If you would care to provide the minute by minute raw data for last nights chart, I will run it through the Modeling Tool as see if it predicts that cycling should occur on your system.
Here it is 24 hours from noon yesterday to noon today:
If you would care to provide the minute by minute raw data for last nights chart, I will run it through the Modeling Tool as see if it predicts that cycling should occur on your system.
Here it is 24 hours from noon yesterday to noon today:
-- Attachment is not available --
Did I miss some data on the heat loss of @cathoderay 's house. I don't recall seeing an experimentally determined value (and I think you know what I think about the accuracy of house heat loss models!). Without a tolerably accurate house heat loss trying to determine whether it will cycle or not is a bit questionable!
This post was modified 1 year ago 2 times by JamesPa
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
Sure, cycling is (usually) caused by over-supply, but the over-supply is (usually) caused because the demand (determined by the emitters/house) is less than the minimum output that the pump can modulate to. Thus it operates at its minimum output and because the demand is less than the supply the RWT and LWT progressively increases. After a while the LWT reaches the upper bound that is programmed into the control system, so the pump shuts off until it has dropped below the lower bound programmed into the control system. Again a basic feedback loop. My boiler does exactly the same.
Concisely put. But cycling is phenomenon that seems to have a number of causes, and this golden standard of continuous, even running seems to be held up as an ideal that users can expect any heap pump to achieve if only we tickle the right couple of parameters in the right way.
But here are a couple of graphs from last friday (different sources so the timeframe is not quite the same) showing my EcoDan 8.5 unit still cycling even as the OAT drops towards 0 and it is working harder than it was at 4pm. The graph showing the power consumption has the very typical profiles that I see with my unit, each cycle rising to a peak before stopping for a break. Compare these to the graphs posted earlier in the thread from KevM's EcoDan – a peak at the start then a longer flatter run. This difference looks to me like the result of decisions made by the running algorithms that are not decipherable.
Do you know it isnt matching demand, so far as I can see your IAT is staying constant which suggests it is
I wasn't suggesting it doesn't match demand in that quote, it fact the opposite, it does match demand in the conditions I was talking about. However, it does fail to match demand at lower OATs, see posts passim. This happens because at those low temps, say -2, the loss is around 12.4 kW, but, in those conditions, my heat pump can only put out, according to the Midea engineering data, about 11.3 kW.
How do you know what your heat loss is at -2. Did you measure it? If not how can we rule out the possibility its oversized?
It was measured, in so far as some measurements were taken of wall and window areas etc, using Freedom's heat pump calculator, which does use individual room heat loss calculations, but as we all know, heat loss calcs are not an exact science. The calculations were done as part of the design process, and as is only to be expected different heat pump suppliers coming up with different figures (some clearly back of the envelope guesses). I accepted 12.4kW as being good enough on the basis of my own calculations, which I had done in the past, when installing previous heating systems, combined with going through the Freedom calculator with a tooth comb (good thing I did, there were some typos that made quite a difference).
I suspect all this got posted and discussed before you joined the forum, but the calculator came up with that 12.4kW figure, and suggested a Midea 14kW unit would be able to meet that demand exactly, ie there was no wiggle/headroom. It then came to light (after installation) that the figures in Freedom's calculator over-estimated the Midea 14kW unit output (12.4 when in fact it is nearer 11.3 at -2 or thereabouts), hence the shortfall at lower OATs.
Both the proper, or best available if you prefer, heat loss calculations and the experimental evidence (plots of IAT against OAT in cold snaps) show the heat pump is under-sized for those conditions. In warmer OATs it is OK.
If you look back I said oversized or poor modulation depth.
I know you did, that's why I ruled out over-sizing, and then moved on to 'poor modulation depth'. I even went on to give an estimate of that 'poor modulation depth'
then thats what the engineering tables are telling you - that it cant modulate down far enough.
That is exactly what I said, maybe not in the same words, but my meaning was just that. The point is that this 'poor modulation depth' forces Midea heat pumps to cycle at a larger range of OATs than I previously thought was the case.
what we don't know is whether you could have got away with a smaller pump.
But we do, see above, I can't see anything that has ever suggested that my heat pump was over-sized for cold conditions (which is when it matters most).
The problem as I see it in the current iteration of discussions is the 'poor modulation depth' of Midea heat pumps, probably across wide range of OATs, though I haven't checked. This inability to lower their output adequately, combined with running hot (fast steep rises in LWT), means they over-shoot the Set LWT and have to resort to cycling to modulate the LWT. We haven't come across this before because we didn't have the data. But now we do. I suspect time will show that my Midea heat pump cycles across the board.
None of the recent discussion here in this thread has considered the equally important, perhaps even more important, question: does this low frequency cycling matter?
