1、2011 ASHRAE 945ABSTRACTBuilding performance models routinely involve tens orhundreds of components or aspects and at least as manyparameters to describe them. This results in overwhelmingcomplexity and a tedious process if the designer attempts toperform parametric analysis in an attempt to optimize
2、 thedesign. Traditionally, during design, parameters are selectedon a one-at-a-time basis and, occasionally, formal mathemat-ical optimization is applied. However, many subsets of param-eters show some level of interaction, to varying degrees,suggesting that the designer should consider manipulating
3、multiple design parameters simultaneously. This paper isdivided into two parts. The first part presents a methodologyfor identifying the critical parameters and two-way parameterinteractions. The second part uses these results to identify theappropriate level of modeling resolution. The methodology
4、isapplied to a generic model for net-zero or near-net-zero energyhouses, which will be used for an early stage design tool. Theresults show that performance is particularly sensitive to inter-nal gains, window sizes, and temperature setpoints, and theyindicate the points at which adding insulation t
5、o varioussurfaces has minimal impact on performance. The most signif-icant parameter interactions are those between major geomet-rical parameters and operating conditions. Increasedmodeling resolution for infiltration and building-integratedphotovoltaics (BIPV) only provides a modest improvement tos
6、impler models. However, explicit modeling of windows,rather than grouping them into an equivalent area, has a signif-icant impact on predicted performance. This suggests thatidentifying and implementing the appropriate level of model-ing resolution is necessary, and that it should be detailed forsom
7、e aspects even in the early stage design.INTRODUCTIONCurrently, net zero energy buildings (NZEBs) are beingcited as an effective solution to pending environmental issues(Griffith et al. 2007). NZEBs are frequently defined as build-ings that export as much energy as they import over the courseof a ye
8、ar, though other definitions exist (Torcellini et al. 2006).Two possible approaches to achieving strong NZEB designsare formal mathematical optimization and simulation-supported design, in which a designer is involved in everydecision. The focus of this paper is early stage design usingperformance s
9、imulation. Put simply, optimization tools outputthe optimal design based on an objective function and a set ofconstraints, offering little insight to what makes a good design.Design tools can provide the means to a designer to exploredifferent concepts and reach the near-optimal design spacewhile ac
10、counting for their experience and preferences. Unlikeformal optimization, design permits the evaluation of unquan-tifiable design traits such as aesthetics and views to theoutside. However, users of design tools are unlikely to arriveat the mathematically absolute optimal solution for largedesign sp
11、aces, though they may come close.To put the current work in context, a software-based solarhouse design tool, which is referenced throughout this paper,is being developed for the early stage design. It will support thedesign of low-energy and net-zero energy houses that includepassive solar, active
12、solar, and energy efficiency features. Thesolar house design tool will replicate what only the mostpatient of designers would do naturally: support conceptgeneration with a series of proper calculations and simulationsthrough many different design options. The tool shouldmanage issues such as approp
13、riate parameter interactions,Parametric Analysis to Support the Integrated Design and Performance Modeling of Net Zero Energy HousesWilliam T. OBrien Andreas K. Athienitis, PhD, PEng Ted Kesik, PhD, PEngStudent Member ASHRAE Member ASHRAE Member ASHRAEWilliam OBrien is a PhD student and Andreas K. A
14、thienitis is a professor and research chair (Tier I in Solar Energy) in the Department ofBuilding, Civil, and Environmental Engineering at Concordia University, Montreal, Quebec, Canada. Ted Kesik is an associate professor atthe John H. Daniels Faculty of Architecture, Landscape, and Design, Univers
15、ity of Toronto, Toronto, Ontario, Canada.LV-11-0282011. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions, Volume 117, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either prin
16、t or digital form is not permitted without ASHRAES prior written permission.946 ASHRAE Transactionsdesign resolution, and modeling assumptions to ensure goodresults from inexperienced energy modelers.Currently, common design practice of solar homesinvolves the expertise of multiple practitioners and
17、 at least asmany building energy simulation programs, some of whichmay be custom-built or modified. However, the savings poten-tial (energy and cost) for small residential buildings on an indi-vidual basis does not justify this type of investment formainstream deployment. Thus, there is a need for a
