Air Source Heat Pump Buffer Vessel Sizing

[chiggs24]

Just for fun. When specifying buffer vessel sizing for an air source heat pump system, what is the biggest single factor you take into account?

In my opinion, when a unit utilises reverse gas defrost, the vessel should be sized to ensure there is enough energy to defrost the unit.... (All other defrost approaches would be sized on minimum water content only)?

What are your thoughts?

[Brian_UK]

A buffer vessel is normally used for a chilled water system if the system volume is below the capacity of the chiller.

When an ASHP defrosts it uses the refrigerant gas within it's sealed circuit.

[chiggs24]

Sorry, I should elaborate. The ice in this example will be on the evaporator coil, where the refrigeration circuit does not circulate. Hypothetically speaking, this would be on a monoblock unit

[chiggs24]

Sorry, I should elaborate. The ice in this example will be on the evaporator coil, where the refrigeration circuit does not circulate. Hypothetically speaking, this would be on a monoblock unit

Sorry, the above was meant to say "Sorry, I should elaborate. If a buffer vessel is not implemented, then heat will be pulled out of the property in order to defrost the evaporator coil. In the middle of winter, surely you want to keep the heat in the property. My thinking is that if a unit which flows at 40 lpm in defrost with a 5 degree differential across the coil, and a 7 minute defrost time in a fairly agressive climate, you would flow a total of 280 litres with a temperature drop of 5 degrees within the heating circuit? This temperature would come from the property if a buffer vessel is not utilized?"

Comments are hugely welcome

p.s. Not quite sure where my head was previous to that previous post.

[frank]

The water in the buffer vessel is not used to defrost the evaporator (outside coil).
The refrigerant gas does this job, so, during a 7 minute defrost cycle, the heat is diverted to the coil instead of into the buffer vessel.

During defrost, the ASHP will stop the circulation pump so that the buffer vessel does not pick up any cooling from the HE

[Bigfreeze]

The water in the buffer vessel is not used to defrost the evaporator (outside coil).
The refrigerant gas does this job, so, during a 7 minute defrost cycle, the heat is diverted to the coil instead of into the buffer vessel.

During defrost, the ASHP will stop the circulation pump so that the buffer vessel does not pick up any cooling from the HE

You're wrong there Frank. The circulating pump would keep running to provide energy to the cooling side so defrost can be completed.

The buffer vessel should be, at a minimum, enough volume to allow a full defrost but you should go large enough that the water temp in the buffer does not drop significantly. The size of the buffer will depend on the size of the heat pump. More cooling power will require a larger buffer. We usually size at 15L per kw heating power.

[chiggs24]

You're wrong there Frank. The circulating pump would keep running to provide energy to the cooling side so defrost can be completed.

The buffer vessel should be, at a minimum, enough volume to allow a full defrost but you should go large enough that the water temp in the buffer does not drop significantly. The size of the buffer will depend on the size of the heat pump. More cooling power will require a larger buffer. We usually size at 15L per kw heating power.

Hi Bigfreeze, what would you anticipate to be the low limit for temperature in the buffer?

[Bigfreeze]

It shouldn't really drop below 10C at the sensor as that would be middle of the tank. Which means the bottom and therefore the return pipe would be cooler. The minimum water temp (without glycol) that should pass through a HX is 4C as the water can crystalize below this temp. Also the lower the temp the longer the defrost will take as there is less energy being supplied to the high side.
You should always build in redunancy beyond the minimum for this reason alone. A defrost should really take between 120 - 180 seconds if all is designed correctly

[chiggs24]

If the heat pump monitors on return temperature, would the 10 degree C still ring true? Would this become a variable dependant upon height of the return tap in on the buffer vessel?

In relation to defrost time, 334 Joules per gram is required to change the state of ice to water and 2.03 joules per gram is required to raise the temperature of ice by one degree.

With this in mind, is it safe to calculate, that hypothetically, if you have around 2kg of ice and the ambient temperature is minus 5degreesC, in order to melt the ice, you require 20,300J to raise the temperature to 0, then 668,000J to complete the phase change to water, which equates to a total of 0.191194kWh.
Also, a 50litre vessel of water held at 50 degrees C indoors (20 degree room temperature) should contain 1.75kWh.
Can you then calculate that the vessel will give up 0.20416kW in 7 minutes, enough to melt the 2 kilos of ice?

I'm no physicist, so if any of the calculations aren't correct, i'm more than open to corrections

[Bigfreeze]

The temp in the buffer would stratify top to bottom according to the temp diff across your condenser. If you had 10K the top of the tank would be 10k warmer than the bottom, it would be 5k warmer if there was 5k difference across the HX. You should always aim to have 5K. As I mentioned above, if this was the case, 10C would be the minimum tank temp and you should always build in redundancy.

Defrosting isn't as simple as calculating the volume of ice to be defrosted. Issues such as the positioning of the ice etc will have a huge influence on how much energy it will take. Say for example that the entire face of the coil is frosted over to a depth of 3-4mm and has not ingressed into the coil. Then simply melting the contact surface will result in the entire sheet falling loose, if the same quantity of ice was in the coil, complete defrosting would take much longer.

