Communal Heating and Hot Water Cylinders
We hope to clearly set out the theories and test results that apply. A forensic examination of the question, as best we can. Also, it is an aim to set out the logic of the control principles that can be deployed. This is not so much a study of what has been done, but what is possible given the use of networked HIUs, with and without storage.
- 1 Why Look at Storage ?
- 2 Background Material
- 3 Topologies
- 4 What would you want in your house ?
- 5 The Design Limits
- 6 Direct Heating, or Indirect
- 7 A Note on Thermal Stores
- 8 Flow & Return Temperatures used for Pipe Sizing
- 9 Coils or Plate Heat Exchanges
- 10 Heat Loss
- 11 Load Calculations
- 12 The Middle Road
- 13 Summary
- 14 Further Documents
Why Look at Storage ?
HIUs are smaller, and cheaper, so why bother looking at storage. If there is potential to save significant energy, and at the same time offset costs in pipework sizes and installation costs, then why not think it through, and apply some reason ? Until recently, storage was far more common, along with a requirement for a backup immersion heater.
Today, in London, you need to be a millionaire to purchase a property of almost any size, and it is becoming expected to deliver luxury levels of hot water supply that match the property prices, and cylinders are still the obvious and normal choice for the big builders.
With the application of modern control, it is possible to co-ordinate storage and significantly drop circuit temperatures for more than 80% of the time, and in the Summer allow them to turn off completely when no heating is required.
This article is not intended as a sales piece, but a technical study, with conclusions shown by evidence. To check our credentials please refer to our history. Historically, and currently, we have supplied both instantaneous HIUs, and a number of storage systems for communal heating projects, so hope to remain impartial in the selection and in advice on systems.
Below is a CPD Approved Presentation we have on Communal Storage, although due for an update.
It is possibly wise to first list the possible types of system, as the debate is not as black and white as initially may be thought.
- Plant Based
- Direct DHW, with all storage and heat exchange in plant
- Dwelling Based
- Instantaneous DHW using PHE
- Unvented hot water cylinder, or thermal store for DHW fed from network via coil
- Unvented hot water cylinder, or thermal store for DHW fed from network via PHE
- Unvented hot water cylinder, or thermal store for DHW fed from internal primary via coil
- Thermal store fed from network via PHE, with heating & instantaneous (via PHE) DHW from store
- Hallway Based
- Instantaneous DHW using large PHE
- Unvented hot water cylinder, or thermal store fed from network via PHE for 5-20 properties.
HIUs for Dwellings
Storage In Dwelling
Storage In Landings
The following schematics cover the most common input arrangements we have supplied. Note, patents apply to some features within these layouts.
Key to components:
1. Unvented hot water cylinder 2. Plate heat exchanger, DHW 3. Plate heat exchanger, CH 4. Recovery pump 5. Recovery thermostatic valve 6. Flow limiter 7. Thermostatic mixing valve for DHW feed 8. Expansion vessel 9. Pressure reducing valve 10. Expansion relief valve 11. Pressure and temperature relief valve 12. ABQM valve with TWA-Z actuator, DHW 13. ABQM valve with TWA-Z actuator, CH 14. Heat meter 15. Heat meter temperature sensors 16. CH circulating pump 17. Immersion heater 18. Cylinder thermostat 19a. Overheat limit thermostat 19b. Economy mode thermostat 20. Recovery hold-off thermostat 21. Primary strainer
What would you want in your house ?
Lets face it, the ability to fill a bath in a couple of minutes, or run taps and showers simultaneously without issues, is what you would expect in a decent hotel. Luxury perhaps, but if it comes at no extra cost, then why not ?
A lot of people swear by their combi-boilers when compared to hot water cylinders. The hot water is always available, where a cylinder runs out of hot water. What if you could have a storage system that manages itself to make sure hot water is always available?
And space. If, as hoped, the cost of running the system is lower, does the additional space of a cylinder out-weight the benefits. By using plate heat exchanger recovery and keeping store sizes smaller, would a wall mounted storage unit, or one the size of a washing machine be satisfactory? I'm guessing it depends on the savings in running costs.
