Effect of Molecular Weight On Compressor Performance 3

[Sawsan311]

Dear All,

I have extensively studying the analysis of the impact of Molecular weight changes on fixed speed and variable speed compressors. Can you provide your feedback on the below simplified summary:

1) Fixed Speed Compressor, fixed compression ratio, speed and head at every flow:

a) With a heavier MW, the mandated head to achieve the desired compression ratio which is fixed would be less based on the polytropic head equation and hence the available head being more will enable driving more gas inside the compressor so operating point will shift to the right along the head/Q curve and can in fact reach stonewall as induced by higher mach number due to less sonic velocity by higher MW. Power consumption increases due to more mass flow.

b) With lower MW, there will be more required head to be developed to achieve the mandated compression ratio, hence the available head would be too low that the flow rate will decrease moving to the left on the H-Q curve with less flow going in the compressor.Power consumption decreases until anti-surge controller activates and opens the anti-surge valve.


Please note for the motor sizing case of new compressor, the governing scenario would be low MW start-up case with fuel gas due to the extremely high head high head required to build up the speed and the discharge pressure and accordingly,the power consumption would be more. Do you agree?

However, for the existing compressor we can see if MW decreases, the fixed speed (fixed head curve by affinity laws) is too low to achieve the constant compression ratio and hence less flow would go in and the point starts shifting to the left along the head curve approaching surge. In other heads, required head would be more and hence less flow goes into the compressor.

The action of the VSD compressors that it automatically reacts to the demanded head for the various MW whilst maintaining the same compression ratio by the performance controller and the integrated anti-surge controller. This in return helps in optimizing the capacity control and power consumption. For a VSD compressor operating under suction pressure primary control, and if MW reduces drastically, the original operating point finds it self at the high head which may cause a reduction in the suction pressure and consequently activating the performance controller to reduce the speed to maintain the suction pressure such that the recycle operation is eliminated/minimized.

Appreciate your views on the above interpretation of the behavior of the fixed speed/ VSD benefits with respect to the impacts of the gas molecular weight changes. Again, during the design stage we generate various cases of gas molecular weight in order to ensure compressor design is covering the envelopes with respect to speed/head/power and accordingly assessing the response of the performance and anti-surge controller for the various cases.

 

[georgeverghese]

For case (a), agreed mass flow increases, but polytropic head is lower, so this should be taken into account in the power demand
Conversely, in case (b), though mass flow is lower, polytropic head to be developed is higher.

"For a VSD compressor operating under suction pressure primary control, and if MW reduces drastically, the original operating point finds it self at the high head which may cause a reduction in the suction pressure and consequently activating the performance controller to reduce the speed to maintain the suction pressure such that the recycle operation is eliminated/minimized. "

In the case of VSD operation, if feed gas mol wt drops, compressor would speed UP to develop more polytropic head in order to maintain required dp across the stage. If operating point moves too far to the left such that it approaches the surge line, the antisurge controller would kick in to recycle gas in an attempt to move the operating point away from surge at the higher speed.

 

[Sawsan311]

Thank you very much georgeverghese.. Our analysis is aligned.

For VSD machines, there can be some pre-determined limit loops of the performance speed controllers which can help in delaying the intervention of the anti-surge controller through speed control i.e. reducing the speed when the suction pressure limit loop is reached, the set point of this limit loop shall be higher than the one activating the anti-surge controller and of course higher than the low suction pressure trip.

For the impact of the MW on power during normal operation, for the same flow, compression ratio for a VSD machine, we can see that the impact on power may cancel out by the combination of high head and low mass flow in case of low MW and the converse goes to MW. This is the main feature of VSD control which optimizes power consumption and absorbs the variations in gas composition.

Do you agree that for a new machine design, the governing case for motor sizing would be however the low MW case (start-up fuel gas case) as the head required is high along with the required flow to achieve the minimum acceptable operating speed?

Thanks

Regards,

 

[georgeverghese]

Yes, the performance / capacity loop controller should kick in almost all times to maintain suction pressure by first modulating speed, and if that doesnt bring the operating point close enough to the setpoint, secondly by modulating the recycle valve.

The effect of polytropic head change always acts in the opposition to mass flow change ( due to changes in feedgas mol wt), so that the net effect on power demand is often minimal - the polytropic eff change also plays a part in the power demand. In your case, the power demand at startup with light fuel gas will depend on where the new operating point lies on the speed-Q-h-eff curve for this compressor - I think it may not be safe to make generalisations about this and instead, extract the results from the compressor curves.