[TiredGeek]

After reading this and considering my own problems, can I just say.......
MY HEAD HURTS

[chiggs24]

The temp in the buffer would stratify top to bottom according to the temp diff across your condenser. If you had 10K the top of the tank would be 10k warmer than the bottom, it would be 5k warmer if there was 5k difference across the HX. You should always aim to have 5K. As I mentioned above, if this was the case, 10C would be the minimum tank temp and you should always build in redundancy.

Ah of course. So a 5 degree delta T, would indeed yield a 50 degree up top with a 45 return lower down the tank. Would this then mean that hypothetically, if you have a metre tall vessel with the flow at the top, the return 500mm further down (45 degrees C), potentially, the bottom of the vessel would be sat at 40 degrees C?

The ten degree minimum makes a lot of sense. Could this be a lower temperature if glycol is used (lower freezing temperature should stave off crystallizing until it reachs its mixture related low limit)?

Defrosting isn't as simple as calculating the volume of ice to be defrosted. Issues such as the positioning of the ice etc will have a huge influence on how much energy it will take. Say for example that the entire face of the coil is frosted over to a depth of 3-4mm and has not ingressed into the coil. Then simply melting the contact surface will result in the entire sheet falling loose, if the same quantity of ice was in the coil, complete defrosting would take much longer.

The above makes a lot of sense also. Because it is so variable, based upon relative humidity, mass air flow across the evaporator, and ambient temperature, is it fair to say then that it is near impossible to calculate how much ice will form, where it will form, and in turn how much energy will be required to melt it.

Referring back to the temperature low limit in the vessel, utilizing the flow rate of the unit, together with the temperature in the vessel, can we than calculate the maximum amount of time available for defrost (assuming a fixed speed pump). i.e. 35 degrees C in vessel with 5 degree delta T and 50 litres a minute flow rate. If the low limit is set at 10 degrees, can we then afford 5 drops of 5 degrees before we hit the low limit? With that in mind using a 50 litre vessel, we would be able to allow a 5 minute defrost. 100 litre vessel would allow a 10 minute defrost and so on.
Also, if the vessel temperature was in fact 50 degrees, that would give us 8 drops of 5 degrees, which would then give us a maximum 8 minute defrost on a 50 litre vessel??

[MikeHolm]

BF, how would the placement of the sensor at the lower part of the tank, up 20cm on a 500l tank for example, effect the system compared to having it in the middle of the tank. This assumes that supply from the HP is just above the return and a "sparge pipe", I believe it is called there, is used. In this case, the heat from the tank as used to defrost will be from the bottom. Not ideal.

[Bigfreeze]

The water at the bottom of your tank is always what will be returned to the heat pump first during defrost. The only difference that having the sensor at the bottom will have is it will give a cooler reading. It won't affect the operation of the machine but keep the cut off point for defrost at about 6-7C or you risk bursting a HX. So if the temp on the return sensor drops below that it will terminate the defrost.

[MikeHolm]

Do you think it would be better to have the defrost heat come from the top via a diverting valve? A bit more material but a faster defrost.

[Jon Glanfield]

BF you have previously suggested 30 litres per kw for buffer sizing, have you had a change of heart? We checked numerous manufacturers and got various responses from 10 through to 30. We aim to do 30 unless space is a maor issue.

[Bigfreeze]

BF you have previously suggested 30 litres per kw for buffer sizing, have you had a change of heart? We checked numerous manufacturers and got various responses from 10 through to 30. We aim to do 30 unless space is a maor issue.

Sorry Jon, I meant to type 30 there. For some reason 15 was in my head.

[Bigfreeze]

I wouldn't complicate it too much mike, its likely the tank will be no lower than 30C anyway. Chiggs was asking for a minimum temp.

[bigred]

Sorry Jon, I meant to type 30 there. For some reason 15 was in my head.

BF do you run into much opposition when sizing on 30 litres per kW? (im presuming ambient 7 and flow 35 for kW output)

Very broadly speaking, if a 3 bedroom house had a heat load of 5 kW, and the unit specced provided say 8 kW at A7/W35 you would then spec the buffer at 8kW giving 240ltrs?

How do you deal with the mitsi's of the world that state you don't need a buffer vessel and as such you wouldn't need to find the extra room in your property for the vessel?
(I agree with the buffer philosophy btw, just interested to find out your tact)

[Bigfreeze]

I explain to them what will happen if you don't install a buffer vessel. People who are interested in efficiency will listen and understand what you're saying has merit. Some people will want to hear a particular thing and when a salesman comes along and tells them it then they'll go with them no matter how crap the system may turn out. But to be honest you can often be better off without these people.

You need to be able to explain what a lack of flow rate will mean for the heat pump and for their pocket.


You may get away without a buffer on a completely open ufh system but you won't where theres stats or where theres rads.

[bigred]

I explain to them what will happen if you don't install a buffer vessel. People who are interested in efficiency will listen and understand what you're saying has merit. Some people will want to hear a particular thing and when a salesman comes along and tells them it then they'll go with them no matter how crap the system may turn out. But to be honest you can often be better off without these people.

You need to be able to explain what a lack of flow rate will mean for the heat pump and for their pocket.


You may get away without a buffer on a completely open ufh system but you won't where theres stats or where theres rads.