The perfect world for a user, is no kit at all. Just hot water on tap, and preferable as much as one desires, as cheap as possible.
The Design Limits
- It is not the place of this article to ascertain the best heat sources. It is assumed for now that flow temperatures of 70C are possible.
- The results of then EST trials are used as a foundation.
- All heat emitters can run at reduced temperatures during warmer weather, although the degree not fixed.
- Calculations will not allow safety margins. The aim is to perform 100%, for 99% of the time, and not wind anyone up too much the remaining 1%.
- 5kW heating loads. Higher loads to also be looked at.
- Primary pipework with a bore of less than 15mm copper is undesirable. It offers no significant advantages, but arguably increases occurrence of blockages and trapped debris.
Direct Heating, or Indirect
For now it is assumed that direct heating is being installed. This is simply to investigate initially the most efficient approach. No doubt indirect systems also have benefits that should be noted, and these will also be looked at.
- Isolation of boiler plant from contamination from radiators within properties. New radiators, system filters & corrosion inhibitor are expected however.
- Where very high pressures or temperatures are deployed, it is often considered unsafe to mix such supplies with radiators. Bleeding a radiator with 6 bar and 90C behind has potential for accidents. In this initial study, we are working within the temperatures and pressure ranges of plastic pipe, and with low water content systems in general.
A Note on Thermal Stores
Thermal stores can be split into three DHW categories:
- No DHW (heating only)
- Those using coils to generate DHW (cheap)
- Those using plate heat exchangers to generate DHW (decent performance)
They can be further split into the following central heating categories:
- No heating (DHW only)
- Those feeding central heating directly using storage water
- Those feeding central heating via a plate heat exchanger
Further, they add more options, especially regarding heat sources. One of the most advanced district storage system we have yet deployed was in partnership with ZedFactory, where the development was fed using district biomass, in conjunction with local wood burners in properties, and communal solar. Believe it or not, as well as in properties for landlubbers, the full system was deployed in the two ZedBoats, moored up in the Thames, and each with four flats on board. The communal boiler was an Ashwell wood chip unit, with small wood burners in flats, each contributing to stored hot water for both DHW and underfloor heating, and solar panels on the roof of the top deck. A carbon neutral floating example of whats possible, and only with thermal stores.
Flow & Return Temperatures used for Pipe Sizing
Reference is made to IEA operating temps report 1999, courtesey of Paul Woods at Cofely UK.
From contributed experience and tests, the biggest savings typically lie with reducing pipe size for capital cost, rather than getting hung up on the different heat losses of differing diameter pipe. In brief, smaller pipe is much cheaper, more responsive due to lower volumes, and easier to handle and install. It also has lower heat losses.
It should be stated from bitter experience that although we are working within the design limits of plastic pipe, the use of plastic fittings that reduce the inline bore of the pipework must be avoided at all costs.
To reduce pipe sizes for a given heat load, one increases the temperature drop, lowering flow rates, and hence pipe sizes. This is limited by the nature of the load, however the principle of keeping temperature drops high is a given.
It therefore follows, that there is no point dropping the supply temperature from a boiler plant, unless it makes significant efficiency gains at source. It thereby makes sense, with gas boilers and CHP, for example, to design for an 80C flow. With CHP you could aim for 90C, for periods of time the pipework allows for.
So, with design flow temperatures at 80C, the return will be what it will be, given the selection of heat emitters. Lets aim for 35C, so size the emitters / radiators for 78-35C.
Coils or Plate Heat Exchanges
There is no doubt that coils are a simple cheap way to reheat a cylinder, and there is also no doubt that a plate heat exchanger, and the associated bronze pump, performs in a completely different league, and offers far more efficiencies to be realised.
The following test is of the recovery of a standard 150 litre unvented hot water cylinder using a standard coil.
The return temperature to the network rises fairly rapidly as the store rises in temperature, and by the time the store has recovered to around 60C, the return is approximately 73C. During reheat, when the store is just been topped up, the return will also be in this region.
