Testing the real heat loss of a Passivhaus building: Can the UK s energy performance gap be bridged?

Transcription

1 Testing the real heat loss of a Passivhaus building: Can the UK s energy performance gap be bridged? Mark Siddall, 1 John Trinick 2, Dr David Johnston 3 1 Northumbria University, Newcastle, UK and LEAP: low energy architectural practice (mark.siddall@leap4.it +44(0) ) 2 Warm, Plymouth, UK (john@peterwarm.co.uk +44(0) ) 3 CeBE, Leeds Metropolitan University, Leeds, UK 1 Introduction In the UK it has become increasingly recognised that the heat loss from new buildings does not accord with current theoretical practices. One of the reasons for this may be attributed to unaccounted thermal bypass (ZCH, 2010 and Stafford et al., 2012). The lead author (Siddall 2009 and 2011) contended that constructing Passivhaus buildings without dedicating due care and attention to thermal bypass mechanisms could result in building performance failures. This paper reviews the test methodology for coheating, current guidance on thermal bypass, and discusses the results obtained from a small number of coheating tests that have been undertaken on Passivhaus dwellings in the UK. The results of the coheating tests suggest that the Passivhaus standard is capable of closing this performance gap, as long as the designers and builders demonstrate an appreciation of how thermal bypass may impact upon a building. Sections 2-3 of this paper examine some causes of energy performance gaps. Sections 4-5 of this paper are based on experience at Ford Close (Trinick, 2012) and highlights some of the challenges and possible solutions for those wishing to conduct their own tests, particularly with Passivhaus buildings. Section 6 offers the results from a number of Passivhaus coheating tests within the UK. 2 Current Passivhaus Guidance and Training The guidance document Protokollband 18 (Feist, 2008) is intended to help support the delivery of appropriate standards of Quality Assurance for Passivhaus buildings. Whilst the Protokollband does discuss a number of factors associated with convective thermal bypass, it has been observed that it only addresses external wall insulation technology and does not consider party wall, timber frame or cavity wall systems, all of which are common in the UK. A further limitation is that the Protokollband is not currently available in English.

2 Thermal bypass mechanisms are not addressed in the Passivhaus Planning Package (PHPP). Certified Passivhaus Designer training does not contain any specific awareness of thermal bypass. However, in the UK the AECB CarbonLite Programme and the University of Strathclyde have supplemented their teaching to address this apparent deficit. Whilst some UK Passivhaus Certifiers and Passivhaus Designers are consciously designing out thermal bypass, the extent to which this is the case is currently uncertain. 3 Lessons Learned from Secondary Research Within the UK it has been recognised that thermal bypass mechanisms can have a significant impact upon delivered performance of the building fabric. Literature reviews undertaken by the lead author (Siddall) in 2009 and 2011 observed the following: 1) Two types of convective thermal bypass exist; open-loop and closed-loop. 2) The concepts of airtightness and windtightness are associated with mitigating the openloop condition. Airtight and wind tight construction protects the insulation layer against air movement whilst also assisting with preventing infiltration and exfiltration. 3) The concept of encapsulation is associated with overcoming the closed-loop condition. Insulation should either be consistently encapsulated by the elements that form the wind and air barriers, or the space between the insulation and wind/air barriers should be filled with solid material; such as a masonry wall. The challenge it would seem is delivering suitable standards of workmanship. Coheating tests, in conjunction with pressure tests and thermography, can assist with developing an appreciation of the delivered thermal performance of the building fabric and enable an evaluation as to the impact of thermal bypass upon performance. 4 How to carry out a Coheating test in a Passivhaus 4.1 The benefit of Coheating tests A coheating test gives the opportunity to measure and compare the actual fabric heat loss with the theoretical figure, which measurement of fuel bills cannot do, as they depend strongly on the owner s occupancy and choices. It can thus be used to feedback on construction quality and promote a better understanding of building physics. 4.2 Coheating test theory A coheating test is a quasi-steady state method of measuring the whole dwelling heat loss attributable to an unoccupied dwelling (Johnston et al., 2012a). It is one of only a few methods (an alternative is the PSTAR method see Subbararo et al., 1988) that are currently available to measure whole dwelling heat loss in the field.

