1、VOL. 5, NO. 3 HVAC one can associate comfort with social, financial, or thermal concepts. However, a hypothesis must be testable; i.e., it must provide some form of verifiable output, in response to some definable input. Consistency between given inputs and outcomes are essential. One or more hypoth
2、eses may be defined as a model, again with the requirement that a given input, or set of inputs, consistently produces a given response. ASHRAE thermal comfort concepts and ASHRAE Standard 55, Thermal Environmental Con-ditions for Human Occupancy, were developed using such models. These are regularl
3、y being modified/expanded.Three basic approaches are used to model thermal comfort: physical, physiological, and psy-chological. ASHRAEs latest research programs attempt to link its thermal comfort concepts to incorporate effects of noise, odor, and other elements in indoor environmental quality (IE
4、Q), and to link comfort with changes in such human behavior as productivity. This editorial reviews such linkages from my personal bias as a physiologist; these views differ from those of many others.ASHRAEs Psychologically Based Thermal Comfort ModelASHRAE and its predecessor societies original app
5、roach (used since the late 1920s) is a psychological model of comfort, based on the responses of subjects to various combinations of six factors. The four key environmental factors involved in thermal comfort include the two dominant factors in cooler conditions1) air temperature Taand (2) air motio
6、n WVand the two dominant factors in warmer conditions(3) humidity (or more precisely, the saturated vapor pressure of air Paat Tatimes the air relative humidity a) and (4) the mean radiant temper-ature (MRT) of the surrounding surfaces. The remaining key factors in thermal comfort are behavioral, na
7、mely, (5) the metabolic heat production M, standardized for almost all thermal comfort studies at 1 met = 50 kcal/m2h or 58.2 W/m2(or for a normal adult male with 1.8 m2 of body surface area M = 90 kcal/h or 105 W) by having the subjects seated at rest; and (6)the level of clothing insulation (clo)
8、(1 clo = 0.155 m2K/W), standardized at 0.6 clo or 0.093 m2K/W (i.e., a long sleeved shirt and pants, worn with light underwear and socks, but no shoes). The level of subjective thermal comfort was measured using the classic ASHRAE seven point Pre-dicted Mean Vote scale (PMV), where 1 = cold, 2 = coo
9、l; 3 = slightly cool, 4 = neutral or com-fortable, 5 = slightly warm, 6 = warm and 7 = hot). This scale, which was used in 1680 by Otto Guericke for a thermal instrument and paralleled the development of the thermometer, appears to be probably too fine for reliable human discrimination. Macintyre, b
10、ased on analysis of data shown in Figure 1, suggested the standard deviation for PMV is 1 full scale unit between expo-sures (i.e., a PMV of 4, has a 95% confidence range from 2 to 6) and is about 0.8 scale units both within and between subjects for hourly readings throughout a single, long exposure
11、.Studies of the data in Figure 1 (from over 3000 subjects in the ASHRAE climatic chamber in Cincinnati, Ohio and in this same chamber after its transfer to Kansas State University) and sup-plemented with studies of preferred temperature (i.e., the Taat which a subject, seated alone in the chamber, n
12、o longer requested a change in Ta) from O. Fangers group in Denmark have lead to the conclusion that the range of air temperature for thermal comfort for individuals of both sexes, world wide, is a 6F (3.3 K) pass band; this band ranges from a Taof 72 to 78F (22 to 1999. American Society of Heating,
13、 Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC if, and only if, WV= 40 ft/min (0.2 m/s), = 40% rh, MRT = Ta, clothing insulation value = 0.093 m2K/W (clo = 0.6), and M = 105 W (1 met).Discussions with other researchers in the field of thermal comfort (R. Nevi
14、ns and A.P. Gagge) and I during the 1970s led to the following suggested trade-offs between these six variables (those identified by * need further validation):1. Each 20 ft/min change in WV() can be offset by a 1F 0.2 m/(sK) change in the Tarange (in the opposite direction) when wearing 0.6 clo, to
15、 a maximum change of 5F (2.8 K). With less clothing, the change is 1F per 10 ft/min WVchange 0.1 m/(sK); for heavy clothing the change may be 50 ft/min per 1F 0.5 m/(sK).* 2. Each 20% change in relative humidity () can be offset by a 1F (0.5 K) change in the Tarange, in the opposite direction; the m
16、aximum change is 2F or +3F (1 K or +1.5 K).3. Each 1F (1 K) change in MRT can be offset by a corresponding change in the opposite direction; with full sun in the desert, the maximum change is about 13F (7 K) with light clothing.*4. Each change of 0.1 clo unit of insulation () can be offset by a 1F (
17、0.028 m2/W) change in the Tarange, in the opposite direction. A pair of shorts 0.2 clo (0.031 m2K/W); the heaviest arctic clothing 4 clo (0.62 m2K/W).5. Each 25 kcal/h change of M can be offset by a 3F Tachange (17.4 W/K) in the opposite direction.*The ASHRAE comfort model generally ignores any evap
