- Email: [email protected]
carbon affordable housing site in the UK
Energy and Buildings 43 (2011) 117–125
Contents lists available at ScienceDirect
Energy and Buildings journal homepage: www.elsevier.com/locate/enbuild
Measured energy and water performance of an aspiring low energy/carbon affordable housing site in the UK Zachary M. Gill a,b,∗ , Michael J. Tierney b , Ian M. Pegg a , Neil Allan b a b
Buro Happold, 17 Newman Street, London W1T 1PD, United Kingdom Industrial Doctorate Centre in Systems, The University of Bristol, Queens Building, Bristol BS8 1TR, United Kingdom
a r t i c l e
i n f o
Article history: Received 23 March 2010 Received in revised form 23 August 2010 Accepted 25 August 2010 Keywords: Domestic Energy Water Building performance Low-energy Post-occupancy evaluation
a b s t r a c t This paper reviews the annual energy and water performance of an aspiring low energy/carbon affordable housing development in southern UK, comprising 25 houses heated by a biomass fueled district heating network. Electrical, heat and water consumption data was collected for all dwellings and more detailed data logging was conducted on a sample of four. The data was analysed to benchmark the overall site performance and compare the individual dwellings. The individual dwellings performed efficiently in terms of electricity, heat and water consumption. Significant variation was identified between maximum and minimum consumers, even when typical performance correlates (i.e. number of occupants, floor area) were accounted for. Water consumption varies by a factor of > 7 and heat and electrical consumption by > 3 which, given the homogeneous design and specification of the buildings, suggested that occupant behaviour was important. Site-wide energy consumption and losses (over and above individual dwelling use) reduced energy and carbon emissions performance by 85% and 70% respectively. Although energy consumption in UK homes can be substantially reduced, the challenge for zero-carbon design is to address issues of total system performance, occupant behaviour and whole-building energy performance. © 2010 Elsevier B.V. All rights reserved.
1. Introduction The demand to reduce energy consumption and carbon emissions in the built environment is well documented [1–3]. Dwellings play an important role within this (contributing approximately 30% of end-user carbon emissions and energy use [4,5]) and correspondingly the UK government has stipulated that by 2016 all new homes will be zero carbon [6]. Whilst the effectiveness of these policy measures has come into question (primarily due to their clarity and implementation costs) [7], in principle low and even zero-energy or carbon homes in the UK1 are achievable through a combination of improved insulation, more efficient mechanical and electrical equipment, higher air tightness (with controlled ventilation) and renewable technologies [8–14]. However, available post-occupancy evaluation (POE) research shows real-world building performance aligns poorly with design expectations and some nominally low-energy buildings perform no better than their more traditional counterparts [15,16]. Computational prediction of
∗ Corresponding author at: Buro Happold, 17 Newman St, London, W1T 1PD, United Kingdom. E-mail address: [email protected] (Z.M. Gill). 1 With a mid-European coastal climate. 0378-7788/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.enbuild.2010.08.025
energy consumption also shows poor correlation to monitored performance even when it is based on as-built construction details and actual building occupancy/utilisation [17]. One must inevitably question the use of computed in situ building performance in the absence of measurement. Measurements in ‘traditional’ dwellings show that characteristics of the building (for instance floor area, number of occupants, terraced or detached) play a role in determining actual energy consumption [18]. Statistically significant variation in heat and electrical energy consumption is noted even in similar dwellings and consumers can be categorised into high, medium and low bands. Furthermore there is an increasing trend of energy consumption over time in the UK contributed to most significantly by ‘high’ energy consumers [19,20]. Achieving low and even zero carbon/energy buildings in the UK is complicated because; design and reality align poorly; zero carbon dwellings are required but are not fully understood by the industry; house characteristics are not the sole determinants of performance; and energy use is increasing over time. This study assesses the performance of an affordable, low energy/carbon site in the UK where the construction and specification of the buildings is homogeneous, and the results from monitoring are presented and analysed. It is intended that this comprehensive measurement of building performance will provide insight into the impact of identified issues on low energy building variants.
118
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
Table 1 Dwelling characteristics. Type
Floor area (m2 )
Bedrooms
Storeys
Number of
Currently occupied
1 2 3 4 (1st end) 4 (1st mid) 4 (Ground)
68.13 74.58 81.11 56.38 49.02 44.70
2 2 3 1 1 1
2 2 3 1 1 1
10 3 9 1 1 2
9 3 9 1 1 2
2. Site description
3. Methodology
The East-Anglian site comprises of 13 two-bedroom and 9 threebedroom houses, plus 4 one-bedroom flats, each constructed to the same design specification and was awarded BRE’s EcoHomes Excellent certification. The site systems exceed the requirements of UK Building Regulations and typical UK dwellings. Low in-use carbon emissions and energy/resource consumption are facilitated by;
During the design, an arrangement was negotiated with the site landlord (a Housing Association) to conduct on-going site monitoring. Consumption monitoring was permitted for all dwellings and a shortlist of potentially willing residents was provided to instigate more detailed monitoring; Four of the five shortlisted dwellings agreed to participate. No energy performance information was fed back to any of the participants during the monitoring period.
• Biomass District Heating Network: A woodchip-fired community heating network (with a gas back up boiler for peak loads) services the heating and hot water demand of dwellings. Pre-insulated pipe work circulates heated water to local heat exchangers (one per dwelling) which extract heat according to demand. Each dwelling is individually controlled by a single programmer (with thermostat) and TRVs on all radiators. • Whole House Mechanical Ventilation with Heat Recovery (MVHR): Replaces the background ventilation usually provided by trickle vents and infiltration. Air is extracted from wet rooms (bathrooms and kitchens) and heat is exchanged to filtered incoming air for living spaces (bedrooms and living room). • High Air Tightness: As constructed the buildings achieve a mean air tightness value of 3.47 m3 /m2 h at 50 Pa (with a standard deviation of 0.367 m3 /m2 h). This is a 65% improvement on the minimum regulatory requirement (10 m3 /m2 h [21]) and 30% improvement on the design target value. • Massing: In groups of three to reduce the area of exposed facade. The two and three storey buildings are grouped to ensure winter solar gain is maximised by staggering the blocks (minimising overshadowing) and by limiting the height of the southern block in each group to two storeys. • Rainwater Harvesting: A header tank (50–60 l capacity) is fitted in each dwelling and supplies filtered, unmetered greywater (with mains water backup) for toilet flushing and garden irrigation. • Optimised Solar Orientation: The main portions of glazing are located on the South facade to enable useful solar gains in the winter. Operable windows and thermal mass in party walls are provided to ensure that the temperatures are stabilised. Initial dynamic thermal modelling in IES VE showed that internal temperatures peak at 25 ◦ C in the summer months, thus avoiding overheating. • Improved Insulation: Installed insulation is better than mandated in the UK Building Regulations [21]: Walls are 0.25 W/m2 K (vs. 0.35 W/m2 K), Glazing 1.9 W/m2 K (vs. 2.2 W/m2 K), Roof 0.12 W/m2 K (vs. 0.25 W/m2 K) and Ground 0.25 W/m2 K (equal to regulations) • Low Energy Lighting: Installed throughout the homes. • Low Flow Fixtures: Installed throughout the homes. There are four dwelling types (Table 1). The site opened in September 2008 and 16 of the 26 dwellings have since been occupied fully. Seven were subsequently occupied from January 2009. Two were occupied in February 2009 and then November 2009. Monitoring data was collected between September 2008 and November 2009 inclusive.
