Optimizing mesh quality and Y+ layer formation on an FSAE car

Here is the dropbox link to the folder. I think the run with 88,725 had the error but im not sure.

Problem is, now that full solution dataset when you download it will always contain time 0 and 1000, so even if you continue that run, the useless 1000 iteration data will always be thereā€¦
Personally, I would start a NEW sim run with 0 start iteration and 5000 end iteration and watch it until 1% ORSI with Moving Average iterations of about 150, which is about 5x the 30 iteration period of your oscillations (even try 250 when it is running)ā€¦ As the value that it is oscillating around gets more stable, you can see how increasing the the MA iterations, filters the oscillations out, giving you a better guess at the value the points are oscillating aroundā€¦
Sorry, when we talked about the way to figure out how many sim iterations to run to for ā€˜quick future runsā€™, I was assuming you were talking about my sim run that already had been run to 3000ā€¦ ONLY AFTER A 1% FULL RUN CAN YOU DECIDE WHAT ITERATION YOU WANT TO USE FOR A FUTURE ā€˜QUICK RUNā€™ā€¦
If I actually realized you were asking about the full car sim run I would have said to use 5000 end iteration and stop it manually when ORSI 1% is reached ā€¦
PS: To save space after solution set down loads, I always delete the useless Time 0 folder which saves a LOT of hard disk ā€¦space

Those are just the 1kb .txt summary filesā€¦ I need the 30mb .csv output file from ParaViewā€¦
I see that you have not setup your windows file explorer to ā€˜show extensionsā€™, it might help you if you had file extensions visible, rather that relying on Windows icon guess for which program you want to use for when you click on a file, I hate it when Windows gets more Macish :wink:

Dont worry i had run the full car simulation before we talked about when to stop / what write interval to use. I had done that sim mainly to see Y+ values so i could know if a full car with 1 layer was landing in the 30-300 range. I now know it is .

So i understand the Moving average input correctly. Is this used to calculate the average Cl or Cd value as its moving along the iterations? So a MA of my current value of 50 would show one full cycle of raw data in yellow? Changing this to 150 or 250 will flatten out the wave values so i can make a better guess of the average value between the cycles correct?

This is very useful thanks!!

I included the 30mb file, it exported as an excel icon for me

Yes, yes, yesā€¦

My latest chart here, has column titles that explain what is happening with ORSI better than the program output screenā€¦ The ORSI value is in fact a Moving Average value at the ā€˜Iteration of interestā€™ and the % is the range of MA values in the previous number of iterations in the ORSI window :wink:

Is it possible that you had that 30mb file in a folder where yPlusHistogram was in the ā€˜Monitoring is ONā€™ green mode?

I think the error comes when my monitoring code opens a yPlusArea.csv file instead of a 2D plot csv file, I can easily plug that error but I did not forsee a person doing thatā€¦

mmm that i cannot say for sure. I have one copy of the histogram.exe in the folder with the case-foam data and another in the downloads directory. Im pretty sure i didnt have it in monitor mode but i can try to replicate this again next time i download a csv file.

OK, I will plug that error anyway in v206 :slight_smile:

PS: I can not seem to duplicate your error on that file when I try to get a yPlusHistogram with it, I get no errorsā€¦ So that bug will float around for a while until it happens to me while I am in debug mode :frowning:

OK so ā€¦ Here is the full simulation run. You have taken the last 500 iterations, which encapsulates one full cycle of the raw yellow data, varying 20% over the 500 iterations, which makes it difficult to estimate the ā€œtrueā€ average of the cycle.

This next picture is showing the 500 iteration section from the entire simulation shown above. It is still fluctuating at 20% but by changing the MA to 2, it takes the average of 2 iterations within 10% fluctuation (because there are no other areas less then 10% fluctuation available) over this 500 iteration range, giving you the closest average of the data set at 0.424 Cd and -0.2367 Cl.

Correct?

Yes I didnt know you needed the 30mb file from the exact moment, I thought it was the other txt file. The one i sent was from the final data set where i got no error. The one with the error is most likely when it was a 25mb file. Next time ill remember to send one with the error.

