The control configuration of crude oil desalting

Pls draw out this control scheme, and also explain why the additional PV is on this exit line to the stab column. It would seem there is no need for the PV - it should be possible for the high set PIC to act on the FCV via an auto signal select.

First there should not be any control valve at the inlet of the desalter to avoid flashing of the crude and breaking the electric field across the electrodes i.e. impeding the coalescence and agglomeration of the separated water droplet from the crude emulsion.

Desalter is meant to operate as liquid full and hence pressure maintained sufficiently above the bubble point through the use of the feed pump.

Level in the separator should be controlled only after passing through the desalter loop otherwise the above concerns may occur. I dont see a reason to have a pressure control at the outlet of the desalter, please clarify.

@george
I think the PV has been applied on the exit line to maintain the desalter pressure above the bubble point; but will fight against the FCV on the same line...right?. FCV is modulated by LIC on the upstream separator.

@Sawsan
I do fully agree with you for avoiding to apply the control valve at the inlet of the desalter.
Also i do agree putting the control valve at the outlet of the desalter instead putting it at the inlet; but do you think the FIC/FCV on the dasalter exit (stab column inlet) should be applied to be adjusted by LIC on the separator or PIC/PCV?

Attached there can be seen three options; which one is applicable and logical?

Option 1 looks okay - no concerns, but may be better with FIC - FCV relocated to desalter exit.
Option 2 - desalter may be prone to flow surges
Option 3 - simpler than option 1 and would be okay also if developed dead head pressure at desalter is within MAWP limitations.

Use a low shear pump to feed the desalter.

Thanks George,
In option 1 considering the desalter is full of liquid,don't the FV and PV fight against each other?
For option 2 please clarify about the flow surge...

e43u8 Yes cascade control based on separator level and flow at the inlet of column can be implemented.

@Sawsan
But i mentioned cascade control based on separator level and pressure at the inlet of column...your opinion on this scheme?

The PV in these options is operated by a low set PIC - it only comes alive when pressure drops below setpoint. During normal operation, it would be dormant, as pressure is higher than setpoint. Yes, the arrangement in option 1 will result in some interaction between the PIC and FIC when desalter pressure drops low.

I would suggest a minor modification of option 3 as the preference. Upstream of the FCV, install a low set PIC. Feed its signal to an auto signal override- select, which also receives the output from the FIC. Feed the output from the signal selector to the FCV. This will enable the PIC to override the FIC and maintain the PIC setpoint if the desalter pressure drops low. ( Otherwise, when desalter pressure is good, the PIC remains dormant). Also enable a flow setpoint max limit stop at the FIC so it does not go to flows which may overload the desalter.

@e43u8 I find it usually difficult to control subcooled liquid system under pressure unless it is an override function as propose by Georgeverghese.

cascade of level with flow is a better reflection of the disturbance of the PV than level with pressure. Low pressure override can be added via a selector relay as explained by Mr.George to ensure maintaining the desalter pressure (although it is genuinely maintained by the proper design of the desalter feed pump) to ensure your pressure is always above the bubble point pressure at the maximum operating temperature.

@Sawsan,
Don't you think considering the desalter vessel(s) is full of liquid and the level of the connection point of outgoing line from desalter to stab column is normally well above the desalter level, basically no need to control the pressure at desalter outlet to avoid the desalter to be subject to bubbling?
Of course in case of low set pressure a low pressure alarm can inform the operator for proper action and seems it would be adequate...right?