Communal Heating and Heat Pumps
This article is a study of how to implement district heating using Heat Pumps.
The key objective are:
- Decent performance hot water and central heating.
- Return temperatures as low as possible, and at least below 30C at all times.
- To also keep the flow temperature to the system as low as possible to maximise heat pump COP.
- Domestic hot water to be protected from Legionella.
- Cost effective.
- Simple to install and commission.
- 1 Plantroom Schematics
- 2 Heat Pump COPs
- 3 Heat Pump Driven Communal Heating vs Ambient Loop Heat Pumps
- 4 Solar Network
- 5 Pipework Losses
- 6 Heat Pump COP
- 7 Commercial RHI
- 8 Heating Startup
- 9 Roof Space
- 10 Cooling
- 11 Borehole Density
- 12 Solar PV vs thermal
- 13 Future-Proofing for Hydrogen Boilers
- 14 Defrost Cycle
- 15 Embedded Carbon Content
- 16 Case Study
- 17 Initial Logic
- 18 HIU Selection
The following generic schematic should be used as a basis for plantroom layouts. Note that the two buffers can sometimes be combined into a single stratified store.
The primary buffer is fed with water heated by the heat pumps, and can be further heated to varying volumes (from the top down) with high grade heat from boilers.
The secondary buffer is only heated by the heat pumps.
Loading Store Layout
The following layout enables heat pumps to recover water from 25C to 55C in 5C or 10C steps, making best use of the lower temperature COPs. Only the final step will be at the lowest COP. Heated water is then discharged into the main thermal store, where it can be stored or used without loss of temperature.
Once a volume equal to or greater than the volume of the loading store has entered the main store, this is moved by a pump to the loading store, where it is then reheated before been fed into the top of the main store, replaced with the next batch of water to be heated.
Such a system has the added benefit of simplicity. There are no control valves required, nothing to modulate, and the logic is simple enough for wired control rather than BMS control (although PLC control is advised).
Heat Pump COPs
Below are COP figures for a Daikin Altherma High Temp Heat Pump.
Calculating COP for a heat network will depend on the buffer arrangement as well as the network return.
With a 5-10C rise through the heat pump, as per the chart, and a network return at 20C or below in Summer, then the water is reheated in stages, from 20-30, 30-40, 40-50, 50-60. To achieve such staged recovery takes two buffer stores, or four series heat pumps, with a valve arrangement to change order at intervals, or cycle on frost levels.
The COP will reduce at each stage but a rough interpolation give a COP for DHW of roughly 4.
Heat Pump Driven Communal Heating vs Ambient Loop Heat Pumps
An ambient loop draws heat from the ground, distributed to properties at low temperatures with individual heat pumps in each property that can achieve 60C, with a hot water cylinder for DHW production.
They work well and are commonly referred to as the latest generation of district heating. However, they do have disadvantages and will not work in some instances.
Advantages over standard district heating:
- Higher COP when running central heating.
- Lower heat loss from pipework.
- Cooling function.
- No energy centre.
- No energy centre. (explained below).
- 100s of individual heat pumps vs one or two large heat pumps. (capital/maintenance costs)
- Minimum density - load per bore hole becomes unsustinable. i.e. better the more horizontal people are distributed.
- Can't be upgraded easily - if you want to pull in a hydrogen boiler then that will never be possible.
- May not cope with higher temperature heating systems.
- Requires a cupboard per property.
- Noise in property.
- Stored hot water - unvented regs / disharge pipes / Legionella
- Requires annual service per property.
Many points are subjective, and will vary from contract to contract.
Having no requireent for an energy centre may sound appealing. Only to estate agents. If you are an engineer responsible for maintaining a building's services, a single goto room where all the important kit is housed makes life much easier than accessing individual properties to service equipment. However, with a standardised product dropped into each property there is less room for both design and installation errors - so you cant end up with a hugely oversized plantroom with expensive BMS, all of which has been designed incorrectly.
