HIU Testing Standards
For too long have we lived with a complete lack of accountability for the use of badly designed or old fashioned HIUs. Until now there has been no standards against which to check an HIUs performance, and the importance of certain features to the efficiency of an entire scheme has been ignored. As a result, we live with a plethora of poorly performing district heating schemes that are so inefficient they cost the end user significantly more.
Imagine a gas boiler industry without standards or any way of gauging efficiency - completely unthinkable, but currently the case with HIUs.
The following consultation letter to the Scottish Government from Ofgem serves to highlight this:
This is set to change, with the first set of HIU tests now completed using a new UK standard derived from the Swedish HIU standards, but flavoured to our own climate. A significant portion of the well known manufacturers have undergone independent testing and the results have now been published.
The tests have been funded by DECC and managed by Fairheat, a company independent from HIU manufacturers. Please see their website for further details at http://www.fairheat.com/hiu-testing/
It should be noted that as yet this standard is not a document of compliance, and there is no legal requirement for manufacturers to submit HIUs for testing, to publish results, or for specifiers and installers to follow the requirements.
We feel however, that clients should be wary of those who do not, as they are the ones left to live with the results of poor performance.
- 1 The BESA Test Regime
- 2 Test Regime, Version 1.9
- 3 About the Tests
- 4 Test Results for Data HIU
- 5 Improvements
- 5.1 Heat Meters
- 5.2 Keep Warm Function
- 5.3 Low Temperature VWART Figure
- 5.4 Tests 1e and 1f: Unbalanced Heating
- 5.5 Reduced Primary Temperature Test
- 5.6 Max DP Test 5b
- 5.7 Self-Learning of HIUs
- 5.8 Photographs
- 5.9 Additional Features
- 5.10 HIU mounting on test rig
- 5.11 Higher DHW Outputs
- 5.12 Test Facilities
- 5.13 Results Repository
- 5.14 Weather Compensation
The BESA Test Regime
Test Regime, Version 1.9
The following document outlines the test regime that has been put in place.
|Modified Swedish Test UK HIU Test Regime_V1-Rev-009|
|VWART Calculation Methodology|
|SP Test Report - DATA HIU|
About the Tests
The tests cover a large range of criteria, all important. The documents themselves provide the best explanation of each test and its reasoning.
The VWART figures are also important, and the key to an efficient system. The lower the better
Volume-Weighed Average Return Temperature, is a figure that represents the average return temperature you can expect under normal operation. Its a simple fact that heat networks and sources work more efficiently with low return temperatures. Furthermore, the higher temperature differences allows networks to deliver more heat.
The published figure takes into account keep warm modes in constant operation. It does not allow for economy modes on electronic HIUs, or for a low temperature approach (see below) and as yet no figures have been published as part of the report. Furthermore, the set temperature of the keep warm mode is not specified in the tests, and as a result the standby losses are not directly comparable.
Test Results for Data HIU
We are please to publish the test results for The DATA HIU.
In brief, it obtained the best efficiency (VWART) figures by a long way.
To view our teaser video please go to https://www.youtube.com/watch?v=HZWCoZigiPM
Test Results Published Seperately
One of the more important tests to be done (1e and 1f) is that of the effect of a poorly balanced central heating system. In other words, a test to see how the units work in the real world. This is a very important test as it highlights why most HIUs installed perform awfully - as radiators are never balanced properly.
4.2.3 Static testing of the space heating capacity given poor radiator set up: Test 1e & 1f Objective: Perform static testing in order to investigate the performance characteristics of the HIU when meeting a heating load across a poorly commissioned radiator heat circuit. In these two tests the HIU pump is utilised, with four radiators sized for 4kW output at 70°C/40°C operation. Test 1e: Test using radiators as heat load. Pump set to speed and/or mode recommended in HIU commissioning guide or maximum fixed speed if no specific guidance given in HIU documentation. Radiator sized for 4kW output at 70°C/40°C operation have lockshield valve fully open. HIU set to 70°C secondary flow rate. kW output, primary and secondary flow and return temperatures recorded for the steady state condition. Test 1f: Some HIUs have added features to improve operation of radiator systems (e.g. responding to high secondary return temperatures and /or the use of different pump settings). Test 1f is the opportunity to put any such feature into action and to test against the poorly set up radiator condition. Test 1f uses the unadjusted radiator setup as per test 1e and the HIU is put in the manufacturer’s recommended mode, as documented in the HIU setup guide and, if requested, the pump speed reduced to lowest speed that still delivers 4kW output. kW output, primary and secondary return temperatures recorded for the steady state condition.
The last page of the report shows a typical hot water draw-off. The following graph shows all tested HIUs side-by-side for comparison. We have marked the points at which the hot out temperature approaches target temperature.
The DATA HIU is shown in red, and as you can see the control is such that the HIU achieved stable temperatures before the tap was even fully opened. This test immediately followed the keep-warm standby test, so one would expect a rapid response.
