Hacker News Reader

Comment by archermarks | Hacker News Reader

Depends on what you mean by order i suppose. The simplest are algebraic (zero-order) models where we prescribe the anomalous transport as an explicit function of other plasma parameters. More complex ones are coupled systems of hyperbolic PDEs, which are typically first-order in space and time.

If you are referring to the order of accuracy of the simulations, that's typically 1st to 4th order, depending on method used.

View on HN

I answered the GP based on my understanding of "closure" meaning different things for moment models vs turbulence. I think they're referring to energy partition scaling in saturated turbulence.

Yes, thank you. It's a closure problem in the former sense.

> Depends on what you mean by order i suppose.

I mean, in the context of Reynolds-averaged Navier–Stokes equations [1], what sort of turbulence model [2] you are using. But apparently you are not working in this context at all, so maybe my question doesn't make sense.

[1] https://en.wikipedia.org/wiki/Reynolds-averaged_Navier%E2%80...

[2] https://en.wikipedia.org/wiki/Turbulence_modeling

Appreciate the clarification. I am familiar with RANS modeling to a reasonable extent but not with a lot of the modern developments. We're still kind of at the stage of trying to find something that works as well as Prantdl's mixing length model, or maybe a k-epsilon model. Additionally, while our processes are turbulent, the turbulence differs significantly from classical turbulence.

For instance, turbulence in Hall thrusters is likely to be governed by an inverse energy cascade, where small disturbances magnify to larger and larger scales and the dissipation occurs via convection at device-scale instead of by diffusion at kolmogorov scales.

Reply on news.ycombinator.com