How to prevent non-convergence and dry plates in Aspen
Thread Content
I have just started learning *Aspen, and during simulated distillation calculations, I often encounter issues with convergence problems and a situation where the calculation \"dries up.\" I’m not sure why this happens; sometimes convergence is achieved after adjusting the feed parameters, while other times it isn’t. I wonder: no matter what the feed conditions are, calculations should be possible, so why can’t it be done? What’s the reason for the dry plate issue? Is this a situation that actually occurs, or is it an error in calculations caused by improper settings? How can one determine the reasons for non-convergence and the dry plate problem? . . . . . . . I earnestly request the guidance of experts from Haichuan. Thank you
I’m not an expert, but I can talk about this issue briefly. It is indeed a very common problem. I remember that when performing distillation simulations, if too much vapor was fed into the system, the entire column could dry out. There are even more cases where convergence does not occur. Here are some methods that might help, for reference only: 1) Since the reboiler is provided at the bottom of the column, and although condensation also occurs at the top, the distillation column still requires external heating. Therefore, it should be fine to have a slightly higher fraction of liquid feed, but it certainly shouldn’t be too high. 2) Whether convergence is achieved depends on the algorithm used. The internal and external circulation algorithm used in RadFrac is similar to the homotopy method, and it requires relatively accurate initial values. These initial values need to be determined based on an engineer’s knowledge of unit operations; if they are set arbitrarily, convergence will often not be achieved. Theoretical knowledge can be used to determine these initial values, and then variable optimization techniques can be employed to find the best solutions. I’ve written some explanations on this topic at http://bbs.hcbbs.com/thread-430336-1-1.html; you might want to take a look. Additionally, in some simulations, the number of iterations often exceeds 25. If you notice that the iteration values are getting smaller but convergence still doesn’t occur, you can increase the number of iterations. I really hope to get points for this answer – I accidentally used Xunlei to download files, and as a result my score dropped significantly
I’m sorry; I’m in a lower grade level and can’t give you extra points. But thank you very much for your help and response. Wishing you quick success in making money
How did a dry plate appear? Very rare to see! ! What are the common causes?
Usually, attention should be paid to the balance between the gas and liquid flow rates; the explanation on the second floor makes some sense.
I’m also just learning it; I don’t understand it very well. I’ll just share my own experience: once, when carrying out the absorption of carbon dioxide using low-temperature methanol, although the flow rates in both the gas and liquid phases showed relatively high values according to the process profile, an error was reported indicating a dry bed. At this point, it is possible to first estimate the temperature and composition of each plate, and enter reasonable values. Because AspenPlus automatically assigns initial values to certain parameters on each tray when performing distillation calculations, if the initial values for temperature and composition are not appropriate, it may result in the disappearance of phases that should be present. Regarding the issue of non-convergence, for highly non-ideal systems, it is difficult to achieve convergence using standard methods; the strong-nonideal method can be employed, or the number of iterations for convergence can simply be increased