My P/T chart doesn't go that high for R134a... lol
Printable View
^ Do you need the Tsat value at the HP? :)
2.757 MPa(g) = 2.858 MPa(a)
T,sat=83.66'C
Those settings sound just about right.
Thanks, Gary.
They are just inside the compressor manufacturer's specification on the HP side, & just sufficient to prevent start-up kick-out on the LP side - well within manufacturer's specification.
The compressor safe performance envelope dilemma
I am facing a real dilemma in regards to selection of the maximum safe operating point for an air-to-water heat-pump compressor. This item comes from one of the foremost name-brand compressor manufacturers, with technology centres in both USA & Europe. The USA operation provides a performance table, while the European operation supplies a detailed perfromance-prediction program, which prints out a detailed performance envelope.
To add to the complexity, the current manufacturing arm of the compressor is in SE Asia. On enquiry, they contacted the US technology & were provided information from experimental results. Refrigerant R-134a.
US technology centre
At rating conditions:
11.1K superheat / 8.3K sub-cooling / 35'C ambient air over
-23'C<Te,sat<10'C
30'C<Tc,sat<65'C
Te,sat/Tc,sat roll-off condition at upper left of table (no explanation)
Maximum continuous rating (experimental results via SEA technical expert/USA):
Safe long-term operation : Te,sat=13'C, Tc,sat=65'C
~ 1000h test data : Te,sat=13'C, Tc,sat=67'C
Danger points : Te,sat=19'C, Tc,sat=71'C & Te,sat=15'C, Tc,sat=73'C
European technology centre - manufacturer's computer performance rating software
At rating conditions:
11.1K superheat / 8.3K sub-cooling / ??'C ambient air over (not mentioned)
-20'C<Te,sat<15'C
30'C<Tc,sat<75'C
Te,sat/Tc,sat upper left roll-off conditions at 25'C suction gas temp & 10K suction superheat
Te,sat/Tc,sat lower right roll-up condition at Te,sat=10-15'C / Tc,sat=30-35'C
-------------
The questions -
Who to believe?
What practical experience do others in this field have to shed light on the dilemma?
Is there any experience database in the field that could assist?
For a heat-pump, the operation is not continuous, but periodic over a range of water temp - hi/low temp limits.
The compression ratios are well within acceptable limits, thus mechanical failure is not the limit in question.
That brings it down to motor failure limit caused by heat, as reflected by high discharge temp and/or high T,comp,b.
I agree with this logic, completely.
I have asked the compressor supplier, in question, to rule on the discrepancy in their operational recommendations.
The real issue at hand is what safe operating window to specify for such a heat-pump, in order to prevent excessive compressor failures in the field, over the long term. It appears that these heat-pumps operate in a very marginal application range, where the manufacturer has little, or no real solid experimental evidence. The US technical operation seems to be taking an ultra-conservative route, possibly to avoid large future warranty claims.
How, as a heat-pump developer, to react wisely, & conservatively in such a situation?
My immediate view is to run the sample test machine up to the European operating envelope boundaries, then manage Tcomp,disch & Tcomp,b carefully according to the operational temp guidelines already put forward by the manufacturer. I'll use this to study long-term effects of the operating limit on compressor reliability. We have made huge progress in this thread already, to addressing compressor reliability protection. Let's continue along this path.
As you'd be aware, this kind of technical decision, is not the kind of thing a designer takes lightly (many sleepless nights).
In the production model, you will of course need to work within the manufacturer's limit in order to qualify for their warranty. We can work with this.
The experimental model is a different story. We can drive it right up to the max.
Now, here's the next part of the equation. Practically, as these are, in the main, export machines, the compressor warranty is not in place outside of the country of manufacture - there's no legal framework to enforce it. As I understand things, there is no 'international warranty' as such. The warranty settlement then rests on me & my main fabricator's shoulders, with referral of the systems to the manufacturer, under failure, to establish root cause. A lot of this settlement is done 'in good faith'.
So, under this scenario, it comes down to an assessment of reasonable financial risk, & good name, in the end. :eek:
Of course. That's what experimental models are there for. This is intended to provide me peace of mind, down the road. :DQuote:
The experimental model is a different story. We can drive it right up to the max.
I have seen a number of references from some of the manufacturers, with water discharge temps quoted in the range of 70-75'C, using R-134a.
Now, this would at the very best, require Tc,sat temps in excess of 70-75'C, with minor recovery from the de-superheating of the incoming hot gas stream. This operating point is right up hard against the European compressor performance envelope & exceeds the US advise.
Practically, with air approach temps in the range 35'C, it will be an ongoing struggle to maintain Te,sat < 15'C.
As we've found, the two viable options for maintaining Te,sat<15'C are:
1. Control fan speed;
2. Use an MOP setting for the TXV.
The first option gives lots of spare room to optimise further, the MOP option concerns me for a number of reasons (maintenance issues, performance issues at start-up?, price/availability).
