http://www.ra.danfoss.com/TechnicalI...1/RD6KA502.pdf
Printable View
^ Nice piece of equipment. I'll check the Danfoss price, but, I'm sure it's going to be incredibly high, with all their castings/forgings & such like.
What about using a sub-cooler for this job? I have 3 on hand of various sizes, using an internally twisted tube.
http://www.vaportec.co.nz/specsheets/subcoolers.jpg
:D
Let me go through carefully which one to fit in the lab machine. It will give us a good idea of what to expect down the track.
While we're at it, are there any other requirments that we should consider:
- minimum height;
- maximum pressure drop allowable (both sides);
- flow velocity;
etc.
Why I ask, is that, I recently commissioned some fin tooling & the first fin profiles could serve this purpose very nicely. I have a second round of tooling in the pipeline & this project could just give me the shove I need to get them cut & sent over. I can then have the 2nd generation of VIC's made up over here to best suit our requirements.
;)
^ Fair-enough. I'll put together some specs in terms of pressure drops, performance & so on - then we can kick it around.
Some ball-park questions (can refine these later):
1. With it being mounted vertically - any idea of say a minimum vertical height you'd like to see? Is 250mm enough, for instance?
2. What about evap liquid retention? Would we want to take up some of the liquid in the VIC during off-time, rather than have it migrate to compressor lower shell?
3. For the suction gas, would we say, for instance, come out of the evap at saturate condition, then superheat all in the VIC (7K)?
4. What kind of target liquid sub-cooling dT would we be looking for across the VIC (say 5K)?
I would like to see the suction line go up near the height of the evap. The VIC needs to fit within that vertical line and be long enough to get the job done.
The VIC will make migration less likely. If you don't have migration problems without the VIC, then you won't have migration problems with the VIC.
Yes.
We want the liquid temp at the TXV inlet to be close to Te,sat.
Some basic heat-balances across the VICQuote:
Gary:Quote:
Originally Posted by desA
4. What kind of target liquid sub-cooling dT would we be looking for across the VIC (say 5K)?
We want the liquid temp at the TXV inlet to be close to Te,sat.
For perfect heat-exchange liquid-to-refrigerant, based on the stream balances & available specific heat property data, the liquid line temperature drop is:
dTliq = (1.122)*dTvap
@ dTvap=7K => dTliq = (1.122)*(7) = 7.86K !!!
These are the physics of heat-transfer, unfortunately.
What is really required in order for the VIC to bring the liquid temperature down further, is to have it take on some of the evaporation duty, off the evaporator.
The other alternative is to simply just use the evap off air to reduce the liquid line temperature, through a finned line cooler.
And with the TXV bulb being mounted at the heat exchanger exit instead of the evap exit, that's what it will do.
Given just sensible heat transfer, trading subcooling for superheat, the heat exchanger is a straight trade-off, with little if any gain.
In moving the TXV bulb to the outlet, the heat exchanger becomes an extension of the evaporator, with latent heat transfer on the suction side.
True. The calculations are showing this to be the case.
The nice things about having the TXV at outlet is that it forces an area exchange in the evap/VIC system to ensure consistent superheat at exit. The check-&-balance.Quote:
In moving the TXV bulb to the outlet, the heat exchanger becomes an extension of the evaporator, with latent heat transfer on the suction side.
Now, the next thing to work on is how much of the evaporator load to shift across to the VIC & then design/size the VIC so that it boils off the refrigerant properly, without flooding, or liquid droplet carry-over. The balance part is no real sweat, given where we want the final liquid temp to head.
The trade is going to be between a pool-boiling design & convective boiling design - I'm going to have to think about this a little more. Actually, if pool-boiling is present, then I'm going to have to think carefully a little more about the liquid routing, as the concept of parallel versus counterflow no longer applies under phase change - only for the sensible heat section of the unit.
The VIC is morphing into VICE (vertical intercooler evaporator)... :)
That's very true - hadn't thought through those implications carefully. :confused:
It will have to be convective at minimum oil carry velocity - 5 m/s vertical seems to be the suggested minimum.