Midea 14kW (for now...) ASHP heating both building and DHW
@harriup - thanks, especially for posting the charts. We now have data from three installations, yours mine and @kev-m's, and they all show some form of cycling, even when the heat pump 'shouldn't' be cycling, ie lower but not defrost range OATs. Your right hand chart is really rather similar to mine posted earlier, slow cycling most of the time at the rate of about once an hour, all happening at OATs that shouldn't trigger cycling.
this golden standard of continuous, even running seems to be held up as an ideal that users can expect any heap pump to achieve if only we tickle the right couple of parameters in the right way.
Indeed, and as I said right at the end of my last post in different words, we need to ask how and why has it become the gold standard. If most heat pumps generally cycle most of the time, but no end users noticed because they weren't doing adequate monitoring, then perhaps it doesn't matter, especially if it is low frequency.
Midea 14kW (for now...) ASHP heating both building and DHW
I suspect all this got posted and discussed before you joined the forum, but the calculator came up with that 12.4kW figure, and suggested a Midea 14kW unit would be able to meet that demand exactly, ie there was no wiggle/headroom. It then came to light (after installation) that the figures in Freedom's calculator over-estimated the Midea 14kW unit output (12.4 when in fact it is nearer 11.3 at -2 or thereabouts), hence the shortfall at lower OATs.
Both the proper, or best available if you prefer, heat loss calculations and the experimental evidence (plots of IAT against OAT in cold snaps) show the heat pump is under-sized for those conditions. In warmer OATs it is OK.
I was faintly conscious that you had come to the conclusion that it was in fact undersized, but not sure on what basis. I wouldn't trust the Freedom Heat Pumps model to discriminate reliably at this level (nor any other fabric model - there are too many fabric uncertainties). Id be surprised based on your previous comments if you would trust a fabric model either. But if you have actually measured it then that's a different matter!
The problem as I see it in the current iteration of discussions is the 'poor modulation depth' of Midea heat pumps, probably across wide range of OATs, though I haven't checked. This inability to lower their output adequately, combined with running hot (fast steep rises in LWT), means they over-shoot the Set LWT and have to resort to cycling to modulate the LWT. We haven't come across this before because we didn't have the data. But now we do. I suspect time will show that my Midea heat pump cycles across the board
Fair enough, that seems to be the case from the data you have produced.
None of the recent discussion here in this thread has considered the equally important, perhaps even more important, question: does this low frequency cycling matter?
Indeed. In principle it means you need a higher FT and thus operate less efficiently. However with the relatively high on to off ratio and the fairly swift recovery I would estimate that the effect is small (it could be calculated, but my enthusiasm for such calculations has limits). So quite possibly not (or not much).
Sure, cycling is (usually) caused by over-supply, but the over-supply is (usually) caused because the demand (determined by the emitters/house) is less than the minimum output that the pump can modulate to. Thus it operates at its minimum output and because the demand is less than the supply the RWT and LWT progressively increases. After a while the LWT reaches the upper bound that is programmed into the control system, so the pump shuts off until it has dropped below the lower bound programmed into the control system. Again a basic feedback loop. My boiler does exactly the same.
Concisely put. But cycling is phenomenon that seems to have a number of causes, and this golden standard of continuous, even running seems to be held up as an ideal that users can expect any heap pump to achieve if only we tickle the right couple of parameters in the right way.
But here are a couple of graphs from last friday (different sources so the timeframe is not quite the same) showing my EcoDan 8.5 unit still cycling even as the OAT drops towards 0 and it is working harder than it was at 4pm. The graph showing the power consumption has the very typical profiles that I see with my unit, each cycle rising to a peak before stopping for a break. Compare these to the graphs posted earlier in the thread from KevM's EcoDan – a peak at the start then a longer flatter run. This difference looks to me like the result of decisions made by the running algorithms that are not decipherable.
-- Attachment is not available -- -- Attachment is not available --
Interesting graph. The early cycles look like they might be minimum output cycling, but after 9pm (and particularly at midnight) its gets a bit more complex because the unit is clearly operating above minimum output yet it still cycles (albeit that at midnight we see only 1 cycle so cant be sure what happens next). As you say could be running algorithms that we don't (and cant) know. One possibility that occurs is that it somehow overcompensates for the drop in OAT (which it can measure) but doesn't allow for the fact that the fabric is nice and warm. Based on a simple 'keep the FT at its design value' this shouldn't happen, but I can see a mechanism where it might happen if the pump is running in 'adaptive' mode. As you say smooth continuous running is an ideal, but in reality perhaps unlikely to happen for very long (in at least some cases) if only because OAT keeps changing.
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
Sure, cycling is (usually) caused by over-supply, but the over-supply is (usually) caused because the demand (determined by the emitters/house) is less than the minimum output that the pump can modulate to. Thus it operates at its minimum output and because the demand is less than the supply the RWT and LWT progressively increases. After a while the LWT reaches the upper bound that is programmed into the control system, so the pump shuts off until it has dropped below the lower bound programmed into the control system. Again a basic feedback loop. My boiler does exactly the same.
