3-stage (4 compressor) cascade using Methane

[Kevin Hotton]

Hi Xtremer's

I'm having fun designing AutoCascade and Cascade cycles. Gary Mole got me interested to see if I could find a cascade solution for Methane using only 3-stages. I have, but it needs 4 compressors (They should be 410a rotary compressors for the higher pressures needed).

The cooling duty chosen is 350W. The average evaporator temperature is -123C (-133.5C to -112.7C). You need to uses two rotary compressors of at least 10.8 cc/rev (good choice Rechi model 44A281B 11.4cc/rev) and two of at least 4.5 cc/rev (good choice Rechi 39A111A 4.75cc/rev) both these compressors have good reserve motor power output and are able to tolerate 34-bar discharge pressures reached on ASHRAE/T standard AC test-cycle.

The refrigerants used were: 410a on warmest stage, 0.90 ethylene 0.10 Methane (mole fractions) on middle stage and 0.90 Methane 0.10 Ethylyene on the lowest temperature stage. Study the diagram of the system and you can see how I exchange heat between the lower two stages.

It was very difficult to bridge the large boiling point gap between Methane and Ethylene, but this cycle arrangement is about the best solution available (I think).

Hoping Gary might just try and build this.

Regards,
Kevin

[Patrickclouds]

hi,

what is that Q-cooler on the ethen stage?
is that a suction line heat exchanger of the 3rd stage(0,9methane, 0,1 ethene)?
will that work? if the 3rd stage isn't running you'll heat the refrigerant instead of cooling when the second stage gas flows trough it.
or did i get something wrong in your drawing?

best regards patrick

[Jack]

It looks good, although you have to 'cheat' somewhat... you are autocascading the 2nd stage to get enough low temp. cooling capacity.
The use of 2 compressors in the 2nd stage is interesting though. Will one bigger compressor not suffice?

And one other observation, the Q-preCond in the 3rd stage, is that an aircooled desuperheater (since its warmth output is not connected to anything)?
Won't the Q-preCond be under room temperature? I see your Material Streams chart says 87.96* C at position 2m, but that seems rather high?

[Patrickclouds]

you get lower pressure ratio when you use two compressors in series. think on the mixture of r1150 and r50. that causes big pressure ratio. too much for one compressor only.

he calculates with a high side pressure of about 32,x bar in the second stage. and low side pressure of 1,5bar.
i think that are the absolute pressures and would cause a pressure ratio of 32/1,5 = 21,x

q-precond seems like a air cooled desuperheater.
position 2m is hot gas temperature. 87,x°C is low
rotarys are build for much higher discharge temperatures of 100°C to 115°C

i don't understand why to use r410 in the first stage. r410 has a high pressure when condensing and you also need a "high" low pressure to increase pressure ratio. you 've choosen 2,7bar and thats only a evaporation temp of -30°C. if you use r404a or r507 you have a lower high pressure and with a strong compressor you can easy evaporate at -40°C. and now that decreases the high pressure in the second stage a lot

[Jack]

Ah yes, I overlooked the 33 bar second stage discharge pressure at position 4. A pressure ratio of 22 might be a bit too much

Suction temp at 1m is -65* C and discharge temp at 2m is 87.96* C, which is a DeltaT of 152.96* C. Is that normal, such a big temp difference? This is why I asked...

[Kevin Hotton]

Ah yes, I overlooked the 33 bar second stage discharge pressure at position 4. A pressure ratio of 22 might be a bit too much

Suction temp at 1m is -65* C and discharge temp at 2m is 87.96* C, which is a DeltaT of 152.96* C. Is that normal, such a big temp difference? This is why I asked...

It is with a Pressure-Ratio of 6 on the methane compressor.

Kevin

[Kevin Hotton]

The more I think about this pseudo-3-stage cascade, the less I like it. If I am needing 4 compressors already, why not just build a conventional 4-stage cascade.

Stage-1 R404a
Stage-2 R23 or Ethane
Stage-3 Ethylene / Methane blend (maybe 0.67 / 0.33 mole fractions -- just a guess)
Stage-4 Methane

This should eliminate most of the +30-bar discharge pressures, thought still likely needing 28-bar (or so pressures).

Kevin

[Patrickclouds]

why don't build a normal three stage but an autocascade in the third stage:

1st stage r404a t0:-38°C
2nd stage r1150 t0:-90°C
3rd stage autocascade with mixture of 0.67/0.33 ethylen/methane blend. if you calculate with a delta T between evaporation temp of 2nd stage and condensation temp of 3rd stage with 10K you'll get a high side pressure of something like 17bara. and the charging of only 2 refrigerants will it make hoepfully much easier than tuning an autocascade with 5gases.

