xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001478806
007 cr mnu|||uuuuu
008 040811s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000452
Total evaporative resistance of selected clothing ensembles
h [electronic resource] /
by Victor Caravello.
[Tampa, Fla.] :
University of South Florida,
Thesis (Ph.D.)--University of South Florida, 2004.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 163 pages.
ABSTRACT: With regard to heat stress, the limiting factor inherent in clothing ensembles is the total evaporative resistance. Clothing with higher evaporative resistance impedes the ability to cool by sweat evaporation. Knowing the evaporative resistance provides a means to compare candidate ensembles. Further, a value for evaporative resistance means that a rational method can be used to assess the heat stress exposure. Evaporative resistance of five clothing ensembles (cotton work clothes, cotton coveralls, and three coveralls of particle barrier, liquid barrier and vapor barrier properties) was determined empirically from wear tests during two study phases. For Phase 1, the metabolic rate was held constant at 160 W/m, and three levels of humidity (20, 50, 70% rh) were explored. Fourteen heat-acclimated participants (9 men and 5 women) completed trials for all combinations of clothing ensemble and environment. In the Phase 2 study, the humidity was held constant at 50% rh, and three levels of metabolic rate (114, 176, 250 W/m) were explored. Fifteen heat-acclimated participants (11 men and 4 women) completed trials for all combinations of clothing ensemble and environment. The data from both phases were analyzed separately using ANOVA. Significant differences were found among ensembles (p < 0.0001). The vapor barrier ensemble had the highest resistance at 0.026 kPa m/W. The liquid barrier was next at 0.018; followed by the particle barrier and cotton coveralls at 0.016. Work clothes was 0.014 kPa m/W. Vapor and liquid barrier ensembles were found to be significantly different from other ensembles. From the Phase 2 study, evaporative resistances decreased with increased activity and ranged from 0.0024 (cotton coveralls) to 0.0094 (vapor barrier) kPa m/W. The higher differences were associated with higher total evaporative resistance. The decreased evaporative resistances in Phase 2 can be explained by the pumping action associated with increased work. The relationship of Re,T to the difference of Pair -- Pskin was explored and found Re,T does not remain constant. Environment appeared to influence this relationship.
Adviser: Thomas E. Bernard.
x Public Health
t USF Electronic Theses and Dissertations.
Total Evaporative Resistance of Selected Clothing Ensembles by Victor Caravello A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Environmental and Occupational Health College of Public Health University of South Florida Major Professor: Thomas E. Bernard, Ph.D. Candi D. Ashley, Ph.D. Philip P. Roets, Sc.D. Skai W. Schwartz, Ph.D. Date of Approval: July 1, 2004 Keywords: clothing, evaporative cooling, heat balance, heat stress, metabolic rate Copyright 2004 Victor Caravello
ACKNOWLEDGMENTS Foremost, I would like to acknowledge th e United States Air Force for providing the opportunity for pursuing my advanced educat ion. I want to thank the senior officers within the Biomedical Sciences Corp and Bioenvironmental Engineering that had the confidence in selecting me for this AFIT assignment which ultimately allowed me to pursue my education interests at the University of South Florida. I would like to acknowle dge the funding support provided through the National Institute for Occupational Safety and Health for without, this research could have not been accomplished. On a more personal level, I am grateful to my family and friends that provided love and support throughout my program. I would like to thank Dr. Candi Ashley, Christina Luecke, Bumni Oladinni, and the Heat Stress Laboratory workers and study participants for their support and time. Las tly, I would like to say a special thank you to Dr. Tom Bernard for his ment oring and support. I choos e to study at USF for the opportunity to learn from him, and thanks to him, I have exceeded my educational expectations.
i TABLE OF CONTENTS LIST OF TABLES................................................................................................................i ii LIST OF FIGURES..............................................................................................................iv ABSTRACT..........................................................................................................................v INTRODUCTION..................................................................................................................1 Heat Exchange..............................................................................................................1 Role of Clothing in Heat Balance....................................................................................5 Insulation..........................................................................................................5 Permeability......................................................................................................6 Ventilation........................................................................................................6 LITERATURE REVIEW.......................................................................................................7 Components of Insulation and Evaporative Resistance.....................................................9 Laboratory Test Methods.............................................................................................12 Hot Plate Method............................................................................................12 Copper Manikin..............................................................................................13 Human Tests for Clothing Thermal Characteristics............................................15 Thermal Resistance Values for Various Work Ensembles...............................................23 Hypothesis..................................................................................................................26 METHODS........................................................................................................................ ...28 Participants.................................................................................................................28 Clothing Ensembles.....................................................................................................29 Protocols.....................................................................................................................30 Trials..........................................................................................................................31 Critical Conditions......................................................................................................34 Statistical Analysis......................................................................................................36 RESULTS........................................................................................................................ .....40 Experimental Data.......................................................................................................40 Total Insulation...........................................................................................................41 Phase 1.......................................................................................................................42 Phase 2.......................................................................................................................47
ii DISCUSSION..................................................................................................................... ..53 Internal Validity..........................................................................................................53 Comparison to Other Studies........................................................................................56 Total Insulation...............................................................................................56 Total Evaporative Resistance............................................................................58 Phase 1.......................................................................................................................59 Phase 2.......................................................................................................................62 CONCLUSION..................................................................................................................... 64 REFERENCES..................................................................................................................... 66 APPENDIX A: PART ICIPANT DATA...............................................................................70 APPENDIX B: EXPERIMENT AL DATA Â– PHASE 1.......................................................72 APPENDIX C: EXPERIMENT AL DATA Â– PHASE 2.......................................................81 APPENDIX D: SAS CODE AND ANALYSIS Â– PHASE 1................................................91 APPENDIX E: SAS CODE AND ANALYSIS Â– PHASE 2..............................................116 APPENDIX F. JMP IN DATA ANALYSIS Â– PROTOCOLS...........................................132 APPENDIX G. JMP IN DATA ANALYSIS Â– M2R5.......................................................143 APPENDIX H. GRAPHS OF Re,T VERSUS P BY ENSEMBLE....................................149 ABOUT THE AUTHOR..........................................................................................End Page
iii LIST OF TABLES Table 1. Comparison of Experimentally Determined Thermal Resistance Values...........................................................................................................23 Table 2. Summary of Partic ipant Characteristics.......................................................29 Table 3. Total Insulation Values for Ensembles.........................................................41 Table 4. Phase 1 Â– Mean Re,T Values with Standard Deviations................................42 Table 5. Phase 2 Â– Mean Re,T Values with Standard Deviations................................47 Table 6. Phase 2 Â– Average Metabolic Rates.............................................................50 Table 7. Statistical Analysis of M2R5 Dataset...........................................................55 Table 8. IT Values from Different Studies..................................................................57 Table 9. Re,T Values from Different Studies...............................................................59 Table 10. Regression Analysis Â– P by Re,T.................................................................62
iv LIST OF FIGURES Figure 1. Typical Time Course For Tre During an Inflection Point Protocol...............18 Figure 2. Phase 1 Â– Mean Re,T by Ensemble.................................................................44 Figure 3. Phase 1 Â– Mean Re,T by Ensemble w/o Ensemble E.....................................44 Figure 4. Phase 1Â—Mean Re,T Values by Ensemble and Environment........................45 Figure 5. Phase 1Â—Mean Re,T Values by Environment...............................................46 Figure 6. Phase 2 Â– Mean Re,T by Ensemble.................................................................49 Figure 7. Phase 2 Â– Mean Re,T by Metabolic Rate........................................................49 Figure 8. Phase 2 Â– Average Metabolic Rates by Ensemble........................................51 Figure 9. Phase 2 Â– Average R e,T by Metabolic Rate and Ensemble...........................51 Figure 10. Comparison of M2R5 Mean Re,T..................................................................54 Figure 11. Comparison of M2R5 Mean Metabolic Demands........................................55 Figure 12. Effect of Environment: Re,T vs. P Â– Ensemble A......................................61 Figure 13. Effect of Metabolic Rate on Re,T...................................................................63
v Total Evaporative Resistance of Selected Clothing Ensembles Victor Caravello ABSTRACT With regard to heat stress, the limiting factor inherent in clothing ensembles is the total evaporative resistance. Clothing with higher evaporative resistance impedes the ability to cool by sweat evaporation. Knowing the evaporative resistance provides a means to compare candidate ensembles. Further, a value for evaporative resistance means that a rational method can be used to assess the heat stress exposure. Evaporative resistance of five clothing ensembles (cotton work clothes, cotton coveralls, and three coveralls of particle barrier, liquid barrier and vapor barrier properties) was determined empirically from wear tests during two study phases. For Phase 1, the metabolic rate was held constant at 160 W/m2, and three levels of humidity (20, 50, 70% rh) were explored. Fourteen heat-acclimated participants (9 men and 5 women) completed trials for all combinations of clothing ensemble and environment. In the Phase 2 study, the humidity was held constant at 50% rh, and three levels of metabolic rate (114, 176, 250 W/m2) were explored. Fifteen heat-acclimated participants (11 men and 4 women) completed trials for all combinations of clothing ensemble and environment. The data from both phases were analyzed separately using ANOVA. Significant differences were found among ensembles (p<0.0001). The vapor barrier ensemble had the highest resistance at 0.026 kPa m2/W. The liquid barrier was next at 0.018; followed by the particle barrier
vi and cotton coveralls at 0.016. Work clothes was 0.014 kPa m2/W. Vapor and liquid barrier ensembles were found to be significantly different from other ensembles. From the Phase 2 study, evaporative resistances decreased with increased activity and ranged from 0.0024 (cotton coveralls) to 0.0094 (vapor barrier) kPa m2/W. The higher differences were associated with higher total evaporative resistance. The decreased evaporative resistances in Phase 2 can be explained by the pumping action associated with increased work. The relationship of Re,T to the difference of Pair Â– Pskin was explored and found Re,T does not remain constant. Environment appeared to influence this relationship.