3 The theory of a coheating test is that the energy input required to maintain a constant and even internal temperature will equal the heat loss over a period of time. Expressed as: electrical heat input + solar heat input = total transmission heat loss + total infiltration heat loss If the electrical heat input, the solar heat input, and the infiltration heat loss can be measured, then the total transmission heat loss (in W/K) can be determined. 4.3 Practical testing Testing involves using electrical resistance heaters to heat a building for a period of at least one week (not including the warm-up period) to around 25ºC whilst logging internal and external temperature, electrical power and solar flux. The external temperature should be low enough to maintain at least a delta T of 10ºC. The ventilation system is sealed to eliminate ventilation heat loss, and the infiltration rate determined either by a recent air pressure test or by a gas decay test. The essential items of equipment used at Ford Close were: A power logger - which usually counts pulses from the electrical meter, alternatively a separate device may be used to convert LED flashes to pulses, before being logged Variable control heaters with accurate thermostats proportional control works best and testing in advance is recommended to ascertain thermostat error and hysteresis Directional circulation fans are required to move heat around a house and prevent hot/cool spots Temperature loggers are required in each room and externally (sheltered & shaded) A Pyranometer and suitable logger are required to measure solar flux, absolute calibration is not essential as regression relies on relative analysis of data The following method can then be used to derive transmission heat loss coefficient in W/K: average the power input and temperature difference in each given period (see 5.1) derive the total input power per degree of temperature difference for each period remove the energy for heating infiltrating air from the overall balance for each period correlate the average W/K for each period to the average solar flux in the same period extrapolate to zero solar flux to find the total transmission loss coefficient The final balance is expressed as: total transmission heat loss = electrical heat input + solar heat input - total infiltration heat loss Leeds Met also publishes a protocol for testing which includes suggested equipment (Johnston et al., 2012a).

4 5 Challenges of a coheating test on a Passivhaus 5.1 Solar gain and thermal mass Solar gain and the effect of thermal mass are common to all coheating tests. The challenge with testing a Passivhaus is the relatively large contribution of solar gains to the overall energy balance, and the lag that is possible between the peak solar gain and the final effect on heat requirement. This is shown in figure 1. It might be suggested that covering the windows externally during the test would remove the solar gain, but this would also affect the heat loss through the glazing. Fig 1: Heat demand (W/K) and Solar Flux (W/m2) time series from test of Ford Close Passivhaus During the analysis various length moving-averages, with and without delays, were tested to represent the effect shown above. However, with input from a conversation with Jez Wingfield (Willmott Dixon Energy Services Limited), the simple approach of averaging both power and temperature difference over 24 hour periods from to 0659 gave the best correlation between heat loss and solar flux, and matched the real effect observed. The result of this method is a heat loss coefficient value for each 24 hour period which can be correlated against the average solar flux for this period see figure 2. Fig 2: Heat demand (W/K) and Solar Flux (W/m 2 ) correlation from test of Ford Close Passivhaus

5 5.2 Infiltration The removal of heat loss by infiltration is also common to all coheating tests. Infiltration is a function of wind speed and temperature, meaning removal based on a regression (as used with solar gain) is difficult. The CO 2 concentration decay method (Claude-Alain Roulet, Falvio Foradini, 2002) was used to give real infiltration measurements shown in figure 3, this confirmed that the infiltration rate was as expected for the pressure test result and gave a partial relationship between wind speed and infiltration. It may be possible to use this relationship to remove the infiltration loss using an average wind speed for a period. However, as insufficient data was available for the relationship, the coheating energy balance was corrected by assuming the infiltration was a constant figure, which was thought satisfactory since this figure was equal to only~ 5% of the total heat loss coefficient. Fig 3: Infiltration results (h -1 ) and wind speed from the test of Ford Close Passivhaus 5.3 Other challenges to be addressed in coheating tests Other challenges that were experienced during the coheating test of Ford Close were: The difficulty in maintaining stable temperature in rooms with south facing glazing (the fixed horizontal shading was not effective in winter) The heat loss due to radiant exchange with the night sky is not related to external temperature measurements and therefore is not allowed for in the analysis 6 Coheating Tests and Results Two Passivhaus dwellings from a recently completed development in Sunderland have also undergone a coheating test. Fig. 4 illustrates the results and serves to demonstrate that the measured and predicted heat loss coefficient for dwelling 1 are in very close agreement with one another (46.7W/K as opposed to the design of 43.4W/K) ), with the difference in heat loss coefficient being well within the range of the measurement error associated with

6 the test (Johnston et al., 2012b). Fig. 5 compares Racecourse dwellings 1 and 2 (far right) with 22 dwellings from the Leeds Metropolitan University coheating database. It can be observed that the performance gap between predicted and measured performance has been effectively closed Corrected data Predicted y = x R² = Predicted Measured Power (W) Heat loss (W/K) Data corrected for solar and wind. Wind added back 2.9ms Delta T (K) 0.00 Fig 4: Solar and wind corrected heat loss data for dwelling 1 Fig 5: Measured versus predicted heat loss coefficients for 22 dwellings from the Leeds Met database In this instance the building envelope performs pretty much as predicted. This raises the question whether similar results can be achieved at other Passivhaus developments. To address this line of enquiry it is necessary to draw upon other coheating tests that have been undertaken on other Passivhaus dwellings. It should be noted that these coheating test were undertaken by parties other that Leeds Metropolitan University and that the coheating methodologies have not yet been normalised; which could lead to uncertainty with regard to the results and their interpretation. These matters notwithstanding, the author has complied data from coheating tests from other Passivhaus dwellings in Table 1. Racecourse Dwelling 1 (Johnston et al, 2012b) Racecourse Dwelling 2 (Johnston et al, 2012b) Larch House (Jenkins, 2011) Lime House (Jenkins, 2011) Ford Close (Trinick, 2012) Predicted (W/K) Measured (W/K) Error compared to target (W/K) Dwelling Type (+7.6%) Endterrace (+4.0%) Midterrace / (+4.1%) Detached / (+10.2%) Detached (+10.5%) Terrace Table 1: Coheating Test results for Passivhaus dwellings in the UK Whilst the availability of coheating data is currently limited, and due to a lack of normalisation the comparability of the results is somewhat questionable, a certain amount of confidence can be found within the studies presented. The measured mean increase in heat loss from the Passivhaus dwellings is 3.16W/K. In contrast, the highest measured heat

