New metrics for communal heating design

By their nature, Communal heating systems are very extensive; they connect the plant room to every customer with flow and return pipes. This inevitably causes significant heat loss from the pipes, either as heat gain into the building, or heat loss to the environment if the pipes are external. Typical distribution efficiencies for existing systems are 68% to 34%. The losses will increase annual heating bills, introduce a summer heat load, and have associated carbon emissions. These are quantified in the paper.

The energy is transported from the heat plant, to each customer via hot flow and return water pipes in the ground (labelled ‘A’). These pipes then pass into each building and rise up the floors to serve customers on each floor, otherwise known as the communal system (labelled ‘B’). Once inside the apartment, heat passes via the heat interface unit ‘HIU’ to the internal heat and hot water pipe work (labelled ‘C’).

By law heat meters are required at the plant room (labelled ‘HM1’), at the entry to the building (HM2), and at each HIU (HM3). These are intended for billing of customers based on heat used. They are also highly valuable for use before occupation - during commissioning of the system - as we describe in the paper.

All of the hot pipes and plant lose heat. The heat loss can be attributed to that lost from the main/underground pipework (labelled ‘A’), that from the distribution within the building (labelled ‘B’), and that inside the customer’s premises itself (labelled ‘C’).

The chart below breaks down the calculated losses from each network component for a reasonably modern and efficient housing scheme (2014). The total calculated distribution loss per customer is 355W. The losses are assigned in the bar chart to pipe work sections A, B & C.

Heat losses as calculated from each network component, for a reasonably modern and efficient housing scheme (2014). The losses are grouped into pipe work sections A, B & C.

Most of the heat losses in the chart above are from sections B&C.

How are these systems currently measured? Document CP1 exists, as a design guide, but the ‘SAP Assessment’ is the only performance assessment tool accepted by the planning and strategic authorities (such as local councils, the Mayor of London, or BEIS). SAP was designed as a compliance tool, not as a design tool. But due to its pervasiveness, it inevitably drives policy, planning decisions, and therefore system design.

The SAP assessment only includes the losses from section A pipework (Williams 2010) (AECOM 2015). This means that only approximately 5% of the total losses are calculated. Or, to put it another way, 95% of the losses are unaccounted for. Whatsmore, there is no "as built" test of actual heat loss. We feel that this process is causing a misrepresentation of the benefits and limitations of district and communal heating as a technology for the greater good. It may also be limiting innovation within the industry.

For a table showing the SAP default value in context, please refer to the final section called “Measured Data”.

Max Fordham have discussed this issue at length for at least six years (FORDHAM 2012). So, we will not develop this any further, except to observe that at the time of writing, there has been no change by the planning and strategic authorities to the SAP procedure.

Inefficiency is often unwittingly built-in at the earliest design stages. Riser routing may have been constrained before the M&E consultant has been appointed, or before they have become familiar with the scheme. So, the opportunity for a lean system arrangement can be missed. During system design engineers will knowingly, or unknowingly, over-size components to minimise risk of under-supply of heat. This leads to excessive losses. As it is uncommon for system performance to be meaningfully measured and shared, these impacts are usually not visible.

The challenge here is primarily about early communication of the design constraints and then quality control over a protracted procurement period. At the first opportunity, the designer must engage with the key design decisions (examples of which are listed in the paper) and communicate the effects of this to the architect and client.

In order to control quality through to completion, a simple reliable test of the as-built performance is necessary. As Casey Cole of Guru Systems has done well to highlight, the means for this should already be supplied as standard. Comprehensive heat metering is already required by law for billing.  If this is properly installed at the right time, it can be used to demonstrate real-life efficiency before project completion, i.e. just in time to be a useful test for the client.

The current, SAP, metric of “distribution loss factor as a percentage of useful heat” is not easily measureable, nor is it useful for real-life design. We propose measureable and achievable KPIs for ‘Good’ CH schemes. These values can be set at the design stage. They can be tested before handover, and checked in use.

The values proposed are to be tested against real schemes:

• <50W per customer continuous distribution loss between plant room and HIU;

• <30m total distribution pipe length per customer;

• <38kW HIU rating for a 1 bathroom property; demand calculated on the Danish, or preferably, the Swedish Curve (e.g. total heat load >3kW per apartment for schemes >60 units);

• VWART<40C; i.e. the Volume Weighted Average Return-water Temperature, which can be calculated with relative ease from metering data over a period, say 24 hours;

• Bypass flow < 5% i.e. flow that is unaccounted for by the HIUs.

These values can be used for feasibility, design, commissioning and ongoing performance monitoring.

It should be noted that the occupants themselves are, understandably, unlikely to engage with any of the above. Their metrics will be: am I warm enough, can I have a hot shower, and does it cost less to run than a gas combi boiler. This is how success will ultimately be judged.

As we mention in the paper, we measured the performance of the London scheme against these metrics before handover. The contractor team were very engaged with the process. The metrics were used by the team to test and improve performance. So, at handover, the client team understood how the system performs, and how this compared to the specification.

Performance data, if it is measured at all, is rarely shared publicly. We feel this is why there is a lack of wider understanding of good and bad design. Below we present the measured data from the London scheme, and compare it to the proposed metrics.

Measured performance of the London scheme against proposed metrics

The table shows that the London scheme complied with three of the proposed targets. The continuous heat loss of 90W is compliant with the project specifications, so the scheme was accepted. However, we note that the measured losses are higher than the target. The specification was defined 5 years ago. Due to the advances in communal heating design since then, we feel that the losses could be brought down, perhaps to 50W. This may include further rationalisation of pipe route, and more recent technology advances in HIUs.

So, how does this compare with other benchmarks? Below we present national averages and specific example schemes against the London scheme, and the SAP default value.

Measured distribution heat losses, from pipe sections A,B&C, for national, example schemes, & for the London scheme. Compared to SAP default value (which is calculated, not measured).

The average heat loss from all schemes in the UK has been inferred from measured data. It is high, at approximately 689 W/customer. This may be distorted by large legacy schemes over 20 years old. The example schemes show a range of measured performance. The London scheme is performing better than all of these.

The London scheme is a relatively small system, at fewer than 40 units. It is a single building and a communal system. As such it will perform slightly better than other schemes with extensive external ‘A’ section pipe work. However, as noted above, the losses from ‘A’ are estimated at around 5% of total, so the total losses presented are still felt to be approximately comparable.

Finally, we have the SAP default heat loss, inferred at around 23W per customer. In the current planning process, this is assumed to be valid for any all-domestic scheme. For a developer or designer to claim this performance, only the most lax design standards and quality control are required. This figure is shown here to be unrealistically low. What's more, current regulations have no requirement for an "as built" test of actual heat loss.

As stated, we feel that this process is causing a misrepresentation of the benefits and limitations of district and communal heating as a technology for the greater good. It may also be limiting innovation within the industry.

We hope that by defining these metrics and method, and by sharing this data, we can start an open conversation on real-life efficiencies. If you are a designer, operator or client with a view on these metrics, please let us know. Even better, if you have measured data of your own, please share it publicly.