District heating inefficiency


I published a series of posts on my blog on the Forever Fuels website, about the real-life inefficiency of the district heating network that supplies heat from the wood pellet boiler to me and my two neighbours.

This has created some controversy. Not everyone is prepared to accept that heat recirculation in summer because of low usage can result in appalling levels of inefficiency. The most recent figures (for May-Aug 2013) indicate that two-thirds of the heat is being lost before it gets to the houses in summer.

This forum topic is for 2 purposes:

  1. Comments or questions about my heating system and the losses.
  2. Tell us about your own experiences of district heating efficiency, good or bad. Have you experienced similar losses, or is your system more efficient? If so, how does it achieve that efficiency?

As in all the posts on the Forever Fuels website, I would emphasise that this appears to have nothing to do with the efficiency or otherwise of wood pellet heating. The district heating losses occur regardless of the heat source. If an oil-fired boiler were producing the heat, just as much would be lost in the same network.

The obvious conclusion if I am right is that the Government was mistaken to encourage "community" schemes, and should instead have encouraged distributed heating (i.e. smaller boilers in individual buildings, rather than one larger boiler sharing its output amongst several buildings). As usual, distributed energy is more efficient than centralised energy. But governments never believe it because they always think that big is better.

I received the following by email from Alexandra Ivanchuk of REHAU.

It contains some useful information, so I asked Alexandra if I could put it up here, which she agreed to:

I enjoyed reading your blog (published 21st August) on district heating pipework and would like to give a few comments raised by the REHAU District Heating team. As a pre-insulated pipe manufacturer, we would like highlight the following:

  1. The installation depth we recommend for (REHAU) pre-insulated pipes is a minimum of 0.6m. The Flexalen pipes you mentioned where installed at 0.3-0.5m, which is in our opinion quite shallow and would definitely increase the heat losses as well as would melt more snow in winter. To say more, Flexalen TI recommends an installation depth of 0.5.m
  2. Pre-insulated pipe is designed to be installed underground, so any heat losses would naturally be higher when installed above ground. According to your blog entry, 30m of pipe was installed under the roof. In this way, they have been protected from UV-radiation but not from the ambient temperature, which we would expect to affect the heat losses.
  3. The twin or DUO pipes are usually more efficient than the single (UNO) pipes so in the aforementioned installation the result would be expected to show better results when the temperature difference between flow and return is as big as for instance 20 degrees. Due to this fact, if the temperature difference is very low, the twin pipes might be less efficient.
  4. Are the heat meters installed before the pipe leaves the boiler room and when it just enters the building? We have lately noticed a common mistake of defining a pipe's heat loss with the heat meters installed far after the building entry with bare uninsulated pipes between the meter and the entry, which increases the heat loss and gives wrong results of heat metering analysis.
  5. Is the system installed with a plate heat exchanger for hot water? With a temperature difference being less in summer the heat losses might increase. In twin pipes, as there are no heat gains from flow to return pipe, the heat would mostly be released to outside, having losses close to 2 x single, both flow pipes. The heat losses should not be as quick as in winter though, when the temperature of the ground is naturally colder and more saturated, so the heat exchange process is expected to happen quicker.

Hope this helps to identify some of the potential improvements to the system.

Many thanks.

Kind Regards/Mit freundlichen Grüßen,

Alexandra Ivanchuk
Applications Engineer - Renewable Energy

REHAU Ltd, Hill Court, Walford, Ross-on-Wye, Herefordshire, HR9 5QN.

See below for the more important points. But to answer Alexandra's questions:

  1. The heat meters are installed within a metre of the point where the pipes enter each property. It couldn't be much less, because one needs a length of straight pipe either side of the meter. Each property is lucky enough to have an external facing service cupboard where the heat pipes enter the property. I accept that mistakes can be made with heat meters, but I have noticed that the industry is much too keen to scapegoat meters for every inconvenient piece of data.
  2. Each property has a hot water tank with built-in heat exchanger. They previously received their heat from hot-water circuits from LPG boilers, and now receive it from the district heating hot-water circuits. I should imagine that the return temperatures are often close to the flow temperatures in summer. But there's nothing much you can do about that. As Alexandra says, if this were mainly a question of the quality of the insulation, you would expect the losses to be higher in winter than in summer. The fact that it is the other way round tells us that this is primarily a problem with recirculation and dissipation, not with the insulation.

