Energy use and thermal comfort in a rammed earth office building
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Energy and Buildings 40 (2008) 793–800 www.elsevier.com/locate/enbuild Energy use and thermal comfort in a rammed earth office building P. Taylor a,*, R.J. Fuller b, M.B. Luther b a Environmental and Energy Manager, Division of Facilities Management, Thurgoona Campus, Charles Sturt University, PO Box 789, Albury, NSW 2640, Australia b Built Environment Research Group, School of Architecture and Building, Deakin University, Waterfront Campus, Geelong, Vic. 3217, Australia Received 2 February 2007; received in revised form 17 May 2007; accepted 21 May 2007 Abstract A two-storey rammed earth building was built on the Thurgoona Campus of Charles Sturt University in Albury-Wodonga, Australia, in 1999. The building is novel both in the use of materials and equipment for heating and cooling. The climate at Wodonga can be characterised as hot and dry, so the challenge of providing comfortable working conditions with minimal energy consumption is considerable. This paper describes an evaluation of the building in terms of measured thermal comfort and energy use. Measurements, confirmed by a staff questionnaire, found the building was too hot in summer and too cold in winter. Comparison with another office building in the same location found that the rammed earth building used more energy for heating. The thermal performance of three offices in the rammed earth building was investigated further using simulation to predict office temperatures. Comparisons were made with measurements made over typical weeks in summer and winter. The validated model has been used to investigate key building parameters and strategies to improve the thermal comfort and reduce energy consumption in the building. Simulations showed that improvements could be made by design and control strategy changes. # 2007 Elsevier B.V. All rights reserved. Keywords: Rammed earth; Thermal comfort; Energy; Office building 1. Introduction Traditionally thermal comfort has been achieved at the expense of significant energy use for heating and/or cooling. However, a A number of ‘green’ buildings have now been constructed in well-designed building should be able to provide good thermal Australia and throughout the world. Many of these buildings comfort, while simultaneously having low energy consump- have won awards. One of these is the two-storey rammed earth tion. The objective of this research was to establish (a) whether ‘Academic Offices Building’ on the Charles Sturt University the office building on the CSU Thurgoona Campus provides a (CSU) Campus at Thurgoona in New South Wales, Australia. satisfactory level of thermal comfort; and (b) if this building This campus has been called ‘‘deep green’’ because of its uses significantly less energy, and thus generates less green- claimed low environmental impact [1] and the Academic Office house gas emissions, than a nearby comparable building of Building has received a number of awards including a special conventional construction and operation. ‘‘jury award’’ from the Royal Australian Institute of Architects The purpose of this paper is to fill the information gap that for an environmentally sound design. However, there has been can exist between an award-winning ‘green’ building design little data presented on the performance of the campus or this and the realities of occupancy and operation. To provide the office building. necessary background to the study, the paper begins with a Thermal comfort is recognized as a key parameter for a description of the rammed earth building and the prevailing healthy and productive workplace. At the same time, lowering climate at its location. The methodology used to assess its energy use in commercial buildings is vital if a significant performance in terms of comfort and energy use is then reduction in greenhouse gas emissions is to be achieved. presented, followed by some results. ATRNSYS model of three of the offices was developed and this has been used to further investigate the building’s performance. The simulation results * Corresponding author. Tel.: +61 2 60242414; fax: +61 2 6051 9894. are discussed and some conclusions are drawn from these and E-mail address: ptaylor@csu.edu.au (P. Taylor). the earlier results. 0378-7788/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.enbuild.2007.05.013