It is important to compare like for like - the latest generation of ambient loop technology, installed to manufacturers instructions, with the equivalent in heat pump driven district heating. This last item may be rare, especially as the rules regarding pipe sizing have been inaccurate for the past 15 years or so, and its only in the past few months the error has been redacted. Furthermore - their are less cowbow manufacturers in the ambient loop industry (to my limited knowledge) than there are in the HIU industry. Heat pumps have always needed proper testing for RHI purposes. HIUs, have only just started getting any testing, and you just need to look at the industry results to see that the average isn't representative of the current technology. Such sites are however coming online, making it possible to go by metered results when comparing methods.
Maintenance costs is ususally the big cost that isnt considered during a tender process - by which much of life if driven. With standard district heating, using direct central heating and a modern DHW only HIU, there should be no need to ever go into properties to service equipment. A visual check every few years may be reassuring, however a remotely connected HIU tells you more than any engineer on site can, and direct heating ensures water quality throughout is under your control (their can be a sting in that tail). So an HIU property can go ten years without needing a service. By comparison, an unvented cylinder, used for DHW in ambient systems, requies annual certification by a registered engineer to be legal. Anyone responsible for arranging access to perform maintenanxce knows the advantage of not having to.
The heart of the comparison however if efficiency. Running at 45C, or at 55C. 2.61 vs 2.77 power inputs for an 11kW output on the heat pump above. 4.21 vs 3.97 COP. A better calculation nis needed, as for DHW storage you need to go above 55C for Legionella anyhow, so for DHW loads the COP may be lower/the same as a 55C network.
To go up to 70C for winter -5C peaks reduces the COP, but not that badly given this is a rare event.
So a centralised heat pump will work with high temperature (70C+) radiators when occasionally needed.
Rougly take plantroom space and cupboard spaces cancel = 75m2. Assume pipework within buildings cancels.
- 100 x Micro Heat Pumps sized to individual peak loads (assumed £375/kW - same as air source below, 3kW = £1100 ??? please advise)
- 100 x Unvented Cylinders sized to individual peak loads (£550 each)
- 100 x Cupboards
- Boreholes to cover full load
£167,500 + boreholes and ground array
- 100 x DHW Only HIUs (e.g.) (£480 each)
- Air Source Heat Pump(s) (£375/kW x 200kW = £75,000)
- 2 x Buffer Stores in series (£5000 each)
- Twin Pump Set (£5000)
- BMS (£5000)
- 1 x Plantroom / Roof Area
Ambient loop would appear to cost £25,000 more in equipment, and then requires a ground array and boreholes.
Are there cases where radiators will suffice? With the high temperature heat pumps, radiators running at 55C will cover all but the coldest days.
(Can anyone help put further prices / corrections to these?)
This is a big one. Its also one the client needs to figure out for themselves.
It is not true that HIUs on a heat network require annual servicing. When central heating is direct that takes care of water quality, and a remote connected HIU tells you more than an engineer can. So, there is absolutely no need to go visit a property - unless something shows up with the data. 'What about strainers' I hear tyou cry. HIUs tell you the DP across the HIU - including the strainer, so its easy to spot remotely. If any checks ever show up debris, then it becomes a different ball game with a service to every system potentially justified.
Heat pumps and unvented cylinders do require servicing, so it is of great benefit to have only a couple of them housed in one room that engineers can get to whenever they need to.
Question: Given an HIU should go at least 10 years before components may want examining or replacing, what is the cost of an upfront 10 year full annual service cover for a heat pump and unvented cylinder?
From a fag packet costing, it appears that for blocks of properties, abmient loop systems are significantly more expensive in both capital and maintenance than an air source driven DH system, restrict one to a single technology moving forwards, may have ground density limits, and imposes stricter limits on central heating peaks and temperatures.
For low rise developments, the thermal losses from widespread DH pipework to the ground make ambient loops more efficient, and having a heat pump in each house adds up. Losses on larger DH pipes feeding large blocks of load have relatively low losses and also make sense.
Ground loops (be they connected to ambient loops of DH heat pumps) are best applied to necessary underground networks. Within a buildings envelope distribution is most effectively achieved through a 55-75C heat heatwork with HIUs, allowing for maintenance and technology upgrade over time without a need to access or upgrade individual properties.
Theoretically the most efficient (low carbon) system would be:
- Low temperature ground array / waste heat loop (low grade heat) - between buildings
- Heat pump in each building
- Oversized solar thermal array (high grade heat) feeding excess heat into ground array
- CHP plant potentially
Such a combination is most effective when the high grade heat from solar thermal can be stored and used for DHW loads without the need of any other energy input.
