From Heatweb Wiki
Jump to: navigation, search

1. Background/Intro

  • It is common practice for pipework feeding HIUs to be grossly oversized resulting on long delays to clear cold water, and hence the need to introduce a permanent keep warm mode to avoid excessive delays of DHW on instantaneous systems
  • Until the recent introduction of the BESA testing regime designers have worked in the dark as to how HIUs perform, in turn leading to a need to ‘assume the worst’, in turn leading to poor design and thermal inefficiencies, with a significant portion (a majority) of heat loss originating from HIUs and terminal pipework.

2. Key Issues/problems in DH (related to the title)

  • The problem we are trying to solve is the blanket application of permanent keep-warm modes on HIUs resulting in poor thermal performance.
  • Network efficiencies as low as 20% are recorded.
  • Keep warm has been viewed as a binary decision, been on or off, rather than calculating the required keep warm temperatures to satisfy the required DHW delays.
  • Excessive heat loss within buildings leads to both poor efficiency (higher running costs) as well as discomfort for occupants, and a need to ventilate heat from buildings. The worst instances of this include the removal of fire doors, and thereby introduce fire safety concerns.
  • The blanket selection of instantaneous DHW HIUs where local hot water storage is the only efficient option, has resulted from the inability of designers to define a point at which storage system become the better option.

3. New Approaches/Solutions/Methods

  • The use of test data on HIU performance, as now provided under the BESA test regime, allows designers to calculate DHW delay times accurately based on any given pipework selection.
  • This paper provides the methodology for calculating delay times by applying known water velocities (from test data) against pipe sizing to determine the length of time for DHW generation.
  • This paper also provides the methodology for calculating the required keep-warm temperature settings on HIUs in order to accelerate DHW production where pipework volumes are too large for pipework to be left to go cold.
  • This paper finally provides a means to determine at which point DHW storage becomes the preferred choice to achieve thermal efficiency.

4. Benefits to sector

  • Unless heat network inefficiency is addressed, the use of DH as a means of providing heat to homes does not add up against the efficiency of individual systems.
  • Embracing the correct design methods targeting efficiency as well as performance will enable the delivery of heat networks that prove themselves to be a sensible alternative to other means of heating properties and generating hot water.

5. Conclusions

  • Following the steps defined in this paper one can be assured that both thermal efficiency and performance are maximised, without sacrificing one for the sake of the other.
  • The availability of performance data is critical to the calculation of real world performance and efficiency.

6. References

  • BESA Test Data
  • CIPHE Plumbing Engineering Services Design Guide
  • Engineering Toolbox


This paper is a study into the known data on the performance of heat networks, with a view to explain the operation and resulting efficiencies through known mathematical functions, and as such provide a basis for calculating the optimal design for a heat network. We look at the system from the demand side, and then into the pipework system supplying that demand on a building level.

It is the aim for this paper to be peer reviewed and fully based on recorded independent test data in conjunction with agreed science.


The purpose of a heat network is to supply the energy (entropy) for hot water and central heating (or cooling) to properties from a centralised plant room. On the assumption a property has a water supply, and heat emitters (radiators/underfloor heating), the purpose of the heat network is to raise (or lower) the temperature on these, in order maintain room temperature at the desired level, and to provide hot water to taps.

The loose constants involved in demand are:

Central heating:

  • Minimum room temperature to protect against frost of 5°C
  • Typical room temperature of 21°C

Hot water:

  • Upper comfort temperature to skin of 40°C
  • Maximum safe supply temperature of 60°C
  • Temperature required to kill Legionella 60°C
  • Time for bacteria at optimum temperature to grow significantly of 2-7 days
  • Minimum temperature to taps for building control of 50°C within 30 seconds

One can do nothing about demand to improve efficiency - it is a fixed, determined by the users. (although with central heating we can decide to extend it).


