2 replies to this topic

[roblar03]

Hello Everyone!

 

I am currently working with a Master Model of an oil & gas plant in HYSYS, and I've been told by one of the seniors I'm working with that it's good engineering practice to avoid use of the number zero when assigning flow rates, and that small numbers approximating zero (e.g. 0,001) should be used to avoid issues with convergence. Is there any hold to this claim, or should I disregard his advice? 

 

Thank you

[latexman]

Since high school, all us STEM disciples know that dividing by zero (0) creates problems in numerical solutions, right?  Your senior sounds like a wise Engineer.

[shvet1]

Convergence algorithms are incredibly intricate, making it challenging to grasp their precise inner workings, particularly those embedded within Aspen. In many instances, these algorithms fail to function altogether. I have personally witnessed this on numerous cases.

Engineers often develop their own biases regarding relevant error prevention and, in some cases, even forcing the convergence to work. This particular bias could be valid or entirely false.

You can conduct an experiment on yourself to verify this.

Edited by shvet1, 16 May 2024 - 09:09 AM.