NOT

A powered-aircraft, wall-resolved, ground-boom validation.

No propulsion or plume modeling
No boundary-layer-resolved prism stack
No certified loudness or community-noise claim

NEXT / AOA CAMPAIGN

Not a sonic-boom prediction.

A better way to ask the next question.

← BACK TOALL WORK