794 P. Taylor et al. / Energy and Buildings 40 (2008) 793–800 Fig. 1. Northern aspect of rammed earth office building (right of figure). 2. Building description Fig. 3. Cross section through the building showing night ventilation path. The CSU office building contrasts sharply with a typical office building in almost every feature (Fig. 1). It is a two- conceal duct work, and the concrete ceilings are exposed. In storey building with load bearing rammed earth external and some places, a corrugated profile has been cast into the ceiling internal walls and there is no steel frame. The windows open to to increase its surface area to enhance convective heat transfer. permit natural ventilation. The building is curved and The floors are carpeted. orientated as shown in Fig. 2. There is a central corridor The windows are single 6 mm glass with wooden frames. running the length of the building on both levels with offices on The offices have double hung vertically sliding sash windows either side. Each office is typically 10.5 m2 in floor area. The with weather strip sealing. In the stair wells and at the end of the circles in Fig. 2 are rainwater tanks. The ellipses are voids corridors upstairs there are manually operated louvre windows. between the ground floor and upper storey giving a light and Window shading has been carefully designed to exclude all ventilation well. Ventilation towers or stacks are situated above direct beam sunlight during the summer months. Each office these wells. has a variable speed sweep fan controlled by the occupier and Hydronic heating and cooling has been installed instead of these fans are seen as an important cooling mechanism. The an HVAC system and there are circulation pipes embedded in building has woollen insulation placed underneath the roof the ground floor slab and the two ceiling slabs. Ninety-eight sheeting rather than on top of the upper ceiling slab. Solid-foam square metres of flat plate solar collectors have also been insulation was installed around the edge of the concrete slabs. installed on the roof of the building. It was anticipated The external doors at CSU all close automatically and seal that these panels would collect sufficient energy in winter to against a wooden frame. significantly reduce gas consumption. By circulating water through the panels at night in summer, it was also 3. Climate expected that radiant cooling would produce a store of cold water to reduce cooling energy requirements on the The region is characterised by long-hot summers and cool/ following day. cold-wet winters, typical of a Mediterranean climate. The mean Cooling is also achieved through using a night ventilation maximum temperature in January is 31.8 8C and the mean purge in summer (Fig. 3). Fresh air is allowed to circulate in minimum in July is 3.1 8C [2]. There is a large diurnal variation through the louvres located under the office windows and out between maximum and minimum temperatures. On average in through the louvres in chimney ventilation stacks. These sets of January, the diurnal swing is 16.6 8C, while in July it is 9.5 8C. louvres are computer controlled, whilst the louvres above the In January, the mean daily average temperature is 23.5 8C. In office doors may be only operated manually. Since a hydronic summer, on average there are 16.6 days when the maximum system is used, there is no need for a suspended ceiling to temperature is expected to be over 35 8C, while in winter there are 31.9 days with a minimum temperature below 2 8C (Fig. 4). Furthermore the daily maximum temperature is below 20 8C for 5 months of the year. In this area of inland Australia the relative humidity in summer is between 30 and 50%. As expected, the skies are less cloudy in summer with an expectation of 10 h sunshine per day, while in winter on average there are 4.5 h of sunshine per day. The global irradiation on a clear day in summer may reach Fig. 2. Schematic plan view of second storey of rammed earth office building. 30 MJ/m2.
P. Taylor et al. / Energy and Buildings 40 (2008) 793–800 795 working environment of their office; this was intended as a general satisfaction question. Section C elicited information on the thermal adaptive strategies of respondents: the use of fans, lights, blinds, and windows and concluded with a series of open questions specifically concerning comfort. Full details of the questionnaire used are available in [6]. An alternative way of measuring thermal comfort is to use one of the various models developed to describe this sensation and then calculate the fraction of occupancy time when these conditions are met by a particular building. The adaptive model of [7] was considered to be most appropriate to a naturally ventilated building such as CSU, where Fig. 4. Local pattern of extreme weather days. occupants have some control over their environment. Having established the neutral indoor operative temperature (‘neu- 4. Methodology trality’) using [7], the summer and winter comfort zone for the location was constructed using the method described by The office has been designed in accordance with a set of Auliciems and Szokolay [8]. Various measured physical