DHW Provision

Where instantaneous hot water HIUs are used on a heat network, where there is no demand to keep pipework hot, bypasses are required to maintain temperatures by allowing some circulation. This is to ensure hot water reaches taps within a reasonable time.

The minimum hot water temperature to taps to satisfy skin temperature is 40°C.

The typical minimum primary supply temperature to achieve this will be 45°C for lower initial flow (full load represents multiple outlets, requiring higher temperature difference). Passing temperatures significantly higher than simply raises return temperatures and flow unnecessarily.

The Peripheral Pipework Region


Thermal losses, within a buildings fabric, increase with the number of branches kept hot at all times, generating excessive thermal losses.

Distinction needs to be made between the main network pipework and the peripheral pipework feeding properties - downstream of any bypasses on the network.

The defining point between the two network regions is the point at which permanent circulation to pipework is hot (>45C) at all times stops, and temporary circulation is allowed, pipework is instead allowed to go cold (typically for up to 48 hours) when not in use - the peripheral region. Where one chooses to locate a thermal bypass (set between 45°C and 50°C typically).

The size of the peripheral region will depend on the type of hot water systems deployed.

  • Where instantaneous heaters (HIUs) are deployed then the requirement to minimise delays in hot water production require the volume of flow pipework in the region to be low enough that the delay is satisfactory.
  • Where hot water cylinders (or thermal storage in any form) are deployed in properties, the size of the region is decided by the size of the cylinder, and the rate of reheat (via a coil and/or a plate heat exchanger), as well as the supply pipe volume/size.


Bypass Positions

The locations of thermal bypasses on a heat network define the peripheral pipework region, beyond the bypasses, where water is not recirculated and is instead allowed to cool down when not in use.

Instantaneous Hot Water HIUs

The rate at which HIUs initially 'flush' the primary pipework of colder water is also a defining factory, with the total time for delivery of heat to the HIU been the volume in the peripheral flow pipework to the most distant HIU, divided by the flow rate of flushing.

The flow rate of flushing can be determined from BESA test data (keep warm start-up test).


The area under the primary flow rate line (dotted red) is equal to the volume drawn, and can be extrapolated to calculate the time to clear a given volume of primary pipework.


An example would be a peak primary start-up flow rate of 15 ltr/mine, as per this test data. Given an acceptable (peak) delay of 60 seconds (to the HIU), the volume of the nearest bypass would be 13.5 litres upstream of the index HIU.

Allowance need to be made for the delay from the HIU to taps, and for the use of timed 'keep warm' on either HIUs or network pipework.

Given that the size of the region is calculated by volume, the diameter of flow (supply) pipework is the determinate factor. Smaller peripheral flow pipes increase the size of the periphery and reduce thermal losses accordingly.


BESA Test Data

The rate that a particular model of HIU can respond to hot water demand (from cold/standby) can be extrapolated from the independent test data for keep-warm hot water response (Test 5b), applying the known primary flow rates to the selected supply pipework volumes. Test data can be located on the BESA website at

Pump Flow

In a plantroom, a bypass may be required to maintain a minimum flow rate through a pump, or to introduce flow to other equipment such as water quality devices.

The two types are:

  • Between pump outlet and the pump inlet, via any local equipment that requires flow.
  • Between the flow pipe to the network, and the network return pipe.

It is generally important for the efficiency of heat generators (or exchangers) to keep return temperatures as cold as possible. Therefore the flow rates passed from the flow to the return should be avoided, or kept to a minimum where necessary.

The use of multiple pumps allows for peak system efficiency by avoid the need for any bypasses for minimum pump flow protection, with the smallest pump in the set capable of matching the no-load conditions, where the only flow to the system is from (thermal) bypasses on the network and local equipment.


Supply of thermal energy (heat) is there to meet demand plus losses.

Supply of flow is there to meet demand plus losses plus parasitic flow via bypasses.

For any system:

  Energy in = Change in internal energy  -  Energy out

Pipework, and instantaneous HIUs, can be considered to have no internal storage. Supply must match demand.

