extrapolate

[smoothdurban]

Hello,

I am trying to find a method to extend a or extrapolate a set of cuves I have in order to display a curves future values.  The curves I am trying to extraplote will be linear.

In addition to this request I was wondering if there is a way to determine the XY intercept(s) between curves.  I have DIAdem 12 advanced edition. 

[Walter_Rick]

Hello Tim Z,

One way is to use the linear regression and a new channel with the interoplation/extrapolation points.The example below requires the DIAdem example data-set.

Call ChnLinGen("/LinearGenerated1",0,50,25,"")
Call ChnRegrYCalc("","[4]/Temperature_3","[4]/LinearGenerated1","/RegressionY","linear")

Greetings

Walter

[smoothdurban]

Hey Walter thanks for the reply unfortunately I do not have the example "data-set" to try your script out on.

If I am reading this right your first command generates a data channel with equalization values. While the other approximates a signal?

I am a little unclear on how to use these commands.  I will attach some data to help clarify.

For instance I have two channels which are plotted WRT each other named CopyYTorque & CopYAngle. at the beginning and end of each plot I grabbed a section of data one in Group Best Fit A and the other in Group Best Fit B,  I then found the line of best fit between BestFitATorque and BestFitAAngle as well as BestFitBTorque and BestFitBAngle.  The result give me the regression of each best fit section.

When I plot these lines they are a little short and I would like to extrapolate these lines as shown in my attached plot.  If you look at the last row of data you'll notice some extrapolated data that I physically input which gives me a desired extrapolation which I calculated by finding the y = mx+b equation of each regression line.

I am looking for a way to automate that extrapolation via a function or a script.

 FOr some reason I cannot add attachments fell free to email for the data and I will send it over at tim.zoldak@magnapowertrain.com

Thanks

[Walter_Rick]

Hey Tim Z,

The example data-set is the file which is install with DIAdem. The file name is "Example.TDM" or "Example_Data.TDM". It contains channel group "Temperatures" with three channel. I took the last channel "Temperature_3" of these group. So I assume you have that data-set and channel too.

You are talking about a attached plot. I don't find this plot.

Greetings

Walter

[smoothdurban]

Ok Thanks.

I loaded the example file and you script creates a channel named LinearGenerated1 in the 4th group but when it tries to run the second line of code it throws thsi error.

Error in <NoName(1).VBS> (Line: 9, Column: 1):

Cannot interpret channel name in ""[4]/LinearGenerated1"" command.

Also if you read my previous post I was unable to attached any files so I left me email address.  If you want to have a lokk send me a an email and I'll forward you the data.

[smoothdurban]

Unfortunately I cannot attach the view file

[Walter_Rick]

I had a look into your data. I assume that you already use the linear regression function of DIAdem. The only thing seems to me that you just need to define a channel for the intersections ("Create evaluation points").

(see picture in attachment)

By the way you are right, the script I posted may not work correct in case of a misspelling problem of the channel name. Please check the channel names you are using the English version the name is "Temperature_3" in the German version it is "Temperatur_3"

Greetings

Walter

[smoothdurban]

Hey Walter thanks for the reply.

I am a little uncertain on what function setting the channel performs when calculating regression.

 I am plotting  [2]/CopyYTorque vs [2]CopyAngle and plotting [4]/BestFitRegressionX vs [4]/BestFitRegressionY

My desired result is to extend the regression line as illustrated.

[smoothdurban]

I got it now.

The generated X axis is what the new extraopolted curve will reference.

No to work on how to find an intersection point between two curves.