Printable View
or use a pipe with a larger diameter outside the phase seperator.Quote:
Originally posted by Gary Lloyd
Also, if necessary, we could soft solder the common suction line along the side of the separator.
Actually I have never constructed an auto cascade so this is a learning experience for me..
Opps, I made a silly error in my sums. Used 2*Pi*r for area instead of Pi*r^2. Im getting tired. Changing that gives the propane an evaporator capacity of 20kW. Using an alternative method based on friction, wall stress and Reynold's number, i got an evaporator capacity of 900Watts. This is closer to the truth I believe.
Code:Roughness 0.0015 mm 0.0000015 meters
High Pressure 19.8 bara 1980000 Pa
Low Pressure 2.8 bara 280000 Pa
Viscosity 0.000125 kg/ms
Radius 0.0155 inch 0.0003937 meters
Density 1.9558 dm3/kg 511.2997239 kg/m3
Length 48 inches 1.2192 meters
Velocity 216.1248958 m/s
Volumetric flow 0.000105241 m3/s
Mass Flow 0.05380976 kg/s
Heat Load 20.12485035 kW Heat of Evap. 374 kJ/kg
Alternative Calculation
group 5568724.589
e/d 0.001905004
Re 33000
velocity 10.24596774 m/s
Volumetric 4.98923E-06 m3/s
Mass Flow 0.002550993 kg/s
Heat Load 0.954071333 kW Heat of Evap. 374 kJ/kg
What about the oil in the system and the two phase issue in the capillary tube. If you factor these in the capacity would be less. most of the oil flow will be through the R290 capillary.
EDIT: Got a Headache yet?
I though you might want to read this:
http://www.russellpetrie.plus.com/Do...alculation.pdf
Also, anyone have or want to buy or help buy this refrigerant property database? Its the NIST database of refrigerant properties and it costs $200, i'd be up for it depending on howmany others are interested. 4 or more maybe. Here is the link.
http://www.nist.gov/srd/nist23.htm
Id be interested in it if I actually ever get round to building my phase-change rig (who knows, if we can sort this autocascade I may build one too)
Still sounds to me like the phase-seperator is the main culprit here. The combination of spraying downwards, the R290 exit being too high and the turbulance being (possibly) too high inside the seperator all sounds like whats causing the bubbles in the sight glass.
Id go with what gary said, first get the seperator intake spraying against the opposite wall, lower the R290 exit and also (possibly) make it larger in diameter to try and slow down the refrigerant (and reduce turbulance)
Heres another thought....
Perhaps you should have the seperator inlet spraying "round" the inside of the seperator, so it tends to swirl round the seperator, this should not only help reduce turbulance (and less liquid R290 being picked up as a "mist" but also give it time to settle.
Also have you checked the R23 exit isnt partially blocked? So far we've chopped the cap tube shorter to increase flow, but what if the R23 exit was partially blocked, forcing some of the R23 gas through the R290 exit? (hence the bubbles in the sight glass)
(On a side note, all you UK types, anyone live in norfolk wanna help me build my rig? :))
Hi Russel,
That was one steep learning curve last night:slobber:
Guess your too tired to progress tonight. Good luck with the next steps mate:toast: . Keep your chin up:D
Regards
John.
The curve is steep so you will slide backward and start over again and not fall off the top into the abyss.
You teach Math? I teach HVAC/R
Hi Chilly1,
:D Yes I teach maths which is much easier to teach than hvac, I deal with perfect systems, these systems are never perfect and there are always lots of dependent variables:mad:
I think an autocascade has to be the most difficult system to tune because of this.
I certainly wouldnt want to attempt this with my present level of knowledge( and thats a degree in physics and a degree in mathematics) And no practical experience, which in imperfect systems is the most important. Theory can get you into the ball park, but practice gets you onto the team or selling the hot dogs:D .
A practical example of this is, I got a briefcase tonight from my wife, three digit combination on each lock, so as a confident mathematician I calculated that there would be 1000000 different combinations assuming a different 3 digit code on each lock. But that isnt the case and there are actually far fewer combinations. Anyone fancy having a guess at how many actual combinations there are. (This taught me not assume anything about my percieved knowledge)
Regards
John.
P.S. Sorry for off topic Russel
I have no degrees but I still teach at the local college and I have been attending college since I was 16. There appears to be something special about machanical systems, They to me have a soft feel, Like math to me is solid well defined provable. Refrigeration systems can be broken down into well defined systems and can be reduced using scientific method, However in the real world the infinite nature of these systems and variables that we need to control lead us to more inferances than most people find acceptable. I enjoy working on these types of systems.
