Interpolation out of table
Feb 21, 2010
03:00 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Notify Moderator
Feb 21, 2010
03:00 AM
Interpolation out of table
Hello
I'm trying to figure out how to interpolate my table cfT from attached mcd-file.
I have there variables b,d and h, which gives me ratio d/b and programmed value depending of h/b- value => lambda.
In given array there is #1 colum for lambda and the #1 row is for d/b-ratio and from this table the value for Cf should be found.
I'm gatefull if someone gives me guidance.
BR Matti
I'm trying to figure out how to interpolate my table cfT from attached mcd-file.
I have there variables b,d and h, which gives me ratio d/b and programmed value depending of h/b- value => lambda.
In given array there is #1 colum for lambda and the #1 row is for d/b-ratio and from this table the value for Cf should be found.
I'm gatefull if someone gives me guidance.
BR Matti
Labels:
- Labels:
-
Other
6 REPLIES 6
Feb 21, 2010
03:00 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Notify Moderator
Feb 21, 2010
03:00 AM
May be this websheet helps you:
http://twtmas.mpei.ac.ru/mas/worksheets/Interp_f_x_y.mcd
Val
http://twt.mpei.ac.ru/ochkov/v_ochkov.htm
http://twtmas.mpei.ac.ru/mas/worksheets/Interp_f_x_y.mcd
Val
http://twt.mpei.ac.ru/ochkov/v_ochkov.htm
Feb 21, 2010
03:00 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Notify Moderator
Feb 25, 2010
03:00 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Notify Moderator
Feb 25, 2010
03:00 AM
Hello and I thank you all who have put your efforts to solve my question.
I must however apologize, not being clear enough first.
The given table is constructed so, that the first uppermost row is value for ratio d/b, and the leftmost column is for value lambda.
These values are not part of data. They are and must be used as parameters to function.
The result must be found from table 1,1 .. 9,4.(Origin 0)
For example when d/b = 7 and lambda is 3, result is 1.38, and if d/b = 1 and lambda = 2, result is interpolated to be 1.33.
So when d/b = 1.5 and lambda=7, I get something like 1.28 by hand calculation.
I interpolated total 8 values, two in 1.5 on rows 3 and 10 and two in 7 on columns 1 and 2.
Then interpolated horizontal, vertical and diagonals values in the middle and finally took mean out of these four.
This programming is what I am going to do. I can easily do it using "basic" and "VB"; "lisp" and "fortran", but this Mathcad is a bit strange still, however I have been reading this forum for quite a long time since joining.
Not even remember any more, when I first wrote here complaining about vlookup-function ;>)
Cheers Matti
I must however apologize, not being clear enough first.
The given table is constructed so, that the first uppermost row is value for ratio d/b, and the leftmost column is for value lambda.
These values are not part of data. They are and must be used as parameters to function.
The result must be found from table 1,1 .. 9,4.(Origin 0)
For example when d/b = 7 and lambda is 3, result is 1.38, and if d/b = 1 and lambda = 2, result is interpolated to be 1.33.
So when d/b = 1.5 and lambda=7, I get something like 1.28 by hand calculation.
I interpolated total 8 values, two in 1.5 on rows 3 and 10 and two in 7 on columns 1 and 2.
Then interpolated horizontal, vertical and diagonals values in the middle and finally took mean out of these four.
This programming is what I am going to do. I can easily do it using "basic" and "VB"; "lisp" and "fortran", but this Mathcad is a bit strange still, however I have been reading this forum for quite a long time since joining.
Not even remember any more, when I first wrote here complaining about vlookup-function ;>)
Cheers Matti
Mar 03, 2010
03:00 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Notify Moderator
Mar 03, 2010
03:00 AM
After considerable effort I have found the reliable regress and cspline methods fail at points between the support points on data like this, and I resorted to a knot cspline prepared on either side of a turning point. I obtained results in agreement with Matti's example.
While the worksheet is user-friendly it would take a lot of preparation if other data was presented. Knot csplines are not common in MsthCad.
This has been a very interesting and challenging post.
Cheers, Theodore
While the worksheet is user-friendly it would take a lot of preparation if other data was presented. Knot csplines are not common in MsthCad.
This has been a very interesting and challenging post.
Cheers, Theodore
Mar 04, 2010
03:00 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Notify Moderator
Mar 04, 2010
03:00 AM
To better illustrate knot splines, data is taken from a Wiki page and the G0 values are changed to the square of the originals. This is a single sample problem on a wide spread of abcissae.
Mar 05, 2010
03:00 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Notify Moderator
Mar 05, 2010
03:00 AM
The SIVUSUH algorithm was more precise but needed much attention from the user. Here the raw data is transferred to a platform more amendable to cspline work and from that another cspline is employed to finish the work as SIVUSUHA.mcd.
Csplines fit at support points, but are notably prone to errors in between. It is best to make sure a cspline fit works by using an intermediate data platform and test.
Csplines fit at support points, but are notably prone to errors in between. It is best to make sure a cspline fit works by using an intermediate data platform and test.