1 Like

No link or pic to look at yet here :wink:

Huh? i copied your pictures to see if im understanding this right, is it not there?

Is it just collapesed?

image

No, pics or collapsed, I have seen this before from you (like the first time you put green arrows at separation point on an image, it wasnā€™t there and still is not I think) , not sure what problem isā€¦

here is the post again
OK so ā€¦ Here is the full simulation run. You have taken the last 500 iterations, which encapsulates one full cycle of the raw yellow data, varying 20% over the 500 iterations, which makes it difficult to estimate the ā€œtrueā€ average of the cycle.

This next picture is showing the 500 iteration section from the entire simulation shown above. It is still fluctuating at 20% but by changing the MA to 2, it takes the average of 2 iterations within 10% fluctuation (because there are no other areas less then 10% fluctuation available) over this 500 iteration range, giving you the closest average of the data set at 0.424 Cd and -0.2367 Cl.

Correct?

I think what is happening is the file isnt working because i edit/add arrows in a google docs drawing, this is making a different file that doesnt open here. Has this been happening the whole time?

Yep whole time, but I just assumed you forgot, sorry, never mentioned before, were you seeing the images and I was not?

EDIT: I do see the images now in your above repost :slight_smile:

yea ā€¦ oh man i had a lot too

EDIT - maybe i should go back and redo all the pictures so it is more helpful for anyone reading

Is that a yea that you saw your images when I did not, now that is really weirdā€¦

I have started using Greenshot for most of my screen captures and note making on them, it lets me capture the exact same screen area repeatedly which is how I got those pixel perfect captures from ParaView windows :wink:

I would do that but I am pretty anal about keeping things organized for those that followā€¦

Not sure we have many people following us right now :wink:

Yes, i see all my images, and i also see them again when you respond to my posts and ā€œquoteā€ things.

I am using the snipping tool to take screenshots. I use google docs so i can have everything online and i can share my documentation easily within the FSAE team.

I donā€™t think you are quite there yet in your understanding of ORSI by Moving Average ā€¦

First, consider to stop using the word average by itself when referring to anything about ORSI by Moving Averageā€¦

Two things happen here, first the data is filtered by taking EACH iteration and calculating a Moving Average for THAT iteration by summing up the results values (the Yellow line) for the previous X iterations (my programs ā€˜# of Iterations in moving average filterā€™) and then dividing that total by X, this is the Moving Average for that iterationā€¦ The ā€˜Black and other non-yellow colorsā€™ line is then a Moving Average lineā€¦

Second, we then look at an ā€˜MA value at any iteration of interestā€™ or ORSI value (the white dot) with the mouse or by clicking on the value in the upper text summary. At this ā€˜Iteration of interestā€™ we then look back Y iterations (my programs ā€˜# of Iterations for ORSI % calculationsā€™ value)ā€¦ Over those Y iterations we look for the Max and Min MA value in that regionā€¦ Then the ORSI % is taken as 100*(Max-Min)/ORSI ā€¦ The horizontal and vertical ā€˜MA graph scaling linesā€™ are set to bracket the ORSI window when a White dot is on the screen (at ORSI iteration, ORSI-X iteration, ORSI window Max Parameter value, and ORSI window Min Parameter value) ā€¦

.
.
.
.

That is not quite a correct description of the 2nd image you have put notations onā€¦
More correct, if you look below at the Blue circled numbers, the extent of iterations shown in the plot lines are from 500 to 2718.
The ā€˜500 section iteration sectionā€™ you mention is for the data at the White dot iteration (currently at 2718i) and in that 500 iteration window the ORSI % is 18%ā€¦ (Cyan circled data)
.


.
.
.

Your ā€˜dashed red pointersā€™ point to a section of iterations where, if you moved the white circle with a mouse to any red iteration, you would have between 10 and 5.7 ORSI %, like this:


.
.
.
.
EDIT: A further note on advanced ORSI by MA plot navigationā€¦ If you move your mouse off the left or right side of the plot (or bump the sides of the plot), the White Dot disappears and the now black ā€˜MA graph scaling linesā€™ bracket the data extent of the current section of MA line that is on the screenā€¦