Given Find Minerr problems

[CRG]

I currently have problems with finding solutions using the solve block, “Given ____ Find.” The problem is when I have equations in the solve block that very far apart in their solutions. Example: a solve block used to calculate flow would have variables with solutions of 10^7 for one variable and other variables with solutions in the order of 10^-6. The existing solve block will not solve these types of equations together in a solve block. The solve block can handle any single equation by it self; however, not when combined into a single block. To be more specific the equations are Reynolds number, Colebrook equation, General Flow Equation for compressible fluids, Velocity, and minor friction losses. I have tried to change TOL, CTOL, and initial guesses with no success. There is only one solution for the set of equations, and I have been manually iterating for the solution of dP or Flow depending on the problem. Just seems a shame that I either do not know how to use the tools of the program or the program will not work as advertised. When I try “Given ____ Minerr”, the program give an incorrect solution.


I was in hopes of not having to write a program to do what a solve block is supposed to do. Do you have any suggestions? I contacted the folks at MathCad and ask if there were any improvements from 2001i to the current version that might help, and they said no.

 

[pstuckey]

There is something you can try before writing your own program. Right-click on "Find", go up to non-linear, then select Levenberg-Marquardt for the solution method. The other solving routines often do not work well even for remarkably simple problems. There is no guarantee that L-M will work either, but it is usually more reliable than the others.

Peter

 

[IRstuff]

Sometimes, you can get the equations to solve if the variables are moved around. I've had cases where moving something from the denominator on one side to the numerator on the other side allows the solution to proceed.

TTFN