1、2008 ASHRAE 483ABSTRACTIn recent years, much attention has been given to the anal-ysis of windows with shading systems. If the solar and thermalperformance of these systems can be accurately predicted,than building designers would be better able to include shad-ing as an effective solar control devi
2、ce.In previous studies, a theoretical model that predicts thesolar performance of louvered shading layers has been devel-oped. The purpose of the present work is to validate thosemodels through the use of a Broad-Area Illumination Integrat-ing Sphere (BAI-IS). Such a device is capable of measuring t
3、hespectral directional-hemispheric properties of an inhomoge-neous material or system (such as a Venetian blind). The exper-iments showed that the zero thickness slat assumption causedsome error when the slat and profile angles are nearly aligned. INTRODUCTIONSolar radiation passing through windows
4、can have asignificant impact on cooling loads. Although a well designedHVAC system is designed to compensate for this, shadingdevices such as awnings, roof overhangs, curtains, and blindsmay be a cost effective strategy to control solar heat gain. Ofall shading devices, louvered blinds are likely th
5、e most popu-lar product due to their versatility in controlling daylight,reducing peak cooling, and protection of privacy.To optimally design energy efficient buildings and main-tain human comfort, architects and engineers need reliableinformation. Specifically, the thermal and solar properties ofgl
6、azing devices are required in order to calculate the heat lossand solar gain through a fenestration system. Unfortunately,building energy simulation programs are currently not capableof handling complex fenestration systems (i.e. shadedwindows) using anything more than a rudimentary model ofsystem p
7、erformance. Typically, two sets of properties are needed during theanalysis of energy performance in the center-glass region of aglazing system: the solar/optical and thermal properties. Thesolar/optical properties of the system are first used to calculatehow much shortwave solar radiation is transm
8、itted, absorbedand reflected through each glazing layer. Then, the thermalproperties (emissivities, convective heat transfer coefficients,and conductivities) can be used to perform an energy balanceat each glazing layer. This process is referred to as solar-ther-mal separation.When considering a sha
9、ding layer, there is some difficultyin performing the aforementioned analysis. Glass layers arespecular in nature, and it is therefore easy to track the directionof energy propagating through the system. When solar radia-tion strikes a blind, however, it is likely absorbed morestrongly than in glass
10、, and transmitted and reflected diffusely.Glass is also opaque to thermal radiation whereas a shade layeris diathermanous (transmits both short and longwave radia-tion). Neither the Solar Heat Gain Coefficient (SHGC) or ther-mal transmission (U-factor) of a window can be calculated bytraditional met
11、hods when a shade layer is included as part ofthe system 1.The present work focuses on aspects of the solar/opticalanalysis. In recent years, much work has been performed toadapt the current window analysis methodology to includeshading layers. Yahoda 2 devised a model capable of deter-mining the ef
12、fective solar-optical properties of the shadinglayer, that allowed for the input of various profile angles andthe material properties of each slat surface. The effective solar-optical property models were derived from a fundamentalValidation of Solar/Optical Models for Louvered Shades Using a Broad
13、AreaIllumination Integrating SphereMichael R. Collins, PhD Tao Jiang, PEngAssociate Member ASHRAEMichael R. Collins is an associate professor in the Department of Mechanical and Mechatronics Engineering at the University of Waterloo,Waterloo, Canada. Tao Jiang is an engineer with Stantec, Vancouver,
14、 Canada.NY-08-0582008, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions, Volume 114, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitte
15、d without ASHRAEs prior written permission.484 ASHRAE Transactionsanalysis of geometry (profile angle, slat width, w, slat space,s, slat angle, and slat curvature, k) and the optical characteris-tics of the slat (beam-to-beam reflection, bb, and beam todiffuse reflection, bd). In comparison to simil
16、ar models byothers 3,4,5, Yahodas model accounts for multiple specularreflections, oblique irradiation, and specular/diffuse materialcharacteristics.When the effective solar/optical properties of a layer areknown, the window system properties can be determinedusing a method devised by Kotey et al. 6
17、. In that work, thesystem is analyzed with respect to specular sources, and beam-to-diffuse sources are identified. Then the system is analyzedfrom a diffuse perspective. Further details of both Yahodasand Koteys models can be found in 2 and 6 respectively.The specific purpose of this work is to val