environmental principles, established prior to the development parameters were used to indicate the percentage of time of the Thurgoona Campus. The principles are divided into three during working hours the indoor conditions in three offices categories; namely low energy, resource management and fell within these comfort zones at those times of the year. The environmental impact [3]. The low energy principle sought to internal and external surface heat fluxes of selected external maximize the passive use of the building ‘‘to collect, store and walls were measured with heat flux transducers (International distribute energy’’, while one of the resource management Thermal Instruments, Model A). Various temperatures in the principles was ‘‘to maximize the comfort of the building offices, namely: at several heights, globe, wet bulb, wall users’’. It was therefore appropriate to choose performance surface and mid-wall were measured with thermistors indicators which were consistent with these design principles. (Betatherm, 10K3A1IA). Reed switches recorded fan rotation The two indicators of performance used in this study are and window status. Data loggers (Campbell Scientific, thermal comfort and energy use. In general, a building that CRX10) stored averaged data at 15 min intervals and a consumes a relatively small amount of energy in its operation weather station (Davies Weather Monitor) collected climatic should be inexpensive to run and will have low CO2 emissions, data. Full details of measurements and instrumentation have an important environmental consideration. Some of the energy been previously reported [9]. used in a building is for the benefit of the occupants to make Both of the above methods were used to examine the thermal them comfortable and therefore productive, but while energy comfort performance of the CSU offices. reduction is important, it must not be at the expense of thermal comfort. The measurement methodology for these two 4.2. Energy use indicators is discussed below. The energy use and greenhouse gas emissions of the 4.1. Thermal comfort Academic Offices Building were compared with another CSU- owned building at the same location, also used for office Thermal comfort is a subjective measurement. Thermal accommodation and a library. Known as the ‘Gordon Beaven comfort is defined by [4] as ‘‘that condition of mind that building’ and constructed in 1987, this building has three split- expresses satisfaction with the thermal environment’’. One way level storeys and is constructed with concrete blocks within a to gauge thermal comfort levels is simply to ask building concrete beam frame. The building is fully air-conditioned and occupants how they experience a building in terms of their maintained in good order. Most of the buildings on the CSU temperature sensations. A number of scales have been Thurgoona Campus have their own gas and electricity meters developed to do this. The survey questions used in this study that are read by the maintenance staff on a weekly basis. were in part taken from the questionnaire developed by Monthly accounts from utility companies are also available. Levermore et al. [5]. However, because an adaptive model was Gas and electricity use in the Academic Offices Building was used to evaluate the building (see below), further questions established directly for 2001 from the meter readings. were added. Comparable data for the Gordon Beaven Building was obtained The final questionnaire therefore consisted of three sections. indirectly by subtracting the energy consumed by the newer Section A requested background information from respondents buildings on site from the overall figure read from the main site (age, sex, type of work, and amount of time spent on leave over meter. Using Australian industry guidelines [10], the useable the summer period). Section B asked a number of multiple- area of both buildings was established either from plans choice questions (on a seven-point scale) to guage a (Academic Offices) or from a list of room sizes (Gordon respondent’s perception of their room with regard to aesthetics Beaven), and was calculated to be 1500 and 2186 m2 and comfort. Section B also asked respondents to rate the respectively.