HIU Differential Pressure Self-Learning Function
The DIGI, DATA and SLIM electronic HIUs, all come as standard with a function to learn the differential pressure of the primary system during operation, and adjust their characteristics to suit.
Differential pressures (DP) can vary significantly across a network. The properties nearest to the plant room will see the maximum pressures generated at the plant, where as the index flat, furthest from the plant, will see much lower pressures - enough to work at full network load. The typical range of pressures is from 0.5 bar up to 1.5 bar, but can be higher on some networks.
The traditional approach is to fit a DP valve, that knocks out excess pressure so the HIU only sees a constant DP. A perfectly adequate solution, however has three disadvantages compared to using an electronic self-learning system:
- They add cost, taking up more space and hence increasing HIU size and heat loss.
- They introduce pressure loss of their own, so that the index flat will in fact need a higher DP to achieve maximum load.
- One more component to go wrong.
The self learning feature works by examining the effect of a valve movement compared to known characteristics, and hence determining what the DP actually is. The self learning feature takes three seconds to perform, and is done every time a tap is opened and temperatures stabilise.
The following graph shows how the self learning feature works during an independent test. The test shows the commissioning process where the unit is turned on for the first time. The preset DP value is 0.6 bar. The actual DP of the system is 2.5 bar. The first test draw-off shows how the system would react without self-learning in place. The hot water temperature overshoots the target on first draw-off, and a spike in DH flow (red dotted line) can clearly be seen. The second draw off starts with heat in the system, allowing advanced functionality to start and enables the system to learn the new DP value. From there on you see no spikes in DH flow or significant overshoot of temperatures when taps are opened.
DP values will vary with load for any one property, so once the initial commissioning has been performed (by running hot taps a few times) then the system is setup to go, and will continue to fine tune itself every draw off.
The following table shows the VWART figures from the Swedish tests.
The following VWART figures are adjusted to provide figures based on a keep warm temperature of 25C, and a return temperature limit of 55C.
|VWART Calculation||Nomenclature||VWART (°C)||Volume(m3 pa)|
|Space Heating||VWART SH||48.7||50.75|
|Standby / Keep-hot||VWART SBY||24.4||3.83|
|Heating Period (including space heating)||VWART HEAT||47.1||11%|
|Non-Heating Period (not including space heating)||VWART NO HEAT||20.2||89%|
The settings for this test were based on a keep warm temperature of 55C, and a return temperature limit of 55C.
|VWART Calculation||Nomenclature||VWART (°C)||Volume(m3 pa)|
|Space Heating||VWART SH||55.0||70.8|
|Standby / Keep-hot||VWART SBY||55.0||108.7|
|Heating Period (including space heating)||VWART HEAT||54.0||82.3|
|Non-Heating Period (not including space heating)||VWART NO HEAT||48.3||124.8|
One can see that by setting the keep warm temperatures lower it has a direct impact on the resulting VWART figures. The key point is that with the DATA you have the ability to set these temperature limits as the installation allows to gain the best efficiencies.
The Swedish tests do not allow for such functionality and made no specification as to keep warm temperatures. Some were set high and some lower, making the published VWART pretty meaningless for comparison. It is more a representation of what the keep warm temperature was set to for the test.
An important point is that of return temperature limitation during heating mode. In the real world, radiators are rarely balanced, and as a result the return temperatures are higher than they should be. Under these conditions, most HIUs will simply pass these higher return temperatures to the plant room and the VWART figures go out the window. The return limiting functions of the DATA will guarantee that the intended VWART figures hold, however poorly radiators are balanced.
A final point covers the real world problems that occur when primary temperatures are dropped on a site to overcome unintended issues such as overheating corridors. With HIUs targeted at a set (and fixed) central heating temperature, the effect of dropping the primary flow temperature is usually that the HIU tries harder to achieve its target - in other words, the valves open further, and return temperatures fly uphill. In the worst instances, the primary flow is dropped as low as the setpoint heating temperature, making it impossible for the HIU to achieve full target. In this case the valves fully open and the return temperature can be approaching that of the flow. All VWART calculations again, go out the window. It is worth knowing therefore that the DATA HIU tracks the primary temperature and adjusts its targets accordingly. The return temperatures will NEVER rise above the figures needed to maintain VWART figures as intended.
|VWART Calculation Methodology|
HIUs were to be submitted with either a specific heat meter (in which case approval only holds if similar a model of heat meter is fitted to the HIU in an installation) or a spool piece that would be replaced during the test with a 25kPa restriction. This would then allow the HIU to be approved for used with any heat meter.
We feel the test should be using a spool piece with a Kv equivalent to an average of the 1.5m3 heat meters available. If the HIU has the option for a 2.5m3 heat meter, then a DP loss test should be run including loss at 30kW, and loss at peak output.
Keep Warm Function
This is possibly the most important test as it determines heat loss and standby return temperatures.