That is exactly the tact that I employ. The opposition that I am finding is that people have been told by one of the inverter compressor air source heat pump manufacturers, that even with a fully zoned system (where there will be next to no system volume when locked off) a buffer vessel is not required due to the fact that the inverter can ramp down its output to suit requirement.

If this is the case, why is everyone not employing inverter compressors. I understand the coeffecient of performance on inverter driven units is less than on fixed speed scroll compressor units, and longevity of the compressor compared to the fixed speed scroll compressor isn't as good, but if the requirement for a buffer vessel was removed (on applications which didnt require use of stratification or other buffer related goodness), then potentially, a whole other market could be opened up for customers who wont or cannot have a buffer vessel.

I guess what I am ultimately asking is, apart from defrost and minimum system volume, why do we ultimately need a buffer vessel?

[mad fridgie]

Why would you use reverse cycle for defrosting, if it not the best way, do we need the buffer at all.
we have unit chuncking away nicely 0C ambient and 35C water, lets say SST-5C SCT 40C ( a pretty good machine)
HOR 4.9kw (heating) and power in 1.3kw (COP3.77), it goes into a reverse cycle defrost, your SST would be high and your SCT would be low, same machine, 7.7Kw cooling, power in 1Kw (COP7.7) so you actual defrost COP ends up being 3.77/7.7= 0.49, you are better using hot gas defrost with an electric boost, defrost COP = 1 less some thermal losses, lets say 0.85 practical defrost COP.

[Jon Glanfield]

I think Big Red makes an important point here, we are accredited Ecodan installers and have installed plenty of their units, but recently have started to question some of the logic and claims touted by Mitsubishi and the general application recommended by Mitsubishi from a hydraulic perspective.

The claimed outputs at low ambients are significantly higher than any other model
They do not routinely specify any min water volume requirement
The flow rate parameters are wide in comparison with other units
Fin spacing is tight
HE occurs outside the thermal envelope

However there just isn't much data to prove that Mitsubishi are wrong either in terms of their data claims or application failures. We know of some where there are defrost issues, probably due to lack of flow rate and system volume, but I think the real issue is in relation to monitoring.

Whilst manufacturer's will bench test as will BRE and so on, we don't routinely install any monitoring kit in the UK in real world installations, a situation that may change as the RHI rumbles on especially with the debate on ASHP in Phase 2. There may be a case of the Emperor's New Clothes in all of this, in that unless a HP is costing significantly more to run than its predecessor the customer wants to be happy with their new toy.

However if an inverter was tested against a fixed speed with a buffer, or even an inverter with a buffer, on identical properties with identical hydraulics and correctly monitored for energy usage and COP it might allow some conclusions to be drawn.

We are starting to install at very least electrical consumption meters dedicated to the HP but I think monitoring should be mandatory and the data lodged centrally, probably via the MCS to provide some clarity to the industry.

Does anyone have any evidence that the inverter driven units employed without buffers are really significantly worse than fixed speeds with buffers? We prefer to install buffers but it is a hard sell against a competitor selling an inverter product or to a customer with limited space.

Jon

[MikeHolm]

How then do air to air units like the Fujitsu HP on my roof (which I didn't install) or mits etc. handle defrost. Do they just take heat from the indoor air and if so what is the COP at that stage? Does it differ from air to water units?

[TiredGeek]

I can't get my head around the conclusion that fitting a buffer tank, say another 800 litres of water in the system, will somehow make my already 200 litre or so system cheaper to run.
How can it be possible that raising 1000 litres of water to 45'c to pump round my rads will be cheaper overall than raising 200 litres to the same temp?

I understand what you guys are saying about having a larger amount of warm water available to use for defrosts in a buffer, but as I can achieve the same flow rates through my Mitsi's as if I had a buffer fitted where does the advantage come from for me? Try as I might I just can't understand it.

I have 200 litres of water in the rads (no thermo valves fitted), the flow through the Mitsi is limited only by the Mitsi internal heat exchanger and is 25 l/min (above the minimum), and yet my units fail to defrost. I fail to understand how a buffer tank would help here.

I'm not trying to be cnut, I'd love to understand the principals
If someone wants to take the time to explain the maths I would be most appreciative

[MikeHolm]

How long does the defrost run for on your machine and have you ever measured the temp going back to the house before the defrost terminates? It seems to me, and we have gone over this before, that the defrost terminates too early given the temp of water provided for defrost. If so, having a larger body of water at the highest temp possible should shorten the defrost time. You may simply be running out of resources and therefore the bigger tank may help.

[Bigfreeze]

Different principal for air to air Mike. Because you're dealing with air there is no issues with freezing etc.

[Bigfreeze]

I can't get my head around the conclusion that fitting a buffer tank, say another 800 litres of water in the system, will somehow make my already 200 litre or so system cheaper to run.
How can it be possible that raising 1000 litres of water to 45'c to pump round my rads will be cheaper overall than raising 200 litres to the same temp?

I understand what you guys are saying about having a larger amount of warm water available to use for defrosts in a buffer, but as I can achieve the same flow rates through my Mitsi's as if I had a buffer fitted where does the advantage come from for me? Try as I might I just can't understand it.

I have 200 litres of water in the rads (no thermo valves fitted), the flow through the Mitsi is limited only by the Mitsi internal heat exchanger and is 25 l/min (above the minimum), and yet my units fail to defrost. I fail to understand how a buffer tank would help here.