1 INTRODUCTION Personal protective clothing has become commonplace in many industries today. While protective clothing provides protection from exposure to chemical and physical agents, it may lead to another health issue Â– excessive heat strain. Heat strain, the physiological adjustment to heat stress, is driven by work demand, environmental factors (such as ambient temperature, relative humidity and air movement), and clothing requirements. Heat strain is marked by increased body temperature, heart rate and sweating. Heat stress has been studied extensively, and one of the critical factors that ties protective clothing to heat stress is evaporative resistance. Heat Exchange To better understand the role of clothing in heat stress, the workplace factors discussed above can be described through a thermal balance model. This model balances net heat gained by the body with the required heat loss to prevent excessive heat buildup; that is, to maintain thermal equilibrium. Thermal balance is frequently described by some variation of Equation 1 [1, 2].
2 S = (M W) + R + C E (1) The heat storage rate (S) represents the net heat gain to or loss from the body. By convention, body temperature increases when S is positive and decreases when heat is lost (S is negative). When S = 0, the body is considered to be in thermal equilibrium. Heat is generated internally by metabolism (M). The rate of metabolic heat gain is determined by the rate and type of external work performed by the body. The total heat generated by metabolic demands from the work is equal to metabolic rate less the rate of external work performed (W). The rate of ra diant heat transfer (R) between the skin and the environment and the rate of heat tran sfer between the air and skin surface (C) collectively characterize dry heat exchange. Positive values for R + C are a heat gain while negative values indicate a heat loss. The term E represents the rate of evaporative cooling due to the evaporation of sweat, which is the primary mechanism for cooling the body. Heat production is determined by the amount of metabolic activity. At rest, the body generates heat from the energy produced to maintain basic body functions such as respiration and heart rate to supply the needed oxygen and nutrients to the cells. Metabolic activity rises as one becomes more active. This rise results with higher demand for oxygen and nutrients accompanied by increased metabolism at the active muscles. With increased metabolism, there is increased heat production in the muscle. The greater the demand, the more internal heat is generated. With the understanding that
3 R + C have a lesser effect of increasing in ternal heat production, the fundamental link between metabolic rate (work demands) to heat storage becomes clear . The minimal effect of dry heat exchange (R + C) as compared to the other terms in Equation 1 becomes evident as these terms are examined closer. Looking first at heat transfer rate by convection (C), as ambient air temperature is raised above skin temperature, the rate of heat gain by convection is increased. Simply stated, C is the difference of the ambient air temperature and the average skin temperature and modified by the rate of air movement over the skin. If the ambient air is cooler than the skin, then heat flows away from the body. The rate and direction of convective heat exchange depends on the temperature gradient between the air and the skin. The rate, but not direction, is also influenced by air motion and clothing. Generally the higher the air motion or velocity, the greater the rate of heat transfer. Clothing provides a barrier to the heat transfer through its insulation, so the more skin that is covered and/or the thicker the clothing, the lower the rate of convective heat transfer. The temperature of surrounding objects affects the radiant heat exchange between the environment and the body. Surfaces of different temperatures have a net heat flow from the hotter to the cooler surface by thermal radiation. The rate of heat transfer by radiant heat (R) depends on two factors. The first is the temperature gradient between the skin and surrounding objects. If the average temperature of surrounding objects is greater than skin temperature, there is a heat gain. Conversely, if the surroundings have a
4 lower average temperature than the skin, a heat loss occurs. The rate of heat transfer is proportional to the temperature gradient. Th e second factor is clothing insulation. As with convection, the thicker the clothing and/or the more skin that is covered, the lower the rate of heat transfer. Also, if the clothing has a reflective surface, the thermal radiation (or heat) is reflected away. Once the environmental temperature exceeds 35F the body can dissipate heat only by evaporation . Evaporation of sweat from the skin is the primary mechanism for losing excess body heat during activity. However, there is a limit to the amount of evaporative cooling that can occur. This limit is due to two factors; a physiological and a physical limit. The physiological limit is the amount of sweat that can be produced over time. The physical limit is the maximum rate of evaporative cooling (Emax) that can occur. Emax is limited by three primary factors. First, evaporation can only occur if the water vapor pressure of the skin (Psk) is higher than the water vapor pressure of the ambient air (Pa). Humidity is the ambient water vapor pressure. As humidity increases, this gradient from skin to air is reduced, and the rate of evaporative heat loss is decreased. The second factor is air movement. As air velocity increases, the boundary layer between the person and environment decreases allowing for an increase in evaporation of sweat (Emax increases). The third factor is clothing. All clothing act s as a barrier to evaporation. The physical characteristic known as the water vapor permeability of the clothing is directly proportional to the ability to evaporate sweat (Emax). Therefore, as the water vapor permeability decreases, so does Emax.
5 Air velocity, generated by body movements and air movement, is important in heat exchange between the body and the environment because of its role in convective and evaporative heat transfer. Increasing the air velocity can increase the convective and evaporative heat exchange by forcing air between the clothing and skin. Role of Clothing in Heat Balance Clothing impedes heat exchange between the body and the environment by limiting dry heat exchange and evaporative cooling. These effects can be described in further detail by looking at three characteristics associated with clothing: insulation, permeability and ventilation . Insulation Insulation describes the resistance to heat flow by convection and radiation. With the environmental conditions being constant, the gradient between the skin and air remains the same, but the as the insulation for an ensemble increases, the heat flow due to radiation and convection decreases. In other words, dry heat exchange through the clothing decreases with increasing insulation.
6 Permeability Permeability is the ability of water vapor to move through clothing. It affects the amount of evaporative cooling that can occur. Clothing with low permeability indicates that evaporation of sweat through the clothing is reduced, resulting in a decrease in evaporative cooling. Protective clothing ensembles can vary over a range from easily permeable to essentially impermeable. Ventilation Ventilation occurs as ambient air moves through the fabric and/or through clothing openings (cuffs, fasteners, and colla r). Clothing that allows air movement increases convective and evaporative cooling. Conversely, if the clothing is designed to limit the movement of air by being encapsulating or tight fitting with elastic cuffs, the convective and evaporative cooling are limited. Although protective clothing ensembles are worn to protect workers from biological, chemical or physical hazards, the barrier posses another hazard to workers by reducing the wearerÂ’s ability to dissipate internally generated heat through sweat evaporation. Depending on the environment and work demands, an excessive level of heat stress may result. Heat stress may cause reduced performance and increased risk of accidents and heat injury. It is imperative to understand the clothing characteristics, with the evaporative resistance being the most important, in order to effective manage the risks associated with wearing the protective clothing.