 

[Sawsan311]

Thank you Mr.georgeverghese

For the effect on polytropic efficiency, I could see from an actual compressor curve the the polytropic efficiency tends to gradually decline with increase in speed of the VSD machine. How can we interpret this behavior and is it correct?

 

[georgeverghese]

This may not always be the case - it can also increase - depends on the locations of the points where you are trending polytropic eff. In any case, the trends for efficency are are similar to those for a VSD pump.

 

[Iomcube]

A good presentation on this subject (highly targeted):

https://www.slideshare.net/sudhindr...ressor-head-impact-of-mw-and-other-parameters

 

[rotw]

Reference to the link, I think the reasoning on peripheral Mach number is more adequate than Molecular weight to explain the impact on the shape of performance curves.
A lighter gas may produce similar effect on curves (reduction of range, curve steepening effect) to that of heavy gas if both are associated to conditions with equivalent Mach numbers.

 

[Sawsan311]

Thank you very much Iomcube for the great slides and thank you Rotw for your feedback..

However, I would like to share one of my observations on the slides with you and rotw,

There is contradiction between slide 8 and 13 since higher MW produces more relative velocity and less tangential velocity so less head for the same impeller speed. I think slide 13 meant compression ratio instead of head on the Y- axis as the head/speed/flow are fixed so what can change is the compression ratio. However, going through the shape of the curve, the slope of the head curve would be less in case of the heavy MW case as mentioned above. Additionally,lighter MW curve is flatter compared to the steep heavy MW curve which makes them more sensitive to changes in flow / differential pressure and more prone to experience surge.
Compressor control vendor develop non-dimensional compressor curves with reduced head and reduce flow such that the effects of changes of gas MW would be normalized to have one overlapping surge limit line among all MW cases. In absence of this non-dimensional curve, we can demonstrate that on P-Discharge-Q the heavier MW will have more elevated surge limit lines than compared to less MW and this is driven by the fact that for the same head/flow point, produced discharge pressure is higher with heavier gas. The design of the anti-surge controller will based on measuring the surge margin based on the difference between the inlet mach number and the mach number at the Surge limit line i.e.how fast the operating point is reaching surge.
For slide-14 can you please explain why the speed curve for the heaviest gas is shown to give higher head than the lighter gases, I think the y-axis should be discharge pressure and not polytropic head for example, see ACFM of 19000.
Note:based on actual compressor curves I worked with, high MW can affect the shape of the surge limit line at high speed ranges for VSD machine as well making forward curvature of the SLL more than light MW and accordingly can contribute in narrowing down the envelope at high speed ranges (despite the non-dimensional SLL generated by reduced head/flow analysis). The above can also be demonstrated by slide 14 where the black line representing the surge point of the heaviest gas to be to the right with respect to the lighter gases.

 

[rotw]

I strongly suggest you to get and read this article by J.R. Gaston :
"Turbocompressor surge control, new solution for an old problem", ASME 92-GT-428, 1992.

Some remarks:
- Flat curves are not a problem for anti-surge control (there are refinery/petrochemicals processes with H2 recycle gas compressors not equipped with anti-surge valves at all; this can be for process safety reasons like being vital to have forward flow, still these machines do work). The pressure ratio rise to surge however implies that a small change in downstream system resistance would mean a high change in flow which may or may not be an issue for the process. In some instances, a minimum pressure ratio rise to surge is required by the specifier.
- A distinction exists between single stage and multistage compressors. On a multistage compressor the stages which initiates surge and choke can change according to the conditions of the gas (temperature, MW, speed) and this defines the final shape of the maps.
- One overlapping surge line among all MW cases can be done but is never 100% perfect. Changes of inlet isentropic coefficient (k), for example, may have effect on accuracy. Two gases with same MW but different compounds composition could have their respective surge lines differ due to this consideration, again as an example. But usually the magnitude of errors can be accommodated or circumvented by proper control strategies.

