1、Authors Collins and Wright are professors in the Department of Mechanical and Mechatronics Engineering at the University of Waterloo, Waterloo, Ontario Canada. Author Huang is a Ph.D. candidate in the Department of Mechanical and Mechatronics Engineering at the University of Waterloo, Waterloo, Onta
2、rio Canada.An Examination of Keyes Universal Chart: 50 Years LaterMichael Collins, PhD Ned Huang John Wright, PhD Member ASHRAE Student Member ASHRAE Member ASHRAE ABSTRACT From the late 1940s to the late 1960s, significant efforts were made by ASHVE and then ASHRAE to evaluate and quantify the impa
3、ct of window shading. In the context of the now defunct Shading Coefficient, well known researchers such as Parmelee, Ozisik, Schutrum, Farber, Yellott, and Keyes laid the groundwork for much of the work that followed decades later. Of particular interest are the efforts of Keyes. In his work, he pr
4、oduced a method of classifying fabric based either on visible inspection, or on property measurements. The result was the Keyes Universal Chart, which was first published in the 1965 ASHRAE Guide and Data Book, and has been part of the Fenestration Chapter of the ASHRAE Handbook of Fundamentals sinc
5、e its inception. The chart compares fabric transmittance, reflectance, and openness. It also permits estimation of these properties by making generalized fabric classifications based on a subjective analysis of how light or dark the fabric is, and how open or closed the fabric weave is. More recentl
6、y, significant efforts have been made to produce window shading models for use in building simulation and daylighting analysis. As part of this research, shading materials have been analyzed using modern and highly accurate spectrophotometric equipment. Unfortunately, that data has revealed inaccura
7、cy in Keyes Universal Chart. The present work examines this inaccuracy. INTRODUCTION With increasing energy demand and dwindling energy supply, the attention given to designing and constructing energy efficient buildings is ever present. Despite all the good things they do, windows are a potential w
8、eak point in any energy efficiency strategy. Thermally, they provide less resistance than wall construction, which is a detriment both in a heating and cooling climate. From a solar heat gain perspective, they have the potential to either offset heating or drive up cooling demand. As buildings becom
9、e better insulated, and as one moves to a more cooling dominated climate, increased cooling demand becomes a serious concern. A window design that is able to transition from high to low solar heat gain would be a great asset. Simple shading devices can be used to make a window switchable. Called a C
10、omplex Fenestration System (CFS), it is well recognized how important these switchable window systems could be. Fittingly, since the mid 1990s, ASHRAE Technical Committee 4.05: Fenestration (TC4.05) and others have paid significant attention to quantifying the benefits of shading devices placed on w
11、indow. Several are worth mentioning. While not part of the TC4.05 efforts, one must include the work of Van Dyck and Konen (1982) whoproduced solar/optical models of shades and CFS for implementation into the WIS software. McCluney and Mills (1993) modeled solar/optical properties of shade materials
12、, and then used this todetermine window system solar/optical behavior. Klems (1994a, 1994b) developed the Matrix Layer Calculation. The method has great potential to accuratelyquantify CFSs both from energy and daylighting perspectives. The complexity of this approach is a problemas it relies on dif
13、ficult to obtain measurements (Klems and Warner, 1995) and is computationally intensive.Still the approach laid the groundwork for the efforts that followed (Klems 2001), and in particular,introduced the use of the Indoor Attenuation Coefficient (IAC). Collins and Wright have made significant stride
14、s towards not only producing accurate models of CFSperformance (Kotey et al. 2009a-d, 2011), but also producing and implementing a methodology that allowedfor these calculations to be included in building simulation software where computational speed is important(Wright and Kotey 2006, Collins and W
15、right 2006, Wright et al. 2008, Foroushani et al. 2015). This mostrecent work was largely supported by ASHRAE Research Project 1311 (Wright et al. 2009).Looking beyond energy considerations: Tzempelikos (Chan et al. 2015a-b) has been carefully studying the complex issue of balancing the combinedimpa
16、ct of windows on energy, daylighting, and comfort.Great strides have been made, and continue to be made in this area, and building designers now have the tools to quantify the benefits of CFSs of many forms. It would be incorrect, however, to assume that all CFS research has occurred over the past 2
17、0 years. In the period from the late 1940s to the late 1960s significant efforts were also made in this area. Parmelee (Parmelee and Aubele 1952, Parmelee et al. 1953) examined the effect of slat type sun shades onheat gain to the indoors using both mathematical analysis and solar calorimetry. Later
