Hi everyone,
I am trying to control a horizontal separator oil box by maintaining steady outflow, and am not too concerned with the level set point. The separator is the second vessel after the slug catcher, which is essentially a surge vessel. The problem we are having is that the way these controllers were tuned originally is to maintain set point.
As an example, the separator oil box level controller 2-LIC-107, has a set point of 2600 mm. There is a flow meter on this oil outlet line that I have converted to an FIC, so we can send a CAS set point. From what I have been reading, the LIC should be proportional only (no integral), so that the response to inflow changes is proportional, and not exaggerated by the integral action trying to maintain set point. Can someone give me some advice as to how we should be controlling this process? Is there other DeltaV functions that could be used in this situation to maintain steady outflow? (I deadband, PIDplus, nonlinear gain etc.)
Thank you!
Jayme
Controls specialist
Andre Dicaire
In reply to Andre Dicaire:
In reply to controls_wl:
In reply to James Beall:
The blue line in the interface control PV.
The green is the separator inlet coming from the slug catcher oil box
The yellow is the separator oil box outlet flow (the one we want to make steady as it then goes from the separator oil box to the treater, which is the final stage of treating so the steadier it is, the better the treating)
The red is the water dump flow (interface)
In reply to LaurentB:
Thanks Jayme!
Any movement of the I/F will "push" more or less oil out. I recommend tuning the I/F level controller for tight level control.
I recommend that you read my ISA article on tuning PID controller on integrating processes, like most level loops, https://www.isa.org/intech/201604basics/ . Note that for Level cascade to flow, you can calculate the process gain and thus tuning without a step test. Please send a sketch of the separator with normal I/F level, normal oil level, and where the I/F and Oil LT's are situated.
I recommend that you read my ISA article on tuning PID controller on self-regulating processes, like most flow loops, https://www.isa.org/intech/201604basics/ .
When using Gap control on integrating process, I recommend a lower, but not 0 gain. A 0 gain for the PID controller (with integral action) on an integrating loop will cause oscillations. If you reduce the gain (1/2 or 1/4) of "out of gap range", you need use the larger reset based on the lower gain, both in and out of the gap.
More later!
James
Hi all , it's a very interesting topic.
Having a look on the trends you post, we can see you have 2 differents issues.
- "High frequencies" disturbances which are process noise ( Pressure, vibrations, liquid moves, and waves ....).This noise does not represents the real quantity inside the separator, but controller tries to compensate these kind of disturbance which induce fast moves on the valves becoming disturbances all downstream process. In adddtion there is a big impact on expected life of valve, due to frictions.
There is 2 ways to eliminate this noise and stop the valve oscillation :
- The wrong way : add filter on PV. This is not a good idea because in case of real disturbance, the response will be to slow with overshoot, and increasing gain to try to fight these disturbance will destabilize the loop
- Adding a dead band , with DeltaV KNL function. Have a look on the picture below to appreciate the result.
- Without KNL the valve ( green) follows the noise ( PV Red) in phase opposition
- With KNL ( dead band tuned with noise amplitude) , the noise is still here , but doesn't matter for your process) , but the valve doesn't move
- With KNL the "real disturbance is well corrected
- Low frequency variation of your level. This low wave frequency level variation is caused too the poor tuning of your pure proportional controller.
As it was said in the different replies, for a separator it's important to have tight level control for process efficiency. Pure proportional controller doesn't allow good performance and lambda tuning suggested by James is a good approach. You have an escellent tools on DeltaV which is INSIGHT which can calculate for you the PID parameters with this lambda method. Insight gives you right parameters very easyly and quickly.
Anyway how to use KNL . As you already read the BOL , I will give you very fast explanations:
- to use it you have to go on PID blocks , and into FRSIPID_OPTS, check "Use Non Linear Gain Modification
- With KNL the proportional and integral action are multiplied by KNL , that means if KNL = 0 , there is no more integral and proportionnal action.
=> set NL_GAIN_MOD to 1
=> set NL_MINMOD to zero : eliminating the controller actions
=> Set NL_GAP to 25% of noise amplitude , put value in EU of PV_SCALE) ,
=> Set NL_TBAND of 75% of noise amplitude put value in EU of PV_SCALE) ,
=> Set NL_HYST of 25% of noise amplitude in EU of PV Scale
Hi Jamye So does 2-LIC-107 gives FIC SP? If it is not the case, please ignore below. If it is the case, I would try to make FIC.OUT set 2-LIC-107.SP. At the same time 2-LIC-107.SP_LO_LIM and 2-LIC-107.SP_HI_LIM can be restricted to avoid empty vessel or overflow.
As you mentioned that you don't car too much about the level's PV, I think there would be acceptable control range of level.
And then 2-LIC-107.OUT can control incoming flow to the vessel.
Regards
JG Yu
+61 428 950 518