7 LITERATURE REVIEW Protective clothing and environmental conditions influence the level of heat stress a worker may experience. Understanding fabric properties may help predict how the environment will affect heat transfer for a selected ensemble. Havenith points out the importance of heat balance when wearing protective clothing . The goal is to maintain the body at around 37C by transferring excess heat from the body to the environment. Heat is produced through metabolic activity and protective clothing may hinder the loss of the heat gained. Some important factors th at affect heat transfer from the body to the environment include the temperature (air, surface, radiant), humidity, wind, movement, and clothing insulation . While all of these factors can affect heat transfer, the primary mechanism the body uses to dissipate heat is evaporative cooling. Therefore it is important to understand the potential barrier an ensemble may pose to evaporation of sweat. As sweat is secreted onto the skin, it should evaporate and cool the body. The rate of evaporation depends on the difference between the water vapor pressure of the skin and the ambient air water vapor pressure as well as the barrier provided by the clothing. This barrier interferes with the ability of water vapor to pass from the skin through the ensemble and into the ambient air. Therefore it is important to be able to
8 distinguish between clothing ensembles in terms of their permeability to water vapor. Permeability is alternatively expressed as total evaporative resistance (Re,T). In addition to evaporative cooling, heat loss from the skin to the ambient air occurs by radiation and convection (dry heat exchange). Dry heat exchange occurs because of the temperature difference between the skin and surrounding air. As with evaporation of water, dry heat also must leave the skin and be transported into and out of the clothing before the heat loss is complete. Therefore, clothing may interfere with dry heat exchange. This characteristic of clothing is referred to as insulation. The total clothing insulation (IT) and the total evaporative resistance (Re,T) are important characteristics to consider when comparing clothing ensembles. IT is an attribute that accounts for a decrease in heat flow due to total insulation provided by the clothing and the air layer between the skin and clothing. The higher the value of IT, the lower net heat flow due to radiation and convection is achieved. Re,T is the clothing characteristic that accounts for water vapor flow due to clothing permeability. The higher the value of evaporative resistance, the less evaporative cooling occurs; hence, the higher the level of heat stress. Although IT is associated with Re,T, the relationship is neither linear nor fixed for all clothing. To complicate matters, dry heat exchange and evaporative cooling are altered with air and body movement. Consequently, as work demands increase, it is possible to
9 see a decrease in the total evaporative resistance and IT. Both terms have static and dynamic values associated with its use. That is IT,stat is associated with the total insulation of clothing absent of movement, and IT,dyn is associated with the total insulation of clothing with movement. The same associations are true for Re,T. Thus, it is important to understand the thermal resistance properties of ensembles and how environment and activity level alters them. Components of Insulation and Evaporative Resistance In 1955 Burton and Edholm introduced the new unit for clothing insulation Â– the clo. One clo of insulation was intended to be equivalent to thermal insulation required to keep a sedentary person comfortable with normal indoor clothing at normal indoor climatic conditions (21C). The purpose of using the unit clo was to remove the awkward physical unit of m2 C/W, so one clo equals 0.155 m2 C/W . Goldman points out the advantage of using the clo as the unit of insulation is that it can be expressed as heat loss that will occur for the average adult male who has 1.8 m2 of surface area, using a simple relationship that such an individual will lose 10 kcal/hr of heat by radiation and convection for every degree (C) difference between the average skin temperature and the air temperature with 1 clo unit of insulation . Therefore 5 kcal/hr will be lost with 2 clo units of insulation.
10 Total clothing insulation (IT) is the combined insulation provided by clothing and the surrounding layer of air. Parsons  describes this relationship mathematically as: IT = Icl + Ia (2) Intrinsic clothing insulation (Icl) is a characteristic of the clothing itself and not the external environment or the body condition. Icl represents the resistance to heat transfer between the clothing surface and the skin. Typical units are C m2/W. Icl values and clo units are still used in several thermal comfort and clothing standards and information on determining Icl from measured values of IT is described in ISO Standard 9920 [7, 8]. Ia describes the thermal resistance or insulation provided by the air between the skin and garment. The properties of this layer are important to heat exchange and can be affected by the external environmental conditions. For an individual wearing an ensemble, the surface area of the individual is increased by an amount related to the thickness of the clothing layer. This new surface area is difficult to determine, but is important for other relationships with heat transfer. A clothing adjustment factor (fcl) is used to account for this new surface area. The term fcl is the ratio of the clothed surface area of the body to the nude surface area of the body.
11 The following equation is an approximation for fcl given by McCullough and Jones (1984) . fcl =1.0 + 0.31 Icl (clo) (3) To determine the intrinsic clothing insulation, IT,stat is measured using a clothed manikin or hot plate as described in the following section. Ia is measured in a similar fashion, but without the fabric sample or clothing. Then, Icl = IT,stat Ia/fcl (4) A more convenient term for measurement is effective clothing insulation (Icle), which is an approximation for Icl for the test conditions. This is described by the following equation: Icle = IT Ia (5) This same principle can be applied for the total evaporative resistance of clothing (Re,T) by dividing it into two components. The evaporative resistance due to the clothing itself (Rcl) and that due to the air layer (Ra) near the clothing or exposed skin. Re,T = Rcl + Ra (6)
12 Values for Re,T and Ra can be determined empirically from variations of the standard tests for clothing insulation using sw eating hot plates or sweating manikins (see next section). In this way, Rcl can be estimated from the following equation. Rcl = Re,T Ra / fcl (7) Again, a convenient approximation is Rcle = Re,T Ra (8) Laboratory Test Methods There are three different methods for determining the thermal properties of a garment. The first method involves the use of a heated plate; the second involves a heated copper manikin; and the third method involves the use of human participants. Hot Plate Method While manikins and hotplates use similar basic principles to determine heat loss and insulation values, they typically have different end goals. A hotplate is designed to provide accurate one-dimensional heat and moisture flow through a fabric sample to
13 determine thermal and water vapor resistance. The goal is to accurately evaluate the material properties for the test environment only . The American Society for Testing and Materials (ASTM) developed a standard method for using a sweating hot plate in met hod F 1868-02 . This test method covers the measurement of the thermal resistance and the evaporative resistance under steadystate conditions of fabrics, films, coatings, foams, and leathers, including multi-layer assemblies, for use in clothing systems. There are several relevant measures from sweating hot plate tests. The most basic measure is the operating heat flux required to maintain a constant skin temperature. In dry tests, it represents the conductive/ convective/radiative heat transfer. In sweating tests, it also includes evaporative heat losses. This sweating test is the most common method used. Copper Manikin A life-sized heated copper manikin can be used in the evaluation of the heat transfer potential of clothing garments. Similar to the hot plate method, the manikin is electrically heated so that the skin temperature is similar to that of people. Depending on number of surface segments the resolution can be adjusted to be sufficiently high to complete the measurement task. Some manikins in use today have 1 zone while others have more than 100 individually regulated segments. By summing up the area weighted heat loss values from the manikin, a total value for whole body heat loss is determined.