18、 stream-lined procedure that reduces the level of expertise, designtime, and number of distinct information sources (CADprograms, textbooks, design guides, etc.). The absence of sucha tool has hindered the widespread adoption of systematicpassive solar design in residential buildings. This gap wasid
19、entified by Athienitis et al. (2006) and in the design ofseveral of the Canada Mortgage and Housing CorporationsEQuilibrium demonstration homes (CMHC 2009). It shouldbe noted that the use of thorough design and simulation isparticularly justified for prefabricated homes or subdivisionsin which a lar
20、ge number of houses is similar or identical.The biggest challenge in designing a low or net zeroenergy house, as with any engineering system, is that thedesigner must make multiple design decisions simultaneouslywith the goal of achieving a high level of performance. Theprocess is not as simple as m
21、erely selecting the best choice formultiple subsystems and assuming that this will yield the bestsystem upon integration. In reality, each subsystem interactswith the others, to some degree. For example, the optimalsouth-facing glazing area for a passive solar house depends onmany other design decis
22、ions, including the level of thermalmass and insulation, as well as the control of solar gains andthe space heating strategy.A frequently cited technique for design is parametric orsensitivity analysis (Hayter et al. 2001). A major limitation tothis method is that parameter interactions are overlook
23、ed. Forinstance, the parametric analysis might be performed on aparameter, during which the other parameters are set to valuesthat do not allow it to be properly characterized. A moreconcrete example is the interaction between thermal mass andglazing area. A house with minimal glazing and relatively
24、constant operating conditions would indicate little benefit inperformance with the addition of thermal mass. Using a one-dimensional parametric analysis, an inexperienced designermight make the generalized conclusion that thermal mass hasminimal benefit for all buildings. Thus, a methodology forquan
25、tifying the significance of parameter interactions isneeded. The focus of this paper is on passive elements of thehouse, though active solar energy collection is a necessaryaspect of NZEBs. The paper includes the description of ageneric model intended for passive solar houses, a one-dimen-sional par
26、ametric analysis, a two-way interaction analysis,discussion about subsystem interactions, and finally, an in-depth analysis about modeling resolution for three modelaspects. MODEL DESCRIPTIONThe underlying model is a rectangular, four-zone house(represented in Figure 1). The purpose of having two ab
27、ove-grade occupied zones is to characterize the possibility of over-heating in the direct gain (south) zone. OBrien et al. (2010)showed that performance is relatively sensitive to zonalconfiguration for passive solar houses. The model wasselected to maximize design flexibility, even if some of thepa
28、rameters and their ranges differ from traditional rules ofthumb for passive solar buildings (though they are constrainedto adhere to the Model National Energy Code for Buildings(a) (b)Figure 1 (a) Isometric view of model and (b) east section view of model. Geometry-related parameters are marked on t
29、hedrawings.2011 ASHRAE 947MNECB NRCC 1997). In the context of the design tool,this design flexibility enables the user to understand allextremespoor and excellent designs.In all, the form, fabric, controls, and operations aredefined by 30 parameters, as listed in Table 1. These wereselected as the m
30、ost influential subset of a larger group of orig-inal parameters. Each parameter can be classified as design or non-design.The non-design parameters are defined as those that affect theservice that the building provides, namely, shelter, space, andprotection from the elements. Put differently, they
31、are likely tobe fixed at the beginning of the design process. The designparameters are defined as those that affect energy perfor-mance, but not the service to the occupants. Continuous parameters can be set to any value within thepermissible range (though they may not all be convenient withregards
32、to available building materials). Discrete parameterscan take on one of a finite set of values. For instance, the mostappropriate way to define different glazing types is to explic-itly model them, rather than have variable optical and thermalproperties, since some combinations of which would not be
33、possible (e.g., high transmittance and low U-factor). The parameters were selected to be designer-friendly. Forexample, instead of defining the major dimensions as lengthand width, the houses floor area and aspect ratio are used. Thereason for this is that the floor space is likely to be fixed (to s
