Can someone explain to me why this ODE solve block procedure (see worksheet attached) is not working with MathCad Prime 1? I did a similar exercise in MathCad 15 (see image) 
using an equivalent formulation and all was OK. What I am trying to do is a sensitivity analysis with the parameter k which in this case is the thermal conductivity of the insulation material. For some reason the variable k does not seem to recognised by the solve block and thus the matrix of functions is not being returned to the variable.
1 ACCEPTED SOLUTION
Accepted Solutions
Mark Buckton wrote:
Also, try as I might, I cannot get rkfixed to accept Tf as a function based on time Tf(t) this is necessary as it represents the temperature rise in a furnace which as you would expect is a function of time i.e. the temperature increased with time.something like Tf(t) = 20+345*LOG(8*t+1) as soon as you do this one gets an error see:
/servlet/JiveServlet/downloadImage/2-173291-27452/1-01-2012+10-33-26+PM.png
See my solution with t/s in the empirical formula for Tf:
Valery or Mike;
Unfortunately this is not the answer that I am looking for i.e. I know I one can solve the ODE for a single value of k, like you I had already done that (noting that this is a simple extract from a larger example). In fact the actual problem requires that the fire temperature is a function of t, and the specific heat of the steel beam is also a function of temperature but this is not the problem as that has been considered in the actual problem. My particular problem, to which my question relates, is NOT how to obtaining a single solution to the ODE, but RATHER I want to solve the ODE multiple times i.e. for successive values of k. then graph each solution vector. (see image below) i.e. each case is a different value of k.
This is why I initially made PSteelTemp a function of k - 
PSteelTemp(k). As I said this is easy to do in MathCad 15 (see initial image on this thread). What I am saying and probably didn't make clear is the same methodology FAILS in MathCad Prime1.0. What I don't know is if this is a limitation in MathCad Prime 1.0 or there is a procedural difference i.e. a work-around way of doing the same thing in MathCad Prime 1.0?
Regards, Mark
Valery Ochkov wrote:
You must use in this case not the function odesolve but the function rkfixed for example and create an user function based on ODE nymerical solution.
I have made this calculation but I see one unit error in Mathcad Prime - we must have K not s*K:
Your result looks to have the correct shape for the curve but it seems a very convoluted way to do this exercise i.e. not very understandable or explicit - to me anyway. I thought MathCad Prime aim was to make things simpler not more complicated, so are you saying that MathCad Prime is not as capable as MC 15 when it comes to ODEs i.e. one cannot pass variables back to the solve block and recalculate solve block function based on a new independent variable and obtain a new solution vector?
Also, try as I might, I cannot get rkfixed to accept Tf as a function based on time Tf(t) this is necessary as it represents the temperature rise in a furnace which as you would expect is a function of time i.e. the temperature increased with time.something like Tf(t) = 20+345*LOG(8*t+1) as soon as you do this one gets an error see:
Thanking you for your forbearance.
Mark
Mark Buckton wrote:
Also, try as I might, I cannot get rkfixed to accept Tf as a function based on time Tf(t) this is necessary as it represents the temperature rise in a furnace which as you would expect is a function of time i.e. the temperature increased with time.something like Tf(t) = 20+345*LOG(8*t+1) as soon as you do this one gets an error see:
/servlet/JiveServlet/downloadImage/2-173291-27452/1-01-2012+10-33-26+PM.png
See my solution with t/s in the empirical formula for Tf:
Mark Buckton wrote:
Thanking you for your forbearance.
Mark
No problem! I work with students 
