I realize that you mentioned gPROME for modeling a bubble bed reactor, however, I would like to recommend another two software platforms for better modeling such reactor. These are HYSYS and Aspen Plus simulation platforms by Aspen Technologies Corporation.

HYSYS and Aspen Plus have much better thermodynamics models for simulating bubble bed reactors and you will get much better results than gPROME. At least your use HYSYS or Aspen Plus to validate the results you get from gPROME.

Hope that you find this information useful.

Professor Yehia Khalil

Yale University

USA

Could you post your model so I may take a closer look?

I suggest you contact my colleague at ISEL: Prof. Rui Filipe (email: [email protected] )

Thank you for trying to help me with the design. According to my supervisor he preferred to build the model on gPROMS and later I should validate the results with Aspen.

I have reviewed all the equations so many times, therefore, I may ensure you that they are correct. So , the problem is on the code.  I have realized that the only way to get results is using the Forward finite difference method for the discritisation. I have also realized that  if I increase the number of steps, the code does not work.

I attached three files, one is the gPROMS code that I wrote, the paper that I am following for the  equations (mass balances and kinetics) and the results of one simulation which seems to be right (however, different initial conditions with respect to the paper). 

Thank you very much for taking the time to help me.

What you descibe sounds more like a numerical problem.

if I increase the number of steps, the code does not work. => numerical diffusion due to rounding errors that increase faster than the increased accuracy due to smaller time or space steps.

I have no experience with gPROMS and came across your question when looking for models that are verified by experiments.

Did you try Gauss- Seidel or Romberg instead of Newton-Raphson? Or leap-frog (additional to the 5star a known upstream value to increase stability) instead of forward finite difference?

Good luck!