I'm not trying to be cnut, I'd love to understand the principals
If someone wants to take the time to explain the maths I would be most appreciative

You can't look at a buffer as additional water that needs to be heated. All heat generated to that tank will be released to the house at some stage so you are not paying to heat unusable water.
The flow rates you are achieving back through the heat pump on your system set up, in my opinion has no chance of achieving the flow rates required to sustain a defrost. If you were to watch a defrost on a good heat pump displaying all the temps across the system you would notice how different the system reacts from when it is in heating mode and how a slow flow rate can be catastrophic to an Air to water system.

A buffer will supply the heat pump with the water it requires at all times and is not hindered by undersized pipework, zones shutting down etc.

[bigred]

I could be massively wrong here, but as the mitsi ecodan uses hot gas from compressor discharge to defrost, how is the buffer utilized in that scenario?

we have unit chuncking away nicely 0C ambient and 35C water, lets say SST-5C SCT 40C ( a pretty good machine)
HOR 4.9kw (heating) and power in 1.3kw (COP3.77), it goes into a reverse cycle defrost, your SST would be high and your SCT would be low, same machine, 7.7Kw cooling, power in 1Kw (COP7.7) so you actual defrost COP ends up being 3.77/7.7= 0.49, you are better using hot gas defrost with an electric boost, defrost COP = 1 less some thermal losses, lets say 0.85 practical defrost COP.

I'm not sure i understand some of your terminology Mad Fridgie. Can you tell me what SCT, SST and HOR mean please

I think Big Red makes an important point here, we are accredited Ecodan installers and have installed plenty of their units, but recently have started to question some of the logic and claims touted by Mitsubishi and the general application recommended by Mitsubishi from a hydraulic perspective.

The claimed outputs at low ambients are significantly higher than any other model
They do not routinely specify any min water volume requirement
The flow rate parameters are wide in comparison with other units
Fin spacing is tight
HE occurs outside the thermal envelope

However there just isn't much data to prove that Mitsubishi are wrong either in terms of their data claims or application failures. We know of some where there are defrost issues, probably due to lack of flow rate and system volume, but I think the real issue is in relation to monitoring.

Whilst manufacturer's will bench test as will BRE and so on, we don't routinely install any monitoring kit in the UK in real world installations, a situation that may change as the RHI rumbles on especially with the debate on ASHP in Phase 2. There may be a case of the Emperor's New Clothes in all of this, in that unless a HP is costing significantly more to run than its predecessor the customer wants to be happy with their new toy.

However if an inverter was tested against a fixed speed with a buffer, or even an inverter with a buffer, on identical properties with identical hydraulics and correctly monitored for energy usage and COP it might allow some conclusions to be drawn.

We are starting to install at very least electrical consumption meters dedicated to the HP but I think monitoring should be mandatory and the data lodged centrally, probably via the MCS to provide some clarity to the industry.

Does anyone have any evidence that the inverter driven units employed without buffers are really significantly worse than fixed speeds with buffers? We prefer to install buffers but it is a hard sell against a competitor selling an inverter product or to a customer with limited space.

Jon

I wholeheartedly agree with you Jon. Electrical consumption meters need to be put in place in order to make air and ground source heat pumps viable. Not only will it give the wider public more confidence in heat pumps, but it will also ensure that customers cannot fall foul of bad installations. Out of interest Jon, which consumption meters are you installing?

I can't get my head around the conclusion that fitting a buffer tank, say another 800 litres of water in the system, will somehow make my already 200 litre or so system cheaper to run.
How can it be possible that raising 1000 litres of water to 45'c to pump round my rads will be cheaper overall than raising 200 litres to the same temp?

I understand what you guys are saying about having a larger amount of warm water available to use for defrosts in a buffer, but as I can achieve the same flow rates through my Mitsi's as if I had a buffer fitted where does the advantage come from for me? Try as I might I just can't understand it.

I have 200 litres of water in the rads (no thermo valves fitted), the flow through the Mitsi is limited only by the Mitsi internal heat exchanger and is 25 l/min (above the minimum), and yet my units fail to defrost. I fail to understand how a buffer tank would help here.

I'm not trying to be cnut, I'd love to understand the principals
If someone wants to take the time to explain the maths I would be most appreciative

Correct me if i'm wrong here, but cost savings will not be system side. If system volume is x amount and the delta T around the system is x, regardless of the size of buffer vessel, only x amount of energy can ever be used? As Mikeholm says, potentially, if you are running say underfloor heating with a 35 degree C store, in an aggressive climate, you could use all the energy up before defrost finishes. I.e. if you have a larger store of a low temperature, you have more energy to complete defrost.

[mad fridgie]

HOR is Heat of rejection, in other words the energy removed by the condenser or your heating capacity. I use data from a compressor, not sales blurb.
SST is saturated suction temperature (pressure/temp relationship at compressor suction)
SCT is saturated condensing temperature (pressure/temp relationship at compressor discharge)
in simple terms SST and SCT will determine how a system works, what duties occur and how much power is drawn.
I could be wrong but I believe the mitsi are reverse cycle defrost, not hot gas.

[MikeHolm]

Different principal for air to air Mike. Because you're dealing with air there is no issues with freezing etc.