Hot water cylinders and thermal stores introduce storage to a system. Supply can be decoupled from demand, using the storage to make up the difference over time. Just as with any battery device, we can fill up our system with energy, and disconnect the supply of energy. Continue use can be achieved by recharging at regular intervals.

Storage allows higher peak levels of demand to be satisfied than the supply network can provide. The supply network can be sized to meet the average demand, rather than expected peak instantaneous demand.

In terms of heat network efficiency, the trade-off is between

  • using storage to reducing peak load, and thereby reduce pipe sizes and resulting surface area for heat loss.
  • using storage in conjunction with pipe sized for instantaneous peaks, to allow rapid and simultaneous recharging of local stores, with network efficiency gained by only running network for short periods.
  • having no storage, with the network maintained hot at all times in readiness for DHW use, thereby saving on the heat loss of storage. Pipes sized for peak instantaneous loads.

In mathematical terms it is about surface area, temperature, and time.

Reducing pipe size is limited by the size needed to deliver central heating - the most steady state load.

Points of Reference

Pipe Sizing

We have two points of reference on pipe sizing.

  1. No storage - pipes sized to meet instantaneous demand
  2. Storage with pipes sized to meet peak heating demand, plus average hot water demand.


We have three points of reference on load.

  1. Peak Winter load
  2. Peak Summer load
  3. Standby

Energy and Power Calculations

Energy and Power are very closely related, with Energy been the integral of Power over Time. In other words, Power is the Rate of Energy Flow.

   Energy = Power x Time

In scientific calculations, energy is measures in Joules, and Power in Watts (1 Watt = 1 Joule/Second). In the HVAC world, we measure energy in a more meaningful unit of measurement, namely kW, or Kilowatt Hours.

Thermal Energy

   P(kW) = q c ΔT / 1000     
   E(J) = m c ΔT     
   E(kWh) = m c ΔT / (3.6 x10^6)     


   E = thermal energy (kWh or J - Joules)

   P = thermal power (kW)
   m = mass (Kg)

   q = mass flow (Kg/s)

   c = Specific heat capacity of water (J/kg K) = 4200 J/kg K 

   ΔT = temperature change (K)

   (K equivalent to °C)

The purpose of this paper is to maximise efficiency. To meet demand with the lowest energy/fuel input. To keep losses to a minimum.

Therefore the understanding of losses is key. We are concerned with the following:

  • Heat loss from supply pipes
  • Heat loss from HIUs or hot water cylinders within properties
  • The level to which the above are useful, so can be dismissed as an inefficiency


The other inefficiency that is not a direct thermal loss, is loss of exergy. Exergy is a measure of how far away from ambient temperatures we are, and more closely linked to fuel efficiency that temperature alone, as it the the job of fuel to move away from ambient temperatures.

With almost any heat source (other than direct electric elements), it takes more fuel to heat to a higher temperature. As such we want to inject the most energy to a system at lower temperatures.

Therefore we wish for the network return temperature to be as low as possible. This goes hand in hand with reducing thermal losses on the return pipework.

We wish for our return to be as close to ambient temperatures as we can, so the flow of energy is one-way.

The best measures we have of return temperature performance for hydraulic interface units (HIUs) is through independent testing, such as the BESA standards. These provide figures for Volume Weighed Average Return Temperatures (VWART) based on test data at various setpoints.

Throughout this paper we shall refer to the test data provided on the DATA HIU (Thermal Integration Ltd.) from BESA testing. At time of writing this article the system has the lowest VWART figure yet seen, and as such provides a reference point for temperatures used in worked calculations i.e. what can be proven to be achievable.

Pumping Power

   Ph(kW) = q(m3/h) ρ g h / (3,600,000)   

   Ph(kW) = q(kg/s) ρ g h / (1,000,000)                          


   Ph(kW) = hydraulic power (kW)
   q = flow rate (kg/s)

   ρ = density of fluid (kg/m3)

   g = gravity (9.81 m/s2)

   h = differential head (m)

When looking at total pump energy we are looking over time.

