EVALUATION OF VENTILATION STRATEGIES TO REDUCE OVERHEATING IN A TYPICAL METAL CLAD BUILDING WITH IN- PLANE ROOF-LIGHTS

Transcription

1 , Volume 8, Number 2, p.37-46, 2011 EVALUATION OF VENTILATION STRATEGIES TO REDUCE OVERHEATING IN A TYPICAL METAL CLAD BUILDING WITH IN- PLANE ROOF-LIGHTS C. Kendrick, X. Wang, N. Walliman and R. Ogden Oxford Institute for Sustainable Development - Technology Group Oxford Brookes University, Headington Campus, Gipsy Lane, Oxford OX3 0BP, England (Received 6 December 2010; Accepted 6 October 2011) ABSTRACT A typical modern portal frame warehouse building with in-plane GRP rooflights was modelled using Tas and Lumen Designer software to assess annual heating loads and summertime thermal comfort in the south of the UK. The effects of rooflight area, ventilation strategy and stratification were assessed. Various combinations of ventilation strategy to reduce internal temperatures were investigated both natural and mechanical. The base case building incorporated 10% rooflights; the effects of 14%, or no rooflights were investigated. Overheating occurred for the unventilated base case but introducing natural ventilation avoided overheating. The areas of rooflight were not a significant factor. Mechanical ventilation was less effective than open cargo doors in conjunction with ridge vents. Buildings with higher internal heat loads risk overheating unless alternative precautions are taken. 1. INTRODUCTION This paper provides and assessment of the exent that overheating is likely to occur in current metal clad portal frame buildings when in-plane rooflights are used. In particular, the temperatures at different levels due to stratification within the occupied space are explored. The effect of various mitigation strategies using natural/mechanical ventilation at high and low level are investigated. Efficient mitigation of global warming requires energy efficient design to reduce CO 2 emissions from combustion processes, proved now to the greatest contribution to global warming. At 46% of the total emissions, buildings are by far the biggest cause of carbon dioxide emissions in the UK, and probably in other densely developed countries as well. A general characteristic of all active buildings is that energy is required to run their heating and cooling systems. For many years it has been predicted that the frequency of overheating in buildings is likely to increase due to the expected effects of global warming [1-3]. Day-lit buildings especially have the potential to be subject to overheating due to the ingress of solar radiation through the windows [4,5]. Therefore, improved passive design techniques will be required to reduce or even eliminate the need for mechanical cooling. Large retail outlets, warehouses and other industrial buildings are found on the outskirts of most towns and cities [6]. They are usually constructed as portal frame sheds with in-plane rooflights (i.e. rooflights that follow the slope of the roof) providing daylight to the interior. Compared to vertical windows in the external walls, in-line rooflights offer a number of advantages: They face the brightest part of the overcast sky which is near the zenith. They are less likely to be obstructed by trees or other buildings. A uniform light distribution is achieved more easily over large areas of floorspace remote from external walls. Their main drawback is higher heat losses in winter and unwanted heat gains in summer. The problem with summer heat gains is due to the nearly horizontal position of the rooflights that admit large amounts of sunlight, particularly when the sun is high in the sky during the summer. This frequently leads to overheating of the internal space and significant discomfort for the staff and clients using the building during the hottest periods. This can lead to the need for mechanical cooling installations with the result of increased power consumption and emissions of carbon dioxide. Currently mechanical cooling is used in supermarkets, salesrooms such as for cars, some retail sheds and warehouses storing heat sensitive goods, and likely to increase in use. However, it is government policy to reduce the carbon emissions from buildings. There is a common assumption that rooflight area is the most important influence on the level of solar overheating that smaller areas 37