18、idate the solar/optical models of Yahoda 2, and to suggest possibleimprovements to his calculation method.METHODOLOGYIntegrating spheres have been used for many years tomeasure the optical properties of materials. An integratingsphere is simply hollow sphere whose entire inner surface isuniformly co
19、ated with a layer of material that has a high anduniformly diffusing reflectance, called a Lambertian surface.Lambertian surfaces are ideal surfaces where each unit areareflects light into all available solid angles with equal effi-ciency. Thus, light entering an integrating sphere is distributedeve
20、nly following only a few reflections, resulting in a uniformfield of light within the sphere. A discussion of the principlesof integrating sphere theory is provided by Labsphere 7.When used as part of a spectrophotometer system, an integrat-ing sphere is capable of determining the total and diffuser
21、eflection and transmission of homogenous samples, spec-trally.A louvered shade is a thick and light-scattering object,and it is therefore difficult to find its transmissivity using atypical spectrophotometer with integrating sphere attachment.Incident light may be scattered inadvertently into, or ou
22、t of,the integrating sphere due to the non-homogenous nature ofthe sample. More importantly, the size of the illuminationsource is smaller than the scale of sample inhomigenaity, andtherefore, sample placement becomes an issue. For non-homogenous sample measurements, the broadarea irradiation techni
23、que has been developed 8. Using thismethod, a large diameter (broad area) illumination source irra-diates a sample section that is greater than the integratingsphere aperture. Ideally, in-scattered light (that enters the aper-ture from outside the measurement area of the sample) andout-scattered lig
24、ht (that doesnt enter the aperture, but haspassed through the measurement area of the sample) will beequivalent. The device is capable of measuring the directional-hemispherical spectral transmittance and reflectance of asample. The integrating sphere spectrophotometer, described byMilburn and Holla
25、nds 8, was implemented for the currentexperiments (Figure 1). There are four subsystems in thisapparatus: the radiant source, the sample traversing mount, thedetector system, and the control system.The radiant source was designed to provide a strong andstable level of irradiance over the entire spec
26、tral measurementrange of the detection system. A 1000 W (3415.2 BTU/hr)Quartz Tungsten Halogen Lamp was used as the basic radiantsource, and a primary ellipsoidal reflective concentrator and asecondary ellipsoidal concentrator were used in order to getthe maximum output irradiation. The lamp output
27、was stabi-lized by an optical feedback control system. Irradiance unifor-mity was accomplished by using a kaleidoscope incombination with a Fresnel lens (16x magnification). An irra-diance level of 800 W/m2(140.9 Btu/hr ft2 oF) over a 50 cm (20in) by 50 cm (20 in) area was achieved at the plane of s
28、phereentrance port.The sample traversing mount is an automated sampleholder which moves the sample to a precise location for bothsample and reference measurements. The system allows thesample to transit both in the horizontal and vertical directionsto within 1 mm placement accuracy. To measure at di
29、fferentincident angles, the sample mount, integrating sphere, and thedetector systems are built on a platform with four castors thatallow the platform to rotate around a vertical axis that passesthrough the sphere aperture. A protractor embedded in thefloor allows for selection of the incident angle
30、 to within 1o.The detector system is built around a 50 cm (20 in) diam-eter integrating sphere, with 5 cm (2 in) diameter knife-edgeaperture. The knife-edge aperture was used to prevent theaperture walls from interfering with the incident light duringoff-normal experiments. A Littrow-style quartz pr
31、ism mono-chromator was used to take the measurement. After passingthrough the entrance slits to the prism, light is refracted andreflected into a complete spectrum. Mirrors are used to reflectthe light to the output slits, where the spectrum is narrowed tonear-monochromatic light. Only the wavelengt
32、h of interest ismeasured by detectors. Two types of detectors were used: asilicon photovoltaic detector (SiPV) and the lead sulphidedetector (PbS). The two detectors are automatically changedat 1100 nm wavelength (SiPV for shortwave, PbS for long-wave). The sensor signal was linked with a reference
33、signalfrom a chopper wheel running at 160 Hz to a lock-in amplifierwhich sends a measurement signal to the data acquisitionsystem.The system was entirely computer controlled through acustom-written control software. Every function from sampleposition, detector type, monochromator wavelength and radi