796 P. Taylor et al. / Energy and Buildings 40 (2008) 793–800 5. Results Thermal comfort during summer in three offices of the Academic Offices Building was investigated over a 28-day period from 13 February to 13 March 2001. Ambient temperatures varied from a minimum of 6.7 8C to a maximum of 36.9 8C with an average of 22.7 8C. Fig. 5 shows the coincidence of the measured psychrometric data and the comfort zones for the three offices and Table 1 presents a summary of the operative temperatures and percentage of time in the summer comfort zone in same three locations. The ground floor office (office A) located on the southeast corner of the building was clearly performing better than the other two offices. Its average temperature was 23.7 8C and was above 25.4 8C for 25% of the time. It was in the comfort zone for 73% of the time, but this was because the office was too cold, rather that too hot. This office was rarely used and therefore had a very low internal heat load. The two upper level offices were on average 1–2 8C warmer than the ground floor office. Office C located on the northwest corner of the building had a median temperature 1.3 8C higher than office B (mid north side), although office B experienced a wider range of temperatures. This was probably because the window of office B was open for 22% of the time, whilst the window of office C was never opened. Offices B and C both recorded temperatures outside the upper and lower limits of the comfort zone. The 75% percentiles for offices B and C were 26.5 and 27.2 8C, respectively. Whilst offices B and C are within the comfort zone for longer periods than office A, it is unlikely that people would complain about office A being too cold during the summer months, while it is highly probable that offices B and C would be experienced as uncomfortably warm. Thermal comfort during winter conditions was analysed over the 38-day period between 30 June and 4 August 2001 (Table 2). The average ambient temperature was 8.1 8C. For 75% of the time the temperature was below 11.1 8C i.e. 7 8C below the lower limit of the comfort zone. Unlike the summer conditions, the winter ambient temperature was always below comfort temperatures. Since the offices were not specifically designed for direct solar gain in the winter, thermal comfort was to be achieved by the hydronic heating system. The psychrometric analysis shown in Fig. 6 demonstrates the Fig. 5. Coincidence of measured psychrometric data with summer comfort effectiveness of this strategy. zone for offices A–C during occupied hours. Fig. 6 indicates that the heating system often failed to provide thermal comfort, particularly in office A, where Table 1 Operative temperatures in offices A–C in selected summer period during occupied hours Temperatures (8C) Ambient Operative office A Operative office B Operative office C Mean 23.1 23.7 24.7 25.8 Minimum 6.7 16.4 15.9 20.1 25% percentile 18.9 22.4 23.0 24.7 Median 22.7 24.1 25.0 26.3 75% percentile 27.3 25.4 26.5 27.2 Maximum 36.9 27.4 31.0 30.3 Comfort zone (%) n.a. 73 76 81
P. Taylor et al. / Energy and Buildings 40 (2008) 793–800 797 Table 3 Energy use and intensity for two CSU buildings in 2001 Building Energy use (GJ) Area Energy intensity (MJ/m2) (m2) Electricity Gas Total Electricity Gas Total Gordon Beaven 1548 567 2114 2186 708 259 967 Academic Offices 246 1024 1270 1500 164 683 847 Table 4 Greenhouse gas emissions two CSU buildings in 2001 Electricity Gas Total (CO2 ! e kg/m2) (CO2 ! e kg/m2) (CO2 ! e kg/m2) Gordon Beaven 207 13 220 Academic Offices 48 35 83 Table 5 Summary of climatic data during model validation periods Summer Winter Ambient temperature (8C) Minimum 14.8 !0.9 Average 25.4 6.8 Maximum 36.3 14.7 Horizontal radiation (W/m2) Maximum 1118 592 thermal comfort zone was achieved for only 13% of the time. Subsequent investigations indicated that the heating system was not working correctly in office A. The median temperatures in the top floor offices were 18.7 and 18.8 8C, approximately 1.8 8C higher than the ground floor office and only just within the comfort zone. A small electric heater was occasionally used in office C, while some of the cooler temperatures in office B are likely to have been due to an open window. Since the offices are naturally ventilated the only way to obtain fresh air in winter (apart from infiltration) was to open the window. Office A never reached 20 8C, while the remaining offices exceeded 19.7 8C for 25% of the time. Therefore these offices would probably have been perceived as ‘just comfortable’ to a well-clothed occupant, but would not have been considered warm. Some Fig. 6. Coincidence of measured psychrometric data with winter comfort zones anecdotal evidence indicates that other parts of the Academic for offices A–C during occupied hours. Table 2 Operative temperatures in offices A–C in selected winter period during occupied hours Temperatures (8C) Ambient Operative temperature Operative temperature Operative temperature temperature in office A in office B in office C Mean 8.1 16.9 18.6 19.0 Minimum !0.9 14.1 12.6 15.6 25% percentile 5.1 16.2 17.5 17.9 Median 8.2 17.0 18.8 18.7 75% percentile 11.1 17.6 19.7 19.7 Maximum 17.4 19.2 22.2 28.1a Comfort zone (%) n.a. 13 69 70 a It is believed that this high temperature was caused by late afternoon direct sunlight falling on the measuring equipment.
