6 replies to this topic

[tomr91]

Hi everyone,

I’m currently working on a simulation project in HYSYS 12 and need to model a distillation column using Koch-Glitsch MINIVALVE trays (Link: https://www.koch-gli...=true&level=3).However, I’m unsure how to properly implement these trays into my simulation.

Specifically, I’m looking for guidance on:

1. Whether there are built-in options in HYSYS to model these trays directly or if I need to create custom specifications.
2. How to input the tray efficiency, pressure drop, and other key parameters relevant to MINIVALVE trays.
3. Any resources or best practices for accurately simulating these trays in HYSYS.

I would greatly appreciate any tips, examples, or documentation links that could help!

Thanks in advance for your support.

[latexman]

IMHO, it's best to work in theoretical equilibrium stages within the simulation, then determine an overall efficiency for your system so you can  translate the simulation's results to reality.

[Pilesar]

For best accuracy, consider using KG-Tower software in conjunction with Hysys. KG-Tower is provided without charge by Koch-Glitsch and is tuned to give good predictions of pressure drop and hydraulics. You would need to provide input from your Hysys model to KG-Tower, then update the Hysys model parameters based on the KG-Tower results. Alternately, ask your KG tech contact about guidelines for developing your Hysys model.

[tomr91]

For best accuracy, consider using KG-Tower software in conjunction with Hysys. KG-Tower is provided without charge by Koch-Glitsch and is tuned to give good predictions of pressure drop and hydraulics. You would need to provide input from your Hysys model to KG-Tower, then update the Hysys model parameters based on the KG-Tower results. Alternately, ask your KG tech contact about guidelines for developing your Hysys model.

 

Many thanks for your response. May I ask you further clarify  what do I need to configure in KG-Tower, and what are the outputs of KG-Tower calculations that I can take into HYSYS?

[Pilesar]

Tower models can range from simple to very precise. Usually, you can get fairly accurate results without putting in excessive details. Fixed valves are a standard tray design. Have you simulated your process in Hysys? Have you ever simulated a distillation tower? Asking about what options there are in Hysys leads me to believe you have not even started the model. I suggest you learn the basics of Hysys and begin! The software company provides training and examples. Look to the web for video tutorials by users. After you get a high-fidelity column converging, then the hydraulics become more critical. Hysys has tools built in. KG-Tower includes many of the same features that KochGlitsch uses when designing trays. KG-Tower should give accurate results for flooding, weeping, pressure drop, downcomer backup which are needed to determine capacity and turndown. The thermodynamic system, physical properties, feed composition and condition, tower operating pressure, tower material balance, etc are essential to a good model. Is your model complete except for hydraulic calcs? Or are you struggling with making a model of your column and don't know how to ask about the basics?

[tomr91]

Tower models can range from simple to very precise. Usually, you can get fairly accurate results without putting in excessive details. Fixed valves are a standard tray design. Have you simulated your process in Hysys? Have you ever simulated a distillation tower? Asking about what options there are in Hysys leads me to believe you have not even started the model. I suggest you learn the basics of Hysys and begin! The software company provides training and examples. Look to the web for video tutorials by users. After you get a high-fidelity column converging, then the hydraulics become more critical. Hysys has tools built in. KG-Tower includes many of the same features that KochGlitsch uses when designing trays. KG-Tower should give accurate results for flooding, weeping, pressure drop, downcomer backup which are needed to determine capacity and turndown. The thermodynamic system, physical properties, feed composition and condition, tower operating pressure, tower material balance, etc are essential to a good model. Is your model complete except for hydraulic calcs? Or are you struggling with making a model of your column and don't know how to ask about the basics?

First of all, thank you for your kind and detailed response. We have built a converging static model of an actual atmospheric column of a CDU. The end goal is to simulate the distillate properties' responses to changes in the feedstock. The tray configuration seems to be essential to simulating the delays due to holdups in the column. We have pretty complete column layout documentation. The column is entirely based on Koch Glitsch trays. However, HYSYS doesn’t support configuring all the properties present in the tray specifications. For example, not sure how MINVALVE and SUPERFRAC trays map into the tray parameters available in HYSYS. The version of the column model with the trays we have configured causes the message “internal calculation completed with errors.”  An example of a specific message: “Trays 5 - 8 are operating above the maximum of 100% jet flood specified. We recommend increasing tray spacing.”

Rather than tweaking the tray settings arbitrarily, we thought it would be right to replicate the settings of the actual column trays as reliably as possible. For this purpose, using the vendor software (KG Tower) seems like a reasonable approach. Our gap in understanding is how to use the KG Tower jointly with HYSYS correctly.

[Pilesar]

Atmospheric crude towers are complicated models! So much depends on feed characterization and selection of pseudocomponents. Trying to match side draws is tricky. Matching the top distillate is more 'that is close enough' instead of 'look how precise a model I made.'

My guess is that inaccuracies in feed characterization will have much greater impact on your model than whether the correct minivalves are used.

I would be surprised if Hysys does not have an option to characterize standard fixed valve trays. Hysys will NOT be able to represent superfrac trays very well. Even KG-Tower will tell you to contact KochGlitsch if you want to model superfrac trays.

If it were me, I would consider estimating performance using sieve trays with a fudge factor if I could confirm the results with plant data. The fudge factor may be as simple as making the column diameter in the model wider where you have superfrac trays. 

KochGlitsch does have technical support. You should be able to send them your converged Hysys results and have them help. They have offered to do that for me but I am hesitant to rely on someone else to do my modeling. What I have done is to send two disparate feed cases to them and then use the answers they send back to adjust parameters in my model to match. Once my model is tuned to give their results, I no longer have to iterate with KG and can adjust the column feed within a reasonable range as long as I keep the tuning parameters constant. For a crude column, some portions of the tower will be constrained more than others. You may need several feed cases to identify the bottlenecks.