ONE QUESTION: regarding compressor 1

[procesdk]

Hi All
There were for around one week ago a discussion regarding inlet suction temperature vs horsepower of compressor.
It was a very interesting discussion and i want to ask how i can find this discussion again. I have tried to search for it without luck....so if anyone could help me finding the discussion or forward it to me I would be very veryhappy

Thanx a lot

procesdk

 

[25362]

I remember having participated in that thread but I cannot find it myself. My viewpoint was that the compression work of a piston reciprocating compressor is insensitive to temperature or density changes in the suction, because being the compressor a constant displacing-volume machine, its work of compression is proportional to pV, for a given gas and equal pressure ratio, and no liquids carried over.

My viewpoint was that mass or density variations due to temperature changes wouldn't affect the work of compression, since T appears in the formula for work of compression multiplied by the mass of gas. Thus, any reduction in temperature would result in a mass increase inversely proportional to the T decrease, the changes actually cancelling themselves out.

Of course, cooling the suction would result in a cooler discharge, that may, or may not reflect, on the intercooler's performance. Anyway, an increased mass flow coming from the first stage, better cooled in the intercooler would, would leave the work of compression practically unchanged.

I think in that thread most experts opposed my position. What is yours ?

 

[jay165]

I wonder what the opposition's argument is - since you're correct. Changing the temperature of the gas and holding all other variables constant does not change the work of a recip compressor. Case closed.

 

[RotaryG]

Depends on what type of compressor you are talking about and what parameters your process is controlling.

For ANY compressor decreasing the temperature would INCREASE the density of the fluid that is entering the compressor. If the volume that is swallowed is the same, as the density is increased so is the mass that is compressed and so is the power. At lower temperatures, it could be more efficient compression – but that will not compensate for increased power due to increased mass flow rate.

Then:

if it is a positive displacement compressor, and if the discharge pressure is maintained the same, then decreased suction temperature would require LESS amount of head to be generated to reach the same discharge pressure as before. Hence the mass factor and head factor would cancel out – thus probably giving you a slight increase in efficiency in compression as the temperatures are lowered.


However, a centrifugal compressor simply generates a predetermined head at a given inlet flow rate. So, in this case as you decrease the temperature and there are no other controls you would be pushing MORE mass while generating MORE pressure (for same head). If system is unable to provide that pressure at the discharge of the compressor then the flow arte need to INCREASE to drop the pressure. Hence in any event as the temperature is dropped, in a CF compressor you would be requiring MORE power.

Hope that the above is clear enough.




Regards,

Guru

 

[25362]

A centrifugal compressor, much like a centrifugal pump, is a dynamic machine in the sense that it provides energy to increase the velocity values of the fluid inside the machine in excess to those at the discharge; the reduction in velocity taking place within the machine is then converted into head.

The theoretical work of compression formula is the same as that given for reciprocating compressors:

kW = (1/9806)*[k/(k-1)]*WRT[sub]1[/sub]*[(P[sub]2[/sub]/P[sub]1[/sub])[sup](k-1)/k[/sup]-1]

W is mass rate, kg/s; P[sub]1,2[/sub] suction and discharge pressures, kPa; R, is the gas constant, 8314/MW of gas, J/(kg.K)

So, what's the change in work of compression coming from ? The answer: from the pressure P[sub]2[/sub] developed.

One cannot change the polytropic head developed by a given centrifugal compressor (CC), rotating at a given speed, by altering the physical properties of the gas being compressed. However one can change the pressure !

By cooling the suction the vapour density is indeed increased. Thus the differential pressure developed by the CC, which is about proportional to vapour density multiplied by the polytropic head, would increase.

Whether the CC is provided with spillback controls to protect it from surge or not, the pressure ratio P[sub]2[/sub]/P[sub]1[/sub], on a suction temperature drop, increases and so does the needed work of compression.

 

[Scipio]

Actually I remember that one too, and was one of the guys opposing 25362's point of view. I just haven't had the time to refresh my thermodynamics to figure out a logical counter-argument My opinion was that, volumetric flowrates and pressure ratios held constant, decreasing suction temperature resulted in a lower horsepower demand. At least for recips, I've never worked on centrifugal comps, so no clue what would happen there. This is based on computer models (with compressor vendor software, not a home-built spreadsheet), and I even saw the reverse effect on a compressor installed in the field since that original discussion - actual suction temperature was higher than designed for, and the compressor didn't have enough horsepower to move the design volume (not mass) of gas.