14 Some performance features of the most commonly used thermal manikins are: simulation of human body heat exchange, measurement of 3-dimensional heat exchange, integration of dry heat losses, measurement of clothing thermal insulation, product development, and providing values for prediction models. Values obtained with different manikins in different laboratories should be comparable and similar within defined limits for the same test conditions. The conditions and requirements for comparable measurements with different manikins and methods are defined in standards. The American Society for Testing and Materials standardized this procedure in ASTM method F 1291-99, and the International Organization for Standardization standardized the procedures in ISO 9920 [11, 12]. ASTM method F 1291 and ISO 9920 are both in the process of being updated to reflect the changes with the new sweating manikin. Similar to the hot plate procedures, these test methods cover the measurement of the thermal resistance and the evaporative resistance under steadystate conditions. With over a hundred different manikins being built and used around the world, it is difficult for any standard to encompass procedures for all types of manikins. Thermal manikins have evolved from its first model in 1941 for testing military clothing, and can be grouped into three categor ies. First are static (non-moving) and nonperspiring units, second are movable (walkable), but non-perspiring ones such as the copper manikin Â‘CharlieÂ’ in Germany used by Mecheels and Umbach in 1977, and third
15 are sweating manikins . To simula te sweating on non-perspiring manikins, many researchers used highly absorbent fabrics on the manikin (under the tested garment) and supplied water to the Â“underwearÂ” by sprinkling or water pipes. The third generation manikins simulate true perspiration and body motion, but again not all sweating manikins are dynamic. Recent innovation at the Institute of Textiles and Clothing, Hong Kong, produced Walter the "sweating" manikin. Walter is made up of water, mechatronics and breathable fabric, allowing realistic simulation of human thermal physiology under various environments. Walter has waterproof but moisture-permeable fabric skin, which can be unzipped, removed and interchanged with differe nt skins, simulating different rates and patterns of perspiration. In addition, Walter is a dynamic manikin that allows researchers to simulate the process of walking. This new technology surprisingly has an affordable price tag associated with it as it may be 90% less than traditional copper and plastic manikins . Human Tests for Clothing Thermal Characteristics Though useful, measurements of clothing thermal properties made on hot plates or manikins do not represent the properties of clothing during wear. The movements of the worker increase the convective air flow both between layers and at the clothing surface, modifying both insulation and vapor permeability. Although recent
16 developments with dynamic sweating manikins provide accurate data on clothing insulation and evaporative resistance, there still exists a difference between manikin and human wear tests. Human wear testing to determine evaporative resistance and clothing insulation values are based on determining the Â“prescriptive zoneÂ” as described by Lind . The upper limit prescriptive zone was defined as the point where heat loss and heat gain were equal (Emax = Ereq). Using the premise of the prescriptive zone, Belding and Kamon proposed a method for determining ambient vapor pressures at 36C for a variety of exercise intensities and air movements . Their method used a time-intensive protocol to determine critical environmental conditions for evaporative heat loss. Kamon and Avellini used the same approach as Belding and Kamon . In their experiments, the participants were subjected to a range of ambient temperatures between 36 and 52C with the water vapor pressure progressively increased at each ambient temperature. It was expected that on the basis of the body core temperature inflection, a line for the safe limit for the psychrometric chart would be empirically identified particularly for the higher temperatures, where individual sweating rate capacity was believed to be the limiting factor. Holmer and Elnas developed a method to determine both the evaporative and sensible heat loss occurring simultaneously . However, this method was difficult because direct measurement of the water vapor pressure gradient between the skin and
17 ambient air was required. Kenney et al. simplified this procedure by minimizing the number and duration of tests necessary to determine these limits . However, these more time-efficient protocols require the ability to systematically change ambient temperature or water vapor pressure. In the Kenney method, the metabolic rate target was 30% of maximal aerobic capacity (MAC). Generally, a person can work at 1/3 their MAC for an 8-hour work day . The testing chamber was controlled in that the dry bulb temperature (Tdb) and wet bulb temperature (Twb) could be closely manipulated. Air velocity was 0.5 m/s or less. The participantÂ’s heart rate (HR), rectal temperature (Tre) and mean skin temperature (Tsk) were monitored. Each participant partook in two trials wearing the garment to be tested. In one of the trials, Tdb was held constant and after 30 minutes for stabilization, the ambient water vapor pressure (Pa) was increased in increments of 0.13 kPa every 5 minutes. In the other trial, the Pa was held constant at a low humidity while the Tdb was increased in 1 C increments every five minutes. The Tre for each participant was plotted and the point of inflection where Tre sharply rose was noted. This inflection point represented the inability of the body to
18 dissipate the heat load and thereafter heat was stored by the body. Data from a participant was plotted, as shown in Figure 1, to illustrate the typical time course for an inflection point protocol. The important points of this figure is the starting Tre with a steady rise until a steady state is achieved, and then the inflection point followed by steep rise in Tre. Using this method, the critical temperature (Tcrit) at a given Pa and the critical water vapor pressure (Pcrit) at a given Tdb can be determined. Figure 1. Typical Time Course for Tre During an Inflection Point Protocol. At the inflection point, the required rate of evaporative cooling (Ereq) was equal to the maximum rate of evaporation (Emax). In other words, the rate of heat storage was zero 36.00 36.50 37.00 37.50 38.00 38.50 0102030405060708090100110120Time (min) Inflection PointRectal Temperature (C)
19 since the evaporative cooling was equal to the net heat gain from metabolism plus the dry heat exchange. Emax = (Mnet) + (R+C) (9) Another condition that existed at the inflection point was that evaporative cooling was at its maximum value. Here evaporative cooling (Emax) was equal to the difference in water vapor pressure between the skin (Psk) and the ambient environment (Pa) divided by the total evaporative resistance (Re,T). This is shown in the following equation: Emax = (Psk-Pa)/Re,T (10) The relationship of Re,T with respect to P (Psk Â– Pa) and Emax is based on the assumption that Re,T remains the same as P and Emax change. However, there are some researchers that question this principle. In his paper, Bernard found that a warm humid environment resulted with a lower Re,T . Theoretically, the maximum sweat rate is proportional to the difference in the saturated partial pressure of water at the skin minus the partial pressure in the air, and the evaporative resistance of clothing worn has no effect. The question about Re,T being a constant, as accepted in Equation 10, regardless of environment has not been evaluated.
20 Without a direct source of radiant heat, the rate of dry heat exchange (R+C) was taken as the difference in Tdb and Tsk divided by the clothing insulation (IT). This is presented in the following equation: (R+C) = (Tdb-Tsk)/IT (11) By substituting Equations 10 and 11 into Equation 9, the following equation results. (Psk-Pa)/Re,T = (Mnet)+(Tdb-Tsk)/IT (12) When the measured values and environmental conditions were placed in Equation 12 for each of two inflection points, there were two equations with two unknowns. This allowed for the calculation of IT and Re,T. At each inflection point, heat gain equals heat loss. This is represented mathematically using the following equation: Mnet + (R + C) = E, (13) where Mnet is the net metabolic heat production (M) corrected for external work (W) and respiratory exchanges due to convection (Cres) and evaporation (Eres). Metabolic rate (M)
21 in W/m2 was estimated from oxygen consumption in liters per minute and the respiratory ratio (R) using the following equation : M = 352(0.23R + 0.77). VO2/AD (14) The Dubois surface area (AD) was calculated for each subject using the following equation : AD = 0.202 W0.425 H0.725, (15) where W was the weight of the body (kg) and H was the height of the body (m). The external work (W) was calculated (W/m2) using the following equation: W = 0.163 mb VW fg /AD, (16) where mb was body mass in kg, VW was walking velocity in m/min, fg was the fractional grade of the treadmill, and AD was the Dubois surface area. Respiratory exchanges, latent respiration heat loss (Eres) and dry respiration heat loss (Cres), were calculated as follows:
22 Cres = 0.0012 M .(34-Tdb) (17) and Eres = 0.0173 M .(5.87-Pdp) (18) The net metabolic rate (Mnet) from Equation 13 can be calculated in W/m2 using the following equation: Mnet = (M W) + Cres Eres (19) Kenney et al. recognized that there may be some heat storage represented by a gradual change in Tre . To account for this, the rate of change in heat storage can be estimated knowing the specific heat of the body (0.97 W h/ C kg), body weight (BW), and the rate of change of body temperature ( Tre / t) before the inflection point was reached . That is, S = 0.97 BW Tre / AD t (20)
23 Thermal Resistance Values for Various Work Ensembles By using a hot plate, copper manikin or human subjects, thermal resistance values for garment ensembles can be quantified in terms of total insulation (IT) and total evaporative resistance (Re,T). Experimentally determined values for select work ensembles are presented in Table 1 for comparison purposes. The IT and Re,T values reported in Table 1 vary within each garment between the method used. Havenith et al. found that heated manikin results for standing/no wind appear to be on the average 0.15 clo (0.023 C m2/W) higher than human subjects results . Table 1. Comparison of Experimentally Determined Thermal Resistance Values. Ensemble Description IT ( C m2/W) Re,T (kPa m2/W) Reference Method Tyvek Coverall 0.070 0.020  hot plate Tyvek Coverall 0.171 0.033  manikin Tyvek Coverall 0.086 0.017  human participant Gore-Tex Outer-wear 0.054 0.009  hot plate Gore-Tex Outer-wear 0.210 0.032  manikin Gore-Tex Outer-wear 0.130 0.028  human participant Cotton, Single Knit 0.079 0.009  hot plate MenÂ’s Summer Casual (short sleeve) 0.201 0.029  manikin Military Fatigues 0.090 0.016  human participant
24 The data in Table 1 shows a larger gap (0.52 Â– 0.72 clo) between human participants and manikin data. Nishi et al.  and Vogt et al.  support HavenithÂ’s findings, but Nielsen at al.  and Olesen et al.  found the human participant values were 0.22 clo lower than manikin data. Havenith et al., found that the permeability index (IÂ’m) changed with wind and movement . In their study, the permeab ility increased three fold with permeable clothing and six fold with impermeable clothing. Additionally, they found that the total insulation was reduced by 32% and that walking at slower rates yielded smaller gains. Breckenridge and Goldman  reported similar findings with an increase in Im by 123% and a decrease in IT by 28%. A few years later, Parsons et al. , Holmer et al. , and Havenith at al.  all find that IT,stat needs to be adjusted for wind and walking. Their findings were adapted in ISO 7933. Havenith et al. reports that a 78% reduction of Re,T with a 50% reduction in IT can be seen  which is similar to the differences seen in Table 1. IT,dyn is converted by multiplying IT,stat by a correction factor (CFcl) as shown in Equations 21 and 22: IT,dyn = CFcl x IT,stat (21) and
252 2094 0 378 0 66 0 398 0 043 0 Walksp Walksp V V clar are CF (22) where Var is the velocity of the air and Walksp is the walking speed. The speed is calculated based on the metabolic demand (M in W/m2) and is shown in the equation below: Walksp = 0.0052 (M Â– 58) (23) Another obvious relationship in Table 1 is that the hot plate values are less than the manikin values for both IT and Re,T. There are a few possible reasons for these differences. First, the values may be lower than manikin data because of the fit or drape of the clothing. Hot plates tend to tested with a tight fit where the manikins and humans use a looser fit. Second, the wetting of the material is likely to alter the values. Manikin values were measured on dry manikins to determine the dry heat exchange while the hot plates were wet. The clothing on human participants was also wet from sweating. Although the hot plate values are not the same as that of the human participants, they are close in two of the garments tested. Wetting of the clothing alters the obtained value by attenuating the resistance . Third, it is difficult to simulate the movements of exercising humans. The presence of body motion aids in the circulation of air through
26 the clothing and therefore also reduces the resistance. The effects of air and body movement on IT and Re,T have been well documented [5, 21, 23, 27, and 28]. Although there is agreement of this needed adjustment, manikin data is not always adjusted for air and movement. The methodology for using human participants in conducting heat stress studies proposed by Kenney et al  has been used by Bernard and Matheen , Barker et al. , Kenney and Zeman , and Malcolm et al  as well as other researchers. Hypothesis A reasonable evaluation of selected protective clothing garments would be a determination of their heat exchange characteristics. The primary purpose of this paper was to explore the methodology for being able to distinguish between garments based on the total evaporative resistance properties within different environments and work demands. The secondary purpose is to challenge the relationship of Re,T with respect to changes in P and Emax. The default assumption is that Re,T remains the same as P and Emax change. There are three null hypothesis to be tested: (1) there are no differences between mean Re,T values among ensembles, (2) there are no differences between mean Re,T values among environments and metabolic rates/demands, and (3) there are no differences
27 between mean Re,T values while P changes within environments and metabolic demands.