7 loss from any new-build dwelling (UK Building Regulations Part L 2006) contained within the Leeds Metropolitan University coheating test dataset was 282W/K, against a predicted heat loss of 225W/K (see left hand side of fig 5). It is noted that 3.16W/K is approximately 1% of the measured heat loss as compared to the Part L 2006 dwelling. 7 Conclusions: Testing Passivhaus buildings may offer considerable learning opportunities and need not be expensive or onerous providing building access is available. Removal of solar gain from the energy balance is best achieved by correlating average heat loss coefficients to average solar flux for 24 hour periods. Removing infiltration heat loss is less significant in a Passivhaus as the loss is small compared to the overall heat loss, and therefore it may be possible to use a constant figure unless extreme winds are encountered. This paper has recognised the impact that thermal bypass mechanisms can have upon building performance and briefly examined relevant Passivhaus design guidance. The lead author has reached the conclusion that this guidance, ideally the Passivhaus Planning Package (PHPP), requires updating so that it recognises the risks that thermal bypass can have the as built performance. It is believed that such measures would strongly support the on going development of the Passivhaus Standard as the worlds leading quality assurance standard for low energy buildings. The certifiers and designers of the UK Passivhaus projects that have undergone coheating tests have been consulted. It has been established that in each instance they have considered the risks imposed by thermal bypass and have then sought to develop their design and construction processes accordingly. At this initial stage, it would appear that Passivhaus Standards of quality assurance, when complimented by an appreciation of the potential impact of convective thermal bypass, can deliver buildings that perform as intended. 8 Acknowledgements Sections 4-6 benefitted from financial support received from the Technology Strategy Board s Building Performance Evaluation programme. Continued assistance and support is also acknowledged to the School of Built and Natural Environment, Northumbria University and the Centre for the Built Environment, Leeds Metropolitan University. References Claude-Alain Roulet, Falvio Foradini, 2002 Feist, 2008 Simple and Cheap Air Change Rate Measurement Using CO 2 Concentration Decays, Feist. W. (2008) Qualitätssicherung beim Bau von Passivhäusern, Protokollband No. 18, Passivhaus Institut

8 Jenkins, 2011 Johnston et al., 2012a Johnston et al, 2012b Siddall, 2009 Siddall, 2011 Stafford et al, 2012 Subbaro et al., 1988 Trinick, 2012 ZCH, 2010 Jenkins. H., Jiang. S., Guerra-Santin. O., Tweed. C. (2011) Coheating test, Future Works, Ebbw Vale v. 0.1, Welsh School of Architecture, Cardiff University Johnston. D., Miles-Shenton. D., Wingfield. J., Farmer, D., and Bell, M. (2012) Whole House Heat Loss Test Method (Coheating). Johnston. D., Miles-Shenton. D., Farmer. D., Wingfield. J. (2012) TSB Building Performance Evaluation, Post Construction and Early Occupation Study Sunderland, Coheating Test Report Siddall. M. (2009) Thermal Bypass: The impact upon performance of natural and forced convection, Proceedings of the International Passivhaus Conference, April 2009 Siddall. M. (2011) Thermal Bypass: The impact upon performance of natural and forced convection, UK Passivhaus Conference Stafford. A., Bell. M., Gorse. C. (2012) Building Confidence - A Working Paper, Centre for Low Carbon Buildings, [accessed 31/08/2012] Subbarao, K. Burch, J. D. Hancock, C. E. Lekov, A. and Balcomb, J. D. (1988) Short-Term Energy Monitoring (STEM): Application of the PSTAR Method to a Residence in Fredericksburg, Virginia. TR Solar Energy Research Institute, Colorado, USA. Coheating preliminary findings report, Unit 4, Ford Close, St Ives ZCH (2010) Closing the performance gap: Building low carbon housing for real. Report of Topic Work Group 4, Carbon Compliance Tool Policy Assumptions Task Group. July 2010, Zero Carbon Hub, Milton Keynes, UK.