DUH.......none the less, if the HP is having trouble defrosting (as some seem to do here) theoretically the system would suck the air temp down till it got close to ambient. Obviously it would either time out or lock out on LP (or some other means). The resource is not infinate.

[Luigi2011]

I think Big Red makes an important point here, we are accredited Ecodan installers and have installed plenty of their units, but recently have started to question some of the logic and claims touted by Mitsubishi and the general application recommended by Mitsubishi from a hydraulic perspective.

The claimed outputs at low ambients are significantly higher than any other model
They do not routinely specify any min water volume requirement
The flow rate parameters are wide in comparison with other units
Fin spacing is tight
HE occurs outside the thermal envelope

However there just isn't much data to prove that Mitsubishi are wrong either in terms of their data claims or application failures. We know of some where there are defrost issues, probably due to lack of flow rate and system volume, but I think the real issue is in relation to monitoring.

Whilst manufacturer's will bench test as will BRE and so on, we don't routinely install any monitoring kit in the UK in real world installations, a situation that may change as the RHI rumbles on especially with the debate on ASHP in Phase 2. There may be a case of the Emperor's New Clothes in all of this, in that unless a HP is costing significantly more to run than its predecessor the customer wants to be happy with their new toy.

However if an inverter was tested against a fixed speed with a buffer, or even an inverter with a buffer, on identical properties with identical hydraulics and correctly monitored for energy usage and COP it might allow some conclusions to be drawn.

We are starting to install at very least electrical consumption meters dedicated to the HP but I think monitoring should be mandatory and the data lodged centrally, probably via the MCS to provide some clarity to the industry.

Does anyone have any evidence that the inverter driven units employed without buffers are really significantly worse than fixed speeds with buffers? We prefer to install buffers but it is a hard sell against a competitor selling an inverter product or to a customer with limited space.

Jon

THis is a very interesting thread with some good topics being discussed here, We have installed inverter driven Mitsi units without buffer vessels on systems with both UFH and radiators and the key is system design, if this is carried out correctly then what the manufacturers such as Mitsi advise regarding buffer vessels is correct, i.e UFH systems pumped in parrellel to primary circuit through LLH or bypass setup to maintain constant primary flow whilst allowing variable secondary flow.

Defrost is reverse cycle but circulator will run and energy will be taken from UFH or radiator ssytem without buffer however this will not affect room temeprature in the 3 minute defrost cycle, more to the point even with a buffer installed the same amount of energy would be required from the water circuit during defrost you would simply have to reheat the water in the buffer as oppose to the emmitters, in practice makes no difference.

For a standard install buffer cylinders on inverters are not necessary as load can be varied direct from the unit itself, the buffer is a trick reserved for fixed speed to get around their inability to do so. also if you fit a buffer on an inverter driven unit would it not have been better (and cheaper) to buy a fixed speed unit instead as in weather compensated mode response time on say 300 litres of water will negate the advantages of the inverter anyway!

With regards to monitoring Mitsi have installed a number of remote web based monitoring systems on Ecodans which they have been carrying out for nearly 3 years now, none of which systems have buffers installed, this allows a seasonal performance to be measured, there are a large number of these sites accross the UK the majority of which are achieving SCOPs of 3 to 4 including DHW in real houses of all types, they have over 20 years of independantly accredited field trial data available. (see extract attached)
Site 2 - 2009 data summary.pdf

[Jon Glanfield]

THis is a very interesting thread with some good topics being discussed here, We have installed inverter driven Mitsi units without buffer vessels on systems with both UFH and radiators and the key is system design, if this is carried out correctly then what the manufacturers such as Mitsi advise regarding buffer vessels is correct, i.e UFH systems pumped in parrellel to primary circuit through LLH or bypass setup to maintain constant primary flow whilst allowing variable secondary flow.

Defrost is reverse cycle but circulator will run and energy will be taken from UFH or radiator ssytem without buffer however this will not affect room temeprature in the 3 minute defrost cycle, more to the point even with a buffer installed the same amount of energy would be required from the water circuit during defrost you would simply have to reheat the water in the buffer as oppose to the emmitters, in practice makes no difference.

For a standard install buffer cylinders on inverters are not necessary as load can be varied direct from the unit itself, the buffer is a trick reserved for fixed speed to get around their inability to do so. also if you fit a buffer on an inverter driven unit would it not have been better (and cheaper) to buy a fixed speed unit instead as in weather compensated mode response time on say 300 litres of water will negate the advantages of the inverter anyway!

With regards to monitoring Mitsi have installed a number of remote web based monitoring systems on Ecodans which they have been carrying out for nearly 3 years now, none of which systems have buffers installed, this allows a seasonal performance to be measured, there are a large number of these sites accross the UK the majority of which are achieving SCOPs of 3 to 4 including DHW in real houses of all types, they have over 20 years of independantly accredited field trial data available. (see extract attached)
Site 2 - 2009 data summary.pdf

Hi Luigi

I hear what you are saying and we too have installed many Mitsubishi and Daikin units without buffers and the majority are fine, 1 though suffers from a frosting issue, which we could foresee and flagged to Mitsubishi pre-installation and they too came back with the recommendations about flow rates. We hit nominal rates without issue, but there is insufficient volume and heat energy in the system and it can't defrost. This is not an isolated issue either, have a chat with Tired Geek and Daikin's appear to be even more susceptible if you have a trawl through the forum.