3.1. Detailed monitoring—4 dwellings Electrical energy consumption was monitored via a single-phase current clamp and data logger attached to the main incoming supply to each building. An instantaneous current measurement was recorded every 10 min at 0.4 A resolution providing a detailed electrical load profile. Small equipment load switching (<100 W) and short equipment use (not occurring at one of the logging intervals or <10 min use duration) was missed and therefore electro-mechanical induction meters reported cumulative consumption (in kWh). Water consumption was monitored via a pulsed output data logger connected to a manufacturer supplied reed switch mounted to the top of the rotary water meter housing. A pulse was outputted for every litre of water, and cumulative consumption was logged at 10-min intervals. Accuracy of the monitoring equipment was checked via comparison to manual meter readings and was better than 0.1%. Heat consumption (supplying heat and hot water demand) was also monitored via a pulsed output data logger recording cumulative consumption in 10-min intervals. The logger was connected via a manufacture supplied pulsed output module connected to the heat meter located at the heating network heat exchanger. The measuring accuracy of the installed heat meter exceeded the requirements of EN 1434 Class 2 and 3 [22]: better than ±2% across it’s operating range. Internal temperature and relative humidity was logged in 30-min intervals using compact HOBO® H08 data loggers. Readings were taken for the summer and winter periods in a frequently used space within the house—the combined kitchen–dining room. All loggers required information to be downloaded manually onto a computer and therefore provided no instantaneous feedback to the residents. Loggers were also concealed to minimise their influence on the occupant’s behaviour. All of the dwellings had three occupants and were end-terraced (2 West and 2 East). Three of the houses were the largest Type 3 whereas one is Type 1. 3.2. Monitoring—all dwellings All dwellings were monitored manually using installed utility meters. Electric and water meter readings were taken periodically via externally located meters. Heat meters were all located within the dwelling on the heat exchanger pipe work, and were therefore less accessible. Readings were taken on the date the occupants moved in and further readings were taken at 6 monthly intervals (April 2009 and October 2009). Dwellings were billed pro-rata for electrical and water consumption. Heat consumption was charged
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
119
Table 2 Comparison of low energy site to UK/local averages, design estimates and a regulation compliant solution. Heat
Electricity
Total energy
Water
kWh/year
kg CO2 /year
kWh/year
kg CO2 /year
kWh/year
kg CO2 /year
litres/person/day
GB average per unit area (m2 ) Mid-Suffolk average per unit area (m2 ) Design solution a per unit area (m2 ) Regulation (TER) /m2 Study site per unit area (m2 ) Site maximum per unit area (m2 ) Site minimum per unit area (m2 )
17,614 191.5 15,826 172.0 8517 121.0 – 6556 92.9 11,752 144.9 2272 46.0
3417 37.1 3070 33.4 213 3.0 – 759 10.8 1360 16.8 263 5.3
4392 47.7 5699 61.9 981 13.9 – 2885 40.9 4481 64.7 990 17.6
1853 20.1 2405 26.1 414 5.9 – 1217 17.3 1891 27.3 418 7.4
22,005 239.2 21,525 234.0 9499 135.0 – 9441 133.8 15,915 196.2 3529 72.2
5270 57.3 5475 59.5 627 8.9 24.8 1976 28.0 3117 38.4 794 15.3
148 – – – – – – 91
Space heating /m2 Hot water /m2 MVHR /m2 Lighting /m2 Other electrical /m2
73 20 – – –
8.6 2.4 – – –
– – 6.7 4.8–9.8 24.6–29.6
– – 2.8 2.0–4.1 10.4–12.5
196 – 28 –
Breakdown
a
Estimated Estimated
Note that design solution assumed 100% space heating and water heating supplied from biomass at 80% efficiency. Interpreted from SAP 2005 results
at a fixed monthly rate, adjusted every 6 months to account for demand variation between dwellings. Seven dwellings were inhabited for only 11 months of the monitoring period and one further dwelling inhabited for only 10 months. For these dwellings annual electrical, heat and water consumption was extrapolated to estimate annual consumption. Electrical consumption was linearly extrapolated based on the average consumption over the monitored period. This method of extrapolation was estimated to have an average error of ±2% based on results from the fully occupied dwellings. Water consumption was extrapolated using the same method and the estimated average error of readings was ±3%. Total heat energy consumption for the sub-sample was scaled up based on a kWh/degree day calculation (excluding hot water consumption). It was not possible to determine the accuracy of this measurement due to limited availability of monthly heat consumption data for all dwellings.
3.3. Whole site monitoring District heating network energy input was recorded via 2 heat meters (biomass and total heat production individually, with the numeric difference being gas heat production) and additionally via the main gas supply meter (recording m3 at NTP). Electricity consumption for the site facilities (predominately plant room equipment and external lighting) were metered via a single meter. Site water consumption was recorded via a single utility supply meter on-site. All meter readings (heat network, electrical and water) were recorded periodically throughout the year.
3.4. Comparison to design During the building design process, accredited modelling software (IES VE SAP 9.80 Calculator and Dynamic Thermal Modelling) was used to calculate the regulated performance of the buildings; the results will be compared against observation.
4. Results 4.1. Annual dwelling energy, carbon and water performance On average, the dwellings were found to perform efficiently in comparison to national and local (Mid-Suffolk region) averages [23,24], low energy benchmarks [25,26] and UK regulations (as modelled according to SAP 2005 [27]). Energy consumption within the dwellings was used primarily for heating and hot water (69%) and the remainder for electricity. Conversely, the associated carbon emissions were dominated by electricity use (62%) with the remainder contributed to by heating and hot water. Compared to national averages, total energy consumption was approximately 56% lower (or 43% lower on a per-unit-area basis) and carbon emissions were 63% lower (52% on a per-unit-area basis), see Table 2. Local averages compare similarly. Water consumption was 39% lower than the national average stated by DEFRA [28]. Note that Great Britain and Mid-Suffolk normalised averages were calculated based on an average dwelling floor area of approximately 92 m2 (adapted from the English House Condition Survey, 2007 Stock Data [29]). Carbon factors were taken from the current (and relevant) version of SAP (2005) [27]; admittedly these values will change in the 2010 revision. Compared against a set of low energy benchmarks, see Fig. 1, heat consumption was 5% lower than the dataset average2 but electrical consumption 11% higher, reflecting the intention to reduce heat consumption by implementation of MVHR. The dwellings compare well in terms of carbon emissions, 15% lower than the dataset average (or 4% when including Passivhaus projects), primarily due to the low heat carbon factor (0.118 kg CO2 /kWh based on a measured 45% biomass utilisation and SAP 2005 carbon factors [27]). Note that carbon emissions for low energy benchmark sites were calculated based on an assumed gas supply for heat unless otherwise stated. Generally, water consumption data was not freely available for the benchmarks. However, it is worth noting that the UK Code for Sustainable Homes outlines 5 levels of water consump-
2 European, Passivhaus projects (Werner and Passive Houses) are excluded from dataset averages.
120
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
tion and the site is close to achieving the second most ambitious target (91.5 vs. 90 l/person/day) [30]. Average dwelling performance exceeded regulatory demands by 25 –37%; these follow from the calculated average target emission rate (TER) listed in Table 2. However, unregulated energy consumption was estimated to contribute 37–45% of the total annual carbon emissions. Note that in defining the figures provided, fan power (for MVHR) was interpreted as total electrical consumption whilst buildings were unoccupied and lighting energy consumption was estimated from 3 sources; the DECADE project [31], Mansouri et al. [32] and SAP 2005 [27]. Further instrumentation or surveying would be required to isolate the electricity end uses more accurately. Annual electrical and heat (space heating + hot water) consumption for all dwellings, in terms of total and normalised energy use, is shown in Fig. 2. Dwelling occupancy is also denoted. Detailed measurements on Houses A–D are considered in 4.2.
Fig. 1. Benchmarked performance of monitored dwellings and other low energy dwellings in the UK.
Fig. 2. Annual energy consumption and carbon emissions per dwelling ranked in order of total normalized energy use (occupancy denoted)—Totals and breakdown per unit floor area.
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
121
Table 3 Heat, hot water and percentage of total water consumption per sample dwelling.
2
Total heat (kWh/m /year) of which: Space heating (kWh/m2 /year) Hot water (kWh/m2 /year) Hot water (kWh/person/year) Total water (litres/person/day) Hot water (litres/person/day) Proportion of hot vs. cold water
House A
House B
House C
House D
105.5 86.4 19.1 517 63.7 21.1 33%
144.9 103.5 41.4 1118 106.6 47.1 44%
95.5 78.3 17.2 465 84.7 19.0 22%
86.8 63.9 22.9 520 69.1 21.3 31%
Table 4 Number of occupants compared to average energy and water consumption. Number of occupants per dwelling
1
2
3
4
5
Overall
Average heat kWh/m2 /year Factor of difference (Max to Min) Average electrical kWh/person/year Average electrical kWh/year Factor of difference (Max to Min) a Average water litres/person/day Factor of difference (Max to Min) Sample size
63.1 1.48 1279 1279 1.03 173.3 1.30 2
83.6 2.04 1128 2257 3.45 95.9 4.73 6
92.4 3.15 1148 3444 2.52 80.8 3.01 11
116.8 1.50 788 3152 1.30 78.2 1.78 4
88.9 N/A 530 2650 N/A 64.8 N/A 1
92.9 3.15 1068 2885 3.46 91.2 7.11 24
a
When measured in kWh/person/year.
4.2. Results from sample dwellings
4.2.2. Space heating The range of thermal energy consumption per unit area, varied by a factor of 3.15 between maximum and minimum consumers across the site, Table 4. House B was the highest heat consumer
throughout the monitoring period, utilising a total of 11,752 kWh/year, 54% higher then the site average and 17%, 27% and 38% higher then Houses A, C and D respectively. The additional heating energy consumption due to hot water demand accounts for a proportion of this increase but space heating behaviour also contributes importantly. The programming of set-point temperatures (Fig. 3) into thermostats was a potential cause of excessive heating demand. Thermostats can be programmed to control the desired temperature at 5 distinct times during the day for weekdays (Monday–Friday) and weekends (Saturday–Sunday) independently. Average set-point temperatures of Houses A–C were approximately equal to 21 ◦ C compared to 19.7 ◦ C for House D, although temperature profiles varied significantly (for instance, a constant set-point in House A and high peak temperatures in House B). Actual internal temperature in all four dwellings across the year had a mean average value of 21.9 ◦ C ±0.3 with an overall range of 16.0–29.9 ◦ C. To put this into context, Summerfield et al. found average internal temperatures to be 19.8 ◦ C at a ‘low energy’ site [20]. Average internal temperatures for Houses B and D (the highest and lowest in the sample) are shown in Fig. 4. Summer temperatures were similar for both dwellings although morning temperatures are higher for B, most likely driven by the high thermostat setting at 6:00am. Winter temperatures differed
Fig. 3. Weekday and weekend thermostat set-point temperatures for sample dwellings.
