14 replies to this topic

[AlexNguyen]

Hello, 
I am a process engineer trying to understand the performance of a VSD centrifugal compressor when there is a demand change. 
The compressor is delivering natural gas. The inlet pressure is controlled by changing the compressor speed. The outlet pressure is allowed to change (not controlled). 

If the system downstream of the compressor requires more flow rate (say a downstream valve opens more), how would the compressor react? Does it
1. stay at the same speed, but runs out on its curve to develop more flow at a lower head?
OR
2. increase its speed to develop more flow at similar head and thus, reach a new performance curve?

My hunch is that most of the time, it will be case #2, but I feel like case #1 could happen as well.
Can you tell me what case will happen and why?
Thank you very much for reading. Sorry if my question seems basic. I searched through many websites and papers but I didn't find one that addressed this question. 

Edited by AlexNguyen, 08 June 2023 - 05:30 AM.

[Bobby Strain]

What you describe seems unrealistic for an application. One can't depict responses without knowing the whole system.

 

Bobby

[AlexNguyen]

Hello, Mr Strain. 
I'm sorry for the lack of information. Let me describe the system. 
So, upstream of the compressor is a KO drum whose pressure is not controlled. 
The gas exits the KO drum and goes into the compressor.
The compressor is driven by a VSD motor, compressing gas from 150 bara to 350 bara.
Normal control is changing the speed to maintain a constant inlet pressure. 
However, when the outlet pressure gets to high, there is a low selector that allows the outlet pressure to take over the control and slow down the compressor. High outlet pressure alarm is 367 bara.
There are flow rate, temperature, and pressure sensors on both the inlet and outlet lines of the compressor for surge control. The surge control valve is on a line between the compressor's outlet flange and a check valve. 
After the check valve is a cooler using cooling water. 
The cooled gas eventually goes to a manual choke valve, and then injected to a well.

Edited by AlexNguyen, 08 June 2023 - 07:47 PM.

[snickster]

Hello, 
I am a process engineer trying to understand the performance of a VSD centrifugal compressor when there is a demand change. 
The compressor is delivering natural gas. The inlet pressure is controlled by changing the compressor speed. The outlet pressure is allowed to change (not controlled). 

 

Why would the main design parameter be the control of the inlet pressure?  This may be true only for HVAC compressors where controlling the suction pressure also controls the evaporator temperature.  

 

If the system downstream of the compressor requires more flow rate (say a downstream valve opens more), how would the compressor react? Does it
1. stay at the same speed, but runs out on its curve to develop more flow at a lower head?
OR
2. increase its speed to develop more flow at similar head and thus, reach a new performance curve?

 

So if you are really controlling the suction pressure (using a PIC controlling compressor speed) based on maybe the need to control the pressure in a reactor vessel on the inlet then you will be trying to maintain a constant suction pressure, in this case: If a valve on the outlet opens then less pressure drop is developed on the outlet and therefore less differential pressure across the compressor, and more flow based on the head versus flow curve of the subject compressor, now considering you are trying to maintain a constant upstream presssure then with a VFD the compressor speed will slow down to a lower RPM so that the flow is the same as before the valve opened (note that at lower RPM the entire head versus flow curve moves down so that you get the same flow at a lower head).  This is because at a given volume of gas upstream, to maintain the same pressure upstream, the volume input must be equal to the volume output by the compressor.  If the compressor flows more at lower differential pressure then the volume input upstream will be less than the volume out as pumped by the compressor.  So the compressor capacity must be reduced by reducing the speed and shifting the head versus flow curve lower.  Remember that you are dealing with a compressible gas and the pressure in any section of the system is determined by the mass in that section.  To increase the pressure more mass must flow in than flow out.  To reduce the pressure more mass must flow out then flow in.  The pressure remains constan in a section of the system when the mass flow in equals the mass flow out.  This is in accordance with the gas equation of state PV=mRT.  So if you have a vessel with V = constant (volume), T= constant, R = cpnstant (same gas so same universal gas constant) then if you increase m you increase P and if you decrease m you decrease P.

My hunch is that most of the time, it will be case #2, but I feel like case #1 could happen as well.
Can you tell me what case will happen and why?
Thank you very much for reading. Sorry if my question seems basic. I searched through many websites and papers but I didn't find one that addressed this question. 

Edited by snickster, 08 June 2023 - 11:09 PM.

[AlexNguyen]

Thank you so much, snickster. I am very grateful that you showed interest and helped me again. 
1. Why would the main design parameter be the control of the inlet pressure?  This may be true only for HVAC compressors where controlling the suction pressure also controls the evaporator temperature.  

 

This compressor is actually downstream of another centrifugal compressor. I do not have any accesss to the control philosophy, just the P&IDs, so I may be wrong. But my guess is that controlling the suction pressure of this compressor is a way of controlling the back pressure to the upstream compressor.

Here is how I understood your explanation, please correct me if I am wrong. I have attached a sketch to assist visualization. 
Let's say the compressor is operating at point number 1 on the performance curve.
Then the outlet valve opens more.
The compressor operation will move to the right at the same RPM, i.e. it now does not need to develop as much head and hence, volumetric flow increases. Let's call this point number 2. 
As the flow increases, the inlet pressure will drop.
Then, the controller, sensing the drop, will decrease the RPM, restoring the volumetric flow rate to the same as point number 1 (to maintain the same inlet pressure), but at a lower head. This is point number 3, and the compressor will operate at this point.

