search

ChristoSilvia / mae6230_cfd

Contains code for homework assignments and projects in the class MAE 6230: Computation Fluid Dynamics at Cornell
0 stars 0 forks source link

Homework 2 #3

Open ChristoSilvia opened 7 years ago

ChristoSilvia commented 7 years ago

In keeping with tradition, I'd like to open an issue for homework 2. It's due on the 28th. I've posted the pdf to the homeworks/2 directory for convenience.

How about we schedule a meeting for the 25th on Saturday. How does that sound?

carmaza commented 7 years ago

Hi Chris! Just wondering if you could get a final form for df/dx in P2? I've got an answer but I was wondering if we could simplify it to implement it on the code. Want me to write it down on the tex file?

ChristoSilvia commented 7 years ago

Huh, I forgot to push my changes! I'll get them up now.

carmaza commented 7 years ago

Thank you! We got the same formula. Now I have pushed the first version of our code.

There are comments for every particular task but feel free to ask anything that might seem obscure. Some comments start with "PRINT", which one can uncomment to print some explicit results. Feel free to delete them if you like. Just in case, those variables declared with an " * " before the name are "pointers". Let me know if you haven't come across them before (and/or check this awesome link that helped me understand pointers). The rest of the syntax should be standard.

P3 is essentially the method dfdx(), which receives the stencil and the point to evaluate the derivative. Note that I am using the most general formula to compute dfdx(x_i) (your Eq. 4). Alternatively, one could separate cases and explicitly write Eqns. (8) and (9), but this would add more code lines.

The rest of the code is an skeleton of what is expected for us to do in P5: given a mesh, compute the coeffs at each mesh point, then compare dfdx for some analytical test function. I tried three quick functions and all seem to give consistent results. The error at each point approaches zero as one increases n and/or the number of stencil points. So far, the test does not include the boundary points, which is something we have to talk about in P5.

What's left is the following: run further tests, optimize, answer P4, apply the code to the test function Olivier proposes, plot the error for different n's (and even better if we also play with the stencil, I guess), and say/do something about the boundary points.

Feel free to change/add/delete anything. I'll be free on Monday at any time if you want to meet.

ChristoSilvia commented 7 years ago

Awesome, sounds good.

carmaza commented 7 years ago

Sounds good. Let me know if you have any issues.

On Feb 26, 2017 12:43, "Chris Silvia" notifications@github.com wrote:

Awesome, sounds good. I found a few optimizations in the later part of 2, which I can add to the code, which reduces the amount of work that the function has to do, so I'll try to add those.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ChristoSilvia/mae6230_cfd/issues/3#issuecomment-282572852, or mute the thread https://github.com/notifications/unsubscribe-auth/AYhbWCsV22_YutSN7xAgpZh86oQyfSf0ks5rgbmpgaJpZM4MBPwy .

ChristoSilvia commented 7 years ago

How would you feel about me doing some of this in python? I can translate things over. The graphing feedback is just so much better there.

carmaza commented 7 years ago

Sure, you can do that. If you find some optimization let me know so I can implement it back on the c++ code.

On Feb 26, 2017 15:09, "Chris Silvia" notifications@github.com wrote:

How would you feel about me doing some of this in python? I can translate things over. The graphing feedback is just so much better there.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ChristoSilvia/mae6230_cfd/issues/3#issuecomment-282583235, or mute the thread https://github.com/notifications/unsubscribe-auth/AYhbWKJCLNUZdzF9IgIAf_jnqhZc7lWVks5rgdvdgaJpZM4MBPwy .

ChristoSilvia commented 7 years ago

I added a section doing 5, and showing the error decreasing.

carmaza commented 7 years ago

Excellent! I'll check what else I can add before printing.

carmaza commented 7 years ago

Hi! Those plots in P5 are awesome! Thank you.

I added the titla and a small text describing how we split up the work. Also, I deleted some comments in the C++ code (although I preserved those who explain what the code does) so we can paste it on the PDF. I would include this code just in case Olivier wants to see (and understand by reading the comments) how this method can be implemented.

So what's left would be: you check (and edit if you like) the description of tasks and fit the C++ code in the PDF. Would you upload the final PDF so I can print it before class?

ChristoSilvia commented 7 years ago

Sure I'll have it for you by 1:30

On Feb 28, 2017 12:19 PM, "Cristóbal Armaza" notifications@github.com wrote:

Hi! Those plots in P5 are awesome! Thank you.

I added the titla and a small text describing how we split up the work. Also, I deleted some comments in the C++ code (although I preserved those who explain what the code does) so we can paste it on the PDF. I would include this code just in case Olivier wants to see (and understand by reading the comments) how this method can be implemented.

So what's left would be: you check (and edit if you like) the description of tasks and fit the C++ code in the PDF. Would you upload the final PDF so I can print it before class?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/ChristoSilvia/mae6230_cfd/issues/3#issuecomment-283105082, or mute the thread https://github.com/notifications/unsubscribe-auth/AHppPHynNMpRVuBvemuL5QYV9JD1PsvZks5rhFcngaJpZM4MBPwy .

ChristoSilvia commented 7 years ago

I've updated the homework and I'm satisfied with the changes. Could you look it over, and if you're satisfied, print it?

carmaza commented 7 years ago

Will do. Thank you!

On Feb 28, 2017 14:04, "Chris Silvia" notifications@github.com wrote:

I've updated the homework and I'm satisfied with the changes. Could you look it over, and if you're satisfied, print it?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ChristoSilvia/mae6230_cfd/issues/3#issuecomment-283131956, or mute the thread https://github.com/notifications/unsubscribe-auth/AYhbWJXFX3-x4a2xnlCq0snbPnRInQPmks5rhG_OgaJpZM4MBPwy .

carmaza commented 7 years ago

Can you compile the tex and generate the pdf? I'm getting errors on this machine that I'm trying to print from.

Thank you!