The oversized thermal array (if possible to oversize) would cover 100% of DHW in summer months, under normal conditions, and provide surplus heat that can be dumped into the ground, which acts as a thermal store. The oversized array will continue to supply DHW in non-heating months.
As low load heating starts up the excess solar thermal will cover this as well as DHW.
Then as is gets colder, the heat pump (and CHP) can top up what the solar thermal does give to drive the heat network.
In winter, the solar with do virtually nothing (less than 10% summer input) so its up to the heat pump. Depending on how oversized the solar thermal is will determine the ground collector temperature, and hence the COP. The geology of the ground will have a large impact. A COP of 5 has been recorded from such an approach in a large domestic property. Such an approach has yet to be seen in a heat network.
Should there not be much room for ground loops, then the heat pumps would be 2-stage, using an air source heat pump to top-up temperatures as the ground array is relaxed to prevent the ground from freezing.
Such a system can be provided in three ways:
- An energy centre (heat pumps and solar storage), with HIUs in properties
- An energy centre (heat pumps) with PHEx/DHW storage in properties
- Solar / ground ambient loop and heat pumps with DHW storage in properties (can solar at 50C+ be transferred through to store without heat pumpp operation using valve arrangement?)
The last two assume you can utilise DHW storage for solar storage, which can get complicated. Otherwise central solar storage in energy centre will be required.
Pipework losses for a building heat network are roughly 6% to 10% when sized to latest guidance, depending on supply temperature, and assuming HIUs without a keep-warm turned on (as most ones we commission are these days).
Using guidance properly (into the realm of realism - size end-runs to 2m/s and decent insulation - and said as an author to the original IoP guidance) and you can get pipework losses to 5% - roughly speaking.
So pipework losses within a building are not significant when compared to COPs of heat pumps for example.
Pipework losses between buildings can be the killer for a standard heat network. In dense areas they make sense, but as density drops and we get to houses rather than blocks, then the heat loss per property can shoot up. This can be mitigated to some degree by DHW storage, and by better insulation on pipes, however the benfits of a low temperature array when underground make sense. Once you get to single dwellings, you need a heat pump per property anyhow - so it becomes an ambient loop by nature.
Heat Pump COP
COP is the king in a low carbon network. The higher the COP, the lower the carbon used, and as COP can vary so much, it is most important to understand the factors that influence it.
- Input temperature
- Output temperatures
- DHW / central heating ratio
- Air temperature
- Ground volume available and geology
With ambient loop systems, or any heat pump using stored domestic hot water, then it is widely accepted that 60C is the required storage temperature for Legionella protection.
HIUs drive hot water instantaneously, with no stored water, and with no recorded cases of infection from an HIU is is widely accepted that they can run at temperatures as low as 45C delivery. The exception is where a keep-warm is deployed that maintains temperatures at 30-40 - this will create a biofilm if left stagnant for long enough, but can be mitigated by occasional strylisation or turning keep warm off when unoccupied.
The COP for DHW production is the heart of heat pump efficiency. Buildings are so well insulated that heating periods are shorter, and the steady state loads are tiny. DHW load accounts for the majority of energy use.
With HIUs, the network runs at DHW temperatures all the time. Peak DHW loads are roughly 8am and 8pm, and it is only during these peak periods that primary temperatures need to be high enough to drive design peak DHW loads. At other times, for example overnight, the network temperatures within a building can drop to 45C at the limit - given skin DHW temperature is 38C, this still leaves a 7C difference to drive heat exchange. The majority of load occurs during the peak times, as well as lunchtime, which is when the network will need to be at 55C typically.
The COP of an ambient loop system will depend on the way the hot water cylinder is reheated, along with the duty and layout of the primary coil. To drive heat through the coil will require a temperature difference, and with a target of 60C that requires the heat pump to get a bit hotter than this. If the difference is small, the transfer is small, and the time taken for the stored water to rise the last few degrees can become unexpectedly long and cause cycling of the heat pump. Plate heat exchangers are used in place of coils on some systems, however they act in single pass (15-60) causing the heat pump to run at high temperature (51-61C) throughout recovery, losing the low temp COP gains.