The operating efficiency of pumps for this paper is assumed to be 80% in worked examples.

Heat Loss

Heat loss is both time and temperature determined.

Worked Example

As an example, consider a system where

   q = 2.783 Kg/s = 10 m3/h = 167 litres/minute
   ρ = 1000 kg/m3 
   g = 9.81 m/s2 
   h = 10m = 1 bar
   => Ph (hydraulic power) = 0.2725 kW 

   Flow temperature = 75C
   Return temperature = 25C
   => ΔT = 50K
   => P (thermal power) =  584.5 kW

Thermal Pipework Calculations


Heat Loss from Pipework

This assumes that the exit temperature from a section of pipe is the same as the entry temperature. In other words, the heat loss is insinificant compared to the energy transported and has negligible effect.

These conditions hold for short lengths of pipework.

For an insulated section of pipe, conductive heat loss can be expressed as

   Q = (ti - to) / [(ln(ro / ri) / 2 π k L) + (ln(rs / ro) / 2 π ks L)]                          


   Q = heat transfer from cylinder or pipe (W)

   rs = outside radius of insulation (m)

   ks = thermal conductivity of insulation material (W/mK)
   k = thermal conductivity of piping material (W/mK)

   L = length of cylinder or pipe (m)

   π = pi = 3.14159...

   to = temperature outside pipe, or ambient temperature (K)

   ti = temperature inside pipe (K)

   ln = the natural logarithm

   ro = pipe outside radius (m)

   ri = pipe inside radius (m)


The following spreadsheet is an example of the use of this equation to calculate rate of loss for 1m of pipe.


Open Google Sheet

Significant Heat Loss from Pipework

Over longer distances of pipework, or where the heat loss is significant compared to the load transported, the exit temperature will be effected by the heat loss, and will be lower. The process of calculating the exit temperature is not a simple one, requiring an iterative calculation to arrive at the solution.

The iterative process requires the balancing of two equations. Firstly the known energy losses for the pipe based on flow rate:

   P(W) =  q c ΔT  = 4200  x  Temperature Drop [C]  x  Flow [l/s]
   q = mass flow (Kg/s)
   c = Specific heat capacity of water (J/kg K) = 4200 J/kg K 
   ΔT = temperature change (K)

The second equation relates to the Logarithmic Mean Temperature Difference (or LMTD) and pulls in the change in heat transfer as the temperature drops along the pipe.

   Q(W) = U  x  Ar  x  LMTD  

   Q = heat transfer from pipe (W)
   U = heat transfer coefficient (W/K/m2)
   Ar = area of heat exchange (m2)
   LMTD = logarithmic mean temperature difference (K)

Note that U x Ar can be calculated from our equation for pipe losses above.

We need to find the solution where,

  Q(W)  =  P(W)


We assume that a tube has two ends (which we call "A" and "B") at which the hot and cold streams enter and exit; then, the LMTD is defined by the logarithmic mean as follows:

   LMTD [K]  =  (ΔTA - ΔTB) / (ln(ΔTA) - ln(ΔTB))

where ΔTA is the temperature difference between the two streams at end A, and ΔTB is the temperature difference between the two streams at end B. With this definition, the LMTD can be used to find the exchanged heat in a heat exchanger:

We can work out the value of U x Ar from a know heat loss for the chosen pipe and insulation.

Calculation of U x Ar

From above, for pipe at constant temperature,

  Q = (ti - to) / [(ln(ro / ri) / 2 π k L) + (ln(rs / ro) / 2 π ks L)]                          

and from heat transfer,

  Q = U  x  Ar  x  LMTD

We can enter a value of one degree temperature difference between the water and surroundings to derive a value for U x Ar...