28 METHODS The primary purpose of this paper was to explore the methodology for distinguishing between garments based on the total evaporative resistance properties within different environments and work demands. The secondary purpose is to challenge the relationship of Re,T with respect to changes in P and Emax. Experimental trials were conducted to determine the evaporative resistance for five clothing ensembles. The protocols included a fixed metabolic demand under three different relative humidity levels for Phase 1, and for three metabolic demands with a fixed relative humidity level for Phase 2. The key to these studies was being able to distinguish the point of transition from compensable heat stress to uncompensable heat stress (Ereq = Emax). Participants Fourteen adults (nine men and five women) participated in experimental trials for Phase 1 and fifteen adults (eleven men and four women) participated in Phase 2. Their physical characteristics are provided in Appendix A and the average and standard deviation of their physical characteristics by gender are provided in Table 2. Following
29 the local IRB procedures, a written informed consent was obtained and all subjects were qualified by a physician. Prior to beginning the experimental trials, participants underwent a 5-day acclimatization period. Acclimatization involved walking on a treadmill at a metabolic rate of approximately 160 W/m2 in a climatic chamber at 50 C and 20% relative humidity (rh). During acclimation participants wore shorts (and sports bra), socks and shoes. Table 2. Summary of Participant Characteristics. Clothing Ensembles Five different clothing ensembles were evaluated in each Phase with only one ensemble being changed for Phase 2. The ensembles included: Ensemble A -work clothes (4 oz/yd2 cotton shirt and 8 oz/yd2 cotton pants); Ensemble B -cotton coveralls ProtocolGenderNum Age (yrs) Height (cm) Weight (kg) Surface Area (m2) Men929.2 6.8183 6.097.2 18.52.18 0.20 Women531.8 9.1161 7.063.5 17.21.66 0.23 All1430.1 7.5175 12.085.2 24.12.00 0.33 Men1128.0 9.5176 11.281.9 11.71.98 0.18 Women423.0 4.7165 6.364.2 18.01.70 0.22 All1526.7 8.6173 11.177.2 15.31.91 0.22 Men2028.6 8.2180 8.689.55 15.12.08 0.19 Women927.4 6.9163 6.763.85 17.61.68 0.23 All2928.0 7.5171 7.676.7 16.351.88 0.21 Phase 2 Metabolic Both Phase 1 Humidity
30 (9-10 oz/yd2 ) and three limited-use protective clothing ensemble: Ensemble C -particlebarrier ensembles (Tyvek 1424 for Phase 1 and Tyvek 1427 for Phase 2), Ensemble D water-barrier, vapor-permeable ensembles (NexGen LS 417), and Ensemble E -vapor-barrier ensembles (Tychem QC). The limited-use coveralls had a zippered closure in the front and elastic cuffs at the arms and legs. All ensembles were worn without a hood and a cotton tee-shirt and/or sports bra and shorts were worn under all clothing ensembles. Protocols Three experimental protocols were followed each Phase. The design for Phase 1 had three environments with a fixed metabolic rate. Treadmill speed and grade were set to elicit a metabolic rate of about 160 W/m2. The first protocol (R7) was a warm/humid environment designed to reduce Emax by limiting evaporation. The second protocol (R2) was a hot/dry environment designed to increase Ereq by increasing radiant and convective (R+C) heat gains. The third protocol (R5) was a moderate environment designed to increase R+C while decreasing Emax. In the R7 protocol, the dry bulb temperature (Tdb) was set at 30C and relative humidity (rh) at 70%. Once the participant reached thermal equilibrium (no change in Tre and heart rate for at least 15 minutes), Tdb was increased 0.7C every 5 minutes. In the
31 R2 protocol, Tdb was set at 40C with rh at 20%. When participants reached thermal equilibrium, Tdb was increased 1 C every 5 minutes. For the R5 protocol, Tdb was set at 34C with 50% rh. Upon reaching thermal equilibrium, Tdb was increased 0.8C every 5 minutes. In Phase 2, the study design called for three metabolic demands: light demand, metabolic rate of 80 W/m2 (M1); moderate demand, metabolic rate of 160 W/m2 (M2); and heavy demand, with a metabolic rate of 240 W/m2 (M3). Actual metabolic rates were calculated using oxygen consumption based on open circuit indirect calorimetry and body surface area. The environment was set at a 50% relative humidity (rh). The starting temperature for the trials was set at 34C, but varied based on the ensemble being worn and individual. When participants reached thermal equilibrium, Tdb was increased 1 C every 5 minutes. Trials The trials were conducted in a Model 7010 climatic chamber designed by Forma Scientific. The chamber was 2.4 m wide, 3.0 m deep, and 2.2 m high (8.0 x 10.0 x 7.3 ft). The range of humidity was 10-90% and the temperature range was 4-60 C (40140 F). Temperature and humidity were controlled according to protocol and air speed
32 was 0.5 m/s. The work demand consisted of walking on a motorized treadmill at a speed and grade set to elicit the desired metabolic rate (80, 160, or 240 W/m2). Heart rate was monitored using a Polar heart rate monitor. Core temperature was measured with a flexible YSI thermistor (401AC) inserted 10 cm beyond the anal sphincter muscle. The thermistor was calibrated prior to each trial using a hot water bath. Skin temperatures were measured with an YSI surface thermistor (409AC) taped to the skin at four points (left chest, right upper ar m, right thigh, and left calf). Average skin temperature was determined using a modified Ramanathan Technique as shown in the following equation : Tsk = 0.3 Tchest + 0.3 Tarm + 0.2 Tthigh + 0.2 Tcalf (25) Assessment of oxygen consumption was used to establish metabolic rate. Participants breathed through a two-way valve connected to flexible tubing that was connected to a collection bag (Douglas bag). Expired gases were collected every 30 minutes during the experiments for 2.5 minutes. The volume of expired air was measured using a dry gas meter. A small aliquot was removed from the Douglas bag and drawn through a drying agent (DriRite) into a Beckman Model E2 Oxygen Analyzer to determine oxygen content. Oxygen consumption (VO2) was calculated according to Equation 25.
33 VO2 = VE O2 CF (26) Where, VE was the expired air flow rate in liters per minute, O2 was the difference in the fraction of oxygen between the inspired and expired air, and CF was a correction factor to convert the volume to st andard temperature and pressure dry (STPD) . During trials, participants were allowed to drink water or a commercial fluid replacement beverage at will. Core temperature, heart rate and ambient conditions (dry bulb, psychrometric wet bulb and globe temperatures) were monitored continuously and recorded every 5 minutes. Trials lasted approximately 120 minutes unless one of the following criteria was met: (1) a clear rise in Tre associated with a loss of thermal equilibrium, (2) Tre exceeded 39 C, (3) a sustained heart rate greater than 85% of the age-predicted maximum heart rate, or (4) participant wished to stop. The order of the ensemble-environment conditions was randomized. Any trial that had to be repeated was repeated at the end. An experimental trial data dictionary is presented with the data for each phase in the appendices. Phase 1 data are provided in Appendix B and Phase 2 data are in Appendix C.
