Turbulence is complex. Analysis shouldn't be.
uPrime is a standalone desktop application for post-processing and turbulence analysis of velocity field data from PIV and CFD. Reads Tecplot ASCII .dat (LaVision DaVis) and MATLAB .mat files. No scripting. No setup. Just analysis.
NEW IN v0.8
- Probability Analysis module: a new window with four tabs covering velocity probability, flow direction, binary space-time maps, and quadrant analysis.
- PDF tab: probability density of
u,v,w,|V|or their fluctuations, sampled at a point, along a line, inside a dragged ROI, or over the whole field, with an optional matched Gaussian overlay and a mean/std/skewness/kurtosis readout. - Flow Direction tab: per-point forward and reverse flow probability maps, with an explicit FFP
P(q > 0)versus RFPP(q < 0)convention carried through the title, colorbar and every export, plus optional Simpson detachment-state contours. - Binary Space-Time tab: the classic space-versus-time map coloured by flow direction along a horizontal or vertical line, with band averaging and normalised axes. Time-resolved data only.
- Quadrant tab: Lu & Willmarth quadrant decomposition of a fluctuation pair, with time- and stress-fraction maps, a joint PDF with the hole hyperbola, Q2/Q4 stress maps, and a hole-size sweep.
- ROI restricts computation: a dragged ROI limits what is read and computed rather than only what is displayed, so a small ROI on a large dataset is correspondingly cheaper. Regions outside the ROI render in a distinct grey.
- Compare Cases viewer (v0.7): standalone window comparing up to six datasets at once, with a 1D overlay plus value table and a 2D Field Compare tiled contour view sharing one color scale.
- DaVis native input (v0.7): read LaVision DaVis
.vc7and.vecfiles directly when the optional lvpyio package is installed separately (pip install lvpyio). Not bundled in the Windows executable; available when running uPrime from source.
Key Features
📊 Convergence verification
Cumulative first, second, and third central moments at any picked point, with automatic N* detection. Know whether your statistics are trustworthy before you publish them.
📂 Two input formats
Reads Tecplot ASCII .dat from DaVis (one file per snapshot) and MATLAB .mat files (full time series in a single file, classic and v7.3 HDF5). Auto-detects variable names.
⚡ Non-blocking computation
All heavy analysis runs in background threads. The window stays fully responsive during POD, DMD, TKE budget, and vortex identification.
💾 Large dataset support
Datasets exceeding 4 GB are automatically memory-mapped to disk. Handle 10,000+ snapshots without running out of RAM.
🎭 Non-destructive masking
Draw masks over walls, shadows, and reflections. Raw data is never modified. Toggle masks on and off instantly.
🔍 Unit auto-detection
Reads mm/m/s units directly from Tecplot .dat file headers and MATLAB variable shapes. No manual configuration needed.
📄 Publication-ready export
PNG (300 DPI), PDF, and SVG with editable text. Compatible with Adobe Illustrator and Inkscape.
🖥️ Standalone executable
Windows .exe with no Python or dependencies required. macOS and Linux users run from source in three commands.
📐 Transform & Alignment
Correct camera misalignment, shift the coordinate origin, and mirror datasets. Built for experimental data where even a 1° tilt can corrupt turbulence statistics.
🧠 Full turbulence pipeline
Reynolds stresses, TKE budget, spectra, POD, DMD, correlations, and vortex identification, all in one environment. No MATLAB or Python required.
📏 Integral scales & correlation
Compute spatial and temporal correlations with automatic integral length and time scale extraction using multiple robust methods.
🔀 Multi-case comparison
Import previous results as CSV and overlay them on top of the current dataset. Per-case styling, manage cases, combined export.
Input Formats
Tecplot ASCII .dat (LaVision DaVis export):
one file per snapshot, select multiple files to load a time series.
Variable names and units auto-detected from the header. Supports 2D2C and 2D3C (stereo).
MATLAB .mat (classic v5/v7 and v7.3 HDF5):
a single file containing the entire dataset. Velocity arrays in shape (Ny, Nx, N_snap);
coordinates as 1D vectors or 2D meshgrid (auto-transposed if needed). Variable confirmation dialog
reads metadata first, so even a 10+ GB file opens the dialog within seconds.
LaVision DaVis .vc7 / .vec (optional):
read DaVis vector files natively when the lvpyio package is installed
separately (pip install lvpyio). lvpyio is not bundled in the Windows
executable, so this input is available when running uPrime from source; once installed,
.vc7/.vec files load through the same workflow as .dat and .mat.