  =>  U  x  Ar  =  1 / [(ln(ro / ri) / 2 π k L) + (ln(rs / ro) / 2 π ks L)]

Iteration of Temperature Drop along Pipe

The following spreadsheet shows the above equations with figures entered for two lengths (10m and 100m) of 22mm copper pipe with 25mm of insulation, a flow rate of 6 litres/minute (0.1 l/s), an entry temperature of 75C, and an ambient temperature of 20C.

Each line represents a 'better educated' guess. Once we have reduced the error on Q=P to below 1W we can stop the iteration.


Google Sheet

Iteration of Flow Rate

Likewise, we can iterate the flow rate to find out what the flow would be given a known temperature drop. An example would be a pipe with a valve fitted at the end that modulates the flow to achieve a set target temperature - such as a thermal bypass.

This example, similar to above except we are targeting an exit temperature of 40C.


Google Sheet

The resulting flow rate is 15 litres/hour.

Iteration of both Flow Rate and Temperature for Supply (Flow) and Return Pipes

Extending the above example, we can calculate the temperature that would arrive back at the source via a similar return pipe. We now the entry temperature is now 40C, and we know the flow rate, so we can iterate the return temperature.


Google Sheet

This example is starting to provide some interesting information. We know the 100m pipe with reasonable flow looses around 900W. With the return pipe close in temperature to the flow, the losses will be similar, so we have a total losses of 1800W approximately.

Now comparing this to the same pipe with the flow modulated to a fixed temperature at the end, the total losses are 623 + 227 = 850W.

In other words, the losses of a pipe with keep warm temperature of 40C will have a standing losses less than half that of a pipe with excessive flow rate, keeping the flow and return at 75C. This is the basis of why a reduced keep warm temperature is so important for overall efficiency


Heat Loss From Pipework

Heat loss from a pipe carrying water at a significant flow rate will show no significant temperature drop along its length and can be calculated by applying a single calculation.

Heat loss from a pipe with low flow will show significant temperature drop along its length and requires the use of iterative calculations to arrive at the final heat loss figure.

Heat Network Thermal Calculations

To apply the calculations on heat loss to a heat network, as opposed to a single length of pipework, requires an examination of the load.

We have three points of reference - peak winter heating load (highest losses), peak summer load, and standby - and we need to perform calculations at these points to give us the goalposts.

Peak Winter Heating Load Conditions

Peak load is fairly straightforward, as the flow rates to properties are significant and the temperature drops will be very small. We can approximate the thermal losses by assuming all flow (supply) pipework is at the same temperature leaving the plantroom, or entering the building we are calculating for.

Return pipework is more complicated, requiring a knowledge of the performance of the interface unit (Hydraulic / Heat / Consumer Interface Unit).

It is a given that the return temperatures from hot water production are lower than those for central heating, so it follows that the highest heat loss on a network will be when all systems are running central heating, but there is no hot water load. Such a condition would exist overnight on a very cold day, where central heating is on, but water use is low or non existent. The time a system actually spends under such conditions is not important at this stage - we are just trying to calculate the peak potential heat loss as an upper goalpost for losses.

So the key figures we need for calculations are the flow temperature, and the return temperature under full heating load. The latter comes from a combination of radiator sizing, supply temperature, and how well a system is balanced. For this study we will assume systems are installed and balanced to design. We will also assume that there are no bypasses open under these conditions - a fair assumption given they are typically thermostatic, and under full load all thermostatic devices will be fully satisfied and draw very little flow.

For example, we may calculate the peak load heat loss by assuming all flow pipework is at 75C, and all return pipework is at 45C. There is no need for iterative calculations under these conditions.

Peak Summer Load Conditions

Peak load in Summer will be purely down to domestic hot water, and the calculation s is the same as for peak winter heating load, except we take the return temperature from the DHW performance of the interface unit. For instantaneous HIUs the figure will typically come from BESA test result, but may also come from plate heat exchanger software calculations.