I received a number of other helpful emails regarding the design and construction of my district heating network. They made similar points to Alexandra. If they agree, I will post some others on here as well, because they contained useful advice.

But their good advice does not negate the point I was making. If one is aware of the scale of the seasonal losses in sub-optimally designed and constructed district heating networks, one can take steps such as these commenters describe to try to minimise those losses. That was part of my objective. But:

  1. Most customers and installers are not aware of the severity of this problem, nor of the measures that should be taken to minimise the losses. So systems will continue to be installed where little thought has been given to these issues, and money has been saved upfront (but with much higher costs over the lifespan of the scheme), by installing under-specified heat distribution networks. And
  2. The recommended measures mitigate the effect of what I describe, but they do not eliminate it. And some of the more significant measures have material impacts on costs and on the quality of life of the occupants (e.g. how long it takes before hot water starts running).

One could take the draconian step of installing a secondary heat-source (e.g. immersion heaters) in each property, and only using the district heating for the months of the year when demand is relatively constant. There will still be a judgment to be made about the spring/autumn periods when heat is circulating more than summer but not enough to avoid some losses. More importantly, the cost of direct electric heating, and the carbon emissions associated with grid-average electricity, significantly undermine the rationale for installing a wood pellet boiler (or any other type of boiler to supply district heating).

The most practical measure, it seems to me, is to install a large buffer tank in each property, and have proper controls on the heat distribution system, so that heat only circulates when the buffer tank is calling. That should significantly reduce the number of times each day that the hot water is circulated, and therefore the number of times that the temperature in the pipe is raised back up to 77°C, before the heat begins to dissipate again. However, it is not credible to imagine that the system will call less than twice a day.

It would be complex to flush the heat distribution network so that hot water is not left standing in the pipes after heat is called, and not easy to see how the flushed hot water could be stored for later. It is very likely that high-temperature water will be left standing in the pipes twice a day, and that the heat in the pipes will dissipate significantly by the time that the buffer tank next calls for heat. The ratio of heat usage to heat loss should be significantly reduced compared to my experience, but it will still be significant.

As I mentioned in my blogs, Austrian and Swedish contacts warned us about this phenomenon. They understand the need for efficient design, and yet they still allow for significant losses in district heating. God design is essential, but it won't do to simply blame the installation or question the accuracy of the metering. If hot water recirculates or stands in the pipes, losses will be much higher than the theoretical calculations, and it is hard to envisage a system where hot water does not recirculate or stand in the pipes for a material proportion of the time in summer.


Thanks for linking to this site. I agree with you that the losses are high. District heating is usually justified on the grounds that the lower capital cost, lower maintenance cost outweigh the distribution losses, and it is often possible to run a central large plant more efficiently than a number of individual boilers. My own experience though agrees with yours, where you have to run heat mains all summer just to feed a small hot water load it does become very wasteful.

On the question of maths though, I just want to labour the point about the actual rate of loss (not the percentage) is all year round, not just a summer problem. Your figures on your blog match mine exactly, but what I have done is take the difference between output and usage and I get this for the last few periods:

Period end        kW
19-Nov-12         5.60
16-Feb-13         6.62
31-May-13         6.48
13-Aug-13         5.42

As you see, the rate of loss is higher in winter than summer, implying that could be all about distribution losses. As long as the pipese are hot, there will be heat loss regardless of whether the pumps are running. In summer there will be a higher return temperature, but the loss will not be significantly higher when there are two pipes at 80/75 than if they are 80/65. In fact using these example temperatures, and asusming duct temperature of 15 winter, 20 summer, I would expect the winter heat loss rate to be 20% higher than summer, which is exactly what my loss figures give.

The pipework suppliers could provide an estimate loss/meter figure for 70 degrees water temperature, you shoudl check and see if this matches the 6kW ball park figure Ive worked out and if it doesnt look deeper. Do you have a buffer vessel in the plant room that is permanently hot as well?