And let's not forget:
3. CPR valve.
Of these three, the MOP would be the least suitable. It reduces the Te,sat and compressor load by reducing refrigerant flow through the evaporator, but this raises the superheat and thus reduces the compressor cooling. The reduced compressor load would probably compensate for the high superheat to control the compressor heat.
The CPR would do the job and in fact would be the simplest. The evaporator pressure increases with the load, but the CPR will allow no more than setpoint pressure to enter the compressor.
The fan control strategy responds directly to the compressor heat and I'm thinking will be the best in safeguarding the compressor.
This is most definitely not the path I'd like to implement.
This must be causing some other imbalance in the system - especially with a constantly-varying heat-load. I will look into this carefully & try simulating the effect.Quote:
The CPR would do the job and in fact would be the simplest. The evaporator pressure increases with the load, but the CPR will allow no more than setpoint pressure to enter the compressor.
I love this strategy for a number of reasons. Tuning the fan response - not only stepping it - at a critical value, has definite system balance/tuning opportunities. This is the preferred strategy, in my view.Quote:
The fan control strategy responds directly to the compressor heat and I'm thinking will be the best in safeguarding the compressor.
Ok, an update on the evaporator fan-control strategy. Back to the drawing board... :(
I had the fellow from Carel come over with one of their products - "FCP - Speed regulator with phase cutting control". The input sensor was rated to high temp up to 120'C. They only have an 8A option out here - but, he promised to look into the 4A model (not sure if still available).
Their chief technician spent all day trying to get the thing to reverse control (decrease fan speed) for a tiny fan. In the end, I was so concerned about the reliability of their offering - the first off was obviously not functional - that I called off the exercise. The price of the unit was in the region of USD 175, at wholesale prices. Frightening. This was not a good day for me, & Carel, in general.
Looked at the Alco pressure control condenser fan system - it is the exact reverse of the strategy we need to implement. :(
I had to spend all day fighting off unworkable solutions - I just want the discharge temp to fan speed control implemented in a simple way - minimum frills.
Now, I need to urgently locate a suitable alternative to the Carel.
For the time being, we can use a simple on/off fan control strategy for testing purposes.
Possibly the most cost effective alternative will end up being a multi-speed fan and multi-stage temp control.
What are your thoughts on frequency inverters?
For instance:
SYSDRIVE 3G3JV Frequency Inverter
http://www.omron-ap.co.th/product_info/3G3JV/index.asp
SYSDRIVE 3G3JX / 3G3MX / 3G3RX Frequency Inverter
http://www.omron-ap.co.th/product_info/3G3JX/index.asp
Not sure about their feedback/input requirements.
Sounds expensive. I suppose it all comes down to cost effectiveness.
Given the steps we have taken to keep the compressor cool, it is entirely possible that we will be able to drive this compressor right up to the max without exceeding our 107C limit, in which case our fan control is a high limit safety control and simple on/off will suffice.
Or possibly we may find that 107C is only exceeded at the highest load conditions, in which case we can drop to an intermediate speed at highest load (keeping just under the 107C limit), using the high speed for anything less than highest load. This could be done with a two speed motor. To this we could add a two stage temp control. The first stage would drop to the intermediate speed and the second stage would stop the motor.
On/off, variable voltage, variable frequency. There are a great many options to meet our needs, but first we need to know just what those needs are.
A proposal - for lab test machine:
1. Use partial blockage of the evap coil face, or
2. Carefully allow air bypass air into the box, to reduce air velocity through evap. Measure evap face velocity, for reference.
The evap face velocity can be adjusted manually, at Tc,sat hold points, until the system stabilises. I'll record the system response & we can then discuss the effects.
Stunning - perfect. Now why, oh why didn't I think of that... :confused: I'll set that up early next week when I get back to the lab.Quote:
I wonder if a light dimmer switch might do the trick.
The hypothesis is very simple.
Most vapour compression cycle prediction, neglects the fact that thermodynamics need infinite time to stabilise to equilibrium. For a heat-pump, the process is dynamic & very far from thermodynamic equilibrium.
The evaporator load begins at maximum requirement at the start of the heating cycle, & ends up being around 70% of Qmax at the end. If the fan speed is not moderated, the evap ends up imbalancing the thermodynamics.
It's that simple, but means a huge, huge amount. This is why the Te,sat rises over the duration of the heating cycle - even though the TXV does its level best to compensate.
I will go so far as to say that every heat-pump that runs a constant fan speed throughout the range, will see increased Te,sat over the range of the heating cycle.
This can actually be shown theoretically by building an equation using the evaporator heat-transfer equation & air-stream heat-balance equation.
Basically, what happens is that the equation looks something like this:
Te,sat = Ta',in - f(Q'e ; m'a ; Cpa ; UA)
As Q'e required drops off, as the cycle advances, with the other terms constant, the last term reduces. With Ta',in relatively constant, Te,sat is forced to rise... :)
The effect can be seen almost the instant that airflow is reduced - say through a blank, or fan speed reduction. Te,sat will drop off over the course of less than 1 minute by some 2-3'C - without even trying.
