[abstruse1]

Greetings. Has anyone worked with the aero drag on a spinning disk with a perpendicular spindle?

I’m trying to minimize the drag on a 3/4” thick x 7” diameter steel disk with a 1” diameter spindle that protrudes 4” from the top surface and 1” from the bottom surface and is concentric with the disk. Disk is spinning like a top at ~17,000 RPM in a 29” vacuum.

So far, my disks have 3/4” thick circumferences, but I’ve been wondering if the edge was cut to a chisel shape it would have a lower drag. Also, there may be other things I could do to lower the drag. So far, all surfaces are smooth.

Any suggestions would be very much appreciated. Thanx.

[aerosayan]

Are you planning to do a full transient simulation with moving mesh for this?
I think you can study only half of the disk, with a flow that is tangential to the disk's outer radius. The grid used could be half of a structured O grid. After analyzing the performance impact of the edge shape, you can try out the other optimizations that you want. Hard disks do run inside partial vacuum with some non-reactive gases to prevent drag.


All fascinating stuff, but I'm afraid this is only how much I know about drag reduction in spinning disks.

[abstruse1]

Sayan, thanx for your suggestions. I have no capability to do any kind of mathematical analyses of this situation. I just have the ability to fabricate these kinds of things and enough awareness of aerodynamics to be pretty sure that the answers to my questions probably is, “It depends.”

I’m hoping to find someone who has enough experience or modeling knowledge to point out what may be obvious changes I could make. Without that, all I can do is follow my nose, try various changes, and keep good notes.

What are the gases you refer to? Do you simply mean gases that are chemically inert? I would think that the density of a gas would be significant as to drag, e.g. helium being the best (aside from hydrogen), but there may be other considerations.

[aerosayan]

Yes I was talking about Helium.
Unfortunately I don't know much about this niche of CFD application. But I can share some information on how I go about doing my research on topics I don't understand much.


To answer the question, I'm currently assuming that you're studying in a college/university. My apologies if your qualifications are higher and I assumed incorrectly.


CFD is vast. Nobody knows majority of the things in it--well maybe except Dr. Roe.
So, I have developed a method to help me understand things more clearly. Since I'm interested in solver-dev and grid-generation-dev for hypersonic flows, my methods might not fully translate well to your task (modelling and solving CFD problems using commercial software). But some of them can be helpful.


i. Select a good reference paper : Try to select a good reference paper or PhD dissertation that is directly related to your CFD problem. Even if your research topic is not about hard-disks, try to see if there's any paper that is explained very well, and you like. This paper will serve as your starting point. This paper will be your Holy Bible.



ii. Try to replicate the research done in the reference paper : Since you currently told that you don't have much experience in solving this disk rotation problem, it would be very helpful for you if you try to replicate the research done in the reference paper. From my own limited experience in solver-dev, I can tell you that there is a lot of nuance involved in understanding the research. Understanding those nuances is basically the experience that sets the experts apart from the beginners.


For example : No expert will solve a big linear system using x = inverse(A)*b. Even though papers write the equation as x = inverse(A)*b.


In your case, try to replicate the research of a spinning hard disks. There are a lot of things involved (shockwaves, resonance frequency, material deformation due to temperature change. material deformation due to hysteresis curve etc.) that you might not be aware of for your problem, but they will be documented well in your reference paper/s.


iii. Try to get better at debugging your current problem : While you try to replicate the research, you will see that for some cases your solution doesn't converge, the solution isn't good, it takes too much time, etc. If you have a reference paper, you will be able to validate your results and this will allow you to be well prepared for your own research.


Since you're planning to do research on a topic that you're not experienced in, you need the practice I mentioned above. This way, you will be able to debug the issues you observe for your research, and you will have more intuition/experience to validate your research.


It takes some homework on your side, I understand. That can be time consuming. But it's necessary.


Imagine if you're a heart surgeon, but don't know how to patch up an arterial bleeding.


Similarly here, your research problem has many variants. There will be many issues with convergence, accuracy, validity. The best thing you can do, is to gain some hands on experience while replicating your reference paper, such that you're well prepared for solving the research problem you want.


I know it will be time consuming. But it's probably the best way to do it. Although I would be happy to know if there's some other way too.


Best regards

Sayan