Many HIUs have a fixed thermostatic keep warm mode, however electronic HIUs have the ability to set a lower keep warm temperature - enough to ensure rapid delivery of hot water, but low enough to keep temperatures in the HIU down considerably. The result has been some considerable variance in standby loss, as we initially left our keep warm temperature at DHW set temperatures of 55C, resulting in poor performance figures, then in the final test we set keep warm just above room temperature (as would be more normal) and obtained the industries best results by a long way.
In the real world, it is normal to set a low keep warm temperature. The point of the keep warm is to ensure there is enough circulation in the network to maintain a hot network, so the delay in delivering heat into an HIU is reasonably short. Most relative heat loss (temperature drop) is in the branches - they go cold first - but there is no need to keep these at full temperature. The keep warm temperature can indeed be set between 25C and 60C to provide a specific delivery time.
The key to a good result on this test currently is to take on board the requirement that DHW needs to be delivered in under 90 seconds to taps following a keep warm period. We do feel that 90 seconds is a very long time, given a keep warm mode in place, and it introduces another question... what is the water content of the supply pipes on the test rig ? If it is possible to run off the water in the pipes in under 90 seconds, then effectively the test does not test a keep warm mode. We feel the water content in the pipes needs to be large enough and the primary supply flow slow enough that the HIU is forced to actually perform keep warm functionality - keeping distribution pipes at a temperature that ensure rapid delivery.
There is no sensor within the HIU to record the actual keep warm temperature making it difficult to ascertain the effectiveness as well as peak temperatures that may affect limescale build up.
The tests currently give no weight to the ability of an HIU to let primary pipework go cold when there is no demand. This has already been seen to improve performance significantly, and as such the tests should provide a clear indication of the savings resulting from this feature within an HIU.
Low Temperature VWART Figure
A VWART figure should be provided for the use of economy keep warm modes.
Tests 1e and 1f: Unbalanced Heating
This test is defined in the testing standard, and would indeed demonstrate why many HIUs do not perform as they say they will in the real world. Yet the figures appear to be missing.
Reduced Primary Temperature Test
A test should be performed where the primary flow temperatures drop to the central heating target temperatures to determine the effect of efficiency. It is very important to know how an HIU will react in these circumstances as this alone is a reason why many sites are less efficienct then anticipated.
Max DP Test 5b
The maximum DP (Differential Pressure) test (on which our DATA HIU failed) does not give the HIU any time to pickup on a high DP value and alter its operation accordingly. In the test our HIU was simply hit with a DP value 2.5 times that it had been working on previously. As such the true operation of the unit was not tested, as on site, units do have time to learn the DP characteristics.
It can be seen on the test graphs for test 5b that following the first two valve opening movements, the system adjusts for the higher than expected DP value. This is most clearly seen on the DH flow line (dotted red), that spikes for the first two movements, but is then corrected for in following draw-offs. Arguably, the addition of a few cycles prior to test start would enable self-learning HIUs to adjust for the rising DP as they would on site.
It is also unclear what the test is for, in that there is firstly the ability of the unit to perform at the extremities of its range, and secondly, the ability to react to unexpected rises in DP. We would suggest that a DP swing test is a separate test to that of testing the unit at extremities - which would allow for the unit to be commissioned accordingly.
Self-Learning of HIUs
Allowance needs to be made for the fact that the latest HIUs learn about their environment, and as such any test should allow for a few minutes bedding in period or few function cycles.
Standby heat loss tests should be photographed to provide a record that all insulation is applied as is should be.
The test should include a report on additional functionality provided but not yet tested for in the current regime, including:
- Legionella functionality
- Pump exercise functionality
- How tamper-proof systems are (from both the public and the installer)
- Commissioning procedures
- Quality of installation documentation
- Security valve options
- Network connectivity (Serial, M-Bus, etc.) and features
HIU mounting on test rig
The current rig layout mounts HIUs to a metal back-plate using metal screws creating a heat bridge. With HIUs designed to avoid heat bridging to the environment, and with heat losses now so low, we believe the test rig could introduce an additional heat loss that can be easily avoided if HIUs are screwed to a wooden back-plate more representative of real world installation conditions.
Higher DHW Outputs
It is a fact that specifiers are asking for outputs over 50kW DHW from a single unit. Even though our HIU can achieve more than 75kW, the test rig was unable to go this high. The rig should be updated to take a minimum of 100kW, and preferably 150kW, to allow results to accurately reflect the full output range of the HIUs.
Allowance should be made to allow for a further test with settings adjusted to accommodate higher than average DHW outputs (>50kW) by including additional cycles in the high DP overheat test 5b.
We would like to see an approved test facility within the UK that can cope with the demands of all manufacturers submitting all their ranges.
Test results for all HIUs should be made available to the latest versions on a web site, as with other recognised standards. This is to ensure that no specifier is left unaware of the performance of equipment they specify, and provides a result against which field units can be held to account.
Given it is now possible to weather compensate using an HIU (indeed in Denmark its a requirement), we feel that in future testing regimes it be considered the effect of this feature on VWART figures. See Weather Compensation Control over HIUs.