I have written code for a for a Heat Exchange Reformer in Aspen Custom Modeler, but it cannot run. The program describes a shell and tube heat exchanger with steam methane reforming reactions occurring on the tube side.
However when it runs, it gives error messages along the lines of:
"Check bounds: Step cut by 2.11607e-02 due to GHR.z("H2",27) below lower bound 0.00000e+00."
and is repeated for other components
Has anyone encountered a similar problem before?
I can share more details if needed.
It is a proprietary software, it could be related to algorithms and formulas or interface with the inputs and executable codes.
It looks like the composition of component 27 (H2) went below 0. It is possible that you set it up so it consumes H2 (to crack heavier compounds to CH4) and the H2 composition went below 0. Add some H2 to the feed to see if that solves the problem. You may need an H2 recycle stream if you are set up to consume more H2 than you have in the feed to crack the feed into CH4.
Thank you for the suggestions. For clarification, the number 27 corresponds to its position along the length of the reactor (there are 60 points in total). It is strange for H2 to decrease as the reactor is supposed to be producing H2. The same message also comes up for CO.
Increasing the mole fractions in the feed of H2 and CO had no effect.
If it wants to consume H2 but you think it should produce H2 it probably is having a problem with unexpectedly low temperatures.
Ok so I tested it at feed temperatures from 300 to 1000 oC and pressures of 1 to 30 bar to no avail.
You can check and try to change the nonlinear solver or its parameter.
Please describe the feed rates composition T and P on both sides of the heat exchange reformer.
The attached image shows the inputs of the tube side stream and shell side stream, with temperature in oC and pressure in bar.
The non-linear solver has 3 modes: Standard, DMO and Open NLA Solver. Within Standard, there is Hybrid, Newton, Fast Newton and Mixed Newto, with either residual and/or variable convergence.
Testing all of these gave the same error as before.
I think the domain definition could also be relevant:
L AS LengthDomain (DiscretizationMethod : "BFD1", HighestOrderDerivative : 1, Length : Hx, NumSections : 1, SpacingPreference : 1/Num_Ele );
Right now, Num_Ele is 4, meaning each unit length is split into 4 pieces. Ideally this should be much higher but increasing it to say 100, leads to much longer run times of several minutes and it shows different error messages that appear to be due to the pressure drop equation for some reason. But the pressure drop seems to work properly on its own at Num_Ele of 4.
Vishnu Pusarapu
The DMO method didn't help you really?!
Because in this method, the bounds of variables are not considered.
- Badges
- Science topic
- Similar topics
- Heat