Again, thank you very much for all that you have done. I truly appreciate your help.




 
 

Attached Files

•  Sketch of compressor's response.jpg   51.93KB   0 downloads

Edited by AlexNguyen, 09 June 2023 - 01:19 AM.

[Bobby Strain]

You need to show controls from the primary source to the terminal destination of the fluid. And, what is the gas? Your description appears to be of gas processing with gas re-injection. Still, you need to show all the controls on the compressor discharge to the wellhead.

 

Bobby

[snickster]

Thank you so much, snickster. I am very grateful that you showed interest and helped me again. 
1. Why would the main design parameter be the control of the inlet pressure?  This may be true only for HVAC compressors where controlling the suction pressure also controls the evaporator temperature.  

 

This compressor is actually downstream of another centrifugal compressor. I do not have any accesss to the control philosophy, just the P&IDs, so I may be wrong. But my guess is that controlling the suction pressure of this compressor is a way of controlling the back pressure to the upstream compressor.

Here is how I understood your explanation, please correct me if I am wrong. I have attached a sketch to assist visualization. 
Let's say the compressor is operating at point number 1 on the performance curve.
Then the outlet valve opens more.
The compressor operation will move to the right at the same RPM, i.e. it now does not need to develop as much head and hence, volumetric flow increases. Let's call this point number 2. 
As the flow increases, the inlet pressure will drop.
Then, the controller, sensing the drop, will decrease the RPM, restoring the volumetric flow rate to the same as point number 1 (to maintain the same inlet pressure), but at a lower head. This is point number 3, and the compressor will operate at this point.

 

Yes Exactly

Again, thank you very much for all that you have done. I truly appreciate your help.




 
 

[snickster]

And your diagram represents exactly what is happening.

[Bobby Strain]

Gas flow to the injection well is not intended to be controlled by varying the speed of the injection compressor. The injection compressor suction pressure control is intended to vary the compressor flow in response to flow change from the upstream compressor, where the injection flow is set. This is the typical control configuration for such an operation. Hence, a change in flow caused downstream of the injection compressor causes the wrong response by the compressor inlet pressure controller.

 

Your analysis is correct, but is not really relevant, and is not helpful to an operator.

 

Bobby

[AlexNguyen]

Hello, Mr. Strain,

 

You need to show controls from the primary source to the terminal destination of the fluid. And, what is the gas? Your description appears to be of gas processing with gas re-injection. Still, you need to show all the controls on the compressor discharge to the wellhead.

 

Bobby

Your description appears to be of gas processing with gas re-injection --> Yes, sir. It is gas processing with gas re-injection. The compressed gas is associated gas that has been separated out, then dried with glycol. 
Normal control is changing the compressor speed to maintain a constant inlet pressure. 
However, when the outlet pressure gets to high, there is a low selector that allows the outlet pressure to take over the control and slow down the compressor.
The discharge gas is then cooled in a cooler (temperature control on the hot gas outlet side to adjust valve on the Cooling Water outlet side). 
Then, the gas goes through a flow transmitter, and a manual choke valve before it is injected to the well. Downstream of the choke valve are pressure transmitters that warn of low pressure, so I am assuming the choke valve helps regulate the injection pressure (I said "assuming" because I don't have access to the control philosophy, just the P&ID).

Gas flow to the injection well is not intended to be controlled by varying the speed of the injection compressor. The injection compressor suction pressure control is intended to vary the compressor flow in response to flow change from the upstream compressor, where the injection flow is set. This is the typical control configuration for such an operation. ---> Yes, you made many good points. I think the control is intended to do what you said. Either the injection flow or the injection pressure is set via the manual choke valve position (sorry I cannot confirm if it is flow or pressure based on the information I have). 

Hence, a change in flow caused downstream of the injection compressor causes the wrong response by the compressor inlet pressure controller. ---> I don't understand what you meant, would you mind explaining it?

Your analysis is correct, but is not really relevant, and is not helpful to an operator. ---> What would you advice to make the analysis helpful to an operator? 

Thank you very much for your time. I appreciate your help. Sorry if I asked too much. I just find it fascinating and am curious to hear what you think.

 

Edited by AlexNguyen, 11 June 2023 - 10:51 PM.

[breizh]

Hi,

Consider this paper to support your work.

Breizh

[AlexNguyen]

Hi,

Consider this paper to support your work.

Breizh

Thank you, Mr. Breizh. The paper has been very helpful.

[CharacterLimit]

Hi,

Consider this paper to support your work.

Breizh

Hello,

 

Was there supposed to be an attachment in this comment?

[breizh]

Hi CharacterLimit,

I share document for a period of time because the storage capacity in this forum is limited. Others may need my support on other topics.

Breizh 

 

[CharacterLimit]

Hi CharacterLimit,

I share document for a period of time because the storage capacity in this forum is limited. Others may need my support on other topics.

Breizh 

oh i see.

 

Thank you i will download this file now.