Standby Conditions

Standby load is far more complex, involving further iteration, but on a larger scale. Manual calculation, as above, becomes almost impossible at this stage reuiring the use of software to perform what can be many thousands of calculations.

Again, we need to start with the nature of the interface units and how they operate under standby conditions. This will vary considerably between differing makes of interface unit and how they are setup.

Possibly the most common setting on an interface under standby is for a thermostatic bypass. This is where the interface will draw water based on an internal thermostatic device. If the temperature drops below setpoint the system will increase average flow, and reduce flow when temperatures climb above setpoint.

The process of calculation is as follows:

  1. Estimate a flow rate drawn by each interface unit
  2. Working back from properties to the plantroom (or entry point) the flow rates are added up to calculate the flow rate in every pipe, and the total flow rate into the network.
  3. Working from the plantroom, an iterative calculation is performed on each pipe to calculate the exit temperature (as outlined above), using the exit temperature of one pipe as the entry temperature for downstream pipes.
  4. Once the exit temperature of the last pipe (the final feed to the interface unit) is calculated, it can be compared to the setting on the thermostatic bypass and the error calculated.
  5. The assumed flow rate to each interface unit can then be increased or decreased, and the whole process repeated.
  6. Once the error to every interface unit is within an acceptable range the flow pipe calculation is finished. This takes a considerable number of iterations, as altering the flow to one unit affects every pipe calculation.
  7. Return pipes are then examined in turn working from the interface back to the plantroom. Where pipes join up, flow rates are added up and the resulting temperature calculated (temperatures in each incoming pipe may differ).
  8. Eventually we will arrive back at the plantroom with the flow rates, entry and exit temperatures, and heat losses known for every pipe in the system.

The acceptable accuracy will make a big difference to the number of iterations required. An accuracy of 2C would be quite acceptable to arrive at a accurate overall figure for heat loss. Attempting to calculate to tenths of a degree will take significantly longer and potentially require more computational power.

The lowest heat losses are achieved where pipes close to properties are not kept hot at all under standby conditions, instead been left to go cold - no heat loss. The calculation process is the same, except we only need to calculate up to the position of thermal bypasses on the network (excluding the peripheral region). A typical example would be where a riser is fitted with a thermostatic bypass at the top, with lateral branches into properties left to go cold.

The two diagrams below represent two approaches to positioning of the thermal bypass, determining the size of the periphery:


Heat Loss From Pipework

The majority of thermal savings available come from peripheral (terminal) pipework in standby condition. Where keep-warm is turned on, return pipework will sit roughly at the keep warm temperature with constant heat loss. Where keep-warm is turned off, peripheral pipework can drop to ambient temperature over time.

Time in State

With the heat losses calculated under these three conditions, it remains to estimate the time the system spends in each state (on average). This is where we depart from hard mathematics and require some guesswork. As yet no studies have been done on a network at large over a whole year in order to arrive at a figure for this.

A crude assumption would be to say that a system spends 1/3 of the time in each of the three states.

The assumption may not be so critical when we are comparing different strategies. Providing the same assumptions are applied in each case we can still estimate the relative efficiency of each.

Hot Water Frequency

The time average flow pipe temperature will depend on the time between periods of loads, when pipes are allowed to cool down to standby temperatures.

Temperature Decay

For a given pattern of load, and hence gaps in load, the average temperature over time is calculated from the time it takes for pipes to cool down to standby conditions. In turn this is a function of mass and hence pipe diameter.

Calculations are performed over time - an integral calculation. For each period of time the rate of heat loss at the current temperature is calculated from the equation described above.