 

[Sawsan311]

Thanks Rotw, I agree with you that a non-dimensional SLL for various MW is not always overlapping, in fact as stated in my post above, higher MW was found to have some curvature in the SLL at some points and decline in slope further at high speeds. I can interpret this effect by lower specific head (as stated by you) as well as higher mach number reached by the heavier gases and accordingly narrowing down the available envelope specially at high speeds.
However, I still want to check slides 8 and 13 in the slide share shared by Iomcube as I think the y-axis should not be polytropic head but compression ratio .

regards,

 

[georgeverghese]

Its better to work with compressor manufacturer generated plots showing polytropic head vs flow at various speeds, rather than with isentropic head for centrifugal compressors. Isentropic head developed is dependent on mol wt , while polytropic head isnt. The plots in the slides show isentropic head.

 

[Sawsan311]

Hi Rotw,thank you for your response,
I was pointing out to the effect of specific heat (typo error) which contribute in the estimation of the polytropic exponent and accordingly the slope of the surge limit line.

I think you are pointing out to the invariant coordinate system where the reduced polytropic head is forced to be independent of the gas composition and where these invariant coordinate curves of reduced head versus reduced flow enable the design of the anti-surge control system as well as the SLL in a way which is reflected in terms of measurable variables (n exponent). However, still various cases of MW are required to be studied when the gas composition changes and hence affecting the value of n which is reflected by the online measurement of suction and discharge temperature and suction and discharge pressure.

The polytropic head in actual compressor curves do change with MW i.e. the higher MW mandates less head required in comparison with the lower MW which requires more head and accordingly the VSD compressor speeds up as per our discussion in the earlier posts.

I think the slides refer to compression ratio as in side 13 it states that high MW gives more tangential velocity and more head which is totally opposite to slide 8

Regards,

 

[rotw]

QUOTED
I can interpret this effect by lower specific head (as stated by you) as well as higher mach number reached by the heavier gases and accordingly narrowing down the available envelope specially at high speeds.
UNQUOTED

Well. To make it a bit gross, when gas MW increases (prefer to think in terms of Mach number but lets stick to MW for simplification), density increases, and when density increases, a % volume flow shift from the design point in the first impeller, is more quickly 'amplified' when flow is moving to the next impeller (the mismatch is even greater). So forth and so on. At some point, one of the impellers (you could not say upfront which one) is going to surge (we can say same about the choking line). All-tough all the other impellers are still operating OK that very impeller is going to 'LOCK' the overall map limits (left side or right side). This is why the surge line evolves quite 'naturally' (parabolic shaped at low Mach numbers/ lower speeds) until it gets abruptly (not smoothly) deflected, and then goes on deflected again by whichever next impeller surging to the point where the surge line becomes virtually almost flat horizontal.
That is the story behind the narrowing down of the envelope at high speeds. Also keep in mind that underlying this behavior is very complex problem, because not all impellers within a machine are of same design (and they are no more off the shelf nowadays), geometry, etc, so its good to understand the bottom line but the best thing to do when there is an inquiry, is clearly to contact the compressor manufacturer.

 

[Sawsan311]

I would like to thank you Rotw for the great explanation,

I am also looking into the impact of MW(mach number) on the shape of the compressor curve, the presentation shared by Iamcube and I read in another article that high MW gases have flat curves. However, I am trying to interpret this with respect to the slope of the tangent line to the curve being of low slope of change in head per unit change in flow. However,I believe this is valid at low speed because as explained by you above there will be deflection in the surge limit line. In other words,deflection and forward leaning of the surge line leads to steeper curves at high speed.

I also read that the invariant coordinates used in the design of the anti-surge control system based on the estimation of the surge parameter Ss= slope of the operating point line to the slope of the SLL. The slope of the operating point would higher for low MW where the head would be high due to high specific heat and accordingly low MW gases tend to enter surge earlier than heavy gases due to the increase in the head.
What do you think?

Thanks alot and Regards,

 

[rotw]

The surge parameter you mention, if I understand correctly is not a method commonly used for centrifugals. While what I was referring to is more common method that is based on the monitoring of the deviation (actual flow vs. surge flow).

In certain applications, especially axial compressors, the operating range is so narrow that if you count 10% standard control line you have already exceeded the choking line. So control is based on monitoring the angle between a line connecting the operating point to the origin and the surge limit line. This technique makes sense when you want to track an operating point that is moving on a narrow operating range in resoect to flow but a much wider range in respect to the head. Steepness of the curves is more a specifics of the axial type of compressor.
Usually axials are applied in Air blower and LNG Refrigerants applications (not limited to but most encountered applications); means the anti-surge control applications are more or less chartered.