It is easy with both approaches to knock overall COP by not allowing the benefits of the lower temperature recovery of stores to be realised - such as by using bypasses to maintain flow rates or low duty coils in cylinders.
Commercial RHI would apply to any form of heat pump driven heat network.
Commercial RHI would usually be based on metered data, taking electrical input vs delivered energy.
A tiered tariff structure operates on a 12 month basis, starting with an installation’s date of accreditation or its anniversary. The regulations specify that during this 12 month period, an initial amount of heat equal to the amount of heat generated by the installation running at its installation capacity for a set periodof time is eligible for a Tier 1 tariff. Additional heat used is paid at tier 2.
For ground or water source heat pumps the first 1314 hours (15% of a year)will be payable at the higher Tier 1 tariff.
The massive 9.56 pence / kWh, for ground source heat pumps, and 10.98 for solar thermal are the gold. Air source receive less than a third of this - why? It could be a reflection of COP from studies on ground source heat pumps show they have 3 times the COP (A COP of 9?). It matters not from a financially twisted technical perspective other than we should prioritise the use of solar thermal and ground source over air source, only using air source to top up when needed.
The RHIs make the use of solar thermal very attractive. The RHI from solar may cover the costs of electricity for heat pumps.
The 'bias' caused by the RHI towards ground source is a driver to maximise the energy input into boreholes maximising the energy that can be extracted. Water is the best source of fresh ambient heat, and if you are sitting on an underground river then thats an RHI diamond crusted jackpot, where giving energy free to your neighbours earns you money.
Electricty is something like 13 pence/kWh. With a COP of 3 applied thats 4.3 pence/kWh.
A 9.56 pence/kWh tariff leaves 5.23 pence/kWh.
Very rough 3000kWh / property / year, for 100 property block, thats over £15,000 / year profit on energy use. Maintenance costs and depreciation may need to be included, especially if the technology isn't expected to last the 20 year RHI contract.
A final piece of advice is to follow old inefficient guidance as Ofgem wont have caught up, apply diversity incorrectly, oversize for safety, use luxury tap fittings, keep insulation to a minimum - anything that can be justified in the tier 1 load (only joking). Such practices inevitably end in problems, with actual in use metered load far lower than design load - payments are on actual energy used, but costs increase because of the inefficiencies. In fact, one needs to be properly on the ball with design and commissioning if calculated profits are to be realised.
Peaks in load generally occur in the morning. Their is an 8am DHW spike, and also a central heating spike.
Central heating systems fitted with timeclosks have the potential to create large surges in load at startup. The graph below shows a development of 100+ systems all turning on together in winter.
Such peaks in central heating load can only be met by central buffer storage, otherwise heating startup periods are considerably longer. Removing timeclocks and user control helps.
Boreholes can suffer from flash freezing if hit with too much load, so some form of limits need to be in place on ambient loops.
Their are two points that a building connects with energy supplies - the ground and the roof.
The roof provides access to solar thermal and air, utilised by solar thermal and air source heat pumps respectively. As ground loop density is limited, making use of both sources to the full makes sense, and may have COP advantages relating to air/ground temperature ratios.
An ambient loop between buildings can be fed to the roof where water-water and air-water heat pump input is provided alonside the solar themal. The lower static pressures on the roof enable the use of lower pressure thermal stores with potential cost savings. And its an obvious location for cold water storage.
It raises the question if the cold returning ambient loop can be used in summer to keep the building core temperatures down, perhaps by venting cold riser air into corridors.
Cooling is nice, especially on the hottest of days, and if you live in a block of flats with the normal (oversized) DH system that cooks buildings from the inside, then it probably feels like you couldn't live withut one (opening windows may simple not be enough to dump heat losses from pipes and hot water systems).
However, its an additional energy use that would normally be used in buildings where opening windows isn't an option in Summer. If you live to a budget then you would normally open a window or turn on a fan, rather than pay for and operate an uneccessary air conditioning system.
From a carbon perspective, cooling introduces the potential for additional energy use to overcome a problem that has more elegant solutions. These may include sizing and insulating pipes properly, turning off uneccessary keep-warm, or utilising riser cooling in the building (i.e. cool down risers and hence the building, rather than overheat risers, corrodors, cupboards and building as we have been).