798 P. Taylor et al. / Energy and Buildings 40 (2008) 793–800 Table 6 General dimensional and operational parameters used in simulations Office Building were appreciably warmer than these offices 6. Discussion during the winter. The energy intensity used to achieve this level of comfort The average energy intensity for Australian offices is and the associated greenhouse gas emissions are shown in reported to be 981 " 62 MJ/m2 p.a. [11] and the Australian Tables 3 and 4, respectively. Building Codes Board (ABCB) design target for new buildings
P. Taylor et al. / Energy and Buildings 40 (2008) 793–800 799 in the Albury-Wodonga region is 640 MJ/m2 p.a. [12]. In terms of total energy use, the Gordon Beaven building is very close to the Australian average, whilst the Academic Office Building uses 13% less than the average. However, since the latter is a relatively new building with a ‘green’ reputation, a reasonable expectation would be that it would use less significantly less energy than the ABCB target. In reality, however, it exceeds this target by over 30%. It appears that the installation of the solar collection system in the Academic Offices has not achieved any reduction in heating energy use compared to a conventional system. The absence of an electrically-driven refrigerative cooling system has resulted in a building that uses only one fifth of the electrical energy of the more conventional building; however, it does not meet comfort standards all the time. Table 3 also shows that gas and electricity usage in the two Fig. 7. Measured and predicted temperatures in office C in summer. buildings is significantly different. The Gordon Beaven building consumes more than twice as much electricity as gas, whilst the Academic Offices Building use four times as when the two temperatures became equal. Since the ventilation much gas as electricity. As a result, total greenhouse gas rate could not be measured directly, the value used was chosen emissions for the former building are almost three times those to obtain a match between measured and simulated results. of the latter ‘‘deep green’’ building (Table 4). Office air temperatures, measured at 1.1 m above floor level, were used to validate the predictions of the summer and winter 7. Simulation models (Figs. 7 and 8). Agreement between measured and predicted values was considered adequate to use the models to The above analysis indicates that Academic Office Building investigate improvements to the CSU building. is not performing well. TRNSYS models were therefore Three strategies were investigated to reduce the office developed to investigate the effect of changes to the control temperatures in summer. The strategies were (a) increasing the system or to the building fabric on the energy use and thermal night ventilation rate; (b) adding insulation to the external walls comfort provided by the building. The first model described the and ceiling; and finally (c) using a hydronic cooling system. top-floor office C and explored a number of options to improve Their effect was not investigated separately; rather the thermal comfort during the summer. The second model was cumulative effect of the strategies was predicted. Progressive used to investigate the energy consumption in the ground floor predictions have been reported elsewhere [13]. The predicted office A in both summer and winter. Five days were selected in effect of all three strategies combined is shown in Fig. 9, both summer and in winter over which time to validate and run compared against the measured temperatures in Fig. 7, and the models. The period in summer had clear sunny days with the statistical impact on operative temperatures is given in Table 8. maximum temperature of 36.3 8C. The winter period was The operative temperature is now below the lower limit of the considered typical of that time of year. There were three sunny ASHRAE comfort zone for 75% of the time, although it would days with cold nights followed by two cloudy days. Table 5 probably be very acceptable by Australian standards. summarises the climatic data during the modelling periods. The validated model of office A was used to try to investigate The general office dimensional and operational parameters two strategies, namely wall insulation and a combination of initially used in the simulations are given in Table 6, while insulation, glazing and reduced infiltration, to reduce the material parameters used are given in Table 7. The variable vent can take values of 0 or 1. The latter indicates the use of night purging, while a zero indicates the contrary. Ventilation of the office occurred when the external air temperature fell 2 8C below the average internal building air temperature, and ceased Table 7 Material parameters Parameter Value Unit Internal film conductance 8.3 W/m2 K External film conductance 3.9 V + 5.62 W/m2 K Wall surface solar absorptance 0.6 n.a. Density of rammed earth 2050 kg/m 3 Rammed earth conductivity 1.0 W/m K Specific heat of rammed earth 600 J/kg K Carpet R-value 0.22 m2 K/W Fig. 8. Measured and predicted temperatures in office A in winter.