Fig. 4. Average internal temperatures for sample dwellings (winter and summer) in kitchen dining room.
For four dwellings, half-hourly (or more frequent) heat, water and electrical consumption and internal temperature were monitored (see Houses A–D in Fig. 2). 4.2.1. Hot water Hot water and space heating consumption were not metered individually. The rate of heating for hot water consumption was taken as the measured consumption (of all heating) between 4th June to 17th August 09, when the heating degree days for the locality were <5% of the annual total [33]. Estimated heat and hot water consumption per dwelling (per person) is itemised in Table 3. House B used roughly twice the heating for hot water as other houses (Table 3). The fraction of total water consumption that was heated ranges from 22% to 44% (assuming 0.067 kWh/lhot water following tests in unoccupied buildings). Washing and cleaning (using hot water) probably contributed to the house-to-house variation in heat demand (indicated in Tables 2 and 3, Fig. 2).
122
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
Fig. 5. Daily electrical load profiles for sample dwellings: (a) 6th and 3rd highest consumers (Houses A, B) and (b) highest consumers (Houses C, D). Dots show daily readings and lines average data over 1 year.
significantly as indicated by approximately 2 ◦ C temperature difference through the afternoon. Admittedly the sample size was small, but nonetheless thermostat settings (Fig. 3) might well contribute to the variation in space heating demand (Tables 3 and 4). Window and door opening behaviour and utilisation of the MVHR to ventilate stuffy air from wet rooms (rather than opening a window) were other factors that could contribute to the significant variation in heat consumption. The annual saving of a mid-terraced dwelling was found to be between 22–29% (heat consumption of 74.9, 95.5 and 105.0 kWh/m2 /year for mid, West and East terraced dwellings respectively).3 This agrees well with Yohanis et al. who measured a 24–30% difference in heat consumption between detached and terraced houses [18]. 4.2.3. Electrical In line with Yohanis et al. [18], electricity consumption per occupant decreased as the number of occupants increased whereas the relationship between total electrical consumption and number of occupants is less clear, and only partially aligns with Yohanis’ assertion that it should increase (Table 4). The factor of difference between maximum and minimum consumers when considering total, per person or per m2 consumption (shown in Tables 2 and 4) was large; between 3.46 and 4.53 respectively. Fig. 5 shows daily electrical load profiles for individual dwellings. Load profiles of Houses A and B were similar, with two broad peaks in consumption. Houses C and D, showed markedly different peak and baseload patterns, albeit similar annual consumption. Energy savings in each of these may require different approaches, for instance a focus on standby and un-switched equipment in D versus equipment efficiency and utilisation in C.
water collection area = 1740 m3 . Pitched roof collection efficiency = 75%, filter efficiency = 90% [35]. Utilisation of harvested rainwater = 100%. An inverse relationship between the number of occupants and the water consumption per person, see Table 4, suggests improved efficiency through common and shared water use (for instance; dish and clothes washing). However, given the measured variation across the whole site and at different occupancy levels, number of occupants was only one determinate of water consumption. The factor of difference between maximum and minimum consumers (measured in litres/person/day) was the most significant of all utilities at 7.11, see Table 4. The daily water consumption profiles for each sample dwelling are shown in Fig. 7, and indicate this variation in 3-occupant dwellings. Typically, a morning and evening peak are found, although the specific time, size and duration of the peak vary considerably. The consequence of this is that strict water reduction targets as stipulated in the Code for Sustainable Homes [30], are unlikely to be achieved through low flow fixtures and rainwater harvesting alone, but require enthusiastic participation from the occupants. 4.3. Site energy performance Ultimately it is the whole site performance that is judged; this depends only partly on building specification and occupant behaviour.
4.2.4. Water Water savings (91 vs. 148 l/person/day UK average) can partially be attributed to the rainwater harvesting, utilised for toilet flushing and garden irrigation. Fig. 6 shows the negative correlation between the estimated harvested rainwater and the average water consumption of the sample dwellings for 10 months between January and October 2009 (when full water monitoring data was available for each dwelling given that data has been downloaded up to November 5th 2009). In 2009 an estimated 614 m3 rainwater was harvested, offsetting approximately 24 l/person/day. Assumptions: Average monthly rainfall equivalent to Met Office data from Cambridge NIAB (<35 miles from site) [34]. Total rain-
3 The 9% difference between East and West end-terraces is not found to be statistically significant.
Fig. 6. Correlation between estimated total harvested rainwater collection and average dwelling consumption (February–October 2009).
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
123
Fig. 7. Average daily mains water consumption profiles for sample dwellings: (a) Houses A, B and (b) Houses C, D. Dots show daily readings and lines average data over 1 year.
4.3.1. District heating network District heating energy was derived from biomass and natural gas. Scheduled meter readings from the district heating network and individual houses (Table 5) highlight the distribution and conversion losses. The fraction of heat from biomass was low but should increase as the site becomes established. If it continues to improve (as expected) then the overall associated carbon emissions of the heat consumption at the site will tend to reduce. Assuming 70% of the heat was supplied by biomass (as per conservative design estimates) then the overall heating carbon factor would reduce dramatically from 0.118 to 0.076 kg CO2 /kWh. Fuel supply and a short maintenance issue with the biomass store moving floor have reduced the use of biomass so far. Average annual heat consumption per dwelling increases dramatically from 93 to 185 kWh/m2 /year when distribution and conversion losses are accounted for (Fig. 8). Correspondingly, average total energy consumption increases by approximately 69% and total carbon emissions by 40%. Nevertheless, total annual heating carbon emission savings are approximately 15% compared hypothetically to gas boilers installed in each dwelling with nominal seasonal efficiencies of 70%. Periodic measurement of heat supplied to the district heating network attributed to biomass and gas, and correspondingly the gas boiler efficiency, is shown in Fig. 9. In the absence of gas use, the efficiency cannot be determined. Gas boiler efficiency fluctuates significantly and was lower than expected with the maximum monthly efficiency measured to be only 77%. The issue of seasonal gas boiler efficiency has been noted by policy makers and accordingly revisions made in SAP 2009 will account for the seasonal efficiency of boilers and discrepancies between manufacturer specifications and actual expected performance (under real operating conditions) [36]. During the summer period the district heating network maintains flow and return temperatures
(90–70 ◦ C) to primarily service intermittent hot water demand. Distribution efficiency then decreases accordingly because less useful heat is extracted but the thermal (insulation) losses remain relatively constant (loop to ambient temperature differences do not change significantly). A more efficient scheduling of plant operation is being investigated to reduce losses during periods of low demand. 4.3.2. Site electrical energy Electrical energy was used for the whole site to operate the district heating network and for external lighting. Annual electrical energy consumption equaled approximately 30,000 kWh (or 21 kWh/m2 /year) bringing total electrical demand to 62 kWh/m2 /year. Fig. 8 shows the impact of the additional electrical consumption in terms of energy and associated carbon emissions. In addition to the previously identified heating system losses (Section 4.3.1), site electrical energy increased average dwelling consumption by a further 9% and total carbon emissions by a further 23%. Pumping power for the district heating network was measured independently during a week in September. According to these measurements, pumping power accounted for approximately 41% of the total site electrical consumption and further investigation is required to assess the impact of other site electrical demands (predominately external lighting, a moving floor in the biomass store, boiler pumps, fans and controls). 4.3.3. Site water consumption Water is used on-site for communal irrigation of the grounds and has also been used to replace water in the district heating network when there was an installation pipe failure. Over the monitoring period, an average of approximately 600 l/day was consumed for site uses although approximately 300 l/day was accounted for due to a major leak in the manually operated play park irrigation system
Table 5 District heating network system efficiencies. All data in kWh unless otherwise stated
01/04/2009
30/10/2009
05/11/2009
Total house consumption Total boiler output of which: Biomass Gas (Gas input to date) Average gas boiler efficiency to date Distribution efficiency to date Proportion of fuel from biomass
140,232 203,704 88,562 115,142 157,932 73% 69% 36%
180,933 297,590 162,703 134,886 197,367 68% 61% 45%
– 300,456 164,373 136,083 199,226 N/A N/A N/A
Notes: Biomass boiler conversion efficiency is not determined as the quantity and quality of wood combusted is not recorded. A linear extrapolation of the meter readings was made (to align periods) based on the daily average heat output over the corresponding period. Adjusted data is italicised. Proportion of fuel from biomass is calculated based on heat input data.
124
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
Fig. 8. Impact of site-systems on average dwelling performance in terms of (a) energy and (b) carbon.
between August and October 2009. Excluding these losses, average water consumption per dwelling increases by approximately 8.5 l/person/day or 9%. 5. Discussion This study focused on a statistically small sample (25) of aspirational low energy dwellings within the UK. Following negotiations four dwellings were made available for intensive data logging, giving an insight into building performance and variation in the context of the whole site. The study contributes usefully to the overall dataset of actual building performance because of the direct comparability of the buildings thanks to their homogeneous design and construction. As suggested in other studies of a similar nature, it is unwise to draw conclusions from these results with respect to the entire domestic stock or carbon policy measures in the UK [37]. However, monitored utility consumption eludes to some poignant issues regarding the reality of achieving low or even zero carbon dwellings.