2 reduce the amount of overheating. The research described in this paper argues that the amount and type of ventilation has a greater influence on reducing overheating than the reduction of the area of rooflights, and can obviate the need for mechanical cooling. CIBSE provides a guide to the standard levels of lighting (illuminances) required within buildings, depending on the tasks and the level of their visual demands [7]. Independent research carried out by the Institute of Energy and Sustainable Development, De Montfort University [8] predicted the levels of solar overheating which will occur inside typical large span buildings. The Part L of the Building Regulations for England and Wales [9] are met if the overall internal gain does not exceed 40 W/m 2 ; they assume an internal gain of 15 W/m 2 ; thus allowing a maximum solar load of 25 W/m 2 ; but the research demonstrates this assumption does not apply to many large span buildings, depending on the building use. 2. THE FORM OF THE STUDY RESEARCH METHODOLOGY In order to carry out the simulation, a typical metal clad portal frame building with in-plane GRP rooflights was designed in line with developments found on the outskirts of most towns in the UK and used as retail outlets, warehouses, workshops and many other functions. The building was drawn up in Autocad (a computer aided drawing program), and the amount of daylighting was modelled using Lumen Designer (an internal daylighting simulation program), and EDSL TAS version (a dynamic thermal simulation program) was then used to assess annual heating loads and summertime thermal comfort in the south of the UK. The effects of rooflight area, ventilation strategy and stratification were assessed. Tas is an industry-standard modelling program that takes into account thermal performance of the building envelope including the windows, doors, rooflights etc. as well as the effect of air leakage and natural ventilation due to controlled ventilation openings, as well as heat gain due to light fittings, people and equipment. Lumen Designer is a modelling tool that calculates daylight factors (as a percentage of a standard natural light level outside) distributed in the form of a grid at, in this case, floor level, depending on the different rooflight areas and distribution. Included in the calculations are the direct and reflected light from the walls and ceilings. Various combinations of ventilation strategy to reduce internal temperatures in the summer were investigated, including natural ventilation through cargo doors and/or ridge vents, and mechanical ventilation. The effect of ventilation was found to be critical, and had by far the greatest effect on reducing the overheating risk. 3. THE MODEL: WEATHER, GEOMETRY AND INTERNAL CONDITIONS 3.1 Weather/Location The building is located in Southern England, so that CIBSE Design Summer Year (London DSY) weather file is used for thermal modelling. 3.2 Geometry The building has two bays, each with a 6 duopitch roof, and is 60 m in length, 40 m in width and 7 m height (floor to eave). There is an office area in one corner of the building with a floor area of 120 m 2 (6 m x 20 m) with two storeys (floor height is 3 m each). Two cases are modelled: 10% and 14% nominal rooflight area. A frame (overlap with opaque cladding) of 12.75% area was assumed. Including frames, the total rooflight area for 10% case is 227 m 2, for 14% case is 319 m 2. There are no rooflights on the roof of the office area. External surfaces are all goosewing (mid) grey, see Fig. 1. Fig. 1: 3-D model of the portal frame building 3.3 Construction The warehouse is of portal frame construction and is clad with profiled metal sheeting insulated as follows: Roof cladding: U-value = 0.25 W/m 2 K. External wall cladding: U-value = 0.35 W/m 2 K. Ground floor (concrete floor): U-value = 0.25 W/m 2 K. Internal wall (office use): U-value = 0.83 W/m 2 K. 38