The following similar test is on a cylinder with a much better coil that is twice the area, and pushed further down into base of cylinder. Return temperatures are significantly improved, as are reheat times as a result. The base of the store is also brought up to temperature and as a result their will be a higher total volume of heated water in the store. The graph also shows how the system recovers with various draw off loads. The return temperatures will never as low as from a cold start, as part of the coil will always be in warm water when the cylinder thermostat initiates. Note the graph highlights a sensor difference (the flow temperature is reading a few degrees too low relative to the store sensors). Tests will be rerun with common types of sensors to ascertain exact temperature sensor accuracy.
The graph to an plate heat exchanger system (our Digi HIU) shows the return temperature performance by comparison, and will also give an indication of the profile with a cylinder recovered using a plate heat exchanger.
The return temperature is typically around 25C under load, and never rises above the DHW set-point of 60C while the system is in keep warm mode. If used with a cylinder, the steady low load would provide a fairly steady return around 20C, while recovering cold water, with return rising to 57C during reheat.
Graphs for plate heat exchanger recover, and high duty coils, coming soon...
Not all coils are the same, and don't forget not all cylinder stats are positioned or set the same.
The performance of indirect coil stores can vary wildly with:
- Surface area (how much coil)
- Flow rate, and hence turbulence for heat exchange
- Number of coils in parallel, and their bore
- Positioning of the coil
- How the coil is formed - coil in coil, Catherine wheel, or standard
- Coil material - copper or stainless - ribbed or smooth
- Is the cylinder thermostat an immersion or a strap on
- Where the cylinder thermostat is positioned
- The hysteresis of the thermostat
- The thermostat setting
- And last but not least, the direction of flow through the coil
Benefits of PHE Recovery
Apart from the low return temperatures, the use of a plate heat exchanger makes other functions possible:
- Instantaneous performance, with hot water available on demand to limits of heat exchanger.
- Storage performance, with an unvented delivering 200kW with 22mm pipes, and considerably more possible.
- Input can be varied within PHE's limits - 3 to 100kW for a typical PHE.
- Variable volumes recovered as the cylinder reheats to full temperature from the top - down. You can just as easily reheat 5 litres, as 150 litres, it will just be that much quicker. A smart system will vary the volume recovered to recent load patterns to minimise standing losses.
- No cold spots in base of cylinder, as PHE circuit can draw from the very base and reheat every drop of stored water
HIU with Storage
The simplest way to achieve the benefits of PHE recovery to a storage cylinder is using an HIU to provide the primary interface, along with a DHW recirculation pump. The pump is connected to the cylinder thermostats initiating recovery of the store, as well as feeding heated water directly into the DHW feed to taps.
The rate of recovery is set by the recirculation rate. This can be set be using one of the following:
- Flow limiters
- Pump speed setting
- Inline valve or fixed orifice
With more advanced control, the pump speed can be altered to suit circumstances.
Orchard Partners' Exergenius
This is a system patented by Orchard Partners that represents possibly the most interesting heat exchange input to a domestic store. The system makes use of heat inputs in series, going from low to high grade. The system allows the proportions of electric and thermal energy used by a system to reheat to be controlled. If electricity is the cheapest energy at a given time of day, then systems can switch to using up any available low cost electricity, reverting to heat when it is more economical to do so. The electric input allows the higher domestic hot water temperatures to be realised with the heat network at lowered temperatures, topping up temperature through the circuit.
An argument often leveled against storage is that of the increased heat losses, but what are the facts ?
A poorly insulated cylinder of old can give off some horrendous heat losses. Modern stores with high levels of insulation can retain heat for days. Likewise, many HIUs are simply radiators, while others (including ours :) give off less than a light bulb. One should compare the best with the best, and factor in managing when the store reheats, and by how much.
It is certainly the case that with modern insulation, experience shows than in most cases the district pipework cools faster than the HIU or cylinder, so reducing losses in pipework (better achievable with storage) may be the deciding factor in the heat loss argument.
With a peak central heating load of 5kW, this equates to 1.7 litres/minute, or 0.1 m3/h - fairly typical from recent experience on sites.