On average, dwellings were found to compare well to recognised low energy and national benchmarks. A combination of improved insulation, air tightness (with controlled ventilation) and optimised orientation contributed to the low annual heat consumption but couldn’t be itemised. Rainwater harvesting and low flow fixtures contributed to low annual water consumption and low energy lighting offset some of the additional energy demand of the MVHR. Identified demand variation (between 3.15 and 7.11 depending on the utility) appears to depend highly on occupant behaviour for both sample sizes (4 and 25). Additionally, unregulated energy use, which is not explicitly accounted for in regulations and targets, is most important in terms of overall carbon emissions. The impact for zero carbon design is that assumptions based on the ‘typical’ or ‘average’ population may lead to misallocation of resources (in terms of design effort, time and money) and solutions which do not provide actual carbon savings. Anecdotally, during a site visit a comment from a resident was: “I should be a low heat consumer because I haven’t turned my heating system on at all during the summer months” (Resident)
Fig. 9. Gas boiler efficiency and monthly heat production in the District heating network.
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
whereas the individual was the second highest heat consumer. Despite an explicit desire to be a low consumer, mis-information (or lack of information) resulted in a severe disparity between reality and expectations. Smart electrical metering should make some progress with regards to encouraging and facilitating conservative electricity consumption behaviour but water and heat remain at the disposal of the uninformed user. Furthermore (and not under occupant control), site-wide energy and water consumption losses and inefficiencies were found to severely reduce the average measured performance; by 85% (energy) and 70% (carbon emissions). Not withstanding monitoring expenses, work shall follow this study using statistically significant samples (selected in a nonbiased way and conducted within the privacy laws and ethical standards) of low energy design variants with different occupant socio-demographics. Complete utility monitoring is essential. 6. Conclusions This is one of the few papers that attempts to measure all available utility inputs (water, electricity, heat) for individual dwellings and associated site uses; this holistic approach was justified. Within this study; • Occupant behaviour, in some houses, undermined overall performance and compliance with standards and design expectations. • The impact of associated site-wide energy consumption severely reduces the average dwelling performance. • Estimated energy use outside the scope of UK regulations, was the most important cause of carbon emissions (37–45%) Confirmation of these findings would complicate the UK mandate of zero carbon new-build dwellings by 2016. Further post-occupancy evaluation work is being conducted to assess the comfort and satisfaction of the occupants, critical performance mandates in a positive uptake of low energy dwellings. Additionally, a behavioural study is being conducted to determine the impact of user behaviour and attitudes on actual energy performance, and to determine any common profligate or frugal patterns of consumption, giving more context to the significant variation in measured energy performance. Acknowledgments Funding from EPSRC and Buro Happold under the Engineering Doctorate scheme is gratefully acknowledged. The authors would like to thank the residents at the housing site for their willingness to participate in monitoring activity, and the housing association for facilitating the work. References [1] Meeting the Energy Challenge. A White Paper on Energy, DTi, The Stationery Office, UK, 2007. [2] B. Metz, O.R. Davidson, P.R. Bosch, R. Dave, L.A. Meyer, Climate Change Mitigation of Climate Change. International Panel on Climate Change, Cambridge University Press, 2007. [3] T. Fox, Climate Change: Have We Lost the Battle, Institute of Mechanical Engineers, 2009. [4] K. King, J. Goodwin, N. Passant, N. Brophy, I. Tsagatakis Local and Regional CO2 Emissions Estimates for 2005–2006 for the UK, AEA report to DEFRA, 2008.
125
[5] J.L. Utley, L.D. Shorrock, Domestic Energy Factfile 2008, BRE, EST and DECC, 2008. [6] Building a Greener Future: Policy Statement, Department for Communities and Local Government, London (2007). [7] M. Osmani, A. O’Reilly, Feasibility of zero carbon homes in England by 2016: a house builder’s perspective, Building and Environment 44 (2009) 1917–1924. [8] L. Wang, J. Gwilliam, P. Jones, Case study of zero energy house design in the UK, Energy and Buildings 41 (2009) 1215–1222. [9] D. Johnston, R. Lowe, M. Bell, An exploration of the technical feasibility of achieving CO2 emission reductions in excess of 60% within the UK housing stock by the year, Energy Policy 33 (2005) 1643–1659. [10] R. Becker, Window air tightness and its influence on energy saving and minimum required ventilation, Building and Environment 14 (3) (1979) 157–165. [11] C. Simonson, Energy consumption and ventilation performance of a naturally ventilated ecological house in a cold climate, Energy and Buildings 37 (1) (2005) 23–35. [12] M. Munzinger, F. Crick, E.J. Dayan, N. Pearsall, and C. Martin Domestic photovoltaic field trials—Final technical report. DTi (2006). [13] Warwick wind trials—Final report. Encraft. Available fro m: http://www. warwickwindtrials.org.uk/resources/, 2009 (accessed 19.02.10.). [14] O. Kaynakli, A study on residential heating energy requirement and optimum insulation thickness, Renewable Energy 33 (2008) 1164–1172. [15] I. Pegg, A. Cripps, M. Kolokotroni, Post-occupancy evaluation performance of five low-energy schools in the UK, ASHRAE Transactions 113 (2) (2007). [16] W. Bordass, R. Cohen, M. Standeven, A. Leaman, Assessing building performance in use, 2. Energy performance of the probe buildings, Building Research and Information 29 (2) (2001) 114–128. [17] M. Ahmad, C.H. Culp, Uncalibrated building energy simulation modelling results, HVAC&R Research 12 (4) (2006) 1141–1155. [18] Y.G. Yohanis, J.D. Mondol, A. Wright, B. Norton, Real-life energy use in the UK: how occupancy and dwelling characteristics affect domestic electricity use, Energy and Buildings 40 (2008) 1053–1059. [19] E. Juodis, E. Jaraminiene, E. Dudkiewicz, Inherent variability of heat consumption in residential buildings, Energy and Buildings 41 (2009) 1188–1194. [20] A.J. Summerfield, R.J. Lowe, H.R. Bruhns, J.A. Caeiro, J.P. Steadman, T. Oreszczyn, Milton Keynes Energy Park revisited: changes in internal temperatures and energy usage, Energy and Buildings 39 (2007) 783–791. [21] Building Regulations Part L1A: Conservation of Fuel and Power in New Dwellings, Office of the Deputy Prime Minister (2006). [22] BS EN 1434-3:2008. Heat Meters Data Exchange and Interfaces. [23] Gas sales and numbers of customers by region and local authority 2005, 2006 and 2007. DECC. Available from: http://www.decc.gov.uk/en/content/cms/ statistics/regional/gas/gas.aspx, 2009 (accessed 4.02.10.). [24] Regional and local authority electricity consumption statistics 2005, 2006 and 2007. DECC. Available from: http://www.decc.gov.uk/en/content/cms/ statistics/regional/electricity/electricity.aspx, 2008 (accessed 4.02.10.). [25] General Information Report 38 (GIR38)—Review of ultra-low-energy homes: a series of UK and overseas profiles. BRECSU (1996). [26] BedZED Seven Years On. Technical Report. Bioregional (2009). [27] SAP 2005: The Government’s Standard Assessment Procedure for Energy Rating of Dwellings 2005 edition, Revision 3. DECC for BRE (2009). [28] Table 5: Resource use and waste. DEFRA. Available from: http://www. defra.gov.uk/sustainable/government/progress/data-resources/national.htm (accessed 18.02.10.). [29] English House Condition Survey, Summary Statistics 2007. Available from: http://www.communities.gov.uk/housing/housingresearch/housingsurveys/ englishhousecondition/ehcsdatasupporting/ehcsstandardtables/ summarystatistics/ (accessed 4.02.10.). [30] Code for sustainable homes: a step-change in sustainable home building practice. Communities and Local Government (2006). [31] B. Boardman, et al., DECADE: Domestic Equipment and Carbon Dioxide Emissions Second year report. Environmental Change Unit, University of Oxford, 1995. [32] I. Mansouri, M. Newborough, D. Probert, Energy consumption in UK households: impact of domestic electrical appliances, Applied Energy 54 (3) (1996) 211–285. [33] Monthly Degree Day Data. Available from: http://www.vesma.com/ddd/ regular.htm (accessed 14.12.09.). [34] Met Office Cambridge NIAB Weather Station Data. Available from: http://www.metoffice.gov.uk/climate/uk/stationdata/cambridgedata.txt (accessed 18.02.10.). [35] Rainwater and greywater use in buildings. Best practice guidance. CIRIA (2001). [36] Draft SAP 2009: The Government’s Standard Assessment Procedure for Energy Rating of Dwellings. DECC for BRE (2009). [37] S. Firth, K. Lomas, A. Wright, R. Wall, Identifying trends in the use of domestic appliances from household electricity consumption measurements, Energy and Buildings 40 (2008) 926–936.
