Message Boards Message Boards

0
|
7359 Views
|
2 Replies
|
2 Total Likes
View groups...
Share
Share this post:

Use FindRoot for the following function?

Posted 6 years ago

In the example below, FindRoot doesn't work with the provided function, calcTresAtTime[mCpRes?NumericQ, mFracClr?NumericQ, timeTarget?NumericQ]. However, no problems are observed when calling the function by itself or from Plot. The documentation mentions that FindRoot first localizes all of the variables, then evaluates f with the variables being symbolic. The examples in the documentation show how to turn this off, by using ?NumericQ.

eq01ResHB = 
MCpRes ures'[t] == 
mCpPump (uclr[t] - ures[t]) + UAambRes (uamb - ures[t]) + 
UAbrg (ubrg - ures[t]);
eq02ClrHB = 
MCpClr uclr'[t] == 
mCpPump (ures[t] - uclr[t]) + UAambClr (uamb - uclr[t]) + 
UAclr (ucw - uclr[t] );
ic = {ures[0] ==  ures0, uclr[0] ==  uclr0};
eqSet = Join[{eq01ResHB, eq02ClrHB}, ic];
vars = {ures, uclr};
KuambRes = 0.025 ;
Kuabrg = 0.236;
KuambClr = 0.0024;
Kuaclr = 0.1;
calcTresAtTime[mCpRes_?NumericQ, mFracClr_?NumericQ, 
timeTarget_?NumericQ] := Module[{TresSolLocal, TclrSolLocal},
parmsRes = {MCpRes -> mCpRes  , UAambRes -> KuambRes, 
UAbrg -> Kuabrg};
parmsClr = {MCpClr -> mFracClr mCpRes, UAambClr -> KuambClr, 
UAclr -> Kuaclr};
parmsBoundary = {mCpPump -> 1, ubrg -> 200, ucw -> 60, uamb -> 70};
parmsInitialCond = {ures0 -> 70, uclr0 -> 70};
eqSetValues = 
eqSet /. parmsRes /. parmsClr /. parmsBoundary /. parmsInitialCond;
{TresSolLocal, TclrSolLocal} = 
NDSolveValue[eqSetValues, vars, {t, 0, 2000}];
N@TresSolLocal[timeTarget]
]
calcTresAtTime[60., 0.4, 300.]
Plot[calcTresAtTime[x, 0.4, 300.], {x, 0, 80}]
FindRoot[ 130 == calcTresAtTime[x, 0.4, 300.], {x, 0, 80}]

Below is the result of the Plot command. So this suggests that the function itself doesn't have any severe problems. enter image description here

Below is the result of the FindRoot command FindRoot output

Have worked through the examples in the documentation, but can't find where I have taken a wrong turn. Any help would be appreciated.

Attachments:
POSTED BY: Robert McHugh
2 Replies
Posted 6 years ago

Yep, that fixes it. Using the line below solves the problem.

FindRoot[ 130 == calcTresAtTime[x, 0.4, 300.], {x, 0.01, 80}]

This is perfectly reasonable for this application. And produces the result

{x -> 63.9137}

Thanks!

POSTED BY: Robert McHugh

The problem is that x == 0. is the starting point you provide and calcTresAtTime[0., 0.4, 300.] sets the coefficient MCpRes and parmsClr to 0., which clobbers the derivative terms in the ODE. Hence the NDSolveValue error. You could start the root search at 1 or higher.

(Plot uses "open" sampling; that is, it does not plug in the end point but use points just slightly inside the interval. It also suppresses some error messages.)

POSTED BY: Michael Rogers
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract