Monday, December 20, 2010

Winter and delta T

19 Dec '10: Before the sunboxes injected summer heat into the ground, in previous seasons, I noticed that the GSHP worked regularly at about 3.6 degC delta-T between outgoing and returning temperature.

I have got worried recently that it is consistently working at about 2.0 degC delta-T.... even in this wintry weather.

Having phoned Ice Energy they tell me that this is not to worry about, it is a sign that it is working well, and finding the heat it needs... doesn't need to drive the outgoing glycol too low. A word from Ivan on Navi tron seemed to confirm that. But both also mentioned Pump Speed, as did Chris Wood commenting below... So I checked pump speed and found that the ground loop pump has been running on the fastest speed setting of the pump. So I shifted this to the middle speed and the delta-T steadied at about 2.6ºC.

Having discussed it with Chris in the comments below, I have been tempted to reduce pump speed further. The result of lowering it seems to be that GSHP needs to refrigerate the outgoing liquid more. Yes, this means the heat pump has to work harder to refrigerate more.... but it reduces pump power consumption, and this could be useful - deepening the outgoing liquid temperature would help the thermostat to be earlier at triggering sunboxes into action when there is any useful heat.  It would make the energy flow recordings of the Sontex a bit more accurate, as a delta-T of 2.0 is too small for that particular model to measure accurately, considering its margin of error. 3.6 would be larger than twice the possible margin of error.
    David Atkins recommends aiming for a delta-T of 3.0-3.5ºC so this would indicate reducing the ground loop pump speed more. So I have tried doing that, in the early evening, and after a while, it settled down to a delta-T of about 3.5º. (So maybe the pump speed got fiddled with in the early autumn, and I forgot.)

If anybody wants to know more about the progress of the chocolate teapots with the diddlysquat  mirrors 'perpetual motion machines' (as they have been derisively referred to on the Navvi troll forum), here is an update.

The Sunday 19th Dec 2010 deep temperature is 10.30ºC.

Note: This has been the hardest winter since 1963 and we have had few hours of sun in the last week, and I have set time-clock to start earlier and end later than in previous winters (because Mrs N-C demands it), and an extra degree on the thermostat. So a much increased heating load, but the GSHP still using less than 30 kWh/day.

One year ago, on the 19th Dec 2009, there had only been two evenings with below zero temperatures in the entire winter (-0.2 and -1.50). It was very mild. I was just beginning to drill holes in the aluminium rails and brackets. The black slabs were stacked in the garage. My daughter and son-in-law were staying having just got back from Istanbul and Athens.
One year ago, on the 19th Dec, the deep ground temperature was 7.10ºC

I now notice that the daily consumption is consistently proportional to the daily temperature (with slight improvement if there is Sun) whereas a year ago it was inconsistent.

In fact, the Weekly Consumption of the house from 12 Dec to 19 Dec makes an interesting comparison.
13-20 Dec 2009: House 233 kWh, GSHP 168.79 kWh, Avg evng temp 1.88º
12-19 Dec 2010: House 226 kWh, GSHP 170.59 kWh, Avg evng temp 0.22º
Something seems to be working here! The GSHP figure includes the circulating pump for the underfloor heating system. It means that at this time of deep winter, we are meeting our Space Heating and DHW requirements with an typical figure of 0.225 kWh /sqm / day.

This seems to me to prove that the earth charging works! But it doesnt necessarily prove that in its first year there is a financial return. If our intention is to prevent chilling over 5 or ten years, it could take that long to prove it. We shall be preventing a major loss of inefficiency in five years, maybe we are merely keeping the machine running as if it was its first efficient year of operation.

   This is like planning for climate change or population control... for a future problem that won't show if the policy worked for the following 25 years.

   So if we prevent chilling in five years time, and the earth does not chill, we will never know what the GHSP would have been performing like in 5 years time without the charging. There might then be a significant cost difference from what it would have been. In fact, after 10 years, we would be blaming the performance on the machine and demanding a replacement - when it is the heat in the ground that needs replacing!

  The fact that during this deep freeze, the power consumption of the GSHP now is about the same per week as the same time last year with mild above freezing weather suggest to me that when temps and sunshine hours improve in the spring, our overall annual saving will be measurable.

