Students Kriging

Jose,

Here are some links that you should take a look at. Discuss any findings with Mateo and Juan. Your first task is to investigate the different approaches to extrapolation that are possible.
The main contenders are:

  • Kriging (Several links to programs to download on the ITC site)
  • ABOS (Program SurGe that can be downloaded from the site)
  • Response Surface (investigate this form of extrapolation. Link to Isotech below, there may be other companies)
  • Others forms of extrapolation that may be useful

LINKS:

http://mujweb.cz/www/SurGe/surgemain.htm
http://www.itc.nl/personal/rossiter/teach/lecnotes.html
http://www.iosotech.com/rs.htm

Please write a report on your investigations so far and also send me a report on progress at the end of each week. You can add a reports page and attach them to this page together with an explanatory title etc.
We will all have a conference call at 10:00 hrs each day on Skype.
I would like you to make a start by analysing the LMP1 aero data that will come from John later this week.

Do Lists (Kriging)

Reports (Kriging)

Attachments (All)Must be under 1MB (pdf preferred)


Scratch Pad (for quick notes and diagrams)


Jose,
Please get yourself An Epsilon email address as soon as possible. I have some LMP1 data that we should look at.


Jose,
Thank you for your fourth report and Excel program. I will study them both in the next few days. You have done some excellent work considering the time you have had available. I will let you have my comments by the start of next week. Look in the "Do Lists" for an update. I am working on the new Michelin tyre data, so unfortunately I cannot devote the time that this project deserves at the moment. I will send you some Le Mans data to look at when you have a secure epsilon email address. As for the slopes at the extremities, I should not worry about them too much. We should just estimate these as equal to the slopes of the immediate neighbours for the time being. These areas are so far away from the operational envelope of the car that this estimation should not have a large impact on the results. Probably more important is an extrapolation into the high ride hight zones, especially under braking when the rear ride height can go outside of the measured range.

Regards, Karl

P.S. Please read my reply to your editing problem. The people at Wetpaint can fix most things. You may have noticed that we are now classed as an "Educational Wiki" and as such the site is now ad-free. Do you like it better this way, or did you find the ads useful?


Jose,
An excellent third report. you are right on track with your assessment of best fit with KED. this is very close to the reality of where the car actually runs, as shown in the diagram below.

Students Kriging - White-Smoke

We must not forget that the aerodynamic maps are a very special case. We should be looking at other applications of Kriging for say tyre data which will have a completely different type of variogram.
For the moment keep all the options open. Next week I should have some LMP1 data for you to analyse. For now stay with the F.R. V6 data so that we can make comparisons. Can you send me a copy of the Excel program you have so far, after the addition of KED? Use the White-Smoke email address for this. I now have some new Michelin tyre data. Next week I will start the optimisation of the weight and aerodynamic distribution. It would be good if we could have some extrapolated maps for this. Watch this space for more information as it becomes available. Is your email address secure for the transfer of this data? Based on the particular spread of data points from the data acquisition. What would be the optimum way to collect the Wind tunnel data? At the moment we have an even grid of ride heights. Would a different type of collection method be more suitable? If so what kind? Always have the "Big Picture" in the back of your mind. I am sure that this technology has many applications. Let me have your thoughts.

Regards, Karl

P.S. Why not add a page for each of your reports under the main heading of Reports (Kriging)? This would make it easier for me to add my comments at the end of each of your reports. In this way it will keep the "Scratch Pad" free for just quick notes. Can you also put a photo or logo on your own profile page so that it is clear who is adding or altering pages.Keep up the good work!




*** SORRY NO INTERNET AT MY HOME FOR SEVERAL DAYS! I AM BACK ON-LINE NOW. ***

Jose,
You have written some very good reports! Keep up the good work. I know that your time to work on this project is limited at the moment, just do what you can. I think that it is a good idea to use Excel as the backbone for the kriging as it has become the race car engineers standard tool for calculations. I will try to get the LMP1 data to you as soon as possible. In the meantime investigate the alternative methods. You can also use the Champ car data that JuanCar and Mateo are using at the moment. this is very good data, and we will all be looking at the same data for validation. Ask them to give you a copy.

Regards, Karl

P.S. you will notice that I have added pages and links. It just keeps things tidy. Please rearrange things as you would like.