18、oration by sweating; although it allows for the 6% rh of normal skin and the 12% of M lost to warming and humidifying the inspired air within the range of comfort Ta. Clothing characteristics other than the insulation value are also ignored; e.g., changes in moisture permeability Imor in intrinsic i
19、nsulation with pumping p by wind or body motion.A Physical Heat Balance Model For Thermal ComfortThe most common physical model for thermal comfort is based on the heat balance equation developed by Winslow, Herrington, and Gagge. It uses a two compartment body model (a core at 37C Treand a shell at
20、 33C Tskin) suggested by Burton as having 1/3 the body mass at Tsand 2/3 at Tre. This model starts with 76% of the metabolic heat production (using 0.76M accounts Figure 1. PMV as a function of TaFrom article by D.A. MacIntyre in 1978 ASHRAE Transactions 84(1).The size of each dot represents the rel
21、ative number of subjects voting a given warmth value (using the ASHRAE seven point scale) when seated at rest, (M = 1 met = 58.1 W/m2), wearing 0.6 clo (0.093 m2K/W), in a cham-ber at the given air temperature, with WV= 40 ft/min (0.2 m/s), = 40% and MRT = Ta. Note that for a given range of Tato be
22、valid, the other five key factors must be specified. 1999. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC i.e., adding clothing or increasing activity. If 0.76M Hr+c the body must deal with this potential heat storage s, i.e., the
23、remainder of 0.76M Hr+c = Ereq, which is the required sweat evaporative cooling. At Ts= 35C, skin begins to produce sweat, with sweat vapor pres-sure Psat this skin temperature = 5.6 kPa (42 mm Hg). Each gram of sweat evaporated from the skin surface removes 0.6 kcal (2.5 kJ) of heat from the body.
24、If ambient vapor pressure (aPa) is well below 5.6 kPa, as it is within the Tacomfort band even at 100% rh, with clothing of 0.6 clo and normal moisture permeability (Imabout 0.4), sweat evaporation can remove 700 W (600 kcal/h) from the body based on its sustainable 1 L/h sweat rate, and three times
25、 this for shorter periods.Again, to date, the physical model of heat balance used by some is limited to sedentary or very light work levels, assumes a comfortable maximum level of 20% skin humidity s(i.e., sPs= 0.20 5.6 kPa), and ignores any evaporative limits associated with clothing entirely.In su
26、mmary, the simplest, physical, heat balance model for thermal comfort:0.76M Hr+c Ereq = 0This equation excludes any s, clamps Tsat 33C and Treat 37C, limits sto 20%, and ignores clothing moisture permeability and wind or body motion induced changes in clothing insulation and permeability.Merging Psy
27、chological (PMV) and Physical (-E) ModelsFanger, A PMV advocate, developed an M + Hr+cmodel to fit PMV and Gagge, an M + Hr+c advocate, developed a PMV model to fit M + Hr+c . They negotiated a mutually acceptable ther-mal comfort model incorporating both approaches after several years of what some
28、members of ASHRAEs Physiology and Human Environments Technical Committee (TC 2.1) termed the “Comfort Wars.” This combined model, reformatted under TC 2.1 sponsored ASHRAE research project RP 781, is now available on CD-ROM as ASHRAEs Thermal Comfort Tool.A Physiologic Model for Thermal ComfortNot u
29、nexpectedly, the same six key factors used in psychological and physical models for thermal comfort also appear in the physiological models. However, some heat storage is allowed; so Tsmay be 33C and Tre 37C; skin humidity scan approach 100% as long as no liquid sweat appears at the skin surface (dr
30、ops appear as s 62% in non-desert environments); the ratio Im/clo, which expresses the percent of the maximum evaporative cooling power Emaxthat can occur from a fully ventilated, 100% wet surface at sPsin a given ambient aPa, is used as the upper limit of evaporative exchange rather than the 700 W
31、(600 kcal/h) limit evoked from maximum sustainable sweat production; a pumping coefficient p, which modify both clo and Im, may also be used. The effects of normal physiological temperature regulatory responses to changes in the four environmental factors (Ta, WV, aPa, and MRT) are also included in
32、the physiological models of thermal comfort. Vasoconstriction, which can induce rapid a fall in the temperature of such thin cylinders as fingers and toes because of their large surface area to mass ratios, is the primary cause of cool perceptions as finger/toe temperatures fall below 25C. This can
33、occur even with torso clothing insulation adequate to negate overall body heat loss. Shivering is uncomfortable per se, and incompatible with thermal comfort; although it can increase M (up to 500 W or 425 1999. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ash