[wdrzal]

Have any of you guys tried using Two stage Compression ?

[Patrickclouds]

yeah in the 5 stage cascade of unrockstar. it works quiet nice, but i don't want 35 bar on those chineese plate hx

[[XC] gomeler]

Have any of you guys tried using Two stage Compression ?

Thought about it, I've seen a few "fallen off the back of the truck" compressors on eBay but I'm still not fond of such high discharge pressures. Haven't had a pressure failure of a joint/component and I'd like to keep it that way.

[wdrzal]

Two stage compression doesn't necessarily mean higher discharge pressures. It does help to achieve a higher mass flow which is needed for the lighter less dense colder gases .

[Kevin Hotton]

why don't build a normal three stage but an autocascade in the third stage:

1st stage r404a t0:-38°C
2nd stage r1150 t0:-90°C
3rd stage autocascade with mixture of 0.67/0.33 ethylen/methane blend. if you calculate with a delta T between evaporation temp of 2nd stage and condensation temp of 3rd stage with 10K you'll get a high side pressure of something like 17bara. and the charging of only 2 refrigerants will it make hoepfully much easier than tuning an autocascade with 5gases.

Hi Patrick,

I am new to autocascade, but when I simulate them this is the behavior I see. The simpler the autocascade (1-phase-separator) the more temperature "glide". So with R1150/R50 autoC the inlet to the evaporator shows -140C but the outlet is -110C. It does solve the high discharge pressure though.

Kevin

[mytekcontrols]

Just a quick thought to what Patrickclouds said about difficulty in tuning an AutoC due to so many gases. Lately I have been using a new method based completely on weight. In the past I would only weigh in the liquids, and then add the gases using an additive pressure process. As I have now been doing charging of all components by weight, I am discovering that each component charged ends up being nearly the same weight at the end of the tuning process. Of course you still need to figure out what your total weight should be for a given sized system. But given experience with building other systems, and then proportioning up or down from this depending on the need, creating a new charge becomes much more easy. I don't know if this makes sense, or if I have sufficiently explained it, but it does seem to take some of the guess work out of the process.

Using this charging system does require a very accurate refrigerant scale, preferably one that has a resolution of at least 2 grams. And I have found it best to use the kilogram units provided on most electronic scales, having at least 3 decimal places to the right of the decimal point (grams).

Edit: All refrigerants being of equal weight does not apply to Argon if it is part of the charge. In most cases it will be a fraction of what you are using for the other refrigerants, and determined by how cold you need to go, as well as how high of discharge pressure you are willing to see.

After analyzing the Polycold charges, it appears that the R23 part of the charge is about 50% as much (by weight) as compared individually with the R123, R123, R14 (HCFC based charge). However the R123, R22, and R14 are very close to the same. This relationship changes when getting into Polycold's HFC charges, with the R236fa being twice the amount by weight as the R125. Although the R23 and the R14 are still nearly the same proportion. So my theory of all the refrigerants in an AutoC being charged at the same weight can be disregarded (must have been hallucinating when I said that). Of course dependent upon which family of refrigerants you are using (CFC, HCFC, HFC, HC), there are most likely repeatable percentages that could be applied irrelevant of system size.

As for needing a high resolution scale. This would only be required for very small volume systems.

[Kevin Hotton]

By weight is how I have been charging all my refrigerants. For the higher vapor pressure gases (ethane and up) I have to transfer from the heavy steel cylinder to a large (tare weight 60-lbs) refrigeration recovery cylinder. Then I need to charge this lighter cylinder (my scale is 50-kg 3-gram resolution) with more mass than I need and finally use the refrigerant recovery machine to charge the proper weight out of the recovery cylinder. The systems I am building are bigger (smallest so far used 12 m3/hr scroll) so total charge has been at least 1-lbm so 3-gram resolution seems adequate.

Kevin

[Kevin Hotton]

I have modelled an autocascade 3rd stage for a conventional 3-stage cascade system. It is very difficult to bridge the large vapor pressure gap between ethylene and methane without the use of R-14 -- this autocascade 3rd stage makes this possible. Average evaporator temperature is -144.6C with 325W duty (at less than 300W -150C easy). Study the schematic of the 3rd stage and the included tables of state-points and composition. I am using ethane, methane and argon. Compressor size needed is about 17.7 cc/rev rotary. I estimated compressor efficiencies as: volumetric 0.88; isentropic 0.70; electric motor 0.85

The first condenser is air-cooled or water-cooled (104C to 30C) the next is the auxilary condenser, and the third is the cascade heat-exchanger which transfer heat to the 2nd stage cascade evaporator (assumed ethylene and at about -90C).