34 Critical Conditions By evaluating the point at which a clear rise in Tre, associated with a loss of thermal equilibrium, the critical condition (Ereq = Emax) can be determined by using the data point preceding this rise. At the point of critical conditions Re,T can be calculated using Equation 12. In this equation there are two unknowns, Re,T,dyn and IT,dyn. The IT,Stat values were calculated from measured insulation values (clo) according to ASTM F 1291, Standard Test Method for Measuring th e Thermal Insulation of Clothing using a Heated Manikin, Option #1 . The insulation provided by clothing (clo) was measured using an electricallyheated manikin in thermal equilibrium with the surrounding environment. The manikin is a full size male with 19 electrically separate segments. The manikin has knee, hip, shoulder, and elbow joints that can be flexible or locked in an immobile position . Measurement and control of the heat supply for each section is achieved by using a digital process computer. Display and recording of the data is conducted by a second computer which is serially interfaced with the process. Temperature readings and power input values for each segment are area weighted when calculating the total insulation value .
35 The insulation value (clo) was measured according to ASTM F 1291, Standard Test Method for Measuring the Thermal Insu lation of Clothing using a Heated Manikin, Option #1 . The chamber had an ambient air temperature of 20C, dew point temperature was controlled at 1C and air velocity of 0.2 m/s, and the manikin surface temperature was set at 33.2C. To test each ensemble, the manikin was dressed in an ensemble with all closures secured. It was hung from a metal stand by a hook in the head. The feet touched the floor with the arms hung at the sides. Equilibrium was maintained for at least one hour prior to testing. Data were collected by computer every 30 seconds for the 30 minute test . The IT,dyn values were calculated for each ensemble by adjusting the IT,stat values for wind and speed as suggested by Havenith et al . Additionally, measured trial data was used to compute other variables ( T, P and M) needed to compute Re,T. The metabolic rate was computed based on O2 consumption using Equations 14, 15 and 26. The equations for differences in temperatures and partial pressures are shown below. T = Tdb Â– Tsk (27)
36 P = Psk Â– Pa (28) 3 237 27 176105 0sk skT T ske P (29) pwb db sk skT T T T ske P067 0 3 237 27 176105 0 (30) where P is the difference in partial pressure of water vapor between the skin and ambient air. Statistical Analysis Statistical analysis included general descriptive statistics and linear modeling. The primary data analysis was conducted with analysis of variance (ANOVA) and verified with the Mixed Procedure. If a significant difference among ensembles was found at = 0.05, TukeyÂ’s Honestly Significantly Different (HSD) was calculated . If the difference between any treatment mean value was greater than the HSD, then the difference was determined to be statistically different. The data were reviewed for outliers defined as data points exceeding the mean two times the standard deviation. A Sharpio-Wilkes statistical test for fit was performed on the ensemble datasets to determine the best fit of the data (normal or log normal). All
37 of the data fit well as being normally distributed. The data (participant, ensemble, protocol, and Re,T) were imported into SAS version 8.2. Since the data were not balanced, the data was analyzed using a mixed linear model as well as the standard liner model (GLM) for comparison. The mixed procedure fits a variety of mixed linear models to data and enables these fitted models to make statistical inferences about the data. A mixed linear model is a generalization of the standard linear model used in the GLM procedure, the generalization being that the data are permitted to exhibit correlation and non-constant variability. The mixed linear model, therefore, provides the flexibility of modeling not only the means of the data (as in the standard linear model) but their variances and covariances as well . The primary assumptions underlying the analyses performed by SAS model Proc Mixed (PM) are as follows: the data are normally distributed (Gaussian), the means of the data are linear in terms of a certain set of parameters, the variances and covariances of the data are in terms of a different set of parameters, and they exhibit a structure matching one of those available in PM . The fixed-effects parameters are associated with known explanatory variables, as in the standard linear model. These variables can be either qualitative (as in the traditional analysis of variance) or quantitative (as in standard linear regression). However, the covariance parameters distinguish the mixed linear model from the standard linear model.
38 The need for covariance parameters arises quite frequently in applications. The most typical scenarios include: (1) the experimental units on which the data are measured can be grouped into clusters, and the data from a common cluster are correlated, and (2) repeated measurements are taken on the same experimental unit, and these repeated measurements are correlated or exhibit variability that changes. PM provides a variety of covariance structures to handle the previous two scenarios. The most common of these structures arises from the use of random-effects parameters, which are additional unknown random variables assumed to impact the variability of the data. The variances of the random-effects parameters, commonly known as variance components, become the covariance parameters for this particular structure. Traditional mixed linear models contain both fixedand random-effects parameters, and, in fact, it is the combination of these two types of effects that led to the name mixed model. Proc Mixed fits not only these traditional variance component models but numerous other covariance structures as well. PM computes several different statistics suitable for generating hypothesis tests and confidence intervals. The validity of these statistics depends upon the mean and variancecovariance model selected. The independent variable was Re,T with three dependent variables (ensemble, protocol, and participant). For the PM model, participant was set as
39 the random-effect parameter, and for three-way ANOVA participant was part of the class statement. Once significance was detected, TukeyÂ’s HSD test was used to test all pair-wise comparisons among Re,T means to determine which ensembles were significantly different. Interaction between two variables was also evaluated (ensemble x protocol). Significance levels were set at = 0.05. Three hypothesisÂ’ were tested: (1) there are no differences between mean Re,T values among ensembles, (2) there are no differences between mean Re,T values among environments and metabolic rates/demands, and (3) there are no differences between mean Re,T values while P changes within environments and metabolic demands.
40 RESULTS The primary purpose of this paper was to explore the methodology for distinguishing between garments based on the total evaporative resistance properties within different environments and work demands. The secondary purpose is to challenge the relationship of Re,T with respect to changes in P (Psk Â– Pa) and Emax. Experimental trials were conducted to determine the evaporative resistance for five clothing ensembles. The protocols included a fixed metabolic demand under three different relative humidity levels for Phase 1, and for three metabolic demands with a fixed relative humidity level for Phase 2. The hypothesisÂ’ tested include: (1) there are no differences between mean Re,T values among ensembles, (2) there are no differences between mean Re,T values among environments and metabolic rates/demands, and (3) there are no differences between mean Re,T values while P changes within environments and metabolic demands. Experimental Data At the critical conditions measured data captured included heart rate (HR), rectal temperature (Tre), skin temperatures (calf, thigh, upper arm, and chest), and environmental conditions (humidity and dry, wet, and black bulb temperatures). Oxygen
41 (O2) consumption was measured and recorded at 30 minute intervals (at 30, 60 and 90 minute point). Using the measured data, other key components were computed. The metabolic rate (M) was computed using Equation 14 and the differences in temperatures and partial pressures ( T and P) were calculated using Equations 27 Â– 30. The data for the critical conditions for all the trials are provided in Appendix B for Phase 1 and Appendix C for Phase 2. Total Insulation Results were reported as IT,Stat values and then converted to IT,dyn by adjusting for wind and movement as suggested by Havenith et al  as shown in Equation 20. Results for both static and dynamic values are presented in Table 3. Table 3. Total Insulation Values for Ensembles. 80 W/m2160 W/m2240 W/m2Ensemble A-(WC) --Work Clothes0.1800.1680.1470.133 Ensemble B-(CC) --Cotton Coverall0.1960.1820.1600.145 Ensemble C-(PB) --Tyvek 14240.1910.1780.1560.141 -(PB) --Tyvek 14270.1900.1770.1550.140 Ensemble D-(WB) --Water Barrier0.1890.1760.1540.140 Ensemble E-(VB) --Vapor Barrier0.1850.1720.1510.137 IT,dynClothing Item IT stat
42 Phase 1 In Phase 1, the primary focus was to determine if the methodologies used can distinguish differences among the five sel ected ensembles (WC Â– work clothes, CC Â– cotton coveralls, PB Â– particle barrier, WB Â– water barrier, and VB Â– vapor barrier) and evaluate how the environment affects Re,T. There were three different environments (R2 Â– hot/dry, R5 Â– moderate, and R7 Â– warm/humid) with a fixed moderate metabolic demand (M2) of 160 W/m2. The average Re,T values and standard deviations are presented in Table 4. Table 4. Phase 1 Â– Mean Re,T Values with Standard Deviations. MeanStd DevMeanStd DevMeanStd DevMeanStd Dev A0.0130.00400.0170.0035 0.0120.00290.0110.0029 B0.0140.00470.0180.0046 0.0120.00350.0120.0035 C0.0150.00520.0200.0042 0.0140.00430.0130.0047 D0.0170.00530.0210.0039 0.0160.00510.0140.0046 E0.0270.00890.0340.0100 0.0260.00510.0210.0065 R7 Ensemble All DataR2R5 A Sharpio-Wilkes statistical test for fit was performed on all datasets to determine the best fit of the data (normal or log normal) All of the data fit well as being normally distributed. The data were analyzed using the mixed procedure and using a three-way analysis of variance (ANOVA). The main effects included three protocols, five ensembles and 14 participants. Not all participants completed all trials and some trials were repeated which resulted in an unbalanced design. Using SAS 8.1, the Mixed and GLM models were used to determine statistical differences for Re,T among ensembles,
43 environments, and participants. Participants were treated as a blocking variable. The SAS code used and data output for Phase 1 is provided in Appendix D. Very significant differences (p<0.0001) were found for ensemble, environment, and participant. TukeyÂ’s HSD test was performed to determine which pairs were significantly different among ensembles and environments. This resulted with Ensemble E being different from all other Ensemble s and Ensemble D being different from Ensembles A and B. This is depicted graphically along with the mean Re,T values for ensembles in Figure 2. The lines below the ensembles indicate ensembles that are statistically similar. Ensemble E was very different from the other ensembles and could have interfered with the ability to differentiate differences among the ensembles. Therefore the data were analyzed again with Ensemble E excluded. In this analysis, Ensemble D was different from all other ensembles and En semble A was statistically different from Ensemble C. These data are presented in Figure 3.