Thanks for the clarification of the potential causes. I'm sure you're right that the problem is that the pipes are hot almost as often in summer as winter and therefore leaking heat as often. My main point is to try to raise awareness and any additional diagnosis that people can provide can only help that.

This is not happening for wont of any controls. If I remember tomorrow, I'll take some pictures and stick them up here. Each property has a valve that shuts off when heat isn't required. Obviously, that doesn't do anything about the heat in the pipes, but at least it stops the heat being replenished as it leaks, until the system calls again. Presumably, it is calling too often.

We have a large buffer tank in the plant room. 2000l if I remember right. Here are a couple of photos from when the system was being installed. I'll try to get a better photo of the working installation tomorrow.

Buffer tank for Leys Farm waiting to be installed Buffer tank at Leys Farm, unwrapped and positioned, but not fully plumbed in

If the rate of heat loss per unit of time is similar winter and summer, and the rate of heat use per unit of time is much higher in winter than summer, it's not surprising that our heat losses per unit of heat used are much higher in summer. That is the figure that affects the economics. The question is what to do about it.

Here are a couple of pictures of the valve in my service cupboard, and the buffer vessel in the boiler room. The buffer tank is 2000 litres, as I thought.

Shut-off valve for district heat circulation at Leys Farm ETA PE-K 70 boiler and buffer tank at Leys Farm



still puzzling over this one from time to time. The heat loss from the 110 meters of pipework should be around 20 watts per meter for 32mm dual pipes, which makes 2.2kW loss. Even if they are not buried to the required depth this loss should still be lower in summer than winter. This seems high compeared to the load, but it may be particularly high because of the layout of your sites.


So it leaves by my figures 3kW approx of loss all year round, on top of the pipework. Do the pipes lose double what they are supposed to, or are there other losses you can reduce? If you want to test this scientifically, one proof would be to run the system  with all the house valves closed, and see what the loss is. And more accurately, to measure the flow and return temperatures of the outgoing pipes and compare this with the figures at the heat meter, which may be before the buffer vessel. Is the plant room nice and hot all year round I wonder?




The plant room is not hot at any stage of the year. It is slightly warmed, as you would expect with the boiler, buffer tank and a lot of piping in there, but I am confident that this part of the installation is to a good quality. Fair Energy did a good job of lagging every pipe in there, and the equipment is a decent standard. I am pretty certain that the losses occur downstream, in the DH pipes and/or in the service cupboards. My service cupboard is not hot but it is significantly warmer than the plant room. The buffer tank in my service cupboard isn't brilliantly insulated, but it's downstream of the meter, so it wouldn't affect the metered loses, just the temperature in the service cupboard.

I had a chat with a customer today who has 60 metres of pipe from his boiler to his home. He has a meter on the boiler and a meter on the house, because the system will also eventually feed a swimming pool which has to be isolated from an RHI perspective. The swimming pool isn't drawing heat yet, so the comparison of the meters gives a pretty straight indication of his DH losses. He reckons he has only been losing 2-3%. He did say that it was driven strongly by the heat load (rather than hot water) in a large, poorly-insulated house with insufficient radiators, so it may be that his losses are more comparable with my winter figures than my summer figures. But 2-3% seemed to me to be as improbably low as my figures are disconcertingly high.

One factor that someone mentioned recently is the ground conditions. The DH pipes to my house run through ground that is saturated in winter. The pipes were laid in a trench with good drainage, but it may be that good drainage is the last thing you want in this circumstance. It may be an effective way of draining away the heat. But this can't be the main factor, because the ground is dry in summer, so this should produce greater losses in winter than summer, not the other way round. I know the level of the water table, because the pipes skirt round a natural pond whose water level rises and falls according to the level of the water table, and it is close to over-flowing in the depths of winter and close to drying up for much of the summer.

The tests you suggest would be interesting, but I need to find someone to run them for me. I don't have time, and I wouldn't be confident I had isolated everything correctly.

In any case, can't we infer the proximate cause from the information already available? The question is not so much whether heat is being lost to the ground, but why to that extent, and why particularly in summer. I don't think those tests will give us an answer to why, will they?