Contents lists available at ScienceDirect
Energy and Buildings journal homepage: www.elsevier.com/locate/enbuild
Measured energy and water performance of an aspiring low energy/carbon affordable housing site in the UK Zachary M. Gill a,b,∗ , Michael J. Tierney b , Ian M. Pegg a , Neil Allan b a b
Buro Happold, 17 Newman Street, London W1T 1PD, United Kingdom Industrial Doctorate Centre in Systems, The University of Bristol, Queens Building, Bristol BS8 1TR, United Kingdom
a r t i c l e
i n f o
Article history: Received 23 March 2010 Received in revised form 23 August 2010 Accepted 25 August 2010 Keywords: Domestic Energy Water Building performance Low-energy Post-occupancy evaluation
a b s t r a c t This paper reviews the annual energy and water performance of an aspiring low energy/carbon affordable housing development in southern UK, comprising 25 houses heated by a biomass fueled district heating network. Electrical, heat and water consumption data was collected for all dwellings and more detailed data logging was conducted on a sample of four. The data was analysed to benchmark the overall site performance and compare the individual dwellings. The individual dwellings performed efficiently in terms of electricity, heat and water consumption. Significant variation was identified between maximum and minimum consumers, even when typical performance correlates (i.e. number of occupants, floor area) were accounted for. Water consumption varies by a factor of > 7 and heat and electrical consumption by > 3 which, given the homogeneous design and specification of the buildings, suggested that occupant behaviour was important. Site-wide energy consumption and losses (over and above individual dwelling use) reduced energy and carbon emissions performance by 85% and 70% respectively. Although energy consumption in UK homes can be substantially reduced, the challenge for zero-carbon design is to address issues of total system performance, occupant behaviour and whole-building energy performance. © 2010 Elsevier B.V. All rights reserved.
1. Introduction The demand to reduce energy consumption and carbon emissions in the built environment is well documented [1–3]. Dwellings play an important role within this (contributing approximately 30% of end-user carbon emissions and energy use [4,5]) and correspondingly the UK government has stipulated that by 2016 all new homes will be zero carbon [6]. Whilst the effectiveness of these policy measures has come into question (primarily due to their clarity and implementation costs) [7], in principle low and even zero-energy or carbon homes in the UK1 are achievable through a combination of improved insulation, more efficient mechanical and electrical equipment, higher air tightness (with controlled ventilation) and renewable technologies [8–14]. However, available post-occupancy evaluation (POE) research shows real-world building performance aligns poorly with design expectations and some nominally low-energy buildings perform no better than their more traditional counterparts [15,16]. Computational prediction of
∗ Corresponding author at: Buro Happold, 17 Newman St, London, W1T 1PD, United Kingdom. E-mail address: [email protected] (Z.M. Gill). 1 With a mid-European coastal climate. 0378-7788/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.enbuild.2010.08.025
energy consumption also shows poor correlation to monitored performance even when it is based on as-built construction details and actual building occupancy/utilisation [17]. One must inevitably question the use of computed in situ building performance in the absence of measurement. Measurements in ‘traditional’ dwellings show that characteristics of the building (for instance floor area, number of occupants, terraced or detached) play a role in determining actual energy consumption [18]. Statistically significant variation in heat and electrical energy consumption is noted even in similar dwellings and consumers can be categorised into high, medium and low bands. Furthermore there is an increasing trend of energy consumption over time in the UK contributed to most significantly by ‘high’ energy consumers [19,20]. Achieving low and even zero carbon/energy buildings in the UK is complicated because; design and reality align poorly; zero carbon dwellings are required but are not fully understood by the industry; house characteristics are not the sole determinants of performance; and energy use is increasing over time. This study assesses the performance of an affordable, low energy/carbon site in the UK where the construction and specification of the buildings is homogeneous, and the results from monitoring are presented and analysed. It is intended that this comprehensive measurement of building performance will provide insight into the impact of identified issues on low energy building variants.
118
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
Table 1 Dwelling characteristics. Type
Floor area (m2 )
Bedrooms
Storeys
Number of
Currently occupied
1 2 3 4 (1st end) 4 (1st mid) 4 (Ground)
68.13 74.58 81.11 56.38 49.02 44.70
2 2 3 1 1 1
2 2 3 1 1 1
10 3 9 1 1 2
9 3 9 1 1 2
2. Site description
3. Methodology
The East-Anglian site comprises of 13 two-bedroom and 9 threebedroom houses, plus 4 one-bedroom flats, each constructed to the same design specification and was awarded BRE’s EcoHomes Excellent certification. The site systems exceed the requirements of UK Building Regulations and typical UK dwellings. Low in-use carbon emissions and energy/resource consumption are facilitated by;
During the design, an arrangement was negotiated with the site landlord (a Housing Association) to conduct on-going site monitoring. Consumption monitoring was permitted for all dwellings and a shortlist of potentially willing residents was provided to instigate more detailed monitoring; Four of the five shortlisted dwellings agreed to participate. No energy performance information was fed back to any of the participants during the monitoring period.
• Biomass District Heating Network: A woodchip-fired community heating network (with a gas back up boiler for peak loads) services the heating and hot water demand of dwellings. Pre-insulated pipe work circulates heated water to local heat exchangers (one per dwelling) which extract heat according to demand. Each dwelling is individually controlled by a single programmer (with thermostat) and TRVs on all radiators. • Whole House Mechanical Ventilation with Heat Recovery (MVHR): Replaces the background ventilation usually provided by trickle vents and infiltration. Air is extracted from wet rooms (bathrooms and kitchens) and heat is exchanged to filtered incoming air for living spaces (bedrooms and living room). • High Air Tightness: As constructed the buildings achieve a mean air tightness value of 3.47 m3 /m2 h at 50 Pa (with a standard deviation of 0.367 m3 /m2 h). This is a 65% improvement on the minimum regulatory requirement (10 m3 /m2 h [21]) and 30% improvement on the design target value. • Massing: In groups of three to reduce the area of exposed facade. The two and three storey buildings are grouped to ensure winter solar gain is maximised by staggering the blocks (minimising overshadowing) and by limiting the height of the southern block in each group to two storeys. • Rainwater Harvesting: A header tank (50–60 l capacity) is fitted in each dwelling and supplies filtered, unmetered greywater (with mains water backup) for toilet flushing and garden irrigation. • Optimised Solar Orientation: The main portions of glazing are located on the South facade to enable useful solar gains in the winter. Operable windows and thermal mass in party walls are provided to ensure that the temperatures are stabilised. Initial dynamic thermal modelling in IES VE showed that internal temperatures peak at 25 ◦ C in the summer months, thus avoiding overheating. • Improved Insulation: Installed insulation is better than mandated in the UK Building Regulations [21]: Walls are 0.25 W/m2 K (vs. 0.35 W/m2 K), Glazing 1.9 W/m2 K (vs. 2.2 W/m2 K), Roof 0.12 W/m2 K (vs. 0.25 W/m2 K) and Ground 0.25 W/m2 K (equal to regulations) • Low Energy Lighting: Installed throughout the homes. • Low Flow Fixtures: Installed throughout the homes. There are four dwelling types (Table 1). The site opened in September 2008 and 16 of the 26 dwellings have since been occupied fully. Seven were subsequently occupied from January 2009. Two were occupied in February 2009 and then November 2009. Monitoring data was collected between September 2008 and November 2009 inclusive.