An interesting thought. Why would this carry through the TXV itself? Wouldn't the flashing through the TXV tend to remain reasonably consistent, with the bulb monitoring superheat at evap exit?
It is entirely possible, that some temp drift does occur - but I wonder about the physics of this.
I'll set up some thermodynamics ideas on this & see what comes out of the numbers. Anyway, the practical tests beat the theory hands down - every time... :D
This would be easy enough to test.
Build a cardboard box around the drier.
At the end of the cycle, fill the box with ice.
If I am correct, as the liquid temp drops at the TXV inlet the Ta,out will drop. If the airflow and incoming air temp are held constant and the air out temp drops, then more heat has been extracted from the air.
Ok. That's a simple-enough test to perform. I'll add that to my list. :)
Now, if the rise in incoming refrigerant temp to the TXV is responsible for some of the Te,sat upward drift, then I'd suggest that a possible solution for this would be a liquid line sub-cooler, or the VIC. Mmhhh...
Are liquid-line sub-coolers available - in a simple form? In other words, say a finned pipe, or something like that?
Obtained 3 off dimmers - max 300W 220V~50Hz. Max fan power on lab machine 210W 220V~50Hz. Should be in good shape.
http://www.imagechicken.com/uploads/...9078894600.png
Water cyclic heating between ~63-68'C.
Observe the temperature drift even at TXV outlet stage.
The white signal is TXV inlet temp - note that the trend of TXV inlet & outlet are NOT the same over time. The TXV outlet temp & evap outlet temp follow similar trends - although offset in temp.
True. The need for an alternative over-arching control system - Tcomp,disc to evap air feed, is becoming very clear.
To be honest, I'm currently more worried about the Te,sat drift upwards to 19'C, during the operational phase. In my view, the inward 'energy pump' via the evap, in excess of what the condenser can/should deliver (from a thermodynamic balance perspective), causes the cycle to lift, since the heat enters faster than it can be extracted.Quote:
I take it the bottleneck in the graph is the off cycle?
The heat-pump cycle seems to be inherently unstable - distinctly different to an aircon system which is inherently stable, if managed correctly. We can talk more about this at a later stage. The evap control provides the necessary decelerator effect.
Give me a few hours - just returned back home from abroad.Quote:
If you could identify all of the lines, we may be able to spot other significant trends.
http://i29.tinypic.com/5wj8ye.png
Hopefully a little clearer picture.
#Number key:
http://i28.tinypic.com/2mev58y.png
Colour key (ignore numeric values, for now, as they are at a specific time instant):
http://i26.tinypic.com/no8rq1.png
Surprisingly, the inlet air temp (orange line) seems to have little effect on the other temps.
The inlet air line is rather interesting - it must be said:
1. It is a bare thermocouple placed slightly upstream of the evap inlet filter - this explains the 'sensor chatter';
2. The hot water vessel, for his particular test (not the lab machine), is an open vessel - located in the test room. The upward drift during the major heating portion of the run, can be explained by hot water steam-off.
Please observe the correlation between the following signals:
#30 - air inlet temp
#27 - out exp
#28 - outlet evap
#29 - suction to comp
#27/#28/#29 track the air inlet temp, with moderation, but, the curve shape is visible - rise (with slight delay), peak, roll-off. The TXV seems to be doing pretty well, but, it simply cannot track closely, at evap fixed air speed.
The drifting air temp has to affect the Te,sat & associated evap temps. This can also be shown from an evap heat-transfer balance.
I've just performed a very crude experiment on the lab machine.
It is conclusive the Te,sat rises in response to evap load push & consequent system imbalance.
I was able to hold the Te,sat value at desired value, throughout the heating cycle, by varying evap fan speed. This is trick stuff...
The question just has to be asked:
Does anyone know of heat-pump designs where the evap duty is managed over the range of the heating cycle?
^ Can I perhaps ask you to set up a rough sketch of this logic? (pencil sketch, scanned up to tinypic.com, would be incredibly helpful).
I'll work on setting that up in the next few days. This is beginning to get very interesting.
http://i26.tinypic.com/20f2tea.jpg
In effect, this gives you a two speed fan, with the second speed being adjustable.
In addition, the fan starts through the temp control contacts, so the dimmer is not subjected to the heavier starting current of the fan motor. This is better for both the dimmer and the motor.
The setpoint and differential on the temp control should be adjusted such that the contacts open at maximum allowable load and close at minimum allowable load.
The dimmer should be adjusted such that the load is maintained just under the maximum allowable load on the hottest day (35C air in).
We could improve upon this with a two stage temp control. With a two stage temp control the second set of contacts could shut off the fan if for any reason the discharge temp continues to rise above maximum.