topic Re: Odesolve problem in Mathcad

Odesolve problem

ValeryOchkov — Tue, 01 Mar 2022 18:35:48 GMT

See Prime 6 in attach (one pendulum with a ribbon)

Re: Odesolve problem

Werner_E — Tue, 01 Mar 2022 21:57:29 GMT

@ValeryOchkov wrote:

See Prime 6 in attach (one pendulum with a ribbon)

Looks like you forgot to attach the file!

BTW, how could L(F):=L.0 be a solution? Given the function L seen in your screenshot this seems to be not possible. F=0 is the only value for F where L(F) would be L.0. Or did you mean something else?

Re: Odesolve problem

ValeryOchkov — Wed, 02 Mar 2022 01:54:22 GMT

Sorry ~ see the attach Mathcad 15 and Prime 6

Re: Odesolve problem

Werner_E — Wed, 02 Mar 2022 13:27:56 GMT

Strange and quite sensible. Usually this could mean "check your equations"?

See here the result with three different end-values and notice the magnitude of the force values.

Force seems to be rising and positive in these scenarios but the error thrown ("number with magnitude greater than 10^307) when calculating L(F(t)) would only occur if F goes below -1413 ?????

Re: Odesolve problem

AlanStevens — Wed, 02 Mar 2022 14:51:57 GMT

Does this make more sense?

Re: Odesolve problem

Werner_E — Wed, 02 Mar 2022 18:12:14 GMT

So the initial force F(0) is zero in your approach?

Valery demanded it to be F(0)=-8.496 but so his first equation x^2+y^2=L(F)^2 is not fulfilled .

It looks like L exceeds the valid range from 0.5 to 1.5 during the solving procedure which results in wrong return values of your function f(L).

This is the reason the odesolve block fails when I replace your interpolating function f(L) by the symbolic inverse

as it would return non-real values which odesolve doesn't like at all.

If I limit the range for L

also the force is limited

Not sure which approach fits Valerys problem better.

Re: Odesolve problem

AlanStevens — Wed, 02 Mar 2022 20:53:45 GMT

Hi Werner, you are probably right. I took a fairly arbitrary approach, as it wasn't clear just what physical system Val was modelling!

Re: Odesolve problem

Werner_E — Wed, 02 Mar 2022 22:48:39 GMT

@AlanStevens wrote:

Hi Werner, you are probably right. I took a fairly arbitrary approach, as it wasn't clear just what physical system Val was modelling!

I don't know either but it may have to do with this thread: https://community.ptc.com/t5/PTC-Mathcad/New-problem-with-circles/td-p/780229

Your answer reminded me that we often had problems with an odesolve block when trying to solve for a function for which there was no ODE (like here with F(t)). It was then necessary to pull out that function from the odesolve like you did and this often meant to use quite "creative" (and sometimes awkward) solutions.

I am not sure why Valery had chosen the initial Value for F(0) the way he did.

I tried with the "correct" initial value F(0)=0 but the solve block still failed. "Correct" according to his initial values for x and y and the first equation.

Maybe the problems occur because his function L(F) isn't an appropriate model for an elastic rubber band (infinite force -> still length 1.5 meter). Maybe it just means that the rubber band breaks 😉

Re: Odesolve problem

LucMeekes — Thu, 03 Mar 2022 14:57:25 GMT

Or there is a 1.5 m titanium (or kevlar, or other strong stuff) rope in parallel to the rubber band....

Luc

Re: Odesolve problem

Werner_E — Thu, 03 Mar 2022 15:37:22 GMT

@LucMeekes wrote:

Or there is a 1.5 m titanium (or kevlar, or other strong stuff) rope in parallel to the rubber band....

Luc

🙂

But then the force applied by the rubber band wouldn't change and again be limited, wouldn't it?