Proper Orthogonal Decomposition(POD), Dynamic Mode Decomposition (DMD) Machine Learning Analysis System VFBasis

The first commercial DMD software in the world.
Automatic extraction of important modes from large CFD datasets based on machine learning.

VFBasis software automatically extracts dominant flow structures from large unsteady CFD datasets using POD/DMD analysis based on machine learning.
JAXA developed the basic theory and VINAS has packaged the technology into VFBasis, enabling a deep understanding of the flow regime leading to innovative and better products.

Step 1 : perform proper orthogonal decomposition (POD).
Step 2 : using POD results as input data, perform dynamic mode decomposition (DMD).
Step 3 : automatic selection of important modes from DMD output based on machine learning.

Important mode selection from the DMD output is performed automatically without the need for experienced analysts.
Faster and uses less memory than other POD/DMD tools.
Overcomes the difficulty of visualizing large CFD datasets.
Enhanced product performance through better understanding.
Works with experimental flow data like PIV (Laser measurement).

In the past, it has not been possible to extract useful information from large unsteady CFD datasets but, with VFBasis, the dominant spatial flow structures and time changes ( frequency, growth-rate ) can be quickly and easily extracted.
VFBasis makes it easy to understand flow phenomena, enabling the analyst or designer to develop more advanced and higher performing products in fluid design and fluid control system design.

Supported CFD solvers

ACuSolve, ADINA-CFD, ANSYS-CFX, ANSYS-FLUENT, CONVERGE, CRUNCH-CFD, FrontFlow/blue, FrontFlow/red, HELYX iconCFD, OpenFOAM, scFLOW, SCRYU/Tetra, STAR-CD, STAR-CCM+, STREAM, VECTIS etc
VFBasis read general format(EnSight Gold, PLOT3D)
Direct Reader: OpenFOAM → respond sequentially direct Reader

Operating System Support

Windows 64bit
Linux 64bit