Kevin

[mytekcontrols]

By weight is how I have been charging all my refrigerants. For the higher vapor pressure gases (ethane and up) I have to transfer from the heavy steel cylinder to a large (tare weight 60-lbs) refrigeration recovery cylinder. Then I need to charge this lighter cylinder (my scale is 50-kg 3-gram resolution) with more mass than I need and finally use the refrigerant recovery machine to charge the proper weight out of the recovery cylinder. The systems I am building are bigger (smallest so far used 12 m3/hr scroll) so total charge has been at least 1-lbm so 3-gram resolution seems adequate.

Kevin

Yes you are correct, in larger systems the resolution can be less. The method that we use for charging Polycold units, is to always use the transfer tank method, utilizing a 0.781 cubic foot (30 lb refrigerant recovery tank). When using this method, we initially charge all the refrigerants (gases and liquids) into the transfer tank and then either use the recovery pump process or the unit's compressor to transfer from the tank into the unit. Larger units require more than one transfer tank to accommodate the unit's volume requirements, so the charge is equally divided between how ever many tanks are required. The great thing about doing it this way, is that field recharges are easily accomplished.

[Kevin Hotton]

This is a suggested 3-stage cascade using all hydrocarbons. I have used a 300W duty target. The first stage is Propene (R404a would work great just raise suction pressure to 2.5 bar and discharge to 15-bar. I could not choose R404a (or blend it) in Hysys). The second stage is Ethane and the third stage is an autocascade design using iso-butane, ethane, and methane.

Of the two stage cascade designs I have seen on this forum nearly all use ethylene for the second stage. The emphasis is the lowest possible evaporator temperature. This goal results in very low suction pressure (sub-atmospheric), extreme pressure ratio (over 15) and small mass flow (little cooling duty).

For my design, I have limited pressure ratio to a maximum of 6. Minimum suction pressure is 2.4-bar (2.5-bar if R404a used). These higher suction pressures result in higher mass flows with smaller volumetric compressor. The cascade specifications are:

Stage 1:
R404a Suction pressure 2.5-bar abs. Evaporator temperature -25C
Discharge pressure 15-bar abs. Condensing temperature 32C
Pressure ratio 6.00
Compressor: Rotary, displacement 12.6 cc/rev (volumetric efficiency 0.83)
Isentropic efficiency 0.70; motor efficiency 0.87
Electric power input: 467 W

Stage 2:
Ethane Suction pressure 2.5-bar abs. Evaporator temperature -70C
Discharge pressure 15-bar abs. Condensing temperature -18C
Pressure ratio 6.00
Compressor: Rotary, displacement 10.0 cc/rev (volumetric efficiency 0.83)
Isentropic efficiency 0.70; motor efficiency 0.87
Electric power input: 405 W

Stage 3: 322W duty
Mixture composition at compressor inlet:
mole fractions; iso-Butane 0.0296, Ethane 0.3003, Methane 0.6701
mass fractions; iso-Butane 0.08, Ethane 0.42, Methane 0.50
Suction pressure 4.5-bar abs. Evaporator average temperature -130C
Discharge pressure 21-bar abs. Cascade condenser temperature -23C to -62C
Pressure ratio 4.67
Compressor: Rotary, displacement 11.0 cc/rev (volumetric efficiency 0.85)
Isentropic efficiency 0.70; motor efficiency 0.87
Electric power input: 945 W
(use R410a model for sufficient motor power reserve)

Note total electric power in 1817 W (less fans) COP = 322 / 1817 = 0.177

This COP is not as good at the 4 compressor 3-stg design (its COP = 0.231 at a slightly warmer -123C). However its highest pressure is 21-bar vs. 33-bar. It really is difficult to work around a lack of R-14. You pay a COP penalty. I also think a good full autocascade design is best if -130C is the goal.

Below is details of 3rd stage autocascade. The first two stages are conventional cascade (with precondenser on ethane stage).

Kevin.

[Jack]

Why do you need the iso-Butane in the third stage?

[Kevin Hotton]

The iso-butane with its high boiling temperature (-11C at 1 atm.) relative to methane (-161C) allows me to drop the compressor discharge pressure. I am investigating no iso-butane mixtures (just ethane and methane). My goal being the highest COP (coefficient of performance -- which is the cooling duty divided by the power input required to produce it).

Kevin

[Kevin Hotton]

To answer Patrickclouds earlier question.

Yes if the 3rd stage is not running the Q-cooler would heat the 2nd stage flow, but ultimately you will be running the 3rd stage -- it's where the final CPU load is. The main reason for the Q-cooler is for oil reasons. I did not want -112C suction temperature at the 3rd stage compressor inlet (might freeze the oil) so the Q-cooler warms it to about -65C (a temperature that on our large 2-stage cascade that I posted has not been a problem).

Kevin