Personally I'm of the belief that the change in compressibility is the source of the change in required power, more than change in gas density, as it effects the volumetric efficiency of each piston stroke, but until I've had the time to go back through the basics I have to admit I couldn't argue the point with math to save my life!

 

[25362]

to Scipio, please tell us whether your field experience was on a multistage unit.

 

[Montemayor]

jay165 & All:

Since I was the original respondent to the query that suggested lowering suction temperature in order to lower power consumption, I feel obligated to follow through and explain how and why I opposed this erroneous engineering statement. Additionally, I find myself bored in what is turning out to be a dull weekend.

Many years ago, in the 1960’s, I actively designed (process-wise) and specified reciprocating compressors for Carbon Dioxide, air, Ammonia, Nitrous Oxide, Nitrogen, Hydrogen, and even Oxygen when I worked in the Industrial gas industry. All of these compressors were multi-stage, horizontal units with most of them having 3 to 5 stages. I also retrofitted and modified many others in the field for various gas services and had to do calculations to justify and backup the change of service – especially with regards to required horsepower and cooling needs. The last ratings I did on reciprocating units were on Carbon Dioxide service in 1988 and on Hydrogen in 2001. The CO2 service involved suction temperatures that ranged down to -70 oF.

I was originally trained and instructed in calculating brake horsepower requirements for CO2 service by using an isentropic process on the T-S diagram which we developed at Liquid Carbonic Corp at that time. Calculations for air and the other gases were a bit more difficult and depended on empirical and other standard published equations. We always found the T-S diagram calculations more accurate than the equations, but years ago T-S diagrams for the other gases were scarce and hard to come by – especially in accuracy. There were no such things as simulators or databases and we relied on the basic slide rule and the enthalpy values in our T-S diagram to calculate accurate estimates. Today, it is not surprising to read the latest 11th Edition, electronic version of the GPSA Engineering Databook (Page 13-2) and find that “the enthalpy change is the best way of evaluating the work of compression. If a P-H diagram is available, the work of compression would always be evaluated by the enthalpy change of the gas in going from suction to discharge conditions.” This is not surprising since most all knowledgeable authors have always expressed similar comments in discussing reciprocating compressors. Field results and measurements have continuously shown that reciprocating compressors closely follow an isentropic compression cycle as traced on a T-S diagram. This follows basic thermodynamic theory.

Having explained my basis for calculating reciprocating compressor horsepower, I will cite a hypothetical case of a single-stage CO2 compressor to illustrate the effect of suction gas density on the power requirements:

Gas compressed = CO2
Volume rate compressed = 100 acfm (at suction conditions)
Suction Pressure = 5 psig
Discharge Pressure = 75 psig
Four different suction temperatures will be analyzed, keeping the above values constant: 100 oF, 60 oF, 20 oF, and -20 oF. NIST enthalpy values will be used for CO2 gas.

100 oF Case:
Density = 0.14518 lb/ft3
Cp/Cv = 1.28934
Mass flowrate = 871.08 lb/hr
Discharge Temp = 309 oF
Enthalpy Change = (267.05 – 222.2) btu/lb = 39,067.9 btu/hr

60 oF Case:
Density = 0.15661 lb/ft3
Cp/Cv = 1.3013
Mass flowrate = 939.66 lb/hr
Discharge Temp = 260 oF
Enthalpy Change = (255.75 – 214.03) btu/lb = 39,202.6 btu/hr

20 oF Case:
Density = 0.17006 lb/ft3
Cp/Cv = 1.31530
Mass flowrate = 1,020.36 lb/hr
Discharge Temp = 211 oF
Enthalpy Change = (244.68 – 206.06) btu/lb = 39,406.3 btu/hr

-20 oF Case:
Density = 0.18615 lb/ft3
Cp/Cv = 1.33194
Mass flowrate = 1,116.9 lb/hr
Discharge Temp = 162 oF
Enthalpy Change = (233.83 – 198.27) btu/lb = 39,716.9 btu/hr

As can be seen by the summarized results, the horsepower indeed does increase in accordance with the increased gas density. In the process of generating the process heat and material balance for a Dry Ice production plant (a US patent that I presently hold), I have not only calculated the effects expected on the brake horsepower demand, but I have also personally witnessed the corroborating field data that demonstrated the accuracy of the horsepower predictions as the temperatures were decreased in the flash CO2 recovery gas. The Dry Ice production cycle is a classic example of the effect of low suction temperatures on reciprocating compressors since the flash gas to be recovered is generated at temperatures that vary between -70 oF and -110 oF. I control the suction temperature with gas-gas heat exchangers.