3 The rooflights consist of two layers of GRP with an intermediate insulating layer giving a U-value of 1.3 W/m 2 K. Optical properties of rooflight samples have been determined by an independent laboratory and combined by methods set out in BS EN ISO 410: 1998 [10] to give properties for the rooflight construction: a visible light transmission of 0.47 and a solar gain factor (g-value) of 0.47 were calculated for the rooflight construction used. 3.4 Internal Conditions Twenty four hour, seven days per week occupation was assumed, with heating to 19 C (no heat mid- June to mid-september). There is no air conditioning. Air infiltration was calculated assuming an air permeability of 7.5 m 3 /h per square metre total envelope area at 50 Pa pressure differential. 3.5 Infiltration Rate Air permeability is the flow rate (Q) at a pressure of 50 Pa divided by the total envelope surface area including the floor. An air permeability of 7.5 m 3 /h/m 2 measured at 50Pa was assumed (10 m 3 /h/m 2 is currently the maximum allowable in Building Regulations Part L). A standard empirical formula was used to calculate air infiltration under normal conditions [11]: 1 20 S Q V S 50 I Inner Volume V = m 3 ; Surface area S (excluding floor area) = m 2 Thus air infiltration I = 0.08 air changes per hour (ach). I = 0.1 is used. 3.6 People 20 people, with heat outputs from CIBSE Guide A as shown in Table 1 [7]. For twenty people, apparent gain = 2600 W, latent gain =2000W. Portal frame building floor area = 2280 m 2. An apparent (sensible) heat gain of 1.2 W/m 2 and latent gain of 0.9 W/m 2 is used. 3.7 Lighting control Auto control/dimmed; Photocell control dimming; Maximum light gain: 5.6 W/m 2 ; Target Room Illuminance: 350 lux. 4. DAYLIGHT MODELLING Lumen Designer was used to investigate the daylighting (illuminance distribution) on the floor. Fig. 2 and Table 2 show the model outputs of the illuminance levels and daylight factor distribution and average, demonstrating the fairly even distribution of lighting over the floor of the building. Table 1: Heat gains from people (CIBSE Guide A [7]) People 20 C 22 C Loads type (in W) Apparent Latent Apparent Latent Light bench work (factory) Medium bench work Fig. 2: Left - Interior rendering by illuminance levels; Right -Daylight factor distribution on the floor (excluding the office at right bottom corner) 39

4 Table 2: Daylight factor for different rooflight percentages Nominal rooflight percentage Average daylight factor (%) 14% % 3.41 There was little difference of average daylight factor across the whole building area (excluding offices) due to different orientations of the building because of the very shallow slope of the roof. Therefore, thermal modelling was based on the long axis running east-west. 5. THERMAL MODELLING 5.1 Method A dynamic thermal model was created using TAS software, integrating daylight analysis results from Lumen Designer. A daylight availability analysis enabled the extent of artificial lighting to be controlled automatically to maintain 350 lux. With realistic levels of interior heat gains from people, lights and equipment, and with solar radiation accounted for, the temperatures can be determined throughout the year at ground and mezzanine floor level. Overheating is investigated by graphs showing percentage of temperature excess hours. Both mechanical ventilation and natural ventilation are investigated in this study. 5.2 Natural Ventilation Natural ventilation includes cargo doors and ridge vents and the combination of these two. The temperature for opening cargo doors is set to 20ºC. The dimensions of natural ventilation openings are summarised in Table 3. It must be acknowledged that closed cargo and personnel doors are important aspects in the security measures for this type of building, so simply opening the doors when temperatures rise may not be acceptable. Grid type doors would allow ventilation to pass through without compromising security. The area calculations in Table 3 were based on 3 cargo doors with a clear opening of 4.20 x 3.91 m and four personnel doors with a clear opening of 2.25 x 1.20 m. The introduction of grids would reduce the clear opening area to an extent depending on the pattern of the grid. The overall sizes of the doors would then have to be increased in order to account for this. The proportion of door opening area in relation to the wall area is given to provide a guide for calculations for gridded doors. 5.3 Mechanical Ventilation Mechanical ventilation is set at 2.5 ach, supplying outside air to the ground floor zone and exhausting from the roof zone. 6. THERMAL MODELLING RESULTS 6.1 Summary of Loads and CO 2 Emissions The heat gain loads for the two cases are summarised in Table 4 and the monthly heating loads for 14% rooflights are shown in Fig. 3. Table 3: Natural ventilation openings Openings Area (m 2 ) Percentage of wall/floor area Cargo doors % (of wall) Personnel doors % (of wall) Ridge vents % (of floor) Table 4: Loads summary (kwh/m 2 per year): 10% (14%) rooflights Heating Solar People/ Lighting Equipment Base case ( (48.09) (10.18) (28.14) Base case without rooflights (16.85) 0.00 (0.00) (10.18) (49.06) Mechanical ventilation (22.79) (48.09) (10.18) (28.14) Natural ventilation (by cargo doors) ( (48.09) (10.18) (28.14) Natural ventilation (by ridge openings) (23.12) (48.09) (10.18) (28.14) Natural ventilation (cargo doors + ridge openings) Natural ventilation (cargo doors + ridge NIGHT) (23.32) (48.09) (10.18) (28.14) (23.33) (48.09) (10.18) (28.14) 40

