Skip to main content
-
-MFra-21-Topaz II
21-Topaz II
April 30, 2018
Solved

Why the solver fails?

  • April 30, 2018
  • 1 reply
  • 1827 views

Hi All,

Why the solver fails if I insert, into the solve block, the constraint highlighted in green ? And how to correct it?

What to do to have the polar plot equal to the complex plot below depicted?

Thank you very much.

Frank

eqdiff system problem.jpg

Best answer by Werner_E

Maybe a problem with MC15 autoscaling

Klick in the plot and at the right you will see the value 2.5x1o^-3. Change it to 0 and see if that helps

 

And yes, the k0 are the same. But one is formatted ad variable and the other as constant. Should not do any harm in MC15.

But the condition you demand k0<u(x)<k1 is a global condition. Can't image that a primarily locally working numerical algorithm would be able to fulfill that.

Its not even possible to demand e.g. u(10^-10)<k1. The error message we get is misleading as so often in a solve block.

1 reply

L
LucMeekes23-Emerald IV
23-Emerald IV
April 30, 2018

You k0 in 'u(0)=k0' is a different k0 when you define it at the top: 'k0 := 2.5 * 10^-3'.

 

And:

LM_20180430_Odesolve.png

Note that the results of Odesolve are ONLY valid between the defined endpoints, 0 and x.end in your case.

 

And your polar plot to correlate with the 'complex plot' is:

LM_20180430_Odesolve1.png

 The argument(s) to the left descibe the magnitude, those to the bottom describe the argument (angle).

 

Success!
Luc

 

 

    -
    -MFra-21-Topaz IIAuthor
    21-Topaz II
    April 30, 2018

    Thanks Luc, but once opened the attached file, with Mathcad 15, the polar plot is this (I've only modified the trace thickness):

    Polar plot.jpg

    that is, there is no change compared to the old one.

    So what to do?

    Furthermore I believe, insted, that k0 is the same!

    W
    Werner_E25-Diamond IAnswer
    25-Diamond I
    April 30, 2018

    Maybe a problem with MC15 autoscaling

    Klick in the plot and at the right you will see the value 2.5x1o^-3. Change it to 0 and see if that helps

     

    And yes, the k0 are the same. But one is formatted ad variable and the other as constant. Should not do any harm in MC15.

    But the condition you demand k0<u(x)<k1 is a global condition. Can't image that a primarily locally working numerical algorithm would be able to fulfill that.

    Its not even possible to demand e.g. u(10^-10)<k1. The error message we get is misleading as so often in a solve block.

    Terms & ConditionsCookie settingsAccessibility statement