18、, Ozisik and Schutrumperformed similar measurements for roller shades (Ozisik and Schutrum 1959) and drapes (Ozisik andSchutrum 1960). Both of these studies were limited to single-glazed windows. Farber et al. (1963) performed a theoretical analysis of solar heat gain through double pane glazing uni
19、ts withboth Venetian blinds and draperies. A parallel experimental study was also carried out to validate thetheoretical treatment (Pennington et al. 1964). Yellott experimentally determined the solar performance of draperies using the ASHRAE solar calorimeter(Yellott 1965). He also measured the sol
20、ar optical properties of fabrics and glass-drape combinations usingcustom-made instruments. In that work, Yellott makes frequent reference to the work of Keyes, and togetherthey propose that fabric properties be rated based on yarn reflectance and fabric openness (the percent openarea between fibers
21、 in a fabric). This approach was dubbed the yarn reflectance-openness system. They also statethat visual estimation of fabric properties is accurate enough for this application. Moore and Pennington (1967) measured the solar optical properties of fabrics, draperies, and glass-draperycombinations usi
22、ng various techniques. They recommended that drapery classifications be designated byfabric solar optical properties using what they called the fabric reflectance-transmittance system, instead of the yarnreflectance-openness system proposed by Yellott and Keyes. They argued that openness needed to b
23、e properlydetermined, and that visual estimation may not be good enough depending on the fabric material, itsthickness, and other characteristics such as color, which may be misleading as to its reflective characteristics.THE DEVELOPMENT OF KEYES UNIVERSAL CHART Although Keyes chart was first publis
24、hed in the 1965 ASHRAE Guide and Data Book (ASHRAE 1965), and referred to in the work of Yellott (1965), Keyes work itself was not published by ASHRAE until 1967 (Keyes 1967). In that work, Keyes not only discussed the solar control abilities of drapes, but also their impact on other factors related
25、 to thermal comfort and daylighting concerns. Further, he reasoned through the usefulness of the yarn reflectance-openness system. He mentioned that if the fabric reflectance-transmittance system proposed by Moore and Pennington (1967) were the only one employed, one would a) have no fundamental und
26、erstanding of what is physically happening between the yarn and radiant input, b) move into complete dependence on instruments, and c) give up the ability to predict other performance characteristics of the drape fabric. He advocated, therefore, that both systems are needed; the fabric reflectance-t
27、ransmittance system for accurate prediction of shading effect, and the yarn reflectance-openness system for approximating shading effect, and for evaluating other fabric characteristics. To develop his Universal Chart, Keyes needed three pieces of information: the fabric reflectance, fabric transmis
28、sion, and openness. He was able to obtain this data for various fabric materials, colors and weaves from four sources: the Yellott Solar Energy Laboratory (Yellott), the University of Florida (Pennington), Pennsylvania State University (Pass), and from the Pittsburgh Plate Glass Company (Schutrum, S
29、tewart, and Keyes) (Keyes 1967). Keyes started by plotting fabric transmittance versus fabric reflectance. To place openness lines, Keyes would plot on this chart all data points within a range of the target openness. For example, all fabrics with openness between 0.15 and 0.25 were plotted. A curve
30、 fit to this data was set to be the 0.20 openness line (Figure 1). As openness and fabric transmission should be nearly the same at zero fabric reflectance, the line was anchored at that point. Note that, even with zero openness (i.e., one cannot see through fabric), a certain amount of radiation ca
31、n still penetrate the fabric by transmittance through “transparent” fibers or by multiple reflections among fibers. In other words, zero openness does not necessarily mean zero transmission. Figure 1 shows this effect as each line of constant openness curves up, indicating increased fabric transmiss
32、ion as fabric reflectance increases to the right. Next, the line connecting 100% reflectance at zero transmittance and 100% transmittance at zero reflectance is a limit, so the plot takes on a triangular shape. Following this, yarn reflectance was included. It was calculated as the fabric reflectanc
33、e divided by one minus the openness. The resulting yarn reflectance lines curve toward the theoretical boundary limit. The stronger the openness lines curve upwards, the more the yarn reflectance lines curve toward the boundary limit line. Note that, again, yarn reflectance cannot be measured, but t
