Centrifugal compressor anti surge

[Indy]

Hi
I posted this thread in the electrical engineering section but it may be better suited here. I’m just trying to understand why compressor Anti-Surge controllers take temperature and pressure signals from the suction side and also the discharge side along with flow to then modulate a control valve to recirculate a portion of the flow back to the suction side to prevent surge occurring.

I suppose I am asking why are the signals not just taken from the compressor discharge side ? Can’t see what I’m missing.

Thanks

 

[crshears]

Hey again Indy,

Here's what I posted there, just in case those who might respond don't monitor that forum:

Maybe I'm full of it, but what's that basic formula again? Something like PV=nrt, which fleshes out to P1V1/T1=P2V2/T2...or something like that; point being that accurately determining the molecular volume of a gas transiting a compressor requires the measurement of volume, pressure and temperature on both the suction and discharge side of the compressor, with the computed difference between the two forming an index of the amount of compression energy being added to the gas, the latter of which could also be cross-computed by plugging [electrical?] readings into an algorithm that plots the efficiency of the compressor against its prevailing loading...

Bottom line is that each and every one of these inputs is required to calculate and then control the required recirc to a nicety so that surging is prevented while holding down to an absolute minimum the amount of energy lost in meeting this objective.

Already scanning the skies for incoming tomatoes...

There aren't any tomato stains on my shirt yet, so I may be on the right track; but that being said, I'm thinking georgeverghese might be the right guy to weigh in on this...

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[rotw]

The anti-surge controller monitors the distance between actual operating point and the surge limit line SLL ( pre-built in the controller).

If distance decreases (i.e. flow reduces) down to threshold value (set point), anti-surge controller is triggered.

Anti-surge controller sends a 4-20 mA analog signal to control valve to open (control element).

Operating point is tracked on an invariant coordinate system: 'reduced head' or 'pressure ratio' vs 'reduced flow'.

SLL is represented in this coordinate system. Actual point distance to SSL is monitored. For example, flow ratio: actual to surge flow .

To determine flow and head of operating point, are needed Press.,Temp. at inlet and discharge and flow; Thus the instrumentation set-up .

With pressure ratio (some machine configurations allow such simplification) the set up is tweaked accordingly.

NB: In practice, distance is not measured against SLL, but against so called Surge Control Line, SCL at safety distance from SLL(~10%).

 

[Indy]

Thank you for both of the replies.

 

[georgeverghese]

Adding to the comments from @crshears and @rotw:

Suction side P / T measurements are required to make the following corrections:
a)The flow readout on the suction side may require corrections for operating temp / pressure / mol wt with reference to the design case P/T/ mol wt used for setting up the FE. Mol wt adjustment signal for flow readout is obtained from (b)
b)To get the corrected operating point on the polytropic head vs corrected acfm curve, adjustments must be made for mol wt.

To get the mol wt, an algorithm is built into the ASC to roughly estimate this based on the readouts for P and T on both suction and discharge.

Obviously, mol wt adjustment is not required if the gas feed chemical composition does not vary for all operating cases, including full recycle at max discharge press at the lowest discharge HX coolant feed temp.

 

[Indy]

Thanks again. I just have one more query. Could you point me in the direction to where I can find the equations that are used for the following:

“To determine flow and head of operating point, are needed Press.,Temp. at inlet and discharge and flow; Thus the instrumentation set-up” .

Thanks

 

[rotw]

I recommend this paper

https://proceedings.asmedigitalcollection.asme.org/proceeding.aspx?articleid=2194519

Edit:
Added also

https://pdfs.semanticscholar.org/7302/59768127200f2f68ce94377c090fadbb145a.pdf

 

[Sawsan311]

The measurements of the suction and discharge temperature are required for calculating the polytropic n exponent which is embedded in the reduced head equation which is part of the invariant coordinates anti-surge controller design. However, care shall be taken for the changes in gas composition which affects the k value (specific heat ratio)embedded in the polytropic exponent. The reduced head /reduced flow relation holds when the ratio of discharge and suction compressibility remains constant which is one of the major assumptions in developing the invariant system along with the need to update the gas composition effect on n value if it is expected to change drastically. For this reason, Engineers preparing compressor's process data sheets shall cover a wide range of gas composition and molecular weight in order for the vendor and the controller designer to be able to cover a wide range of Z compressibility and n exponent. The polytropic efficiency also changes with the gas composition.

 

[Phani Y]

Hi all,

I have a doubt. I have compressors whose ASVs are continually 30% open for very long. Is it due to improper valve coefficient of ASV? I mean can this issue be resolved by setting proper Cv for ASV?

In case the information I provided is insufficient, please quote an example and answer the query.

