The Best DH System we can think of

From Heatweb Wiki
Jump to: navigation, search

So, given we have the most enormous range of kit, and there are so many options when it comes to district heating design, it's about time we sat down and picked what we feel to be the best overall solution.


  • The lowest possible return temperatures.
  • The ability to turn off the heat network for considerable periods without affecting performance.
  • Pipework as small as possible.
  • No Legionella risk.
  • High levels of DHW performance.
  • Simple to install
  • The minimum space within properties
  • Offer as small a range of options as possible keeping kit standardised.


Stored DHW in properties may offer the best solution, but does introduce Legionella implications, as well as potential for extra heat loss and wasted stored heat.

Storage however is pretty crucial if we wand to let networks relax. The best method we have yet seen is to place storage in a communal place central to between 3 and 10 properties. This battery of heat will allow us to satisfy the majority of low level DHW demand rapidly, without placing a demand on the network at large. The use of a single store reduces the space requirements, reduces heat loss, and reduces CAPEX.

The store is also of a primary nature - a thermal store rather than a DHW store. This allows spikes in central heating demand to also be buffered, as well as overcoming Legionella implications.


The DHW still needs to be generated, and can either be done at the thermal store, or within each property.

The policy on hot water temperature is the lower the better for network efficiency, but we need to keep users happy, and need to keep the more risk adverse among us happy that Legionella can't crop up. The answer is to set DHW temperature to 50C, and ramp it up to over 55C for an hour per day to sterylise. This at least keeps everyone happy.

We do feel with less than 3 litres in DHW pipework downstream of the HIU, the German regulations would apply removing the need to sterylise and allowing DHW temperatures to be dropped to 45C (a bit low for some people's taste, and given they are paying for the heat have some say in the quality of that heat).

With DHW generated at the store, there will be no need to kit within properties, other than another water meter. It does however increase the volume of the DHW pipework to over 3 litres, again introducing Legionella implications. We need to introduce a secondary return loop, so there are now five pipes to each property - CWS, DHW, DHW return, heating flow, heating return.

Given however that we do need some kit in the properties - to control central heating - we have a box and can locate the DHW plate heat exchanger here, reducing DHW volumes to below the 3 litre threshold, and also reducing the number of pipes to three.

So there are two options... DHW generated at the store, or DHW generated within properties.


At this point we have a thermal store (say one per landing) fed from a central heat network, and feeding heat to small HIUs in each property than manage the central heating and generate DHW.

We can heat up the stores from the network, and turn the network off leaving the stores to cope with low level demand.

When there is some heating demand, we can bring the network on, at weather compensated temperatures to feed heating while leaving the higher temperature storage in reserve for DHW demand.

And when stores are getting low on temperature, or when we anticipate a predictable spike in demand (typically three times a day), we can ramp up network temperature to recharge stores or drive higher temperature central heating if its cold.

Direct or Indirect

At this point the options come in. Whether to do direct, with the water in the network, buffer, and heating circuits been the same, or to use a plate heat exchanger to separate the main network from the buffer.

There is no right or wrong way about this, except that direct connection is more efficient. it is certainly also possible to close the connections to the heat network fairly rapidly if a loss of pressure is detected locally.