For hot water, we are working towards a total recovery period of 4 hours per day, as a starting point. This is to at first explore the benefit of reduced pipe temperatures most of the time, without any hot water load. This period will be split over the day to best deliver peak demands.
As a worst case, for sizing, 300 litres per day, at 55C, bunched into two heavy hours, of 150 litre draw-offs. This also ties in with the EST data.
If using plate heat exchanger recovery then return temperatures of 25C can be assumed. Heat exchanger selection will easily allow this. Therefore sizing to deliver 150 litres at 55C over one hour, equates to 8kW. 4kW if over the full two hours, but draw-off times vary around the normal. A precise study of the stats to follow these assumptions.
At an 8kW load, with primaries at 78-25C, the flow would be 2.2 litres per minute, 0.12 m3/h.
Running various load profiles through the Waterload Tool on the Heatweb Flash Tools Page, 100 litres of storage would deliver some very decent loading, if drawn during peak times (when primaries are raised in temperature for hot water recovery).
In total, you're looking somewhere between 3 and 4 litres per minute to satisfy both hot water and heating at peak times, for someone who uses 300 litres of water per day.
Looking first to the branches to each property. Distance of 5m each way, 10m total run, would lead to a pressure loss in 15mm copper tube for 4 litres per minute of 3kPa.
How many flats would a 15mm pipe service? Allowing a pressure drop on common run of 25kPa, and a star formation, with 5m branches as above, and starting at 4 properties. 3.5 x 4 = 14 litres/minute. You lose 1.5 kPa per metre of pipe. So, that gives you 16m of pipe, a run of 8m, before branching off to each property.
Performance Compared to Instantaneous HIU
For HIUs we need to apply the diversity. The Danish standard would ask for 40% for 4 properties, or 1.6 properties at full DHW load. Allowing 35 kW per property gives 56kW. At a temperature drop of 78-25C, this equates to 15 litres per minute.
Two of the heating loads should shut off if using HIUs with hot water priority, resulting in a heating load to add of 3.4 litres/minute.
Total peak flow to the 4 properties would be around 19 litres/minute.
Versus the 14 litres/minute for storage, under these conditions, the HIUs need approximately 30% more flow.
The instantaneous HIU figures, can't drop from here. We have sized at 56kW, for four properties, and given a cylinder feeding a bath tap could deliver over 80kW alone, 56kW is probably low enough. Also, if the HIUs do not have hot water priority, the flow increases to 24 litres/minute, almost twice that of the storage option.
At this point it may be worth noting an HIU function ready to field test whereby HIUs cooperate in groups regarding hot water priority. In this example, the heating in all four properties could shut down, when there are taps running in one or two. This would drop peak flows back to 15 litres per minute - much closer to storage.
With storage, there is however still room to maneuver. Allowing the stores to recover for longer, staggering recovery periods and possibly sizing slightly larger, we could drop to 1 to 2 litres per minute for all four properties (still delivering over 250 litres/property/day), dropping the total peak flow to 8 litres/minute for both HW and Htg.
If peak heating loads are higher than the 5kW, 1.7 lpm, used, the relative advantage of storage will increase as one approaches pipe limits. Certainly if the heating is driven through a heat exchanger, temperature drops will be lower, and flow will be higher.
If primary temperatures drop from 80C, the relative advantage of storage will increase.
As you look to more and more properties, the diversity of hot water use brings the two approaches closer and closer in performance for given primary pipe sizes.
The Middle Road
All systems have storage.
So far the argument has been limited to whether its is in the plant room or in the property.
There is a middle-ground to also be considered where properties in groups of up to 20 are fed from a local storage (S)HIU feeding domestic hot water. This system has been deployed with great effect at Oxford Brookes University, with a 300 litre thermal store connected to the heat network feeding 10 dwellings. Central heating was fed direct, and a Pandora thermal store was used to firstly avoid the Universities demand for no stored DHW for Legionella implications, and secondly to overcome the need for any discharge pipes. With the direct heating, this meant no discharge pipes at all.