 

[rotw]

- Possibly your process conditions / gas composition have shifted, and control system is RE-positioning the operating point in envelope; What is 30% opening as we do not even know if the valve is equal percentage or linear characteristics or whatsoever;
- Possibly your compressor has degraded and its operating map has changed (wear and tear, fouling, etc.);
- Possibly your valve/actuator is mechanically deteriorated and/or its controller is at fault;
- Possibly operator changed set point of anti-surge control system via a manual override without you knowing;
- Possibly duty required capacity controller (if there is one) to change set point of anti-surge controller so to work on continuous recycle;
- Possibly you have load sharing controller and master controller responsible for this behavior;
- Possibly the calibration (ranges) of your instrumentation is at fault;
...etc..

I am afraid we can go on and on, forever...

 

[Sawsan311]

Phani,please check the surge control margin assigned since the commissioning stage of the compressor. Higher surge control margin assigned during the early surge test of the compressor may have contributed in narrowing down the operating envelope and hence more recycling.

As stated by Rotw, your surge limit line may have shifted due to ageing of compressor or fouling. The configured surge limit line in the system was based on different gas conditions then the current actual conditions.For example,the design map considered heavier gas with a deflected MW, where the actual gas is lighter causing frequent recycling at high speeds if your machine is VSD. The same can be caused by your surge detection system which may have initiated the adaptive response of shifting your SCL further to the right because your compressor is actually surging due to the possible reasons explained by Rotw above.

I recommend to conduct a surge test to re-validate the surge signatures and the SLL again and re-assign the surge control margin based on the actual gas in process.

 

[rotw]

If you post the compressor operating map and the process parameters actually measured, I could take a quick look.

 

[georgeverghese]

30% open on the recycle valve is a lot of gas recycling - the driver will have to speed up to maintain required capacity, increasing utility power demand. Another guess would be that this maybe done deliberately in order to maintain heat extraction at the downstream waste heat recovery unit, in case this is a GT driver.

 

[Indy]

This might be a silly question in regards to my original post. There is a compressor inlet pressure transmitter measuring in mBar, an outlet pressure transmitter measuring in Bar and an outlet flow meter measuring in mbar. Why is the inlet P and outlet F measured in mBar while the outlet P is in Bar?

Thanks

 

[crshears]

Gauge or PT units are typically chosen on the basis of expected operating range; use too small a unit and either the needle on the gauge will go over scale or the PT will hit its limit and no longer transmit proportionally to its measured variable; choose too large a unit and the accuracy of gauge readings, and the output signal accuracy of a PT will be sub-optimal.

You've got me as to why flow would be measured in mPa; perhaps you can explain to me how a flow of so many physical units per unit time can be measured in a force per unit area value.

Clearly I'm missing something...

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[rotw]

This might be a silly question in regards to my original post. There is a compressor inlet pressure transmitter measuring in mBar, an outlet pressure transmitter measuring in Bar and an outlet flow meter measuring in mbar. Why is the inlet P and outlet F measured in mBar while the outlet P is in Bar?

The flow is measured (most often) by means of an orifice device (plate) or Venturi. These devices when installed on the flow line create a pressure drop which can be correlated to the flow. The pressure drop is generally relatively small so it is often measured in kPa or mbar. Based on the calibration curves of the orifice, the measurement of the pressure differential, calculation method and compensation, it is possible to infer the flow. For purpose of anti-surge control, it is not really required to know the flow as defined in classical physical sense as a volume or mass per unit of time (except of course for display and monitoring purpose). In fact, the invariant parameters (reduced flow) can be worked out directly from pressure differential. The two papers quoted above give sufficient the details about this.

 

[crshears]

Meaning it is not actually flow being measured but the delta P between the upstream line & the point of minimum pressure due to venture constriction? In that case the correlation is roughly exponential, IIRC...

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Indy]

Thanks. There is an orifice plate on the outlet flow measurement (from the drawing) but ( from the drawing) it appears to be just a pressure transmitter on the inlet. Not sure why this is different from the outlet ?

Thanks

 

[rotw]

Meaning it is not actually flow being measured but the delta P between the upstream line & the point of minimum pressure due to venture constriction? In that case the correlation is roughly exponential, IIRC...

Losses are quadratic function of flow.

Thanks. There is an orifice plate on the outlet flow measurement (from the drawing) but ( from the drawing) it appears to be just a pressure transmitter on the inlet. Not sure why this is different from the outlet ?

In some cases, the flow transmitter can be located at the outlet. As standard, such measurement shall be compensated by pressure and temperature sensors at outlet. On the inlet there should still be a pressure transmitter. In fact, exact configuration depends which variant / set up is opted for anti-surge control.

 

[georgeverghese]

Flow measurement with an orifice plate is calculated from the raw readout obtained with a field mounted dp (differential pressure) cell - these have 2 taps on the line, while a pressure transmitter has only 1 tap. ISA symbology for a dp cell in flow measurement service is typically FT.