34、uitthe needs of the occupants), while the aspect ratio may be flex-ible, depending on building lot constraints. The nominal settings, denoted in Table 1, were selected tobe representative of a good passive solar house design ratherthan average values within the range. Unless otherwise noted,the mode
35、l uses these values except for the parameters that arebeing explored.Table 1. List of Model Parameters and Corresponding Values No. Abr. Name Definition Min Max Nominal Unit Design? Discrete?1 IN Infiltration rate Air infiltration rate (constant) 0.025 0.075 0.05 ach 0 02 IG Internal gains Internal
36、(sensible) heat gains scheme1 3 2Class number10 03 HS Heating setpointMinimum temperature to which zones are controlled during the day (7 a.m. to 10 p.m.)16 (61)22 (72)22 (72)C (F) 0 04 HSNNighttime heating setpointMinimum temperature to which zones are controlled at night (10 p.m. to 7 a.m.)17 (63)
37、22 (72)18 (64)C (F) 0 05 CS Cooling setpointMaximum temperature to which zones are controlled during the cooling season22.5 (73)27 (81)26 (79)C (F) 0 06 FA Floor areaTotal conditioned floor area (including basement)100 (1076)300 (3228)200 (2152)m2(ft2) 0 07 ST StoriesNumber of stories excluding base
38、ment1 2 2 1 0 08 AR Aspect ratioWidth (oriented nearest to E-W) to length (oriented nearest to N-S) ratio (assumed rectangular form)0.5 2 1 1 1 09 OR OrientationHouse orientation; angle between Wall 1 normal and south ( CCW)45 45 0 degrees 1 010 WR Wall resistanceThermal resistance of all above-grad
39、e (opaque) walls from surface to surface4.4 (25)12 (68.1)6 (34.1)m2K/W (hftF/Btu)1 011 CR Ceiling resistanceThermal resistance of the ceiling from surface to surface8.8 (50)15 (85.2)10 (56.8)m2K/W (hftF/Btu)1 0948 ASHRAE Transactions12 BSBasement slab resistanceThermal resistance of all basement sla
40、b (or slab on grade) from surface to surface1.6 (9.1)3 (17)1.6 (9.1)m2K/W (hftF/Btu)1 013 BWBasement wall resistanceThermal resistance of all base-ment wall from surface to surface3.1 (17.6)6 (34.1)3.1 (17.6)m2K/W (hftF/Btu)1 014 WT1 Window type 1Type of window for south-most window(s)1 5 3Class num
41、ber21 115 WT2 Window type 2Type of window for east-most window(s)1 5 3Class number21 116 WT3 Window type 3Type of window for north-most window(s)1 5 3Class number21 117 WT4 Window type 4Type of window for west-most window(s)1 5 3Class number21 118 FT Frame typeFrame type for all windows on house1 3
42、2Class number31 119 WWR1Window-to-wall ratio 1Window-to-wall ratio for south-most window(s)0.05 0.8 0.4 1 1 020 WWR2Window-to-wall ratio 2Window-to-wall ratio for east-most window(s)0.05 0.5 0.1 1 1 021 WWR3Window-to-wall ratio 3Window-to-wall ratio for north-most window(s)0.05 0.5 0.1 1 1 022 WWR4W
43、indow-to-wall ratio 4Window-to-wall ratio for west-most window(s)0.05 0.5 0.1 1 1 023 CI Air circulation rateAir circulation rate between zones (assumed constant while on); turned on if T 3C0400 (847)200 (423)L/s (cfm)1 024 OH Overhang depthOverhang depth to window height ratio0.001 0.5 0.3 1 1 025
44、BLSShades close solar thresholdBlinds/shades are closed if both of these conditions are exceeded01000 (317)150 (47.5)W/m2(Btu/hft2)1 026 BLTShades close temperature threshold15 (59)40 (104)20 (68)C (F) 1 027 TMSThermal mass on south zone floorThickness of concrete on south zone floor0.001 (0.0030)0.
45、2 (0.61)0.1 (0.31)m (ft) 1 028 TMVThermal mass on vertical wallThickness of concrete on interior vertical surface0.001 (0.0030)0.2 (0.61)0.1 (0.31)m (ft) 1 029 RT Roof type Roof typology 1 2 1Class number41 1Table 1. List of Model Parameters and Corresponding Values (continued)No. Abr. Name Definiti
46、on Min Max Nominal Unit Design? Discrete?2011 ASHRAE 949The house was assumed to be occupied between 6 p.m.and 8 a.m. by a family of four. Fresh air was introduced to meetminimum requirements according to ASHRAE (2004), usinga heat recovery ventilator (HRV) with an effectiveness of 60%.Infiltration
47、(IN) in the nominal model was modeled at a fixedrate in air changes per hour (ach). The effect of a moreadvanced infiltration model is examined in a later section. Temperatures were controlled using ideal controls andequipment, in which heat was added or removed from thezones at a rate required to m
48、aintain the zone air temperaturewithin the temperature control setpoints. In this way, the heat-ing and cooling energy are determined. Heating was allowedyear-round, while cooling was allowed from May 1 toSeptember 30, representing the typical cooling season forToronto (the location for which the si
49、mulations wereperformed). The heating setpoint (HS) was defined duringtwo control periods per day: day (7 a.m. to 10 p.m.) and night(the remainder of the day). Windows were modeled by grouping all windows on eachsurface as a single window and with a frame that surroundsthem. Windows were assumed to be congruent to, andcentered within, their parent surfaces. The effect of explicitlymodeling individual windows (instead of grouping them) isexamined in a later section. Several strategies were implemented to minimize over-heating and improve thermal