It definitely isn't black and white though, but it makes no sense that nominal flow rates can be maintained through an auto bypass, which contains a spring and plunger which will restrict flow. We have used low loss headers and seen the contents mix due to the low volume, poor stability of the stratification and strength of circulators used on the primary side of heat pumps. We have also seen this occur on small 100 litre buffers.

Yes inverters will vary load, but what happens when the min operating frequency is higher than the demand from the space heating and there is no mass volume to dump the heat to? Run times are also affected when loads are low, because start and stop frequency will increase wearing the compressor.

Whilst the water in the buffer has to be replenished post defrost, the point is that the heat energy is there in the first place en masse to allow this to occur. With a thermostatic controlled hydraulic it may not be in the absence of a buffer. Daikin at least specify a min water content that has to be available via the bypass but we have never been able to obtain a volume from Mitsubishi.

In terms of the WC, the HP in a buffered system will look after the buffer alone, it will therefore target buffer temp in line with the WC slope. WC exists on inverters and fixed speeds, it just means that as set point is approached on an inverter unit it backs off and uses less energy and doesn't overshoot.

Please don't get me wrong, as I said before this is not black and white and my contention is that the inverter driven design parameters ride rough shod over years of proving and refining in northern europe where this technology was forged. Largely it is successful although I would conjecture that we don't know if it could be more efficient if it was blended with the traditional approach of using buffers too.

The other really important point I believe is that there is little empirical historical data to go at in terms of longevity of systems in the UK climate where humidity levels play havoc with the defrost requirements unless anyone can point me to it.

Jon

[Luigi2011]

HI Jon,

Thanks for the reply, I dont think either the bypass or buffer method is 100% ideal but the bypass will open and satisfy a minimum flow rate requirement while allowing water to circulate to say one or 2 zones and it does not require the extra space and to be honest in reality its never really a problem.

What i was pointing out is that with large buffer vessels say 30L/kW as has previously been suggested on this thread you will have a very slow response time if say outdoor conditions change and you need to alter flow temperature by a couple of degrees but to do so via 300L+ of water, surely being able to quickly alter flow temperature directly from the unit is the point of inverter technology and the millions spent researching it by the big players in the air source industry.

When demand is less than minimum capacity (typically 30% rated capacity on Mitsi units) its simple the unit switches off, however it then depends on how well designed the system is especially in respect to the buildings heat loss. It is very easy to calculate the point at which this will happen with regards to outdoor temperature if heat losses are known, if this point is reached when temp difference between the building and outdoor is small and hence so are heat losses then the system is off and you dont need heating on anyway so there is no problem. This will only ever become an issue when the ASHP is massively overside and this tipping point between minimum capacity and building thermal performance is reached at a point whereby you require the heating on constantly, i.e when it is very cold outside, this is why an inverter driven unit must be correctly sized, something that is emphasised by the manufacturers who understand this such as mitsi and daikin, a fixed speed unit on a buffer tank gets around this in a very crude and basic manner from a controls point of view.

[mad fridgie]

a couple of points,

We are taliking about water heating units that are used for static heating devices (rads and underfloor) so by pure nature is slow reacting, so quick response in temperature is just not required, I really can see a point for an inverter for this applicaqtion, at 10 starts an hours 6 mins off period on stastic heat exchanger will not change the feel of a home. (air to air systems with forced draft heating, inverter is required 100%)
Buffer tank for defrost may or may not be required, but if this is the reason for installing a tank, then to use the argument correctly, then water flow should be reveresed during defrost, water to heat pump drawn from the top of the cylinder andreturned to the bottom

[Bigfreeze]

Luiggi, I think Jon laid out the problems that arise very well above.
Your contention that a 300L buffer would take ages to respond is nonsense in fairness. A 300L buffer could be reduced in temperature in a matter of minutes, it will not affect the weather compensation in terms of delivering a certain temperature of water. How quickly do you think flow temp needs to be altered?

If you simply use a bypass, you are almost immediately returning the water you've just heated/cooled back ontop of the heat pump which exacerbates problems in both the heating and defrost cycles. You won't really see this with a Mits or Daikin as the information they give you is minimal. Install a top of the range unit with sensors in all critical positions and it will be obvious to you from the first few minutes of running how essential perfect water flow is to heat pumps.

I deal with two companies based in europe with 65 years of experience in heat pumps between them, both with modulating ASHP's and both state that it is essential the use of buffer in any situation where you are not feeding a fully open UFH system and tbh I'd take their advice over Mitsubishis or Daikins any day of the week.

[Luigi2011]

I think having just read some of your comments on different threads on this forum it is clear you dont fit or have an extensive knowledge of inverter controlled systems. Your knowledge of heat transfer and the impact of temperature difference seems flawed, its laughable that you think that a large volume of water with a temp difference of say 1 degree between input and target will respond in a few minutes...to do this you will have to have a far greater temperature difference, heres a couple of equations you should maybe familiarise yourself with if theory is something you aspire to understand. Q = m cp dt and E = cp dt m

If the information that Mitsi and Daikin give is minimal then maybe you could point me in the direction of a company which can substantiate their factory data and performance claims with such a vast array of independantly accredited field trial data measuring all important aspects of the system performance on systems setup as described in real houses, which to be honest is what matters at the end of the day? which units do you fit as a matter of interest?