I hope the above serves to explain in clear detail how the effects of lowering the suction temperature to a reciprocating compressor will increase the power demand – not decrease it.


Art Montemayor
Spring, TX

 

[Scipio]

25362,
Two stage machine, sour natural gas. *chuckle* I think I see a few hours with my head bent over texts on fundamental compression and thermodynamics before I can convince myself on this one now

 

[25362]

One cannot and shouldn't argue against facts: Montemayor's rich experience in compression of many gases, especially when he says that practical experience has corroborated the fact that enthalpy differences on isentropic compression reflect the work of compression. The opposite Scipio's experience that on cooling a sour natural gas (mixture) the work of compression was actually reduced, if true, should be analysed.

Refinery people, and I counted myself among them, think that compression work in piston reciprocating compressors is quite insensitive to relatively small changes in suction temperature, or density, or MW of gases, keeping all other factors constant.

This is based on the assumption that real gases behave in practice as ideal gases. It is true that temperature and density changes cancel out in narrow ranges, but this is not the case with regard to k=Cp/Cv. One should remember that the theoretical adiabatic work of compression is proportional to
[k/(k-1)]*(r[sup](k-1)/k[/sup]-1). (r is pressure ratio).

Most real gases show an increase in k with a drop in temperature. This fact alone, using the above formula, would explain the small increase in work of compression when cooling the suction.

The change of k values for air or nitrogen at atmospheric pressure below 300 K is quite small thus, based on this fact, the increase in work of compression by pre-cooling them would be indeed small.

Hydrogen and helium have a peculiar behaviour, helium doesn't change its k value on cooling at atmospheric pressure, hydrogen's k drops with temperature, a fact that would result in a reduction of power on cooling the suction. (Art Montemayor: this happens for 100 and 60 deg F, compressing pure hydrogen between 300 psia and 750 psia, even when using enthalpies and mass flow rates following your method, please comment).

Other indirect factors may affect power consumption. Let's mention among them, valve flutter, oil viscosity, and valve leakage. Would a lower temperature affect them, and as a result, change the compression work? Would compressing a cooler natural gas affect the lube oil's viscosity by dissolving light hydrocarbons, the friction at the compressor moving parts would drop, thus reducing power consumption? I wonder.

We are dealing with real-life machines. The actual compressor's internal pressures aren't those measured upstream or downstream, but lower and higher, respectively. The expansion and compression steps aren't truely adiabatic.
If, for example, the compressor is partly unloaded by a head-end clearance pocket, what would be the real temperature of the gas undergoing compression?

The estimates presented by Montemayor show small percentual increases -though not negligible- in power consumption on cooling the suction to CO[sub]2[/sub] reciprocating compressors. The smallness of these variations may be one of the reasons for the generally accepted viewpoint held by refinery people that the work of compression (mainly for for air and hydrogen-rich gases) in reciprocating units is practically insensitive to suction vapours' density or MW while all other process parameters are kept constant.

It would still be of much interest to hear other opinions.

 

[Scipio]

Actually I think Montemayor and I are arguing a similar point from different angles.

His cases are based on maintaining fixed suction & discharge pressures and actual volumetric flowrate with falling suction pressure. Under these conditions, a falling temperature results in a higher mass flow rate, and consequently a higher horsepower demand.

What I was thinking, is that if we were holding the mass flow rate, rather than actual volumetric flow rate, constant, less horsepower would be required at lower temperature.

 

[25362]

It is apparent that when pressure and volume displacement are constant, the way to keep mass flow rate constant when dropping temperatures at the suction side of a piston reciprocating compressor is by reducing the molecular mass of the gas.
This means increasing the number of moles flowing, since pV is about proportional to nRT, where n, the number of moles, equals the mass flow rate divided by the molecular mass.

 

[Scipio]

Actually I was thinking more along conventional lines of slowing down the compressor or opening variable volume pockets to maintain a fixed mass flow rate.

 

[25362]

Those could indeed be ways of reducing the work of compression. I seem to remember that Montemayor said once he likes the variable pocket unloading method to regulate the compressor's operation.

 

[25362]

I came back to this thread because it seemed to me that when the questioner referred to cooling the suction of a hydrogen-rich gas reciprocating compressor he didn't specify where is this cooling done.

If the cooling is carried out just ahead of the compressor the same gas may become denser as everybody said.

But if the cooling is done ahead of the gas-and-heavier separating vessel, the gas may become lighter (of lower MW) and the conclusions on the compressor's power consumption may differ.