3.1. Detailed monitoring—4 dwellings Electrical energy consumption was monitored via a single-phase current clamp and data logger attached to the main incoming supply to each building. An instantaneous current measurement was recorded every 10 min at 0.4 A resolution providing a detailed electrical load profile. Small equipment load switching (<100 W) and short equipment use (not occurring at one of the logging intervals or <10 min use duration) was missed and therefore electro-mechanical induction meters reported cumulative consumption (in kWh). Water consumption was monitored via a pulsed output data logger connected to a manufacturer supplied reed switch mounted to the top of the rotary water meter housing. A pulse was outputted for every litre of water, and cumulative consumption was logged at 10-min intervals. Accuracy of the monitoring equipment was checked via comparison to manual meter readings and was better than 0.1%. Heat consumption (supplying heat and hot water demand) was also monitored via a pulsed output data logger recording cumulative consumption in 10-min intervals. The logger was connected via a manufacture supplied pulsed output module connected to the heat meter located at the heating network heat exchanger. The measuring accuracy of the installed heat meter exceeded the requirements of EN 1434 Class 2 and 3 [22]: better than ±2% across it’s operating range. Internal temperature and relative humidity was logged in 30-min intervals using compact HOBO® H08 data loggers. Readings were taken for the summer and winter periods in a frequently used space within the house—the combined kitchen–dining room. All loggers required information to be downloaded manually onto a computer and therefore provided no instantaneous feedback to the residents. Loggers were also concealed to minimise their influence on the occupant’s behaviour. All of the dwellings had three occupants and were end-terraced (2 West and 2 East). Three of the houses were the largest Type 3 whereas one is Type 1. 3.2. Monitoring—all dwellings All dwellings were monitored manually using installed utility meters. Electric and water meter readings were taken periodically via externally located meters. Heat meters were all located within the dwelling on the heat exchanger pipe work, and were therefore less accessible. Readings were taken on the date the occupants moved in and further readings were taken at 6 monthly intervals (April 2009 and October 2009). Dwellings were billed pro-rata for electrical and water consumption. Heat consumption was charged
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
119
Table 2 Comparison of low energy site to UK/local averages, design estimates and a regulation compliant solution. Heat
Electricity
Total energy
Water
kWh/year
kg CO2 /year
kWh/year
kg CO2 /year
kWh/year
kg CO2 /year
litres/person/day
GB average per unit area (m2 ) Mid-Suffolk average per unit area (m2 ) Design solution a per unit area (m2 ) Regulation (TER) /m2 Study site per unit area (m2 ) Site maximum per unit area (m2 ) Site minimum per unit area (m2 )
17,614 191.5 15,826 172.0 8517 121.0 – 6556 92.9 11,752 144.9 2272 46.0
3417 37.1 3070 33.4 213 3.0 – 759 10.8 1360 16.8 263 5.3
4392 47.7 5699 61.9 981 13.9 – 2885 40.9 4481 64.7 990 17.6
1853 20.1 2405 26.1 414 5.9 – 1217 17.3 1891 27.3 418 7.4
22,005 239.2 21,525 234.0 9499 135.0 – 9441 133.8 15,915 196.2 3529 72.2
5270 57.3 5475 59.5 627 8.9 24.8 1976 28.0 3117 38.4 794 15.3
148 – – – – – – 91
Space heating /m2 Hot water /m2 MVHR /m2 Lighting /m2 Other electrical /m2
73 20 – – –
8.6 2.4 – – –
– – 6.7 4.8–9.8 24.6–29.6
– – 2.8 2.0–4.1 10.4–12.5
196 – 28 –
Breakdown
a
Estimated Estimated
Note that design solution assumed 100% space heating and water heating supplied from biomass at 80% efficiency. Interpreted from SAP 2005 results
at a fixed monthly rate, adjusted every 6 months to account for demand variation between dwellings. Seven dwellings were inhabited for only 11 months of the monitoring period and one further dwelling inhabited for only 10 months. For these dwellings annual electrical, heat and water consumption was extrapolated to estimate annual consumption. Electrical consumption was linearly extrapolated based on the average consumption over the monitored period. This method of extrapolation was estimated to have an average error of ±2% based on results from the fully occupied dwellings. Water consumption was extrapolated using the same method and the estimated average error of readings was ±3%. Total heat energy consumption for the sub-sample was scaled up based on a kWh/degree day calculation (excluding hot water consumption). It was not possible to determine the accuracy of this measurement due to limited availability of monthly heat consumption data for all dwellings.
3.3. Whole site monitoring District heating network energy input was recorded via 2 heat meters (biomass and total heat production individually, with the numeric difference being gas heat production) and additionally via the main gas supply meter (recording m3 at NTP). Electricity consumption for the site facilities (predominately plant room equipment and external lighting) were metered via a single meter. Site water consumption was recorded via a single utility supply meter on-site. All meter readings (heat network, electrical and water) were recorded periodically throughout the year.
3.4. Comparison to design During the building design process, accredited modelling software (IES VE SAP 9.80 Calculator and Dynamic Thermal Modelling) was used to calculate the regulated performance of the buildings; the results will be compared against observation.
4. Results 4.1. Annual dwelling energy, carbon and water performance On average, the dwellings were found to perform efficiently in comparison to national and local (Mid-Suffolk region) averages [23,24], low energy benchmarks [25,26] and UK regulations (as modelled according to SAP 2005 [27]). Energy consumption within the dwellings was used primarily for heating and hot water (69%) and the remainder for electricity. Conversely, the associated carbon emissions were dominated by electricity use (62%) with the remainder contributed to by heating and hot water. Compared to national averages, total energy consumption was approximately 56% lower (or 43% lower on a per-unit-area basis) and carbon emissions were 63% lower (52% on a per-unit-area basis), see Table 2. Local averages compare similarly. Water consumption was 39% lower than the national average stated by DEFRA [28]. Note that Great Britain and Mid-Suffolk normalised averages were calculated based on an average dwelling floor area of approximately 92 m2 (adapted from the English House Condition Survey, 2007 Stock Data [29]). Carbon factors were taken from the current (and relevant) version of SAP (2005) [27]; admittedly these values will change in the 2010 revision. Compared against a set of low energy benchmarks, see Fig. 1, heat consumption was 5% lower than the dataset average2 but electrical consumption 11% higher, reflecting the intention to reduce heat consumption by implementation of MVHR. The dwellings compare well in terms of carbon emissions, 15% lower than the dataset average (or 4% when including Passivhaus projects), primarily due to the low heat carbon factor (0.118 kg CO2 /kWh based on a measured 45% biomass utilisation and SAP 2005 carbon factors [27]). Note that carbon emissions for low energy benchmark sites were calculated based on an assumed gas supply for heat unless otherwise stated. Generally, water consumption data was not freely available for the benchmarks. However, it is worth noting that the UK Code for Sustainable Homes outlines 5 levels of water consump-
2 European, Passivhaus projects (Werner and Passive Houses) are excluded from dataset averages.
120
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
tion and the site is close to achieving the second most ambitious target (91.5 vs. 90 l/person/day) [30]. Average dwelling performance exceeded regulatory demands by 25 –37%; these follow from the calculated average target emission rate (TER) listed in Table 2. However, unregulated energy consumption was estimated to contribute 37–45% of the total annual carbon emissions. Note that in defining the figures provided, fan power (for MVHR) was interpreted as total electrical consumption whilst buildings were unoccupied and lighting energy consumption was estimated from 3 sources; the DECADE project [31], Mansouri et al. [32] and SAP 2005 [27]. Further instrumentation or surveying would be required to isolate the electricity end uses more accurately. Annual electrical and heat (space heating + hot water) consumption for all dwellings, in terms of total and normalised energy use, is shown in Fig. 2. Dwelling occupancy is also denoted. Detailed measurements on Houses A–D are considered in 4.2.
Fig. 1. Benchmarked performance of monitored dwellings and other low energy dwellings in the UK.
Fig. 2. Annual energy consumption and carbon emissions per dwelling ranked in order of total normalized energy use (occupancy denoted)—Totals and breakdown per unit floor area.
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
121
Table 3 Heat, hot water and percentage of total water consumption per sample dwelling.
2
Total heat (kWh/m /year) of which: Space heating (kWh/m2 /year) Hot water (kWh/m2 /year) Hot water (kWh/person/year) Total water (litres/person/day) Hot water (litres/person/day) Proportion of hot vs. cold water
House A
House B
House C
House D
105.5 86.4 19.1 517 63.7 21.1 33%
144.9 103.5 41.4 1118 106.6 47.1 44%
95.5 78.3 17.2 465 84.7 19.0 22%
86.8 63.9 22.9 520 69.1 21.3 31%
Table 4 Number of occupants compared to average energy and water consumption. Number of occupants per dwelling
1
2
3
4
5
Overall
Average heat kWh/m2 /year Factor of difference (Max to Min) Average electrical kWh/person/year Average electrical kWh/year Factor of difference (Max to Min) a Average water litres/person/day Factor of difference (Max to Min) Sample size
63.1 1.48 1279 1279 1.03 173.3 1.30 2
83.6 2.04 1128 2257 3.45 95.9 4.73 6
92.4 3.15 1148 3444 2.52 80.8 3.01 11
116.8 1.50 788 3152 1.30 78.2 1.78 4
88.9 N/A 530 2650 N/A 64.8 N/A 1
92.9 3.15 1068 2885 3.46 91.2 7.11 24
a
When measured in kWh/person/year.
4.2. Results from sample dwellings
4.2.2. Space heating The range of thermal energy consumption per unit area, varied by a factor of 3.15 between maximum and minimum consumers across the site, Table 4. House B was the highest heat consumer
throughout the monitoring period, utilising a total of 11,752 kWh/year, 54% higher then the site average and 17%, 27% and 38% higher then Houses A, C and D respectively. The additional heating energy consumption due to hot water demand accounts for a proportion of this increase but space heating behaviour also contributes importantly. The programming of set-point temperatures (Fig. 3) into thermostats was a potential cause of excessive heating demand. Thermostats can be programmed to control the desired temperature at 5 distinct times during the day for weekdays (Monday–Friday) and weekends (Saturday–Sunday) independently. Average set-point temperatures of Houses A–C were approximately equal to 21 ◦ C compared to 19.7 ◦ C for House D, although temperature profiles varied significantly (for instance, a constant set-point in House A and high peak temperatures in House B). Actual internal temperature in all four dwellings across the year had a mean average value of 21.9 ◦ C ±0.3 with an overall range of 16.0–29.9 ◦ C. To put this into context, Summerfield et al. found average internal temperatures to be 19.8 ◦ C at a ‘low energy’ site [20]. Average internal temperatures for Houses B and D (the highest and lowest in the sample) are shown in Fig. 4. Summer temperatures were similar for both dwellings although morning temperatures are higher for B, most likely driven by the high thermostat setting at 6:00am. Winter temperatures differed
Fig. 3. Weekday and weekend thermostat set-point temperatures for sample dwellings.