Concisely put. But cycling is phenomenon that seems to have a number of causes, and this golden standard of continuous, even running seems to be held up as an ideal that users can expect any heap pump to achieve if only we tickle the right couple of parameters in the right way.
But here are a couple of graphs from last friday (different sources so the timeframe is not quite the same) showing my EcoDan 8.5 unit still cycling even as the OAT drops towards 0 and it is working harder than it was at 4pm. The graph showing the power consumption has the very typical profiles that I see with my unit, each cycle rising to a peak before stopping for a break. Compare these to the graphs posted earlier in the thread from KevM's EcoDan – a peak at the start then a longer flatter run. This difference looks to me like the result of decisions made by the running algorithms that are not decipherable.
I think your 'cycles' after 8pm or so are defrosts. The cycling you see in my recent posts is because the IAT target is being reached and the ASHP switches off, lets IAT drop and restarts after a while. The third type of cycling, based on flow temperatures rarely happens to me now because the Ecodan AA algorithms try their best to prevent it. I think that's what @cathoderay system is doing.
I think your 'cycles' after 8pm or so are defrosts.
I'm not sure they all are. The hallmark, at least on my system, of a defrost cycle is the LWT falls below the RWT (if you think about it, it does make sense). On @harriup's right hand chart this only happens during the 23:30 cycle, and there is another clue, the OAT jumps up, because of the warmer exhaust air. That in turn leads me to suggest that this means the heat pump is drawing in perhaps quite a lot of exhaust air - how otherwise would we get that OAT spike?
Midea 14kW (for now...) ASHP heating both building and DHW
I was faintly conscious that you had come to the conclusion that it was in fact undersized, but not sure on what basis. I wouldn't trust the Freedom Heat Pumps model to discriminate reliably at this level (nor any other fabric model - there are too many fabric uncertainties). Id be surprised based on your previous comments if you would trust a fabric model either. But if you have actually measured it then that's a different matter!
Of course I am sceptical of heat loss calculators - after all, they are classic whatiffery models! What if the walls are made of this and are that thick, what if there are two air changes an hour, what if the roof space is only partly insulated, the whole thing is pure whatiffery. But we also have to have something to to help us choose the right size heat pump, and I took a pragmatic view that a spreadsheet based heat loss calculation would have to be 'good enough', otherwise i might as well roll the dice twice and add to results together to get my heat loss in kW.
Real heat loss measurements are of course not easy to do. I seem to recall there was a discussion some time ago on whether you could use the energy input to the building (ie energy out from, not in to, the heat pump) when both IAT and OAT were stable as a guide to if not measurement of the heat loss in those prevailing conditions, and it might just be possible to extrapolate based on an assumption of a linear relationship between heat loss and OAT given a stable IAT. Interestingly, my IAT has been steady all day at about 18.5 to 18.8, the current Midea OAT is 6, and the predicted loss (from the Freedom calculator) at that OAT is 7.5kW, and the actual current 'ability' value from the wired controller over modbus is 7.44kW.
Midea 14kW (for now...) ASHP heating both building and DHW
If you would care to provide the minute by minute raw data for last nights chart, I will run it through the Modeling Tool as see if it predicts that cycling should occur on your system.
Here it is 24 hours from noon yesterday to noon today:
I put the data into the Modeling Tool and it does indicate the your heat pump will probably cycle until the OAT sensor indicates a temperature of 4C, though I doubt that this is the correct ambient air temperature. Close examination of the raw data shows this to be the case in the real World.
I think that I have discovered the primary reason why your heat pump cycles at temperatures above 4C, according to the OAT sensor, the root cause is the Plate Heat Exchanger, which I believe that I am correct in saying is still installed within your system.
The secondary reason is the cold air circulating, which I believe is being made worse by the primary reason.
I think that I have discovered the primary reason why your heat pump cycles at temperatures above 4C, according to the OAT sensor, the root cause is the Plate Heat Exchanger, which I believe that I am correct in saying is still installed within your system.
It is still installed. Could you please elaborate on how why the PHE is the root cause? I don't have an absolute objection to removing it, but there are also some good reasons for keeping it. I also want to be sure about the relative importance of various factors. For example, I have today put forward a suggestion that maybe almost all heat pumps cycle at most temperatures to varying degrees, we just didn't know before because we didn't have good enough data. Certainly @harriup's right hand chart shows clear cycling in an Ecodan that that has LWT/RWT that looks remarkably similar to mine posted earlier, occurring over a range of temperatures. If cycling is inherent to heat pumps, then that might be the root cause, with the PHE perhaps aggravating it. I can't ignore the observation that all three (OK n=3 is small, but it is larger than zero) heat pumps we have data for show cycling of various forms across a range of OATS. Someone needs to pull the black swan out of the bag!
Midea 14kW (for now...) ASHP heating both building and DHW
Thinking about installing a heat pump but unsure where to start? Already have one but it’s not performing as expected? Or are you locked in a frustrating dispute with an installer or manufacturer? We’re here to help.