In a situation where boreholes are not sufficient to cover load, it is believed that the introduction of solar themal arrays to reheat the ground during summer make for a more efficient system.
It is equivalent to providing greater ground surface for solar heat collection - where all the heat comes from in the first place (unless there is water flow). A better study will be required as solar thermal collectors collect heat at around 90%, where the ground may be concrete that will radiate heat back to the atmosphere far more than a thermal panel.
Solar panels are not free however, and existing recreational ground provides a free collector.
Solar PV vs thermal
If you can get the COP higher than 3 on a heat pump then PV adds up initially, however where ground loop density is lacking the heat from the ground is more valuable than the sun (as its always there) so it will not be so straight-forward. An air source heat pump with a COP of 3 would still make sense, however RHI tarriffs may twist this. As you get into oversized solar thermal arrays, then it will become a balancing act, requiring a calculation to determine the best mix of pv/thermal to get the lowest carbon content - i.e. minimising use of external electricity supplies to drive heat pumps, which only PV (or hydro) can do.
With a heat pump COP of 3, and a ratio of Solar thermal to PV efficiency ratio of 3 as well, the same amount of grid electricity is used. If the COP is above 3, then PV will generate more thermal heat overall.
If there is roof space for more input than is required, over a 24 hour period - then the question of surplus heat comes into play. Solar thermal can dump into the ground, potentially to be reclaimed (with losses) via a heat pump. Solar PV can be fed into the electricity grid.
As the Feed In Tarriff (FIT) for Solar PV has ended, Solar Thermal would seem to be the way to go - see https://www.moneysupermarket.com/gas-and-electricity/business-energy/a/feed-in-tariffs/
Wind turbines appear to be the best use of land/sea for renewable electricity generation, so its easy to see why PV has lost the FIT now there are cheaper options. Heat is a different matter. Heat is used locally (for hot water and heating) so it continues to make sense to incentivise local renewable heat generation based on metered use.
The only way to financially gain from surplus energy generation (excluding selling to neighbours) is to store it longer term - seasonally - and that takes massive storage, such as you get with ground source arrays - batteries don't come in these sizes without making the news. Surplus energy dumped into the ground will not register on the metered energy use, so no payment, however the increased ground temperatures make the heat pump COP greater, so the original solar energy does eventually contribute financially.
Excess solar thermal can be dumped directly into a ground array. Excess PV could be dumped via an air source heat pump - the COP could potentially be much higher than 3, however it requires both air source and ground source heat pumps.
40m x 8m area of solar thermal, at 150kWh / m2, offers 48,000 kWh per year. Thats £5000 or equivalent to 30 litres of DHW (at tap temperature) per proprety per day, however a portion will be dumped to ground reducing net gains but extending ground source functionality. A proper solar calculation needs to be done, including allowance for type of panels, solar irradiation, orientation, elavation, shading factors, and operating temperatures.
Future-Proofing for Hydrogen Boilers
If, as it would appear, that hydrogen is a viable fuel source for boilers, then that could change the perspective on the selection of district heating.
Hydrogen is a high grade heat source capable of running radiators, however one would expect such an emerging technology to improve over the initial decades of use. A central energy centre would be the obvious choice for managing the introduction of such alternative fuel supplies in the future.
Embedded Carbon Content
Should one examine the carbon footprint of multiple individual heat pumps over larger (diversified) central heat pumps ?
We all know the answer - yes if its significant over the lifecycle of the equipment. But its a another (groan) calculation to be done, just to know if its significant.
Due diligence and all that. Leave it for someone else to think about (hopefully).
Existing block of flats, with a small amount of external ground, but considerable flat roof space.
Currently fitted with radiators on a heat circuit, with a DHW supply and return, all fed from an energy centre fitted with two gas fired boilers with integral DHW storage.
Radiators need to run fairly hot (65C) to achieve heating loads - windows have already been double glazed.
Flat roof area provides space for solar thermal arrays, although some shading may reduce effectiveness.
The ground is too small for boreholes to drive load, however there is potential for a borehole to cover part-load.