Fig. 4. Average internal temperatures for sample dwellings (winter and summer) in kitchen dining room.
For four dwellings, half-hourly (or more frequent) heat, water and electrical consumption and internal temperature were monitored (see Houses A–D in Fig. 2). 4.2.1. Hot water Hot water and space heating consumption were not metered individually. The rate of heating for hot water consumption was taken as the measured consumption (of all heating) between 4th June to 17th August 09, when the heating degree days for the locality were <5% of the annual total [33]. Estimated heat and hot water consumption per dwelling (per person) is itemised in Table 3. House B used roughly twice the heating for hot water as other houses (Table 3). The fraction of total water consumption that was heated ranges from 22% to 44% (assuming 0.067 kWh/lhot water following tests in unoccupied buildings). Washing and cleaning (using hot water) probably contributed to the house-to-house variation in heat demand (indicated in Tables 2 and 3, Fig. 2).
122
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
Fig. 5. Daily electrical load profiles for sample dwellings: (a) 6th and 3rd highest consumers (Houses A, B) and (b) highest consumers (Houses C, D). Dots show daily readings and lines average data over 1 year.
significantly as indicated by approximately 2 ◦ C temperature difference through the afternoon. Admittedly the sample size was small, but nonetheless thermostat settings (Fig. 3) might well contribute to the variation in space heating demand (Tables 3 and 4). Window and door opening behaviour and utilisation of the MVHR to ventilate stuffy air from wet rooms (rather than opening a window) were other factors that could contribute to the significant variation in heat consumption. The annual saving of a mid-terraced dwelling was found to be between 22–29% (heat consumption of 74.9, 95.5 and 105.0 kWh/m2 /year for mid, West and East terraced dwellings respectively).3 This agrees well with Yohanis et al. who measured a 24–30% difference in heat consumption between detached and terraced houses [18]. 4.2.3. Electrical In line with Yohanis et al. [18], electricity consumption per occupant decreased as the number of occupants increased whereas the relationship between total electrical consumption and number of occupants is less clear, and only partially aligns with Yohanis’ assertion that it should increase (Table 4). The factor of difference between maximum and minimum consumers when considering total, per person or per m2 consumption (shown in Tables 2 and 4) was large; between 3.46 and 4.53 respectively. Fig. 5 shows daily electrical load profiles for individual dwellings. Load profiles of Houses A and B were similar, with two broad peaks in consumption. Houses C and D, showed markedly different peak and baseload patterns, albeit similar annual consumption. Energy savings in each of these may require different approaches, for instance a focus on standby and un-switched equipment in D versus equipment efficiency and utilisation in C.
water collection area = 1740 m3 . Pitched roof collection efficiency = 75%, filter efficiency = 90% [35]. Utilisation of harvested rainwater = 100%. An inverse relationship between the number of occupants and the water consumption per person, see Table 4, suggests improved efficiency through common and shared water use (for instance; dish and clothes washing). However, given the measured variation across the whole site and at different occupancy levels, number of occupants was only one determinate of water consumption. The factor of difference between maximum and minimum consumers (measured in litres/person/day) was the most significant of all utilities at 7.11, see Table 4. The daily water consumption profiles for each sample dwelling are shown in Fig. 7, and indicate this variation in 3-occupant dwellings. Typically, a morning and evening peak are found, although the specific time, size and duration of the peak vary considerably. The consequence of this is that strict water reduction targets as stipulated in the Code for Sustainable Homes [30], are unlikely to be achieved through low flow fixtures and rainwater harvesting alone, but require enthusiastic participation from the occupants. 4.3. Site energy performance Ultimately it is the whole site performance that is judged; this depends only partly on building specification and occupant behaviour.
4.2.4. Water Water savings (91 vs. 148 l/person/day UK average) can partially be attributed to the rainwater harvesting, utilised for toilet flushing and garden irrigation. Fig. 6 shows the negative correlation between the estimated harvested rainwater and the average water consumption of the sample dwellings for 10 months between January and October 2009 (when full water monitoring data was available for each dwelling given that data has been downloaded up to November 5th 2009). In 2009 an estimated 614 m3 rainwater was harvested, offsetting approximately 24 l/person/day. Assumptions: Average monthly rainfall equivalent to Met Office data from Cambridge NIAB (<35 miles from site) [34]. Total rain-
3 The 9% difference between East and West end-terraces is not found to be statistically significant.
Fig. 6. Correlation between estimated total harvested rainwater collection and average dwelling consumption (February–October 2009).
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
123
Fig. 7. Average daily mains water consumption profiles for sample dwellings: (a) Houses A, B and (b) Houses C, D. Dots show daily readings and lines average data over 1 year.
4.3.1. District heating network District heating energy was derived from biomass and natural gas. Scheduled meter readings from the district heating network and individual houses (Table 5) highlight the distribution and conversion losses. The fraction of heat from biomass was low but should increase as the site becomes established. If it continues to improve (as expected) then the overall associated carbon emissions of the heat consumption at the site will tend to reduce. Assuming 70% of the heat was supplied by biomass (as per conservative design estimates) then the overall heating carbon factor would reduce dramatically from 0.118 to 0.076 kg CO2 /kWh. Fuel supply and a short maintenance issue with the biomass store moving floor have reduced the use of biomass so far. Average annual heat consumption per dwelling increases dramatically from 93 to 185 kWh/m2 /year when distribution and conversion losses are accounted for (Fig. 8). Correspondingly, average total energy consumption increases by approximately 69% and total carbon emissions by 40%. Nevertheless, total annual heating carbon emission savings are approximately 15% compared hypothetically to gas boilers installed in each dwelling with nominal seasonal efficiencies of 70%. Periodic measurement of heat supplied to the district heating network attributed to biomass and gas, and correspondingly the gas boiler efficiency, is shown in Fig. 9. In the absence of gas use, the efficiency cannot be determined. Gas boiler efficiency fluctuates significantly and was lower than expected with the maximum monthly efficiency measured to be only 77%. The issue of seasonal gas boiler efficiency has been noted by policy makers and accordingly revisions made in SAP 2009 will account for the seasonal efficiency of boilers and discrepancies between manufacturer specifications and actual expected performance (under real operating conditions) [36]. During the summer period the district heating network maintains flow and return temperatures
(90–70 ◦ C) to primarily service intermittent hot water demand. Distribution efficiency then decreases accordingly because less useful heat is extracted but the thermal (insulation) losses remain relatively constant (loop to ambient temperature differences do not change significantly). A more efficient scheduling of plant operation is being investigated to reduce losses during periods of low demand. 4.3.2. Site electrical energy Electrical energy was used for the whole site to operate the district heating network and for external lighting. Annual electrical energy consumption equaled approximately 30,000 kWh (or 21 kWh/m2 /year) bringing total electrical demand to 62 kWh/m2 /year. Fig. 8 shows the impact of the additional electrical consumption in terms of energy and associated carbon emissions. In addition to the previously identified heating system losses (Section 4.3.1), site electrical energy increased average dwelling consumption by a further 9% and total carbon emissions by a further 23%. Pumping power for the district heating network was measured independently during a week in September. According to these measurements, pumping power accounted for approximately 41% of the total site electrical consumption and further investigation is required to assess the impact of other site electrical demands (predominately external lighting, a moving floor in the biomass store, boiler pumps, fans and controls). 4.3.3. Site water consumption Water is used on-site for communal irrigation of the grounds and has also been used to replace water in the district heating network when there was an installation pipe failure. Over the monitoring period, an average of approximately 600 l/day was consumed for site uses although approximately 300 l/day was accounted for due to a major leak in the manually operated play park irrigation system
Table 5 District heating network system efficiencies. All data in kWh unless otherwise stated
01/04/2009
30/10/2009
05/11/2009
Total house consumption Total boiler output of which: Biomass Gas (Gas input to date) Average gas boiler efficiency to date Distribution efficiency to date Proportion of fuel from biomass
140,232 203,704 88,562 115,142 157,932 73% 69% 36%
180,933 297,590 162,703 134,886 197,367 68% 61% 45%
– 300,456 164,373 136,083 199,226 N/A N/A N/A
Notes: Biomass boiler conversion efficiency is not determined as the quantity and quality of wood combusted is not recorded. A linear extrapolation of the meter readings was made (to align periods) based on the daily average heat output over the corresponding period. Adjusted data is italicised. Proportion of fuel from biomass is calculated based on heat input data.
124
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
Fig. 8. Impact of site-systems on average dwelling performance in terms of (a) energy and (b) carbon.
between August and October 2009. Excluding these losses, average water consumption per dwelling increases by approximately 8.5 l/person/day or 9%. 5. Discussion This study focused on a statistically small sample (25) of aspirational low energy dwellings within the UK. Following negotiations four dwellings were made available for intensive data logging, giving an insight into building performance and variation in the context of the whole site. The study contributes usefully to the overall dataset of actual building performance because of the direct comparability of the buildings thanks to their homogeneous design and construction. As suggested in other studies of a similar nature, it is unwise to draw conclusions from these results with respect to the entire domestic stock or carbon policy measures in the UK [37]. However, monitored utility consumption eludes to some poignant issues regarding the reality of achieving low or even zero carbon dwellings.
