Jump to content



Featured Articles

Check out the latest featured articles.

File Library

Check out the latest downloads available in the File Library.

New Article

Product Viscosity vs. Shear

Featured File

Vertical Tank Selection

New Blog Entry

Low Flow in Pipes- posted in Ankur's blog

Hysys Recycle Convergence Issues


This topic has been archived. This means that you cannot reply to this topic.
5 replies to this topic
Share this topic:
| More

#1

  • guestGuests
  • 0 posts

Posted 20 October 2009 - 11:30 PM

Hi,

I am simulating a process flowsheet with 3 recycle loops (this is already the optimal tearing and the number of tear streams cannot be reduced further), and Hysys is having some convergence issues, the flowsheet is very hard to converge ( i am using default sensitivity of 10). I am wondering what's typical sensitivity values for temperature/pressure/composition used in hysys?

thanks in advance!

Andina

#2 Srikrishna Chaitanya

Srikrishna Chaitanya

    Gold Member

  • Members
  • 56 posts

Posted 21 October 2009 - 05:02 AM

Hi Andina ,

Its very difficult for the people to understand if you ask generally , pl elaborate your question , which flowsheet you are simulating(reaction , distillation etc) ? , you have 3 recycles in the same system or whole PFD , if possible attach your simulation file so people here try solve it quickly.

#3 AlertO

AlertO

    Gold Member

  • Members
  • 194 posts

Posted 22 October 2009 - 05:04 AM

I think it's not quite easy to solve 3 recycles simultaneously. For first iteration, the default, sensitivity of 10 is OK to find a rough result. How ever, try to reduce these number to achieve an accuracy results especially flow and composition.

Good luck

#4 joerd

joerd

    Gold Member

  • ChE Plus Subscriber
  • 371 posts

Posted 23 October 2009 - 09:35 AM

It all depends (of course).
First of all, check if you have feasible specs. Make sure that all components have someplace to go (for example, if you have a trace component X in the feed, but all your unit operations are set up so that X is not present in any outlet stream, your flowsheet cannot converge). Observe the behavior of the recycles while solving (use the trace window, or open all three recycle windows while the flowsheet is solving) and see which variable(s) cannot converge.

If you have nested loops, or independent loops, convergence should not be too hard. You may specify the inner loop recycles with a higher tolerance, or fewer iterations, to minimize disturbance to the outer loop, as long as you check the final error when the flowsheet is solved. You can also play with the damping parameters or the solving method to see if it improves the situation.

If you have interfering loops (or whatever these are called), then convergence will be a lot harder. The solution of Loop A will upset Loop B, and vice versa, so you may end up in an endless oscillation. Try to think hard if you can eliminate the interference by moving the recycles around - even if you have to add one more, it may be advantageous.

Lastly, if you have Adjust operations in your flowsheet, these count as recycle loops as well, so take a good look if you can do away with them, or at least ignore them initially and switch them on later.

This is pretty general, if you want more specific advice, please post your simulation, or a detailed description of what you have done.

#5 daryon

daryon

    Gold Member

  • ChE Plus Subscriber
  • 131 posts

Posted 25 October 2009 - 12:17 AM

Just a couple of things to add that may help:

The Calculation Level for a Recycle (accessed under Main Properties) is 3500, compared to 500 for most streams and operations. This means that the Recycle is solved last among unknown operations. You can set the relative solving order of Recycles by modifying the Calculation Level. This may help solve the flowsheet.


HYSYS Internal Tolerances are as follows:
Vapour Fraction 0.01
Temperature 0.01
Pressure 0.01
Flow 0.001**
Enthalpy 1.00
Composition 0.0001
Entropy 0.01
**Flow tolerance is relative rather than absolute

For example, the internal tolerance for temperature is 0.01 and the default multiplier on a recycle is 10, so the absolute tolerance used by the Recycle convergence algorithm is 0.01*10 = 0.1. Therefore, the assumed temperatures and the calculated temperature must be within 0.1°C (where C is the internal units) of each other if the Recycle is to converge. HYSYS always convert the values entered to the internal units before performing calculations. A multiplier of 10 (default) is normal, and is recommended for most calculations. Values less than 10 are more stringent; that
is, the smaller the multiplier, the tighter the convergence tolerance.

#6 Padmakar Katre

Padmakar Katre

    Gold Member

  • ChE Plus Subscriber
  • 992 posts

Posted 06 November 2009 - 08:16 AM

Dear,
Good Explanation. Just adding one point as you already mentioned the Flow Convergence Tolerence. Yes it is too low as in steady-state simulation you will never able to close your material balance i.e. net inlet = net outlet. In case of transient state its ok as there are hold-ups. So always start with by default values as 10 for all parameters and then try to minimize it inorder to get the real feel and an accuracy.




Similar Topics