I think Jon had some good points for discussion based upon his experiences with similar equipment, and based upon our experiance what i was saying is that I have not come across the issues he has raised so to simply say it is nonsense really isnt very constructive. You go ahead and fit your buffer vessels but if you buy the right kit and know how to design/install it then there is no need for this extra cost/issue with packaging.

[MikeHolm]

We seem to re-hash this discussion a lot so it is obviously a contensious issue.

We have the exact same problem with boilers in regards to minimum heating outputs and low heat loads so it should be old hat. Unless you use a buffer you will have this issue, a 30% minimum capacity on an inverter is just like Baxi, Viessmann etcs' minimum fire rate. Most of my service calls are in the shoulder periods and the cycling reduces the longevity of the product.

I would think, for longevity sake, that a buffer should be used regardless of what Mitsi and Daikin say and it should take out half the nusance call backs.

There is still the difference in opinion between BF and MF regarding whether the defrost heat comes from the top of bottom of the tank and that would be a good discussion. IMO, if you have a high stratification in the tank, the defrost should come from the top and if not...take it from the bottom and save equipment. Taking MFs much earlier advice I have installed a sparge pipe (no idea where that name came from) in the bottom of one tank so the only output from the top of the tank is to the loops.

In Canada we often have mechanical rooms as big as UK living rooms so room is less of an issue.

[Luigi2011]

We seem to re-hash this discussion a lot so it is obviously a contensious issue.

We have the exact same problem with boilers in regards to minimum heating outputs and low heat loads so it should be old hat. Unless you use a buffer you will have this issue, a 30% minimum capacity on an inverter is just like Baxi, Viessmann etcs' minimum fire rate. Most of my service calls are in the shoulder periods and the cycling reduces the longevity of the product.

Capacity of fossil fuel boilers as oppose to ASHP is generally greater and hence if the minimum burn is a percentage of maximum output as with inverter driven units then it is likely that at a given ambient tempertaure there will be requirement for a buffer vessel as the heat losses of the building have diminished to a point whereby the minimum burn is too high respectively. When designing for an inverter controlled ASHP system with no buffer so long as you can ensure your sizing is correct to the point that peak losses can be met and that minimum step occurs at an ambient temperature at which heating is not required due to the thermal performance of the building and diminishing heat losses then there is no problem and the extra equipment and labour becomes surplus to requirements, its actually not that difficult and i suspect the manufacturers who design and develop this equipment and know a good deal more than us have realised this before giving otherwise irresponsible advice. I agree that if you were to apply fossil fuel boiler sizing philosophy to an inverter driven unit without a buffer vessel then the problems you describe would likely be present and in my experiance this is often the issue with nusance call backs you refer to, but that is to be expected if the 'using a sledgehammer to crack a walnut philosophy' is adopted. Having never had such issues on systems designed and installed correctly i think that speaks for itself, not to mention the proof of years of sophisticated monitoring data to substantiate.

[Jon Glanfield]

I can understand what you are saying Luigi in terms of the sizing, which is crucial, but what about flow rate. I have never been comfortable with the range variation that Mitsubishi permit and I have not seen figures to confirm what the impacts are of the flow rate being at a certain lpm. In addition the flow rate will vary on a thermostatically controlled system as stats open and close, which again has to have an effect on performance at some level, though I have not seen it quantified.

Jon

The other point about flow rate is that I am not sure that we have hit nominal that many times on Mitsubishi systems, it's usually mid way between the min and max levels. When we use a buffer though we hit max straight away. Interestingly too we usually install a variable flow setter on all units, and despite large bore pipework between a buffer and the HP, I don't think we have ever had to restrict flow or adjust pump speed because it has exceeded nominal.

You are also not impacted upon by incompatible delta t design on the heating side, flow restrictions, excess loop lengths and all the other curve balls encountered on existing and sometimes newly designed systems, which in an ideal world you would design out, but this just isn't the reality of the situation.

We do like some aspects of the Mitsubishi but it does not lend itself to a traditional buffer installation, in that a 240v signal is required to fire it, if this were in the form of a pipe stat, the set point becomes fixed not floating and WC is moribund. Other units will simply look after the buffer via a sensor and without needing a 240v signal and space heating can be drawn off the other side using simple conventional controls e.g. pump and room stat.

I guess in the end everyone has a different take and if you have not encountered issues then great, have you tried units from any other manufacturers at all for comparison?

Jon

[mad fridgie]

I neither use inverters, nor do I use a thermal buffer, but more than likely a larger low loss header than most.
I do not have problems with defrost or any issues with controlling comfort levels.
When using reverse cycle defrost (as most manufactures do, as i proved above is a poor method for this application), on systems where zones are controlled, I simple install an overide relay which opens the zone valves during defrost.
As most systems use static heat exchangers and relatively low flow temperatures, sudden but short changes in temperature is going to make little or no effect on household comfort. Neither is short term losses of heat (on and off of a fixed spped heat pump)
Inverters are great technology, but should be used when direct reaction speed of the enviroment is required, basically on forced draft heat exchangers.
Inverters also generally have poor power factor, so not always as good for the enviroment as a fixed speed.
I think most systems are just overkill, use the KISS principle.