The 2 locks are separate systems, there are 1,000 combinations on the first and 1,000 combinations on the second. 2,000 in total. Am I right? Took me a couple minutes to get that after thinking, "How on earth can it not be 1,000,000?". I suppose that there are 1,000,000 combinations of codes, but when you unlock one at a time it breaks down to 2000 :D
Chiily1, did you get a chance to read the article I posted. I was going to run the model through excel but i dont have surface tension figures for R290. The model takes into account the 2-phase flow in the tube. As for the oil, if we install an oil separator we can minimise this error :)
Damn, that article was what I was looking for a long time. Now all I need is the NIST database.
For propane, check this out:
http://webbook.nist.gov/chemistry/fluid/
I just did the sums. Using the article model, I estimate with the current cap tube size assuming 100% R290 is traveling through. I get a cooling capacity of 763Watts with 4feet of 0.031" tube.
NIST database ;)
That sounds WAY too much for your autocascade. But i'll keep quiet and let you and Gary figure things out.
Hey Dabit. I want to test out the database with a R404a model. What operating parameters do single stage R404a direct die systems run at? Then I will use the accepted cap tube lenght of 10feet of 0.028" and see what the numbers tell me.
Im looking for suction pressure, degree of subcooling(cap tube inlet temp), cap tube inlet pressure (high side pressure)
It is way too much. The entire exercise was thrown off by the lack of separation. Once we can effectively separate, we need to start over.
Here is an interseting article about oil/refrigerant transport in capillary tube. The high oil viscosity in relation to refrigerant viscosity has the effect of reducing mass flow rate through the capillary. So if we remove the oil before it gets to the capillary then we should be able to get greater capacity out our systems. You can read it for yourselves here:
http://www.russellpetrie.plus.com/Do...ry%20tubes.pdf
Did my sums for R404a, heres what I got for an adiabatic capillary tube:
Condensing Temperature: 30ºC
Subcooling: 8K
Evaporating Pressure: 1.013 bara
Capillary Tube: 10 feet of 0.028" ID
Liquid Fraction: 0.53
This gives an estimated cooling capacity of 115Watts.
If you want me to try the model with different figures, then I need to know the following parameters:
Refrigerant:
Condensing Temperature:
Degree of Subcooling:
Evaporating Pressure:
Capillary Tube dimensions, or heat load:
The only problem with those formula is that they don't assume 2 phase flow In this instance the problem is more complex we have 2 parrellel circuits with multiple fluids (5) and the surface tension changes with temperature of both the oil and the refrigerant. . It will put you in th eball park though.. Bubbles in the cap tube will decrease its capacity while a tube full of refrigerant wil increase its capacity Sorry got this backwards....
From what I got from the article, Pi5, Pi6 and Pi7 dimensionless groups account for the 2 phases in the capillary. When you talk about 2-phase flow, do you mean the point in the capillary when the refrigerant starts to evaporate and we have a vapour and liquid phase? Or are you talking about oil and refrigerant? You say a cap tube with bubbles will give a greater capacity, from what I have read the more liquid in the tube the greater the mass flow rate thus greater capacity. Have I missed something?
You will never have a cap tube with a total liquid column you will always have bubbles in it. The gas bubbles will cause a rise in velocity in the last section of the capillary tube. This was how it was presented to me. and if we can get rid of the oil in the cap tube we will have greater capacity. The issue of bubbles in the cap tube can be resolved by subcooling it, and I did get that backwards.
I can only see one reson why the bubbles would cause an increase in velocity in the later sections of the cap tube, that is if the bubbles expand. Do you know why the bubbles would expand later in the cap tube? The pressure would be the same so that would suggest the temperature has increased, is that what happens?
Regards
John.
The pressure drops all the way across the capillary tube. And it begins absorbing heat at the inlet...
I can see why the temp would increase now, after thinking about it, but I cant see why the pressure would drop. Enlightenment please:D
Regards
John.
The tube has a high pressure at one end and a low pressure at the other end although the pressure does rise abruptly at the outlet it also rises across the length of the tube. Like with a water hose with one end open and a flow of water if you measure the water pressure at the hose bib and half way down the hose the pressure would also be less.
If you have a capillary tube that is not subcooled you will be able to measure a drop in temperature across its entire length this temperature drop will be to below ambient is some cases, for this to happen there must be a pressure drop.