34、rae.org). Published in HVAC (2) deliver increased oxygen to the work-ing muscles and remove CO2produced; and (3) transfer heat from the working muscles to the skin for potential loss by Hr+c and E. While an increase in heart rate of more than 30 beats per minute is undesirable, failure at Task (1) r
35、esults in heat exhaustion collapse; failure in Task (2) can cause muscle cramps and damage; and failure in Task (3) can lead to fatal heat stroke.Gagge and Stoljwik synthesized the M Hr+cwith a multicompartment body model that incorporated an extensive array of sophisticated physiological factors; h
36、owever, this still treated clothing primarily as insulation and was limited in considering combinations of M and clo that generated much sweat. A physiological model was developed with my group at the U.S. Army Research Institute of Environmental Medicine. I subsequently expanded it to incorporate t
37、he psychological PMV model, using the more sophisticated values from the physiological model for Hr+c , Ereq, and Emaxbased on clo, Im , and p.Rohles, in his work for the U.S. Air Force, considered more severe environments; he extended the PMV scale to a 9 point scale, with 4 = very cold and +4 = ve
38、ry hot (using Fangers rotation of the PMV axis around 0 = comfortable). Working for the U.S. Army, I found it useful to allow calculated PMV values to run beyond this 9 point scale; a calculated value of +5 appears to correspond to “I quit,” since the subjects frequently opted to discontinue further
39、 exposure under such conditions. Calculated values from +6 and +10 appear to correlate with conditions that produced heat exhaustion collapse, while I believe calculated values +10 repre-sent conditions with high heat stroke risk.In summary, physiological models offer greater possibilities to includ
40、e a full range of work, clothing, and environments in considering conditions consistent with thermal comfort. I model thermal comfort as conditions that the demand D is less than 20% of the bodys capacities C for temperature regulation. Conditions resulting in D/C between 20% and 40% may not be comf
41、ort-able (producing both some thermal discomfort and also “arousal”), but may also result in greater productivity. D/C from 40% to 60% are associated with performance decrements, from 60% to 80% with limited tolerance times, and 80% with increasing risks. Such models may also be more appropriate for
42、 incorporating productivity concerns than the simpler psychological or physical models of thermal comfort.A Model for Human BehaviorStudies of human behavior and productivity are known from Sumer over 4000 years ago. However, the framework for a general theory of behavior began with the 1932 physiol
43、ogical (fight or flight) studies of W.B. Cannon It must add from the “Hawthorne Effect” studies of the same era by E.B. Mayo, and draw from the studies in the 1950s of stress by Hans Selye, on stud-ies from 1980 on by Alluisi and others, and on work in the 1990s by Conroy, Fineberg, Deitch-man, and
44、others. These studies exposed the crucial need to develop and understand a general theory of behavior, and to make an explicit model for it. Such a model had to have both a taxon-omy (the science, principles, and laws of a classification system) and a syntax so that the model is more than just a ser
45、ies of labels with definitions. It links a behavior to antecedents and pro-ceeds through the individual, and his groups preparation for each step of the required tasks, includes modifying mediating variables and defines the consequences of the behavior in terms 1999. American Society of Heating, Ref
46、rigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC the constants or parameters of the trained network can then be transferred to the user who can calculate the per-formance of the heat exchanger under any other flow rate or inlet temperature conditions.ANNs are an es
47、tablished technique; see, for example, Haykin (1994) for an account of the history and mathematical background. There have been a few applications to heat transfer prob-lems. One is in the area of liquid crystal thermography to determine heat transfer coefficients (Jambunathan et al. 1996). The gene
48、ralization capacity of the ANN has been used to design a finned heat exchanger (Lavric et al. 1994, Lavric et al. 1995). Huang and Nelson (1994) also applied this technique to determine the delay time for a HVAC plant to respond to control actions. Ding and Wong (1990) controlled a simulated hydroni
49、c system using an ANN. In a pre-vious paper Daz et al. (1996) applied ANNs to analyze experimental heat-exchanger data.The goal of the present study is to represent heat exchangers using ANNs. The procedure used to set up and train the network is described first. Then a series of problems of increasing com-plexity are formulated to facilitate understanding. These problems are: one-dimensional conduc-tion, convection with one heat transfer coefficient, convection with two heat transfer coefficients,