[MikeHolm]

MF, are you saying that you would rather use hot gas on a system with a single zone and no buffer (big LLH only) in other words is there a place for the big buffer tank, reverse cycle on the same type of system, and if so when?

[mad fridgie]

MF, are you saying that you would rather use hot gas on a system with a single zone and no buffer (big LLH only) in other words is there a place for the big buffer tank, reverse cycle on the same type of system, and if so when?

Does not have to be single zone, I do not see any major disadvantages with buffer tanks just not alot of advantages either, Unless is really big and a true thremal buffer
Reverse cycle is the industry standard, just showing is actually the wrong defrost method, do not expect to covince many here that I am correct,

[robb1303]

From reading this thread it seems at the very least everyone is agreed that flow rate is critical to the correct operation and to reaching the claimed output of the heat pump.
Can you installers let me know how you are checking that the flow rate is at the manufacturers recomended level, is there any specific equipment or techniques you use to prove this?
If the flow rate is below the recomended level, what would be your next course of action assuming the circulating pump supplied by the heat pump manufacturer was already at its highest setting. Would you change the the heatpump circulator for a larger size of fit an additional circulator in the system to run in parallel with the first, or maby go about rectifying system resistances in the pipework?

With regards to an inverter driven heatpump without a buffer then as long as some circuits in the house are open and its not just feeding water back on itself then there will be enough warm water to sustain a defrost, lets not forget whatever means of transfering heat into the house will also run in reverse to take heat from the room. It would be reasonable to assume that is the heatpump is requiring a defrost then the ambient will be low, the house will be requesting heat and the heating circuits will be open, I dont know when you would get a situation where the heatpump needed defrosting and the heating circuits in the house were all bypassed?

Additionally a inverter driven heatpump running at 30% isnt cycling in the way a gas boiler on low burn is operating, I think someone was refering to an increase in failure rate due to low running which shouldnt be the case as long as the inverter is carefully matched to the compressor to ensure the correct oil delivery to the moving parts isnt compromised, the new generation for digital scrolls from one big manufacturer do however seem a waste of time.

[Jon Glanfield]

Hi Robb I know this thread is a bit aged but just been re-reading it, in terms of defrost and open circuits, the problems can arise in say a holiday let over winter, where internal ambient is set low, so minimal heat demand, but if and when a defrost is required there is not sufficient to complete an effective de-frost. If the HP is only heating the buffer, this will have a store of heat ready to defrost the unit regardless of internal ambients.

Flow rate can be measured empirically with a calc using delta t's that I can dig out if you want it, but we tend to install a flow meter with a balancing adjuster to be certain.

If you have got less than required flow rates you would have to take an holistic view of the installation, but this to me supports the argument for buffers, because you more or less guarantee rates in advance and no future issues on the house side will impact, like customer fiddling, sludging etc, whose effects may take time to reveal themselves on the HP, but only when the damage has been done.

[fixmyheatpump]

All good stuff here very interesting...... has anybody got any information or actual monitored data on how the inclusion of a buffer vessel in an inverter driven heat pump primary circuit effects annual COPs or more importantly running costs please? After all that is the reason people are buying these systems from us so it should be the first question on the agenda. I've not seen any evidence but logic says to me that constantly heating a larger volume of water than is actually required will increase the amount of energy used annually, and the argument of whatever is put into the vessel will be used in the house isn't good enough we need to see hard proof before they can be specified. In the meantime we will continue to install as per the manufacturer's recommendation and to date we've had no freezing issues on our installations whatsoever, although I have seen a few in a service capacity ;-)

[Jon Glanfield]

I haven't seen any data but was talking to manufacturer who had some measurement which suggested that short cycling was not a huge issue in terms of efficiency, but the wear on the compressor is and might not hit running costs but could affect life cycle costs if a replacement is needed as a result of a premature failure.

You are right though measurement is needed to evidence COPS/SCOPS but the wear issue may be more difficult to plot.

We obtained a formula recently for the calculation of when an inverter driven unit will start short cycling and on average with a well known Japanese brand it tended to be at about 10 degrees ambient, which we believe is bit too low. This is with a ramp down to 30% of operating capacity, with others that only ramp to 50% the ambient is higher and the issue is compounded further if the heat pump has to be oversized to suit the load due to gaps in the product range e.g. a 14kw unit on a 9kw load.

Jon

[mad fridgie]

I think you have to choose your word carefully, short cycling can be seen in different ways, to me short cycling means very rapid "on's" and off''s" and yes can cause major problems, how ever controlled cycling less so.

[Jon Glanfield]

I think the difficulty here is though that the cycling we are talking about is unpredictable and will be dependent on variables o/s of the installers controller, e.g. overall system design, resistance of a an auto bypass (if fitted), number of loops available on a zoned system, no of TRV's shut down, end user stat settings and so on.

With a buffer there is less need for the user to be involved with the heat pump side and this should mean less potential for issue and the heat pump will avoid any kind of cycling if the buffer is large enough.

[mad fridgie]

Because of the nature of the cycle (slow acting) a simple minimum run and minimum off timer configuration is more than sufficient.