26 Dec '10 Postscript: The ground temperature is on 10.0º, despite this continuing long cold season! Just holding in there, in double figures, and I hope for some sunny days as the days draw out into Spring, and we get a bit of recharging from winter sun. The deep ground temp a year ago was 7.50º.


  1. David,

    The delta T between the outgoing and returning glycol temperture is mostly determined by the flow rate - the heat pump does not adjust its glycol temperature to 'find' heat. Its a themodynamic process, as is a consequence of the evaporating pressure falling due to the temperature of the ground loop falling - I wont explain here. Suffice to say that if the heat pump extracts a fixed amount of heat then the only other variable to affect the delta T is flow rate. A smaller delta T = faster flow rate. (is this occuring because you have two pumps running inline?).

    So if you extract a constant heat (say 6kW) then the delta T will be the same for the same flow rate. This delta T between flow and return glycol temp will not vary (much) with ground or glycol temp (subject to exclusion of caveats below). So if you have the same flow rate all season, you should see the same delta T all season - all that happens is that the average of these two temps falls throughout the season.

    In reality there will be a slight change in delta T between flow and return due to reasons below but the only variable, which could shift a delta T by up to 50% is the flow rate.

    ( A point from the discussion above- the heat extraction is not entirely constant because as glycol temp falls, so does COP and heat output, hence the heat extraction from the glycol, will fall... but not massively over the range we are talking about... and also we could argue the fact that colder glycol is more viscous would slow the flow rate down... We will ignore these points here for the sake of simplicity in the discussion above)

  2. Hi, I am not aware of changing the pump speed, but then I remember discussing it with you and David A. Next time I open the front, I will try to remember to check it.
    He has also mentioned that. As I have the immersion heater (additional heat0 function disabled, isnt it better to maintain a good pump speed, to avoid the GSHP getting 'desperate'....

  3. makes no odds - the plate heat exchanger is designed for a certain temperature approach and delta T... its probably quite broad but between 2 and 5 deg C is typical, so if you are 2, or 3.5, or 5 it probably makes little measurable difference.

  4. I have modified the pump speed to middle, but am tempted to reduce it further. The result of lowering it, is that it refrigerates the outgoing liquid more. This could be useful as it would be quicker at triggering sunboxes to action is there is any useful heat. I will ask David A too, although Chris, you might have a verdict on this hypothesis?

  5. Chris, I have now modified the pump speed to the slowest, which causes it to settle back to a delta of about 3.5 degC working at this time of year. A larger delta-T will also improve the accuracy of the metering, as the Flowmeter becomes less accurate if the flow and return temps are too close.

  6. I think that 3.5deg C is an appropriate delta T. but note that the heat pump does not work harder to produce this greater delta T - as the heat pump is extracting heat at a constant rate and the glycol flow is slower, it just means that the fluid per unit volume is cooled more as it takes longer for that 'unit' of fluid to move through the evaporator.

    Think about it like this .... for a given ground tempertaure and a certain heat extraction then there will be an average temperature that the glycol will fall to. For instance on a particular day, the average may be 1degC with a flow of 0degC and a return of 2degC (i.e. deltaT is 2degC.). Now take exactly the same working conditions and slow the pump speed by half. The average glycol temp is still 1degC but the deltaT is 4degC.. so therefore the flow would be -1degC and the return would be 3degC. All that has happened is that the temperature spread is equal either side of the average.

  7. Thanks Chris,
    at this time, the boxes arent doing anything, so i am paying more attention to the GSHP itself.
    Hopefully, there will be sun tomorow.. it is forecast.
    with the slower pump speed, I think we have the risk of more red light alerts. That may have been why it got speeded up, a couple of months ago.

  8. exactly David - a consequence of the lower speed is the greater deltaT hence a wider spread about the mean, thus you will have a lower flow temp and this is what the heat pump will measure as a precaution from freezing the glycol etc / low evaporating pressure etc.. so yes the sensor may cause the alert.


Comments will be moderated before showing. Please make them relevant to the subject of the posting. Comments which advertise commercial products will usually be deleted.

Popular Posts