34、-ant source on/off was commanded through the software inter-face. Experimental results from data or parameter file accesswas also recorded and processed by the software.The integrating sphere was calibrated via a number ofmethods. ASHRAE Transactions 485 The light source was aligned by placing a mir
35、ror in thesample mount. The light source placement was adjusteduntil the beam came back to the diode.Holmium Oxide is widely used as a reference materialin spectrophotometry for wavelength calibrationbecause it has a number of well defined absorptionpeaks. A smooth and homogeneous Holmium Oxideglass
36、 reference sample was measured 5 times to cali-brate spectral transmittance measurements obtainedfrom the BAI-IS. Measurements between wavelengthsof 0.4 to 2.5 m were made at normal incidence.The absolute calibration was accounted for via the testprocedure.Test MethodThe procedure for measuring the
37、spectral total transmit-tance measurement of the louvered blinds samples is asfollows:1. Rotate the whole platform to a desired angle (profileangle) using the protractor embedded in the floor as refer-ence, and secure the platform. 2. Turn the power on for the equipments at least 10 minutesprior to
38、use. Make sure the radiant source and chopper areoperating in a stable condition.3. Install no sample the traversing mount. 4. Setup the scan parameters. Care must be taken whensetting the sample X position especially in the case of adeep profile angle. Make sure that only shading from thesample cas
39、ts shadow on the entrance port of the sphere (asopposed to its framework).5. Run the test and record the experimental data as the 100%baseline measurement (100).6. Install an opaque sample on the traversing mount. 7. Run the test and record the experimental data as the zerobaseline measurement (Z).8
40、. Install the sample on the traversing mount.9. Run the test and record the experimental data as samplemeasurement (T). The spectral transmittance measurement () of a sampleis calculated using(1)Total Solar Property DeterminationTo evaluate total solar properties (as opposed to the spec-tral propert
41、ies), the 50-point selected ordinate method, wasapplied 9. By this method, the irradiance distribution overthe total solar spectrum is divided into n wavelength intervals,where each of them holds 1/nthof the total irradiance energyof the reference spectrum. An air mass 1.5 solar spectral distri-buti
42、on covering from 0.305 to 4.0 m was adopted 10. Thespectral transmittance of the sample blinds is evaluated at thecentroid iof each interval, and then the total solar transmit-tance is calculated form Equation 2 9.(2)where sis the total solar transmittance, and n is the numberof selected points. The
43、 range of the wavelengths investigated in this studywas from 0.4 to 2.2 m due to equipment limitations. Themissing wavelengths included in the reference spectrum 9only contribute 3% of the total energy in that spectrum. Trans-mittances at wavelengths above 2.2 m were evaluated as thesame value at 2.
44、2 m, those below 0.4 m were evaluated asthe same value at 0.4 m.Test SpecimensSlats from a commercially available mini-blind wereemployed for this experiment. The slats were 15 mm wide (0.6Figure 1 The integrating sphere spectrophotometerdescribed by Milburn and Hollands 8.T Z100 Z-=s1 n i()i 1=n=48
45、6 ASHRAE Transactionsin), 0.17 mm (0.007 in) thick, and white in color. If consideredto be a perfect arc, the slats had a radius of curvature of about33 mm (1.3 in) and included angle of 26o. Half-inch slats werechosen because of the size of the slat in relation to the inte-grating sphere aperture.
46、The more slats in the aperture area, themore homogenous the sample becomes. To cover a broaderrange of cases, a set of flat black slats were produced by spray-ing black paint over the original white slats. The black slatsrepresent an extreme condition in nature, and it was expectedto test the capabi
47、lities of the numerical model. A dual beam Cary 5000 UV/VIS/NIR spectrophotometerwith an integrating sphere attachment was employed to obtainthe spectral properties of the blind slat material a wavelengthrange extending from the ultraviolet (0.250 m) to the nearinfrared (2.500 m) at normal incidence
48、. The total specularand diffuse solar properties of the slats were found to be s=0.3and d=12.3 for the black slats; and s=1.9 and d=65.4 for thewhite slats. Test specimen holders were constructed that accuratelyoriented the slats (Figure 2). The test specimen holders werea set of rigid rectangular f
49、rames constructed from Acrylic. Theholders were all 202.7 mm (8 in) height x 228.1 mm (9 in) widex 18 mm (0.7 in) long. This size was chosen so that during off-normal experiments, when the projected area of the sampleperpendicular to the light source was small, the light would notstrike the support frame. On each side of the frame, 16 slots(2.54 mm deep x 1.2 mm wide x 15 mm long) (0.1 in x 0.05in x 0.06 in) were accurately carved by an engraving machine.These slots exactly held the slats at a sp