   Q [Watts] = (ti - to) / [(ln(ro / ri) / 2 π k L) + (ln(rs / ro) / 2 π ks L)]                          

For a know size of pipe the mass of the water can be calculated from:

   m {Kg) = π r2 L ρ

   m = mass of water (kg)
   π = Pi = 3.14159
   r = pipe radius (m)
   L = pipe length (m)
   ρ = density of fluid (kg/m3)

The change in temperature can be calculated from:

   ΔT = (Q x t) / (m x c)     

   Q = heat loss (W)
   c = Specific heat capacity of water (J/kg K) = 4200 J/kg K 
   ΔT = temperature change (K)
   t = time period (s)
   m = mass of water (kg)

Placing these into a spreadsheet allows one to calculate the drop over time as follows (some columns hidden):


Google Sheet


Google Sheet


Pipework Temperature Decay

As pipes increase in size, so does the time taken for them to cool down. Above a certain size, the rate of cooling will be so slow that pipework cannot significantly cool between uses, resulting in constant heat loss regardless of load. If thermal efficiency savings are to be made, terminal pipework sizes need to be kept to a minimum to allow the pipes to cool down and for heat loss to reduce accordingly.

Heat Loss Decay

Heat loss is relative to the temperature difference to the surroundings, or 20C in the above example. The following graph extends the above example to show this.



Google Sheet

The area under the graph is proportional to the thermal losses from the pipe over time.


Pipework Heat Loss Decay

Taking both pipe size and time into consideration, the size of pipework makes a very significant difference to overall heat loss. Smaller pipes may well spend the majority of their time with relatively small heat loss, where larger pipes will always loose heat.

Useful Losses

It is also important to consider the benefit of some of the losses. For example, heat lost from distribution pipework in a well insulated building will contribute directly to central heating. How much heat loss reaches properties will depend on the architecture and construction of a building.

Heat Network Calculators

The following link provides access to online software for performing the calculations outlined above over a riser:

More information on the calculator can be found at:

Excel calculation examples can be found at:

Standard Pipe Sizes and Capacities



Standard Tube Sizes on Terminal Branches.

With a permitted water velocity between 0.5m/s and 1.5m/s, 22mm pipe can typically supply between 32kW and 85kW instantaneous hot water, and as such is suitable for all normal property types. With diversity applied (as per Danish standard DS 439: 2009), 22mm pipe can supply over 5 properties with instantaneous hot water.

By comparison, 15mm pipe will supply up to 46kW, making it suitable for individual dwellings below a certain size, but not for multiple properties.

Storage Systems


The gains of local hot water storage are derived from:

  • The method of reheating, that determines the time weighed return temperature. A coil will have a higher average return temperature compared to a plate heat exchanger circuit.
  • The time that network pipework is allowed to cool down, either to ambient, or to weather compensated central heating temperatures (the minimum temperature needed to drive all ongoing central heating loads).
  • The ability to time reheat of independent systems, shifting the time of load (diversity) to coincide with periods when higher temperature heat is available.
  • The ability of the network (or sections thereof) to flush hot water at end of high temperature operation, pushing thermal energy into storage rather than leaving it in the network to become thermal losses.
  • The size of the storage. The larger the store (prepared hot) the longer primary supply pipework can remain cold before storage somewhere starts to run out and heat is required to maintain ongoing domestic hot water.
  • Reduced losses from smaller pipework, sized to average load rather than instantaneous peaks.
  • The installation cost savings from smaller pipework (material &labour).
  • Housing density, where long runs of pipes to (many) individual dwellings will suffer proportionately higher network losses. This is exacerbated where the additional pipework is external to the thermal envelope of habitation, and losses are written off.


The additional losses are derived from:

  • Time weighed standing heat loss of the hot water cylinder, outside of heating season.
  • The additional cost of the cylinder, with associated discharge pipe and labour.


The field of storage calculations is more complex than for instantaneous systems, with many more variables, and 'abilities' requiring a suitable level of control.

