5 replies to this topic

[sivaji1233]

I have an HYSYS simulation file where Distillation column is simulated and converged using the theoretical number of trays and with an efficiency of 100%. The theoretical loadings data was mentioned Tray datasheets, now the client has asked for loadings based on actual tray numbers. I tried to simulate the previous file where theoretical trays are converged by changing the numbers of trays based on efficiency and providing the same efficiency on all trays. I was unable to converge. and "FATAL ERROR -ZERO DIAGONAL IN COLUMN MATRIX" pops up. I tried changing feed tray location but no results. I do not have any luxury to change the specs as the whole design will change. Kindly give any suggestions. 

 

[Bobby Strain]

Your client employee is not very knowledgeable about how we design fractionation equipment. And you seem hesitant to disagree with him. Never in 50 years have I used anything other than theoretical tray loads and fluid properties to specify loads and properties for packing and trays. And provide actual tray count for the expected tray efficiency. I would have a talk with my leader and let her convey the proper information to your client. But, you may be taking instructions from a client employee. That is not the same as your client. This sort of thing happens frequently, so don't be overly concerned.

 

Bobby

[colt16]

I have an HYSYS simulation file where Distillation column is simulated and converged using the theoretical number of trays and with an efficiency of 100%. The theoretical loadings data was mentioned Tray datasheets, now the client has asked for loadings based on actual tray numbers. I tried to simulate the previous file where theoretical trays are converged by changing the numbers of trays based on efficiency and providing the same efficiency on all trays. I was unable to converge. and "FATAL ERROR -ZERO DIAGONAL IN COLUMN MATRIX" pops up. I tried changing feed tray location but no results. I do not have any luxury to change the specs as the whole design will change. Kindly give any suggestions. 

 

The theoretical trays loading should be used in tray specification. If the client is asking for the Tray datasheet loading to be changed to actual trays then this is a mistake and it may lead to a a tray flooding result even when it is not the case. If the client is asking the simulation to be done with actual trays with the efficiency keyed in then there is no difference with your original file as the loading calculated will be the same. It is a waste of time to key in the efficiency.

 

I am puzzled by why your simulation did not converge though. Have you tried removing the tray efficiency and see if it would converge?Theoretically this situation will give you a lower duty and hence lower loading (it is the same as using more trays in a tower). Please ensure correct feed location.

 

If my suggestion converges, then perhaps you need to check your efficiency input. Most of the time a simulation does not converge because of some error in the input.

[solnoid]

 

I have an HYSYS simulation file where Distillation column is simulated and converged using the theoretical number of trays and with an efficiency of 100%. The theoretical loadings data was mentioned Tray datasheets, now the client has asked for loadings based on actual tray numbers. I tried to simulate the previous file where theoretical trays are converged by changing the numbers of trays based on efficiency and providing the same efficiency on all trays. I was unable to converge. and "FATAL ERROR -ZERO DIAGONAL IN COLUMN MATRIX" pops up. I tried changing feed tray location but no results. I do not have any luxury to change the specs as the whole design will change. Kindly give any suggestions. 

 

The theoretical trays loading should be used in tray specification. If the client is asking for the Tray datasheet loading to be changed to actual trays then this is a mistake and it may lead to a a tray flooding result even when it is not the case. If the client is asking the simulation to be done with actual trays with the efficiency keyed in then there is no difference with your original file as the loading calculated will be the same. It is a waste of time to key in the efficiency.

 

I am puzzled by why your simulation did not converge though. Have you tried removing the tray efficiency and see if it would converge?Theoretically this situation will give you a lower duty and hence lower loading (it is the same as using more trays in a tower). Please ensure correct feed location.

 

If my suggestion converges, then perhaps you need to check your efficiency input. Most of the time a simulation does not converge because of some error in the input.

 

 

Can you elaborate on why you would provide the vendor with theoretical loads? I would think that one would want to adjust for the efficiency of the trays to get the desired product specs (perhaps more reflux is required with the actual number of trays). Does providing theoretical loads apply to both a new column design and an existing revamp too? Thanks.

[colt16]

 

 

I have an HYSYS simulation file where Distillation column is simulated and converged using the theoretical number of trays and with an efficiency of 100%. The theoretical loadings data was mentioned Tray datasheets, now the client has asked for loadings based on actual tray numbers. I tried to simulate the previous file where theoretical trays are converged by changing the numbers of trays based on efficiency and providing the same efficiency on all trays. I was unable to converge. and "FATAL ERROR -ZERO DIAGONAL IN COLUMN MATRIX" pops up. I tried changing feed tray location but no results. I do not have any luxury to change the specs as the whole design will change. Kindly give any suggestions. 

 

The theoretical trays loading should be used in tray specification. If the client is asking for the Tray datasheet loading to be changed to actual trays then this is a mistake and it may lead to a a tray flooding result even when it is not the case. If the client is asking the simulation to be done with actual trays with the efficiency keyed in then there is no difference with your original file as the loading calculated will be the same. It is a waste of time to key in the efficiency.

 

I am puzzled by why your simulation did not converge though. Have you tried removing the tray efficiency and see if it would converge?Theoretically this situation will give you a lower duty and hence lower loading (it is the same as using more trays in a tower). Please ensure correct feed location.

 

If my suggestion converges, then perhaps you need to check your efficiency input. Most of the time a simulation does not converge because of some error in the input.

 

 

Can you elaborate on why you would provide the vendor with theoretical loads? I would think that one would want to adjust for the efficiency of the trays to get the desired product specs (perhaps more reflux is required with the actual number of trays). Does providing theoretical loads apply to both a new column design and an existing revamp too? Thanks.

 

 

No, when you use actual tray numbers in simulation, you will get less reflux due to lower heat and condensing duties because actual tray in the tower will be more than the theoretical trays.Therefore this value given to the vendor to evaluate their tray hydraulics will not be correct.

 

In simulation, we always use theoretical stages (which has less trays) compared to the actual. 

 

If the simulation is performed with actual trays with efficiency keyed in individually then it will be the same as doing the simulation with theoretical trays only. 

 

Theoretical tray loads apply always to column design/revamp/troubleshooting.

[solnoid]

My mistake. What I actually meant was actual trays with efficiency keyed in. However, I assumed that it wouldn't be the same as the theoretical load at 100% efficiency. Thanks for the explanation.

 

sivaji1233, I hope my inquiry didn't detract from your question. I can see that you still haven't resolved your convergence issue.