With a single unit covering 10 properties, the capital costs per dwelling plummet. It allows you to afford to implement proper storage control, using large surface area plate heat exchangers to both draw heat from the network into the store, and to drive DHW demand. The units are located in utility cupboards and require no annual maintenance, although a manufacturer service contract is in place. In this example, the supplies were not metered, however if such an approach was used on private dwellings, there would be a need for a hot water meter as well as a heat meter for each property. These can go either in the utility cupboard, or in the properties, with M-Bus to read.
The system is capable to meeting normal demand at peak load times, as well as exceptional demand 80% of the time when load is not peak. It means that end user, if they time their baths correctly, could fill it in under a minute (pipework and installation limits allowing). A further advantage to the end user is they have no kit to maintain. In fact overall there is less kit by a factor of 10 to maintain.
Heat losses from such a system are arguable the lowest of all, with the losses from only one store, as opposed to 20 HIUs or smaller stores, while the main network can drop to lower temperatures or turn off.
Local vs Plant Storage
Once should ask, whats the point of storage in a riser cupboard, when you may as well have all the storage back at plant. From a maintenance point of view, the manager is taking on responsibility for maintaining units in the field, where if storage is in plant there is none.
In response, each property in the field will typically be metered - and will require checking and recallibrating every 3-5 years. Arguable this will be at the heat providers expense. So there is already a maintenance requirement in place. The potential to have meters fitted in a cupboard does make the prospect of a meter replacement project (recallibration costs more than a new meter) that much easier to achieve.
If systems are not metered - local storage does introduce maintenance, but its straightforward, and if access to cupboards is a given, then a whole block of flats can be done in a day, with say 10 units in total, feeding 100 properties, 30 minutes per unit, one man. Further, on systems feeding data out, servicing would be more responsive, attending systems that flag something is drifting from normal operating parameters (pump for example), and could potentially go for up to 5 years unserviced with a thermal store and suffer no technical problems.
The big attractions of local storage are:
- Electric backup of hot water. This can also be achieved with cylinders in plant also.
- Steady hot water load on network - or shifting the load to times when energy is cheap.
- Lower peak network flow reducing pipe sizes
- The network can go off at times low demand - and no heating demand - saving considerable network losses.
- At times of low demand and heating, the network can be weather compensated.
- Lower network temperatures enables use of low grade heat sources.
- The ability to deal with exceptionally high loading without any fear. The network can be sized accurately to the known peaks, in the safe knowledge that locally, the water systems can cope with any estate agents scrutiny.
Potential for Reduced Plant
Given local storage results in flat-line DHW loads from plant, and given central heating load can be allowed a reasonable period to startup, is it not possible to avoid the need for any storage in the plant? If stores reheat in co-operation, then boilers or other heat generators can simply feed heat in at their most efficient level, then either modulate down if heating is required, or turn off, once all stores are satisfied.
Such steady demand could further enable a more modular approach to heat generation, with small local boilers or renewable heat sources feeding into the network. With managed storage, it is straightforward to reserve storage for peak supplies of renewables, such as solar input when the sun shines, or heat pumps run on cheap rate electricity.
No one solution fits all. Storage will definitely out-perform HIUs on delivery to taps, and will provide extra benefits such as electric backup, the ability to attach wood burners, and allow primary pipework to run colder, but they cost more and take up a lot more room. Each design needs working through to see if the performance benefits outweigh the extra unit costs, and loss of space.
If you are going to compare standing heat losses of a store with those of an HIU, you need to include savings from a network left to go cold under storage, and as yet we do not possess the site data to analyse. When comparing costs you need to include reduced pipe sizes from storage, and reduced storage at plant.
The Danish, and others, may have spent nearly 40 years perfecting mechanical HIUs, but it is worth knowing that in the UK we have also been perfecting the use of small scale storage over a similar timescale, as well more recently the benefits to be reaped from digitally controlled and connected systems. Historical experience should not tarnish what is possible today.
But only real world data will decide, and currently HIUs have the edge simply down to the lack of trials and data on advanced storage.