To establish a potential point in design where storage efficiency versus instantaneous efficiency are equal, we need to set a point of reference for storage that maximised efficiency, as follows:

  • pipework sized to central heating peak potential load, without any diversity,
  • plate heat exchanger recovery of hot water storage reheated to 60°C,
  • reheat from demand triggered with volume hysteresis (i.e. latched cylinder thermostat/sensor positions, minimum as per HWA specification for thermal storage, section 3.4)
  • reheat from standing losses triggered at 40°C on lowest thermostat on store, allowing a 20°C hysteresis,
  • reheat completes when primary supply temperature drops below 62°C. At all times the primary supply remains hotter, stores will maintain a full charge in readiness.
  • peak hot water load periods starting at 07:30 (to 08:30) and 19:30 (to 20:30)
  • when one system is in need of heat outside of peak, all systems take heat (i.e. reheat hysteresis latches are released on both volume and temperature).
  • peak load primary flow temperature of 70°C (this is variable, with a ideally >63C primary to achieve 60C hot water storage).
  • time weighed average reheat return on DHW temperature of 30°C,
  • peak load central heating return of 50°C (20°C drop across heating circuit).

The operation more closely resemble an electricity network, where utilities providers may turn on storage heater and cylinder power supplies at cheap rate times (Economy 7 / 10 / 2000).

Pipe Capacity for Storage Systems

One can achieve 35kW load on a 22mm pipe running central heating (70-50°C).

One can achieve 70kW load on the same pipe running hot water storage reheat (70-30°C), however a store heat exchanger will typically achieve 20kW peak.

A property may use 7kWh of domestic hot water per day, and 70kWh of central heating. A ratio of 1:10 on energy use, however times when hot water is used will remain highly diversified.

A 35mm pipe will drive 22 properties with peak heating load of 5kW, while a 15mm pipe will serve up to 3 properties.

If network supply temperatures are capable of been driven to 90°C, then more than double the load can be driven through the same pipework without exceeding 1.5m/s velocity. The final limit on output will be heat emitter (e.g. radiator) size.

Additional Information

See also:

Comparison of Instantaneous Systems to Storage Systems

To get a rough feel for how the thermal losses of a system using instantaneous hot water stacks up against storage, a few ball-park assumptions need to be made:


  • Pipe size is 22mm when feeding one to five HIUs.
  • Intermittency is such that on a typical system there is an 8 hour gap overnight for pipes to cool, but no period long enough in the day for pipes to cool significantly. During the day we will assume 30 minutes is the average time from last draw-off.
  • No permanent keep-warm.


  • Pipe size is small enough it cools down in <1h (i.e. 12 or 15mm).
  • Cylinder heats up twice per day at set times – pipes hot for 4 hours (2h + 2h) average.
  • Standing heat loss of 50W.

From these assumptions one can say that an instantaneous system would see terminal pipework cool down for 7 hours a day. Storage systems would by comparison allow pipes to cool for 19 hours per day. So, roughly the pipework losses from a storage system are 1/3 those of the instantaneous system when looking at peripheral pipework.

If we assume 10 metres of peripheral pipework per dwelling, then per property we are, for 22mm pipework looking at 8W/m, or 80W total when at temperature. For the storage system with 15mm pipework, we are looking at the same heat loss at temperature.

  • For instantaneous, total 24 hour losses = 80W x 17 hours = 1.36 kWh
  • For storage, total 24 hour losses = (80W x 5 hours) + (50W x 24 hours) = 1.6 kWh

However, the storage system will enable much larger sections of the network to go cold as there is no need for any thermal bypasses as response time is not a problem.

Now, if one were to lower the housing density, so we have 20m of pipework per property, then the situation reverses:

  • For instantaneous, total 24 hour losses = 160W x 17 hours = 2.72 kWh
  • For storage, total 24 hour losses = (160W x 5 hours) + (50W x 24 hours) = 2 kWh


Thermal Losses of HIUs vs Cylinders

Below a certain housing density, as the lengths of pipework associated with each property increase, hot water storage will result in greater thermal efficiency (lower losses) than the use of instantaneous HIUs. In dense housing developments, the relatively short length of pipe per property make HIUs more thermally efficient.