5 Fig. 3: Monthly heating loads for different ventilation schemes: 14% rooflights It can be seen that the ventilation scheme has little effect upon heating loads, and any variations can be ascribed to lack of optimisation of heating control (e.g. some heating may be applied at the same time as ventilation). 6.2 Overheating (Ground Floor) Figures 4 and 5 show ground floor overheating and the percentage of occupied hours over certain temperatures with the building with 14% rooflight area. The base case shows significant risk of overheating in the base case when there is no ventilation. Elimination of rooflights does reduce the risk of overheating, although this would result in a significant increase in energy use, running costs and resultant CO 2 emissions and would have a serious effect upon the internal environment for occupants. Introduction of natural ventilation through open cargo doors has a much greater effect and ensures lower risk of overheating. This example building with 14% rooflights would not exceed 28 C for more than 1% of occupied. Ridge vents are less effective on their own, but would also substantially decrease overheating risk. However, ridge vents in conjunction with cargo doors allows buoyancy-induced flow and gives the best results. 6.3 Overheating (Mezzanine) The effect of stratification was investigated by analysing temperatures at a notional mezzanine level (Fig. 6). The model does show higher temperatures at mezzanine level, indicating some thermal stratification. However, thermal comfort remains close to an acceptable level for the cases when natural ventilation is allowed via open cargo doors, and particularly when combined with ridge ventilation. 6.4 Temperatures for Three Warmest Days Mechanical ventilation is most effective at reducing temperatures overnight. However, using natural ventilation (cargo doors and ridge vents) keeps daytime peak resultant temperatures as low as achieved by mechanical ventilation. The unventilated building (with or without rooflights) gets significantly hotter than any of the ventilated cases, demonstrating that ventilation strategy is critical, and is of greater significance than rooflight area (up to a point) in avoiding overheating. The example building with 10% rooflights, ventilated with cargo doors and ridge vents, does not exceed 25 C resultant temperature of any of the three warmest days. A building with 14% rooflights does not exceed a resultant temperature of 26 C on any of the three warmest days (see Fig. 7). 7. COMPARISON BETWEEN DIFFE- RENT ROOFLIGHT PERCENTAGES In order to simplify the comparative analyses between 10% and 14% rooflight cases, only the results from no ventilation at all, and from natural ventilation through open cargo doors, are compared. 41

6 Fig. 4: Ground floor overheating: 14% rooflights Fig. 5: Percentage of occupied hours over certain temperatures: 14% rooflights 42

7 Fig. 6: Percentage of occupied hours over certain temperatures, mezzanine: 14% rooflights Fig. 7: Resultant temperatures on ground floor for three warmest days: 14% rooflights 43

8 7.1 Overheating Figures 8 and 9 clearly show that use of natural ventilation has a very high impact, even when ventilation is only through cargo doors, whilst a change in rooflight area from 10% to 14% has a less dramatic effect upon overheating risk. The modelled building with 10-14% rooflights and effective natural ventilation has a low risk of overheating to uncomfortable levels. The model shows marginally higher temperatures at the mezzanine level, indicating some thermal stratification. 7.2 Temperatures for Three Warmest Days To best illustrate the findings about the impact of ventilation in the prevention of overheating in this type of roof-lit building, the conditions on the three warmest days in the year in London were modelled. The result is shown in Fig. 10. This graph shows that at the hottest point on the three warmest days of the year, unventilated buildings may overheat (reaching temperatures of over 30 C) but use of natural ventilation, if only through cargo doors, limits peak temperatures and significantly reduces overheating risk. Rooflight area (10-14% investigated) is a secondary effect for the example building studied, with an increase from 10% rooflights to 14% rooflights making less than 1 C difference to peak temperatures in the ventilated building. Fig. 8: Ground floor overheating Fig. 9: Percentage of occupied hours over certain temperatures at ground floor level 44

