Join the 0.9 beta
OxiPhys 0.9 is free to use and checked against published results, and it is still changing. Try it on your own problems and report what goes wrong.
The beta is a test version. For everyday work, the stable 0.9.0 is on the install page.
Who the beta is for
Engineers, researchers, teachers and students who can tell when a result looks wrong: anyone able to check a stress, a frequency or a pressure drop against a hand calculation or a reference. There is no sign-up and no cost. Download it, use it, and write in when something is wrong or missing.
What the beta adds to 0.9.0
- Beams and shells, from Python.
- Contact between parts with Coulomb friction, in the app and from Python.
- Transient structural analysis and large deflection in the app.
- SST wall functions and conjugate heat transfer for flow.
- Compressible gas flow and two-phase (volume of fluid) flow, from Python.
- OxiTunnel vortex shedding, validated at Re = 40 and 100 as well as 20.
- A direct solver on every core: a STEP part with 24,484 quadratic tetrahedra solves in about 13 s on two cores instead of about three minutes.
How to take part
Download the beta
Choose the 0.9 beta bundle for your system below. Linux is out now; Windows and macOS follow.
Install it
Unzip it and run its setup program, exactly as the install guide describes for 0.9.0. Installed into the same folder, the beta replaces 0.9.0. To keep both, choose another install folder: the desktop and menu shortcuts then open the one installed last.
Run your own cases
Start from a template or your own model, and compare the result with a hand calculation, a standard or a published value. Validation lists the cases already checked, so you can see where your problem sits.
Report what you find
Email beta@oxiphys.com: a wrong number, a crash, something confusing or something missing. Each report is read.
Update, or go back
New betas are listed under Updates; run the new setup program with the same folder to update. To go back to 0.9.0, run its setup program with the same folder.
Download the 0.9 beta
Free to use under the OxiPhys licence, like 0.9.0. Python, the internet connection during setup and the other requirements are the same as for 0.9.0: see the install page.
What is ready, and what is a first version
Validated parts have been compared with exact solutions or published data. First versions pass their own checks but have not been tried on enough cases to be trusted without a second look.
| Area | State | Checked against |
|---|---|---|
| Statics, modal, buckling, transient, thermal and incompressible CFD | Validated | Exact solutions, observed orders of accuracy and published benchmarks; every case is listed under Validation. |
| Beams and shells (Python) | First version | Beam formulas reproduced exactly; plates within 0.33 % of Timoshenko and Woinowsky-Krieger; Scordelis-Lo roof +0.40 %. No buckling or large deflection of beams and shells yet. |
| Contact between bodies, with friction | First version | Sliding friction equal to μN to 1e-6, Hertz peak pressure within 0.08 %. Partial slip (Cattaneo-Mindlin) is not yet validated, and sliding must stay small. |
| SST wall functions | First version | Bulk velocity in a channel within 1.3 % of DNS on 4 to 12 cells across. The thermal wall function has not been compared with data. |
| Conjugate heat transfer | First version | Laminar cases within 0.07 % of Shah and London. Turbulent conjugate heat transfer is not validated. |
| Compressible flow (Python) | First version | Sod shock tube; Mach 2 wedge pressure within 0.007 % of oblique-shock theory. 2D only, no turbulence model. |
| Two-phase flow, volume of fluid (Python) | First version | Sloshing period within 0.53 % of linear theory on an 80 x 80 mesh; mass conserved. No surface tension, no 3D case. |
| OxiTunnel | Validated | The 2D cylinder at Re = 20, 40 and 100 (Strouhal number within 0.35 % at Re = 100). Not validated in 3D above Re = 20, nor for turbulent bluff bodies. |
The flow solvers still run mostly on one core.
What to put in a report
- The version, from Help, About OxiPhys, and your system (Windows, macOS or Linux, and which version).
- What you did, step by step, what you expected and what happened instead.
- The full text of the Messages panel (F10) when something failed.
- The project file (
.oxiphys) if you can share it. If the model is confidential, say so: a smaller model that shows the same fault is just as useful. - For a result you think is wrong: the value you compared it with, and where that value comes from.