44 0.000 0.005 0.010 0.015 0.020 0.025 0.030 ABCDEEnsembleRe,T Figure 2. Phase 1 Â– Mean Re,T by Ensemble. 0.000 0.004 0.008 0.012 0.016 0.020 ABCDEnsembleRe,T Figure 3. Phase 1 Â– Mean Re,T by Ensemble w/o Ensemble E.
45 Interaction between ensemble and environment was tested and found to be significant (p=0.0187) with Ensemble E in the mix and not significant (p=0.8820) when analyzed without Ensemble E. The interaction between environment and ensemble can be seen graphically in Figure 4. 0.010 0.015 0.020 0.025 0.030 0.035 ABCDEEnsembleRe,T R2 R5 R7 Figure 4. Phase 1Â—Mean Re,T Values by Ensemble and Environment. The statistical software JMP-IN 5.1 was used to analyze the mean Re,T values for each environment within an ensemble. This resulted with all ensembles being very significantly different within each environment (p< 0.001).
46 Analysis of Re,T by environment resulted with the environment being significantly different. TukeyÂ’s HSD detected all pairs to be very significantly different (p<0.001). Figure 5 presents the mean Re,T values by environment and indicates that all environments are statistically different from the others. 0.000 0.001 0.002 0.003 0.004 0.005 0.006 R2R5R7EnvironmentRe,T Figure 5. Phase 1Â—Mean Re,T Values by Environment. The statistical software JMP-IN 5.1 was used to analyze the mean Re,T values for each environment within an ensemble. This resulted the Re,T values being significantly different between environments for all Ensembles (p<0.001).
47 Phase 2 In Phase 2, the primary focus was to verify the methodologies used can distinguish differences among the five sel ected ensembles (WC Â– work clothes, CC Â– cotton coveralls, PB Â– particle barrier, WB Â– water barrier, and VB Â– vapor barrier) and evaluate how the metabolic rate affects Re,T. There were three different metabolic rates (M1 Â– light work, M2 Â– moderate work, and M3 Â– heavy work) with a fixed moderate environment (R5) at 50% rh. The average Re,T values and standard deviations are presented in Table 5. Table 5. Phase 2 Â– Mean Re,T Values with Standard Deviations. MeanStd DevMeanStd DevMeanStd DevMeanStd Dev A0.0110.0020.0110.0020.0130.0030.0110.001 B0.0120.0030.0140.0030.0120.0020.0110.003 C0.0130.0030.0150.0040.0120.0020.0110.001 D0.0150.0040.0180.0050.0150.0020.0120.002 E0.0240.0060.0280.0050.0240.0040.0190.003 M3 Ensemble All DataM1M2 The data were reviewed for outliers and 20 out of 226 data points exceeded the mean two times the standard deviation. A Sharpio-Wilkes statistical test for fit was performed on all datasets to determine the best fit of the data (normal or log normal). All of the data fit well as being normally distributed. The data were analyzed using the mixed procedure and using a three-way analysis of variance (ANOVA). The main
48 effects included three protocols, five ensembles and 15 participants. Not all participants completed all trials and some trials were repeated which resulted in an unbalanced design. Using SAS 8.1, the Mixed and GLM models were used to determine statistical differences for Re,T among ensembles, metabolic rates, and participants. Participants were treated as a blocking variable. The SAS code used and data output for Phase 2 is provided in Appendix E. Very significant differences (p<0.0001) were found for ensemble, metabolic rate, and participant (p<0.0001). TukeyÂ’s HSD test was performed to determine which pairs were significantly different among ensembles and environments. This resulted with Ensembles D and E being different from all other ensembles. There was no statistical difference detected when analyzing the data with and without outliers, therefore the complete dataset was used for all data references. Figure 6 depicts the mean Re,T values by ensembles graphically. The lines below the ensembles indicate ensembles that are statistically similar. Analysis of Re,T by metabolic rate resulted with the environment being significantly different. TukeyÂ’s HSD detected all pairs to be very significantly different (p<0.0002). Figure 7 presents the mean Re,T values by environment and indicates that all environments are statistically different from the others.
49 0.000 0.005 0.010 0.015 0.020 0.025 ABCDEEnsembleRe,T Figure 6. Phase 2 Â– Mean Re,T by Ensemble. 0.000 0.004 0.008 0.012 0.016 0.020 M1M2M3Metabolic DemandRe,T Figure 7. Phase 2 Â– Mean Re,T by Metabolic Rate.
50 All trials were designed for the participants to elicit a desired metabolic rate based on varying the speed and grade of the treadmill for each individual. During Phase 2 there were three desired metabolic rates 80 W/m2, 160 W/m2, and 240 W/m2. The average metabolic rates by protocol are provided in Table 6 and shown graphically in Figure 8. Table 6. Phase 2 Â– Average Metabolic Rates. EnsembleM1M2M3Avg A121175250183 B118177241178 C108178251177 D111177259182 E114176249181 Average114176250 Interaction between ensemble and metabolic rate was tested and found to be very significant (p<0.0001). The interaction between metabolic rate and ensemble can be seen graphically in Figure 9.
51 100 120 140 160 180 200 220 240 260 280 ABCDEEnsembleMetabolic Rate (W/m2) M1 M2 M3 Figure 8. Phase 2 Â– Average Metabolic Rates by Ensemble. 0.010 0.012 0.014 0.016 0.018 0.020 0.022 0.024 0.026 0.028 0.030 ABCDEEnsembleRe,T M1 M2 M3 Figure 9. Phase 2 Â– Mean Re,T by Metabolic Rate and Ensemble.
52 Looking for where the interaction might occur, the statistical software JMP-IN 5.1 was used to analyze the mean Re,T values for each metabolic rate within an ensemble. This resulted the Re,T values not being significantly different between metabolic rates within Ensemble A (p=0.0717) and B (p=0.0610), and very significantly different for Ensembles C, D, and E (p<0.001). The complete JMP-IN analysis of Phase 1 and 2 (protocols) is provided in Appendix F.