34、he openness concept offers a way to estimate its value. As a final step, Keyes produced general fabric classifications on this chart. Fabrics were classified by weave as Open (I), Semi-open (II), and Closed (III), and by color as Dark (D), Medium (M), and Light (L). Table 1 summarizes the classifica
35、tion system. Figure 2 shows the Keyes Universal Chart along with lines of fabric reflectance and examples of fabric classifications. The process of developing the Keyes Universal Chart is fully described in Keyes (1967). Figure 1 The development of Keyes Universal Chart (Keyes 1967). Table 1. Fabric
36、 Classifications Outlined by Keyes (1967) Classification Region on Keyes Universal Chart ILOpenness 0.25, Yarn Refl. 0.50 IIL0.07 0.50 IIILOpenness 0.50 IMOpenness 0.25, 0.25 0.25, Yarn Refl. 0.25 IID0.07 Openness 0.25, Yarn Refl. 0.25 IIIDOpenness 0.07, Yarn Refl. 0.25 Figure 2 Keyes Universal Char
37、t and fabric classifications (ASHRAE 2013). This Keyes Universal Chart, has remained virtually unchanged for 50 years, and has been published in the Fenestration Chapter of every Handbook of Fundamentals. The only significant change was the replacement of shading coefficients with IACs in the 2001 H
38、andbook of Fundamentals ASHRAE 2001. Issues with the Universal Chart ASHRAE Research Project 1311 recently developed new solar optical and thermal models, and a new solution methodology by which window and shade combinations can be modeled in building simulation software (Wright et al. 2009). As par
39、t of that work, the link between the solar optical properties of shades and the solar optical properties of the materials from which they were made was required. In several cases, those models had already been developed. For example, drapery models for determining the solar optical properties of a d
40、rapery layer based of fabric properties had already been developed by Ozisik and Schutrum (1960), and by Yellott (1965). The accuracy and limitations of these models needed to be established, and in the case of drapery, this required the measurement of solar optical properties of several new fabrics
41、. In total, Kotey et al. (2009a) examined 9 fabrics, representing 8 of the 9 Keyes fabric classifications, and a sheer. A IIIDsample was not included. He did so by using a highly accurate UV/VIS/NIR spectrophotometer. First, the specular (beam-beam) transmission, or openness, was measured. Then, wit
42、h the help of an integrating sphere attachment, the total (beam-hemispherical or beam-total) reflectance and transmission were determined. Complete details of the measurement methods were documented in that work, and the results are reproduced in Table 2. While the validity of the older drapery mode
43、ls proved to be very good, the same could not be said for Keyes Universal Chart. For each classification shown in Table 2, one could plot three different points on the chart: OpennessReflectance, OpennessTransmittance, and Reflectance-Transmittance. These three points would form a right triangle on
44、the chart with the reflectance-transmittance point located at the right angle. Given the Keyes had to approximate the openness curve from several points representing a range of measured openness values, it was not expected that these 3 points would overlay one another. They should, however, be in cl
45、ose proximity: the triangles should be small. Furthermore, it was expected that in the absence of bias, that upward pointing and downward pointing triangles would be seen. This, however, was not the case. Not only are some of the triangles large, indicating chart inaccuracy, but all of the triangles
46、 point in the same direction, indicating bias. Also concerning is the fact that some points lie significantly beyond the physical limits of the chart. Table 2: Fabric Properties. Reproduced from Kotey et al. (2009a). Classification Openness (Beam-Beam Transmission) Fabric Reflectance(Beam-Total Refl
47、ectance) Fabric Transmittance(Beam-Total Transmittance)Sheer 0.45 0.19 0.80IL0.26 0.42 0.56IIL0.01 0.56 0.43IIIL0.01 0.68 0.30IM0.33 0.23 0.64IIM0.02 0.32 0.28IIIM0.01 0.38 0.20ID0.23 0.15 0.32IID0.05 0.21 0.23Figure 3 Universal Chart (ASHRAE 2013) including data from Kotey et al. (2009b). It is unl
48、ikely that one could definitively show the origin of this error. Keyes (1967) obtained data from four sources, but at no point provides a detailed listing of the numbers or types of samples used. Concerning the methods by which each measurement was obtained; he refers to Pennington et al. (1965) for
49、 determining total reflectance and transmittance, and describes a custom built apparatus consisting of a slide projector, collimating tube, and photocell for measuring the openness (Figure 4). A photocell reading is taken both with and without the sample in place, and the ratio of these readings is the openness. Reliance on the referenced data sources is also not helpful. In the paper by Ozisik and Schutrum (1960), 9 samples are described, but no mention is made of the openness. The tests are only briefly described as being done outdoors using a pyroheliometer. In Pe