Thinking more on the VIC(E), I'm wondering about the suitability of a tall, slender PHE, for this application. I actually have enough to ask my supplier to give me an idea on this, if it seems sensible.
Now that I think more on this, I seem to also remember that there may be something in the Alfa-Laval technical book on similar applications, where this could suit the VIC(E).
Condenser sub-cooling re-visited
Some pages back, an idea was proposed to develop a ratio of condenser sub-cooling to TD, much along the lines of the Magoo evap tuning rule [SH,e=(0.6-0.7)*TD].
Would you perhaps have any thoughts on a suitable ratio for the condenser SC/TD?
From the current lab experimental data, this value (tube-in-tube) seems to shift from start-up (e.g 30%) to hot condition (e.g. 40%), but remain in a reasonable range, if the system charge is correct.
Have we tried different charge levels to see what subcooling works best?
^ I can add, or remove charge, as we require - that's the easy part. :)
The difficult part will be once the machine configuration looks to be reasonably settled, then performing a full charge versus performance study, with varying charges.
What I was looking for is a yardstick, around which to begin.
As an example, on one machine, with a different condenser type, in the range of calculated refrigerant charge, the SC/TD ratio seems to be around 0.3-0.4. This pretty much corresponds to the original theoretical design estimates.
Now, for the current lab machine, with a tube-in-tube condenser, I'm not sure if this type typically responds with a greater allowable SC/TD, by nature of its long condensing length.
I'd always been under the impression that tube-in-tube condensers typically allow more SC than for instance plate condensers.
^ From what I'm seeing so far, it seems that the SC/TD ratio will probably vary between condenser families, as well as rise some % from start to end of heating cycle.
The correlation to TD looks to be fairly sound, so far.
All that will go out the window with the addition of the water regulating valve, because any measure of subcooling assumes full water flow.
We may end up putting a sightglass on this system.
Here is how I would determine the ideal charge:
Assuming the primary purpose of this system is to fully heat feedwater before sending it to the storage tank:
Further assuming that we want to keep the Te,sat at 13C and the Tc,sat at 75C:
I would empty the tank and hook up the feedwater supply to the condenser water inlet.
Start the system and adjust the water outlet valve to where it maintains Tc,sat=75C.
Adjust the fan speed to where it maintains Te,sat=13C.
Every 15 minutes add refrigerant, adjust Tc,sat and Te,sat if needed, then record all data.
The ideal charge is that which gives us the lowest condenser approach temp (hottest leaving water).
We can then see what condenser outlet subcooling works best for the condenser.
If the water flow is controlled to maintain Tc,sat=75C, the TD when there is cold water at the condenser inlet is going to be huge. We may not be able to use a charging formula based upon SC as a percentage of TD... but we need to test this.
I would be very reluctant to put a sightglass on this system because service techs everywhere are going to charge to a clear sightglass plus, and will totally ignore the weigh-in charge.
Thanks for this. It seems to be a very cunning way of getting the system settled on a once-through (direct) basis.
I'll have to adjust my current water piping, to implement this (currently a pump-around loop), but this is no problem, as it really only needs a tee, a valve & a few connectors to set it up. I'll get the bits in today & modify to suit.
This will get us set for the direct heating option. I predict a very low water flowrate for this option.
Pump-around loop
How would we go about determining correct system charge under a pump-around scenario, with typical dTw in the region of 2-3K? Water lifts from around 30'C to around 65'C exit in final pass. Total circuits around 11-12. This is the common scheme - alternative to the direct method.
Since the secondary purpose of this system is to maintain the tank temperature, once the tank is full we can connect the tank back to the water inlet and see if the charge is right for this purpose as well. Again adjusting the water flow for Tc,sat=75C and the fan for Te,sat=13C, the best charge is that which gives us the highest dTw.
If these two ideal charges are different, then splitting the difference gives us the optimum charge for the pump around loop system.
IOW, if one charge is ideal for say W,in=30C and the other charge is ideal for say W,in=60C, then we would need to charge it in between for going from 30C to 60C.