Target is carbon neutral.
A simple approach may be to fit high temperature heat pumps, and offset the electrical use by paying for another wind turbine somewhere.
Carbon off-setting, as it is called, is an obvious choice, moving the energy collection to a more suitable location. Such schemes often fall into disrepute, as it al comes down to justifying the spend actually gets net carbon reductions. Is there a simple .gov run scheme for extending wind farms through heat network carbon off-setting? e.g. 1 turbine = £x = y kWh/year
Then throw as much solar thermal in as feasible to reduce the off-setting required.
Heat pump efficiency benefit from lower (input and output) temperatures, with a higher COP (Coefficient of performance).
Any heat exchange process introduces a temperature drop, so, to minimise temperature drop between supply and load, direct heating is beneficial.
To further minimise temperature drop between supply and load pipes selected for low thermal losses are beneficial.
Heat exchangers for domestic hot water (DHW) should be taken to 60C for an hour to kill Legionella, should the system be inactive for a period. This allows lower temperatures to be used for the majority of the time to improve efficiency. How low network temperatures can go for DHW depends on the end users acceptance. 50C would generally be desirable, while 45C would be an absolute minimum, as outlets require 40C, and a heat exchanger requires at least 5C drop. These two goalposts result in network temperatures ranging from 50C at low load, 60C at peak summer load, and 65C at higher loads, and should result in a heat pump COP between 4 and 5.
Given central heating temperature requirements for much (if not all) of the year will be lower than DHW, hot water storage local to properties would potentially enable the network to be run at higher temperature for DHW for as short as time as possible - just long enough to satisfy total DHW loads. This would however require reheating of stores to happen at the same time, which can easily be accomplished through controls strategy. Heat exchange for cylinders using coils requires a high temperature difference, takes longer to recover, and has high return temperatures. As such, it is beneficial to reheat storage cylinders via a plate heat exchanger. A standard HIU can be used to reheat a hot water store in this way, with a re-circulation pump.
Plastic network pipework should be used, as the low temperatures translate to longer pipework lifetime, and allow supply temperatures to be cycled over 24 hours to match load profile, with highest temperatures at the 8am/pm peak load times. Actual timing would vary based on live data (from plantroom heat meter). Plastic pipework is typically tested to rapid cycling from low to high temperatures (e.g. WRAS pipework approvals).
Flow pipework should be sized to the limit (small as possible = size to upper velocity limit), with margins for extra load taken up by varying supply temperatures. Return pipework is not critical due to its low temperature and heat loss. One exception may be a main supply pipe that needs to be oversized to allow for expansion of a site in future.
Pipework sizing must not be done on a general pressure drop per metre for all pipework. Flow pipework needs to be treated differently to return pipework, and pipework near properties should be treated differently from longer distance transport pipework.
External design temperatures used for sizing pipework and heat emitters should be linked to supply temperature, weather compensating the site. Where topping up boilers are provided, it should be noted that network supply temperatures can be raised to over 80C to accommodate one in ten year weather events (-15C outside). This keeps pipework sizes and heat loss to a minimum.
All flow pipework within buildings should have a minimum of 40mm insulation. Return pipework within buildings can be left uninsulated, or a minimum of insulation, as heat loss for DHW load on the return is negligible (return may even be colder than ambient), and heat loss in Winter will contribute to heating the building fabric.
As the return temperature performance is critical to the operation of a heat pump system, and given return temperatures are a function of HIU and heat emitter performance the following steps should be taken to ensure efficient operation:
- Any HIUs should be independently tested and their VWART figures used in the selection process. See https://www.thebesa.com/ukhiu. Note we have (and always have had) the lowest approved VWART figure in the industry.
- Keep warm modes should not be employed on HIUs, with thermal bypasses fitted on the distribution network, set to typically 50C, and located as far from properties as response times allow. See PAPER: HIUs.
- Heat emitters (radiators) should sized generously, fitted with preset TRVs that can accommodate the very low flow rates required for typical loads, and protected with suitable differential pressure control, typically to 20kPa. (note look at Danfoss AB-PN valve for heating control).
The most suitable HIU for such an application is the SLIM-EXTRA HIU.
Where underfloor heating is to be used the MIX HIU should be used.