On average, dwellings were found to compare well to recognised low energy and national benchmarks. A combination of improved insulation, air tightness (with controlled ventilation) and optimised orientation contributed to the low annual heat consumption but couldn’t be itemised. Rainwater harvesting and low flow fixtures contributed to low annual water consumption and low energy lighting offset some of the additional energy demand of the MVHR. Identified demand variation (between 3.15 and 7.11 depending on the utility) appears to depend highly on occupant behaviour for both sample sizes (4 and 25). Additionally, unregulated energy use, which is not explicitly accounted for in regulations and targets, is most important in terms of overall carbon emissions. The impact for zero carbon design is that assumptions based on the ‘typical’ or ‘average’ population may lead to misallocation of resources (in terms of design effort, time and money) and solutions which do not provide actual carbon savings. Anecdotally, during a site visit a comment from a resident was: “I should be a low heat consumer because I haven’t turned my heating system on at all during the summer months” (Resident)
Fig. 9. Gas boiler efficiency and monthly heat production in the District heating network.
Z.M. Gill et al. / Energy and Buildings 43 (2011) 117–125
whereas the individual was the second highest heat consumer. Despite an explicit desire to be a low consumer, mis-information (or lack of information) resulted in a severe disparity between reality and expectations. Smart electrical metering should make some progress with regards to encouraging and facilitating conservative electricity consumption behaviour but water and heat remain at the disposal of the uninformed user. Furthermore (and not under occupant control), site-wide energy and water consumption losses and inefficiencies were found to severely reduce the average measured performance; by 85% (energy) and 70% (carbon emissions). Not withstanding monitoring expenses, work shall follow this study using statistically significant samples (selected in a nonbiased way and conducted within the privacy laws and ethical standards) of low energy design variants with different occupant socio-demographics. Complete utility monitoring is essential. 6. Conclusions This is one of the few papers that attempts to measure all available utility inputs (water, electricity, heat) for individual dwellings and associated site uses; this holistic approach was justified. Within this study; • Occupant behaviour, in some houses, undermined overall performance and compliance with standards and design expectations. • The impact of associated site-wide energy consumption severely reduces the average dwelling performance. • Estimated energy use outside the scope of UK regulations, was the most important cause of carbon emissions (37–45%) Confirmation of these findings would complicate the UK mandate of zero carbon new-build dwellings by 2016. Further post-occupancy evaluation work is being conducted to assess the comfort and satisfaction of the occupants, critical performance mandates in a positive uptake of low energy dwellings. Additionally, a behavioural study is being conducted to determine the impact of user behaviour and attitudes on actual energy performance, and to determine any common profligate or frugal patterns of consumption, giving more context to the significant variation in measured energy performance. Acknowledgments Funding from EPSRC and Buro Happold under the Engineering Doctorate scheme is gratefully acknowledged. The authors would like to thank the residents at the housing site for their willingness to participate in monitoring activity, and the housing association for facilitating the work. References [1] Meeting the Energy Challenge. A White Paper on Energy, DTi, The Stationery Office, UK, 2007. [2] B. Metz, O.R. Davidson, P.R. Bosch, R. Dave, L.A. Meyer, Climate Change Mitigation of Climate Change. International Panel on Climate Change, Cambridge University Press, 2007. [3] T. Fox, Climate Change: Have We Lost the Battle, Institute of Mechanical Engineers, 2009. [4] K. King, J. Goodwin, N. Passant, N. Brophy, I. Tsagatakis Local and Regional CO2 Emissions Estimates for 2005–2006 for the UK, AEA report to DEFRA, 2008.
125
[5] J.L. Utley, L.D. Shorrock, Domestic Energy Factfile 2008, BRE, EST and DECC, 2008. [6] Building a Greener Future: Policy Statement, Department for Communities and Local Government, London (2007). [7] M. Osmani, A. O’Reilly, Feasibility of zero carbon homes in England by 2016: a house builder’s perspective, Building and Environment 44 (2009) 1917–1924. [8] L. Wang, J. Gwilliam, P. Jones, Case study of zero energy house design in the UK, Energy and Buildings 41 (2009) 1215–1222. [9] D. Johnston, R. Lowe, M. Bell, An exploration of the technical feasibility of achieving CO2 emission reductions in excess of 60% within the UK housing stock by the year, Energy Policy 33 (2005) 1643–1659. [10] R. Becker, Window air tightness and its influence on energy saving and minimum required ventilation, Building and Environment 14 (3) (1979) 157–165. [11] C. Simonson, Energy consumption and ventilation performance of a naturally ventilated ecological house in a cold climate, Energy and Buildings 37 (1) (2005) 23–35. [12] M. Munzinger, F. Crick, E.J. Dayan, N. Pearsall, and C. Martin Domestic photovoltaic field trials—Final technical report. DTi (2006). [13] Warwick wind trials—Final report. Encraft. Available fro m: http://www. warwickwindtrials.org.uk/resources/, 2009 (accessed 19.02.10.). [14] O. Kaynakli, A study on residential heating energy requirement and optimum insulation thickness, Renewable Energy 33 (2008) 1164–1172. [15] I. Pegg, A. Cripps, M. Kolokotroni, Post-occupancy evaluation performance of five low-energy schools in the UK, ASHRAE Transactions 113 (2) (2007). [16] W. Bordass, R. Cohen, M. Standeven, A. Leaman, Assessing building performance in use, 2. Energy performance of the probe buildings, Building Research and Information 29 (2) (2001) 114–128. [17] M. Ahmad, C.H. Culp, Uncalibrated building energy simulation modelling results, HVAC&R Research 12 (4) (2006) 1141–1155. [18] Y.G. Yohanis, J.D. Mondol, A. Wright, B. Norton, Real-life energy use in the UK: how occupancy and dwelling characteristics affect domestic electricity use, Energy and Buildings 40 (2008) 1053–1059. [19] E. Juodis, E. Jaraminiene, E. Dudkiewicz, Inherent variability of heat consumption in residential buildings, Energy and Buildings 41 (2009) 1188–1194. [20] A.J. Summerfield, R.J. Lowe, H.R. Bruhns, J.A. Caeiro, J.P. Steadman, T. Oreszczyn, Milton Keynes Energy Park revisited: changes in internal temperatures and energy usage, Energy and Buildings 39 (2007) 783–791. [21] Building Regulations Part L1A: Conservation of Fuel and Power in New Dwellings, Office of the Deputy Prime Minister (2006). [22] BS EN 1434-3:2008. Heat Meters Data Exchange and Interfaces. [23] Gas sales and numbers of customers by region and local authority 2005, 2006 and 2007. DECC. Available from: http://www.decc.gov.uk/en/content/cms/ statistics/regional/gas/gas.aspx, 2009 (accessed 4.02.10.). [24] Regional and local authority electricity consumption statistics 2005, 2006 and 2007. DECC. Available from: http://www.decc.gov.uk/en/content/cms/ statistics/regional/electricity/electricity.aspx, 2008 (accessed 4.02.10.). [25] General Information Report 38 (GIR38)—Review of ultra-low-energy homes: a series of UK and overseas profiles. BRECSU (1996). [26] BedZED Seven Years On. Technical Report. Bioregional (2009). [27] SAP 2005: The Government’s Standard Assessment Procedure for Energy Rating of Dwellings 2005 edition, Revision 3. DECC for BRE (2009). [28] Table 5: Resource use and waste. DEFRA. Available from: http://www. defra.gov.uk/sustainable/government/progress/data-resources/national.htm (accessed 18.02.10.). [29] English House Condition Survey, Summary Statistics 2007. Available from: http://www.communities.gov.uk/housing/housingresearch/housingsurveys/ englishhousecondition/ehcsdatasupporting/ehcsstandardtables/ summarystatistics/ (accessed 4.02.10.). [30] Code for sustainable homes: a step-change in sustainable home building practice. Communities and Local Government (2006). [31] B. Boardman, et al., DECADE: Domestic Equipment and Carbon Dioxide Emissions Second year report. Environmental Change Unit, University of Oxford, 1995. [32] I. Mansouri, M. Newborough, D. Probert, Energy consumption in UK households: impact of domestic electrical appliances, Applied Energy 54 (3) (1996) 211–285. [33] Monthly Degree Day Data. Available from: http://www.vesma.com/ddd/ regular.htm (accessed 14.12.09.). [34] Met Office Cambridge NIAB Weather Station Data. Available from: http://www.metoffice.gov.uk/climate/uk/stationdata/cambridgedata.txt (accessed 18.02.10.). [35] Rainwater and greywater use in buildings. Best practice guidance. CIRIA (2001). [36] Draft SAP 2009: The Government’s Standard Assessment Procedure for Energy Rating of Dwellings. DECC for BRE (2009). [37] S. Firth, K. Lomas, A. Wright, R. Wall, Identifying trends in the use of domestic appliances from household electricity consumption measurements, Energy and Buildings 40 (2008) 926–936.