Good comment.Quote:
Gary:Quote:
Originally Posted by desA
Pump-around loop
How would we go about determining correct system charge under a pump-around scenario, with typical dTw in the region of 2-3K? Water lifts from around 30'C to around 65'C exit in final pass. Total circuits around 11-12. This is the common scheme - alternative to the direct method.
Why would anyone let their tank temp drop down to 30C?
I think it might be very helpful to have a piping diagram of the "typical" system you are referring to. My idea of typical may be entirely different from what is typical in your area.
An example:
Take an establishment which only uses most off its water (21 m3) during a peak time of 19h00 - 21h30. Hot water storage tanks are extremely pricey in this region - more than the heat-pumps, for instance.
Current system as follows:
1. Small inventory of hot water, via small fuel-fired boiler - pump-around. Start of cycle 30'C - end ~65'C - hold when hot. Enough for off peak usage.
2. At peak usage, hot water storage not enough, so heat mains water directly from 30'C to 65'C on the fly - for use during peak time only.
3. Cost of fuel incredibly high, need to go to alternative energy source.
Proposal (by consultant):
4. To install a heat-pump in either of two configurations:
4.1 Direct heating of small volume flow of water, through heat-pump (direct), into hot-water storage tanks - sufficient for peak usage. Heat-pumps operate off-peak, taking advantage of off-peak electrical rates. Fuel-boiler on make-up call if water storage begins to run short.
4.2 Pump-around heating of larger volume flow of water, through heat-pump, into hot-water storage tanks - sufficient for peak usage. Heat-pumps operate off-peak, taking advantage of off-peak electrical rates. Fuel-boiler on make-up call if water storage begins to run short.
Now, which option is better to use - bearing in mind that the COP,hp for the direct system at Tc,sat~70'C of 3.08, whereas the average COP,hp over the range Tc,sat=40'C to 70'C is close to 4.96 (theoretical)? The COP's improve further if the hot water is held at say 55'C, or 60'C instead of 65'C.
The average power output of the heat-pump over the range Tc,sat=40'C to 70'C is around 9.1% higher than that at Tc,sat~70'C.
--------
I'd be very interested in how things are done in your part of the world.
As I understand things, many operations have the water storage tanks on the roof & then gravity feed down to the lower floors, except for the booster pump down to the upper few floors.
Some of the other places may be different.
I'd like to know more about how the pressurised systems work, as I've heard a fair bit about them in Europe, Australia etc.
Can you explain this a little more? I'm not a building/piping engineer, to be honest. :confused:Quote:
Are there pumped loops to keep the water hot at point of use?
Remembering that in Asia, ambient temp is often around 25'C at night.
Thanks - nice article.
Now, the tank-type heater takes care of its own internal circulation - whether just natural convection - some are also pump-around, to get better heat-transfer.
The loop they seem to show is more on the water usage end, & would be independent of the storage pump-around loop, which is there to continuously pass water through the heat-source (heat-pump) & back to the tank, slowly raising its temp.
Heat-pumps are 'slow heaters' & either one has ridiculously low water flow-rates on a one pass system - then store at high temp, or use a pump-around system which gradually brings up the whole mass water temp, over time.
The water makeup - even as for the Grunfos example - just mixes inside the hot storage tank.
The direct systems seem to like to use a series of storage tanks, linked top-to-bottom, making use of the natural buoyancy of the water, to allow the hot water to rise to the top, ready for instant hot water draw-off, on demand. I've seen some this in Japanese designs & some European concepts. Apparently it is difficult to NOT mix the water contents & so these systems have limitations, it seems.
The thermodynamics for the direct system, have the Tc,sat~70'C - this is at lowest COP,hp for the system ~ 3.08.Quote:
Gary:Quote:
Originally Posted by desA
Now, which option is better to use - bearing in mind that the COP,hp for the direct system at Tc,sat~70'C of 3.08, whereas the average COP,hp over the range Tc,sat=40'C to 70'C is close to 4.96 (theoretical)? The COP's improve further if the hot water is held at say 55'C, or 60'C instead of 65'C.