9 Resultant Temperatures (C) London DSY 10% Base case 14% Base case 10% with cargo doors open 14% with cargo doors open Ground Floor Hours (Days July 11-13) Fig. 10: Resultant temperatures on ground floor for three warmest days 8. CONCLUSIONS The analysis shows there is an overheating risk with buildings that are totally unventilated. Although both the ventilation strategy and rooflight area affect the risk of overheating, the ventilation strategy was found to have by far the most dominant effect. In a large open span metal clad building, cargo and personnel doors can be opened to provide effective cross ventilation as required. Opening cargo doors is common practice and is effectively self-regulating. Roof vents allow air to escape at high level and increase the effect of ventilation. They may also be left open at night to give secure night ventilation. It was found that the introduction of some natural ventilation (by opening cargo doors) has a significant effect, and ensures there is no overheating risk for buildings of this type incorporating 10% or 14% rooflights. The example building with either 10% or 14% rooflights with ridge vents and open cargo doors did not have total solar internal gains exceeding 35 W/m 2, and the internal temperature did not exceed 28 o C for more than 1% of occupied hours. Modelling also showed that resultant temperature is not predicted to exceed 26 o C on the three warmest days of the year. Investigation of stratification shows this has some effect, particularly at higher rooflight areas, but the modelling suggest this is not a very significant factor. The risk of overheating can be reduced by eliminating rooflights altogether, but this would result in significant increases in overall energy use for artificial lighting, and therefore running costs and CO 2 emissions, and would give a significantly poorer internal environment inside the building. The introduction of rooflights saves approximately 40% of artificial lighting energy needed, with only a small heating penalty, provided that appropriate controls are fitted to enable dimming in response to available daylight. Although this work demonstrates that there is no overheating problem at 14% rooflight area for the warehouse considered, if other internal gains are higher (for example from significantly increased occupation density, or retail display lighting), it is possible that overall heating conditions may be exacerbated leading to increased overheating. These results are of significant interest to designers of portal frame warehouse type buildings who are concerned to minimise the use of energy for cooling their buildings while maintaining sufficient levels of natural lighting from in-plane rooflights. The simulation model, based on a typical sized building, provides a reliable indication of energy saving design and operational options for this ubiquitous building type. 45

10 ACKNOWLEDGEMENT Brett Martin Daylight Systems funded the research on which this paper is based. REFERENCES 1. U. Amato, B. Coluzzi and V. Cuomo, et al., Effects of thermal control and of passive solar elements on the dynamic behaviour of a building, Applied Energy, Vol. 17, No. 4, pp (1984). 2. A. De Herde and A. Nihoul, Overheating and daylighting in commercial buildings, Renewable Energy, Vol. 5, No. 5-8, pp (1994). 3. A. Dorata and C. Chwieduk, Recommendation on modelling of solar energy incident on a building envelope, Renewable Energy, Vol. 34, No. 3, pp (2009). 4. H. Bodmann, K. Eberbach and P. Reuter, Rooflighting and sun protection, Energy and Buildings, Vol. 11, No. 1-3, pp (1988). 5. X. Wang, C. Kendrick and R. Ogden, Influences of lighting control and natural ventilation on energy use and overheating for a day-lit industrial building, Proceedings of Eleventh International IBPSA Conference, Glasgow, July 2009, pp (2009). 6. M. Geary, Retailers move out to the shed: Martin Geary looks at the trend toward volume selling from warehouse outlets, The Independent Sunday, 7 March (1993). 7. CIBSE, Guide A, 7th edition, London, CIBSE Publications (2006). 8. J. Mardaljevic and K. Lomas, Daylighting and Solar Analysis for Rooflights: Resolving the Perceived Conflict in Part L, Recommendations, Leicester, De Monfort University (2003). 9. Building Regulations Approved Documents for England and Wales, Part L, London, HMSO (2006). ng/buildingregulations 10. BS EN ISO 410: 1998: Joint ISO/CIE Standard: Spatial Distribution of Daylight - CIE Standard General Sky. 11. CIBSE, Testing Buildings for Air Leakage, TM23 (2000). 46