Hi Bruno, we are installers, and although we have not put in large heat mains, I have read your posts with interest and wanted to offer some points up for discussion.

It would appear that the intermittent nature of the summer heating cycles means that the length of the dormant periods result in cooling of the fluid in the pipework. When the system re-engages, the volume of fluid in the pipes has to be re-heated, but also as it circulated through the buffer/cylinder in each property it will heat strip, cool and destratify the contents of that store, and depending on the pipework layout may also impact on the main thermal store in the same way.

The intermittent nature of summer heating will increase the delta T between the water in the pipework and the water in the stores due to longer decay times. In winter with more frequent calls for heat the decay will be less and the impact lower therefore.

Could a potential improvement to this be a 3-port diverter valve that directs the cold pipework water back to the boiler to be heated in a bypass until it hits a certain set point and is allowed to enter any of the stores, similar to a laddomat on a log gasification boiler.

It should or could be then feasible to install a subsidiary bypass ciruit that feeds a towel warmer or drying room radiator or similar, which is timed or temperature controlled to distribute the heat out of the pre-insulated pipework, when the boiler cycle finsihes, so that the heat serves a useful fucntion as opposed to being allowed to decay.


That sounds an interesting idea. I've got a suitable towel-warmer circuit in my house. Mine probably offers the greatest benefit, as it's the longest run of pipe. On the other hand, in summer, you usually don't want lots of heat dumped into part of the house, even if it's only the bathrooms.

The test for me is whether I can find someone who can implement a solution that they are prepared to guarantee will justify the cost. Many of the recommendations I have seen will probably have some benefit, but it is hard to quantify how much, and the capital cost is usually significant. Would you be confident that the improvements in efficiency from this type of solution would more than justify the capital cost?

Hi Bruno I would need to know more about the schematics, layouts, buffer and cylinder sizes, pipework volumes and so on to start looking at quantification, for me of the ideas I have posted the first is the most important to avoid destratifying and cooling any of the buffer stores that then have to be recovered.

In terms of the towel warmer circuit, you are right the heat is not always needed but it has to be better than heating the ground up and if it could be directed to a drying area, you can offset tumble drier costs to improve the benefit then.


What would be really tidy would be to distribute the heat at around 40°C, and then upgrade it with a heat pump. I already have a GSHP in my service cupboard, as it was the previous way I heated my house (or thought it was, until I discovered after I removed the "backup/peaking" LPG boiler that it had been doing most of the heavy lifting). I could switch the heat source for my heat pump from the ground loops to the district heating main, and have a super efficient system. But would the RHI let me do that? Would it heck! Only inefficient solutions will do for government incentives.

Re offsetting tumble drying costs, we don't need much tumble drying in summer. We can dry most things by hanging them outside. Using the heat that way is only a fig leaf to pretend that the heat isn't being wasted.

Tasha Kosviner has put together a great blog post on this issue at YouGen, followed by some discussion. It seems that DECC and Ofgem have their heads stuck firmly in the sand on this issue.

In my experience control of the heat main and set up of the building CH is key to achieving high efficiency. I have been retaining 86% of the boiler heat meter figure over 3 end point heat meters. This also includes the heat loss of a 4000 litre buffer and pipe work in the plant room which shouldn't be underestimated. I would think it could easily account for a third of the total distribution losses. So to only lose approx 10% in the ground I consider a success! The heat main has to be weather compensated to each heating zone eg I am flowing 24/7 heat over the main and only at the temp required by each house, matching the heat load for the required house temperature. Forget traditional CH controls eg the room stat is always in demand but TRVs are there for control so the heat main is set to achieve the warmest desired temp. I rarely put more than 55c flow into the main and with hindsight would oversize the radiators by a larger degree than is the case now which would enable even lower flow temps. I have experienced 30% summer efficiency just running DHW and now try to keep non heating season use to a minimum as PV largely covers that requirement.

James Neale

Add new comment

Filtered HTML

  • Insert Google Map macro.
  • HTML tags will be transformed to conform to HTML standards.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
This question is for testing whether you are a human visitor and to prevent automated spam submissions.
Enter the characters shown in the image.