The average power output of the heat-pump over the range Tc,sat=40'C to 70'C is around 9.1% higher than that at Tc,sat~70'C.
... at what W,in temps?
I'm thinking the COP for a direct system will be much higher with cold feedwater entering the condenser.
For the indirect system, the maximum COP,hp is at start of cycle, something around 6.92. This reduces as the Tc,sat rises with the incoming feedwater temp, until eventually it hits the 3.08 mark at Tc,sat=70'C.
Now, the efficiency of heat-transfer in the condenser is a different story, for the direct, versus pump-around systems.
The VIC(E) re-visited: (I'm presently attempting to model this)
Can we review the sequence of connections for the VIC(E) system?
T,liq,hot -> VIC -> T,liq,cool -> TXV,in -> TXV,out -> evap inlet -> Tevap,out -> VIC -> Tvap,sh -> capillary measure
In other words, evap output vapour (partially evaporated) moves to VIC, where it is completely vapourised, then superheated.
Is this the correct sequence?
How will this look on the pressure-enthalpy diagram?
CPR re-visited
Are there any electronically-driven pressure regulating valve systems that can be used in place of a CPR?
These CPR's are substantially-priced items, but, of course are mechanical & if treated properly, will probably be expected to last way beyond an electronic lifetime.
It seems very strange that the COP would be identical for both strategies at Tc,sat=70C despite the difference in Tw,in.
And what is the cause of the increased COP at start of the cycle for the indirect system?
I'm thinking the answer must be the liquid temp at the TXV inlet... in which case the VIC will change everything.
Ok, thanks.Quote:
Gary:Quote:
Originally Posted by desA
The VIC(E) re-visited: (I'm presently attempting to model this)
Can we review the sequence of connections for the VIC(E) system?
T,liq,hot -> VIC -> T,liq,cool -> TXV,in -> TXV,out -> evap inlet -> Tevap,out -> VIC -> Tvap,sh -> capillary measure
In other words, evap output vapour (partially evaporated) moves to VIC, where it is completely vapourised, then superheated.
Is this the correct sequence?
How will this look on the pressure-enthalpy diagram?
I'm not sure what you mean by "capillary measure" but otherwise seems right.
I meant measurement by TXV capillary bulb - tried to shorten the wording. Apologies. :o
Now, are there examples of a VIC-type system in applications?
Why I ask, is that, I'm currently running the VIC setup on a simulator I use to model all kinds of stuff, including the refrigerant & heat-pump circuits. No matter how hard I try, the thing will not converge cleanly. The last time I had this was in simulating CO2 circuits. When I eventually researched a little deeper, I found out that CO2 circuits exhibit 'bi-stability' or even 'multi-stability'. In other words, there are different stable solutions & the solver jumps between these - sometimes settling on one, or the other.
Practically, this would seem to be taking place due to the change in solution modes as follows:
1. Evap fully evaporates refrigerant, adds small amount of superheat. VIC then adds further superheat.
2. Evap fully evaporates refrigerant, adds no superheat. VIC then adds superheat.
3. Evap partially evaporates refrigerant - exits wet. VIC then continues evaporation & adds superheat (ie. is an evaporator).
Using the sequence as outlined above, one single stable solution does not seem to exist, on a simulator, at least.
This could actually mean, in practice, that an evap-VIC combination could possibly dance around between stable solutions, if not carefully managed - in theory. :confused:
This is why I would really like to know how these may have shaped up in the field.
COP is determined thermodynamically from Te,sat & Tc,sat temps.
In our case, Te,sat is relatively fixed - will be more so when using the CPR. :)
When Tc,sat is low (at start of cycle), then COP is high, since motor power is low & Q'cond high, at start. When Tc,sat is high (at end of cycle), then COP is low, since motor power is high, & Q'cond is lower at end.
For the direct flow through the condenser, we push Te,sat to the highest Tc,sat~70'C, with Te,sat~12.5'C. Thermodynamically, this is the same as that for the hot (end) pump-around case. For thermodynamcs, the condenser (waterside) doesn't feature at all in the COP calcs. :)