800 P. Taylor et al. / Energy and Buildings 40 (2008) 793–800 Board although the low electricity use does give low- greenhouse gas intensity. It is clear that the installed solar system did not perform as intended. Thermal comfort was not adequately achieved during occupied hours. In the summer evaluation period, two out of the three offices measured had operative temperatures outside the comfort zone for over 20% of the time. In the winter, the situation was worse with the two top floor offices providing thermal comfort for only 70% of the time and the ground floor office for only 13% of the time. The computer modelling has indicated the value of increasing the R-value of the external walls by external insulation. Model validation indicated that there was a relatively high infiltration Fig. 9. Predicted combined effect of increased night ventilation, additional insulation and hydronic cooling on air temperatures in office C. rate. This unwanted air exchange lead to cool internal temperatures in the winter. However, in the summer the required air exchange used for the night purging has been shown to be too Table 8 low to cool the building for daytime use. Wall conduction rate, Effect of strategies on office C air temperatures in summer infiltration and night purging are all examples of systems that Statistic Measured (8C) Predicted (8C) Operative (8C) allow, either intentionally or unintentionally, heat to be Minimum 23.0 18.5 18.8 transferred between the building and its surrounds. A key aspect 25% percentile 25.8 22.3 22.0 in the design of a ‘low energy’ building is the control of this type Median 26.7 23.2 22.9 of heat exchange so that a large input of heat is not required in the 75% percentile 27.7 23.8 23.4 winter and the building does not overheat in summer. The results Maximum 30.5 24.9 24.4 of the thermal modelling have shown how to manage the heat flows so that the building performs better. Table 9 Energy flows, peak loads and heating energy requirements for existing and modified east-facing external wall in winter References Variable Existing Insulated Insulated wall + wall wall insulated slab + [1] L. Johnson, The view from Australia—green limits in the land of plenty, low-E window + Architectural Design 71 (4) (2001) 52–59. reduced infiltration [2] Bureau of Meteorology, http://www.bom.gov.au/climate/averages/tables/ cw_082056.shtml viewed July 28, 2003. Heat from external wall (MJ) !121.4 !26.2 !31.2 [3] Webster-Mannison, M. Charles Sturt University, Thurgoona Campus, Peak heating load (kW) 3.0 2.5 1.0 New South Wales, Australia. Alternative Construction. Contemporary Total heating energy (MJ) 390.5 249.0 62.3 Natural Building Methods, in: L. Elizabeth, C.Adams (Eds.), Wiley and Sons, New Jersey, 2005, (Chapter 2). [4] American Society of Heating, Refrigerating and Air-conditioning Engi- heating energy and improve thermal comfort in winter. The neers Inc., ASHRAE 55—Thermal Environmental Conditions for Human hydronic heating in the ceiling was switched off for the study so Occupancy, Atlanta, GA (2004). that the model could predict the mechanical heating required. [5] G.J. Levermore, D. Lowe, J. Ure, Occupant feedback questionnaire Under these conditions, 390.5 MJ was needed to heat the office producing a fingerprint and a score, ASHRAE Transactions 105 (2) (1999) 661–670. for the entire 5 days, with a peak requirement of 3 kW (Table 9). [6] P. Taylor, Energy Use and Thermal Comfort in a Rammed Earth Office Almost one third of this energy (121.4 MJ) was lost through the Building, Thesis for M. Architecture, Deakin University, Geelong, Vic., external wall (indicated by a negative heat flow). Australia. External wall insulation alone reduced the heating require- [7] R. de Dear, G.S. Brager, Developing an adaptive model of thermal comfort ment in winter to 249 MJ, with the heat lost through the external and preferences, ASHRAE Transactions 104 (1a) (1998) 145–167. [8] A. Auliciems, S.V. Szokolay, Thermal Comfort, Design Tools and wall reduced by 78%. If other energy savings measures e.g. floor Techniques, Passive and Low Energy Architecture International, in insulation, low-E glazing, etc. had also been adopted, even greater Association with Department of Architecture, University of Queensland, energy savings could have been achieved (Table 9). Although the Brisbane, Qld, 1997. simulations indicate that the heat lost to the external wall would [9] P. Taylor, M.B. Luther, Evaluating rammed earth walls: a case study, Solar increase slightly because the office was warmer, the heating Energy 76 (1–3) (2004) 79–84. [10] Property Council of Australia, Method of Measurement for Lettable Area. demand has dropped to only one sixth of its original value. No Sydney, Australia, 1997. heating is required at night to maintain the required temperature, [11] P. Bannister, Australian Building Codes Board: Class 5 Benchmarking, as the office air temperature did not fall below 18 8C. Exergy Australia Pvt. Ltd., Belconnen, ACT, Australia, 2004. [12] Australian Building Codes Board, Building Code of Australia, vol. 1, 8. Conclusions Section J, Canberra, ACT, Australia, 2006. [13] P. Taylor, R.J. Fuller, M.B. Luther, in: Proceedings of the Solar 2004 Annual Conference of Australian and New Zealand Solar Energy Society The building has been shown to have larger energy intensity on Validated Model and Study of a Rammed Earth Wall Building, than the target specified by the Australian Building Codes Murdoch University, Perth, WA, December 1–3, 2004.
You can also read