Arrhenius data curve fitting in Mathcad 13

[bpeterson272]

I have been attempting to use a solve block to fit my kinetics data using the Arrhenius equation predicted value, subtracting |predicted - actual|, and finding the best fit through minimizing the sum of the differences. In Excel this is tedious but achievable; in Mathcad I am stumped.

My process logic is roughly:
1. import data into array: time, weight, weight change, temp; determine rows in file
2. start solve block with given
3. assign guess values for A, E, m, and n
4. use loop to calculate 2D vector of weight and weight change based on guess values for A, E, m, and n.
5. calculate 1D vector of |predicted - actual|
6. sum the 1D vector as "S"
7. Use Minerr(S,A,E,m,n) to minimize S.
8. Plot two graphs - dW/dt actual and dW/dt predicted

What actually happens is there are no iterations of A, E, m, or n occurring, and therefore no curve fitting. I am confident my equations are correct; if I use the proper guess values I get the proper plots. It is abundantly clear I am clueless about the logic in solve blocks. The examples provided in Mathcad are understandable but not obviously applicable to my particular problem.

For those unfamiliar with the Arrhenius equation, it has the general form:

dW/dt = -A*exp(-E/RT)*(W)^*(1-W)^(m)

 

[IRstuff]

You should navigate to the Mathcad Collab and post your file:

http://collab.mathsoft.com/~Mathcad2000

TTFN