53 DISCUSSION The focus of this study was to conduct experi mental trials to explore two research areas. First, trials were conducted to di stinguish among garments based on the total evaporative resistance propert ies between different environments and work demands. Experimental trials were conducted in each pha se to determine the evaporative resistance for five selected clothing ensembles. Th e protocols included a fixed metabolic demand under three different relative humidity leve ls for Phase 1, and for three metabolic demands with a fixed relative humidity leve l for Phase 2. Second, the data from the experimental trials were used to discern weat her or not the generall y accepted theory that Re,T remains constant. Internal Validity Phase 1 and 2 both had one protocol that had the same design -a moderate work rate (M2 Â– 160 W/m2) and a moderate environment (R5 Â– 50% rh). Ensemble C (PB) was changed between Phase 1 and 2, but the other ensembles remained the same. Some of the same participants from Phase 1 we re used again in Phase 2, but most were different. Comparing the moderate work ra te and moderate environment (M2R5) data from both phases provided internal va lidity to the methodology. The mean Re,T values for
54 ensembles for Phase 1 and 2 were plotted and presented in Figure 10. The average metabolic rates for Phase 1 a nd 2 are provided in Figure 11. 0.000 0.005 0.010 0.015 0.020 0.025 0.030 ABCDEEnsembleRe,T Phase 1 Phase 2 Figure 10. Comparison of M2R5 Mean Re,T. In Figure 10 the grouping of data points for each ensemble appears to be tightly correlated with the exception of Ensemble C. As discussed previously, Ensemble C was changed from a Tyvek 1424 for Phase 1 to Tyvek 1427 for Phase 2. Statistical analysis using JMP-IN 5.1 was used to analyze the M2R5 data to compare mean Re,T values within ensembles. This analysis resulted with only Ensemble C being significantly different (p=0.0349). The statistical results (p values) and mean Re,T values are provided in Table 8. The JMP-IN analysis is provided in Appendix G.
55 140 145 150 155 160 165 170 175 180 ABCDEEnsembleMetabolic Demand Phase 1 Phase 2 Figure 11. Comparison of M2R5 Mean Metabolic Demands. Table 7. Statistical Analysis of M2R5 Dataset. Phase 1Phase 2 A0.76920.0130.013 B0.47410.0130.012 C0.03490.0150.012 D0.25770.0170.015 E0.16530.0260.024 Mean Re,TEnsemblep Value At first glance, the metabolic rates seen in Figure 11 appear to be significantly different. However, the scale is compresse d making the small differences (< 10%) seem larger. The differences in Phase 1 and 2 we re not enough to change the conformation of internal validity.
56 Comparison to Other Studies Total Insulation In order to determine the evaporative resi stance, an understanding of the ensemble properties must be understood. The clothing properties were de rived from manikin experiments conducted at the Institute fo r Environmental Research, Kansas State University by Dr. Elizabeth McCullough. Using her manikin, and following ASTM F 1291, she was able to determine th e total clothing insulation (IT,stat), intrinsic clothing insulation (Icl) and the clothing area factor (fcl) for the six ensembles used in the experimental trials. As reported by Havenith et al.  heated manikin results for standing/no wind appears to be on average 0.023 C m2/W higher than human participants. While some studies support this claim, other studies find the manikin data as being lower than human participants. After adjusting for wind and movement the IT,dyn values were compared to other studies that had similar ensembles. The studies included Barker et al. , Kenney et al. , and Bernard and Mathee n . After adjusting the IT values, the values used in this study are clearly higher th an other studies for similar ensembles. The reported IT values are shown in Table 9.
57 Table 8. IT Values from Different Studies. EnsembleCurrentBarkerKenneyBernard A0.1470.0840.050 B0.1600.1070.0560.107 C0.1560.0860.059 D0.1540.0860.050 E0.1510.0860.035IT (m2 K/W) The Barker et al. and Bernard and Matheen studies reported IT values that were 33 Â– 45% lower, and the Kenney et al. study repor ted values that were 62 77% lower. A primary difference in all of these studies is the adjustment for wettedness. Although there isnÂ’t a set standard for adjusting for clothing wettedne ss thus far, many researchers use a 50% default adjustment. The Barker at al. and Bernard and Matheen studies both used a 45% adjustment for wettedness. Had the current values been adjusted for wettedness, the IT values would match the Barker et al. study very well. On the other hand, Kenney et al. used a simultane ous derivation method to compute IT. Using his methodology to compute IT with the Phase 1 data resulted in too much variation in IT to make it useful. On the face of it, having a good estimate of the IT is important because it is used to compute Re,T. However, Barker et al. demonstrated that relatively large changes in IT result in minor changes in Re,T. Therefore the manikin data adjusted for wind and movement is sufficient for determining the IT of the ensembles used.
58 Total Evaporative Resistance In 1993 Kenney at al. , building from previous research, setup the framework for conducting human experiments in a climate controlled heated chamber. The methodology they established is still in use. Using the principle of the prescriptive zone as established by Lind , the determination of the inflection point is esta blished by selecting the point preceding a rise in Tre. At the inflection point, critical conditions exist where S = 0 and Emax = Ereq. From these conditions the basic heat balan ce equation can be manipulated by substituting terms for Emax and Ereq and solving for Re,T. Similarly to the IT values, mean Re,T values were compared to the same studies Â– Barker et al. , Kenney et al. , and Bernard and Matheen . Although there were large differences (33 Â– 45% lower IT values reported by Bark er et al. and 62 Â– 77% lower values reported by Kenney et al.), the Re,T values had less differences as shown in Table 10.
59 Table 9. Re,T Values from Different Studies. EnsembleCurrentBarkerKenneyBernard A0.01330.01310.0092 B0.01400.01590.00960.0155 C0.01540.01630.0112 D0.01740.01760.0123 E0.02730.01360.0344Re,T (kPa m2/W) Barker et al. reported three Re,T values that were within 6%, one at 14% and one at 50% (vapor barrier suit). Kenney et al.Â’s values ranged from 26 Â– 32% difference, and Bernard and MatheenÂ’s reported valu e was 11% higher. BarkerÂ’s IT values were close to the ensembles used in this study with the ex ception of not adjusti ng these values by 45% for account for wettedness. However, even with the 45% difference in IT, there is minor differences in Re,T. Barker et al. had previously repor ted this relationshi p, and this study supports it. Based on the Barker et al. and Bernard and Matheen studies, the Re,T values calculated in this study appear in line with other research. Phase 1 The methodology used was able to disti nguish effectively be tween the selected ensembles as illustrated in Figures 2 a nd 3 showing significant differences among Ensembles D and E. Since Ensemble E is a vapor barrier suit, it is expected to be different from ensembles that do not prev ent vapor transmission such as cotton and Tyvek coveralls. Similarly, Ensemble D is a li quid barrier suit, so it is also expected to
60 be different from particle barrier and co tton clothing (Ensembles A, B and C ) with respect to evaporative resistance. In Figure 4 there is a clear difference between environments. The differences between the mean Re,T values remains the same within Ensembles A Â– D, but increases for Ensemble E. The increase difference accounts for the interaction between the environment and ensemble and is verified by not seeing an interaction when the data is analyzed without Ensemble E. The fact that there is a difference between environments is by itself an important finding. The relationship between Re,T and P with respect to Emax (Equation 10) has generally been accepted that Re,T remains constant as P and Emax change. This relationship is alluded to in ISO 7933 a nd discussed by Parsons (2003) . Again, Figure 4 clearly shows that Re,T is not the same as the environment changes. Using the data from Phase 1 and Phase 2, P was plotted against Re,T to test this theory. The data were plotted for each ense mble across all protocol s and a regression line was calculated. All of the gra phs are presented in Appendix H, and the graph for Phase 1 Ensemble A is shown in Figure 12. It is obvious that Re,T does not remain constant as P changes, and the rate of change (slope) appe ars constant within Phase 1 for the different ensembles suggesting that environment is a f actor. However, the Phase 2 graphs do not show any consistent effect for activity. Th e Phase 2 data is confounded by the metabolic
61 rate and therefore there isnÂ’t an expected effect. Regression analysis performed on the data resulted with Phase 2 ensembles ha ving an R-square value of 0.002 Â– 0.115, while the Phase 1 ensembles ranged from 0.388 Â– 0.617. Phase 1 Ensemble A0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 11.522.533.544.55 PRe,T Figure 12. Effect of Environment: Re,T vs. P Â– Ensemble A. Table 10 presents the slopes and inter cepts of the regression analysis. By reviewing the slope va lues, environment (Phase 1) a ppears to have consistent and significant slope values across the ensembles. However, the values for Phase 2 suggest the slopes are neither cons istent nor significant. A lthough Ensemble A in Phase 2 presents a good positive slope, all of the other ensembles do not.
62 Table 10. Regression Analysis Â– P by Re,T R2SlopeInterceptR2SlopeIntercept A0.6170.00010.00040.1150.02440.0001 B0.4530.00010.00250.0180.39510.0007 C0.4770.00010.00370.0010.84780.0001 D0.3880.00010.00020.0010.84200.0009 E0.5340.00010.74740.0050.64300.0006 Phase 1Phase 2 Ensemble Phase 2 Similar to Phase 1, the methodology used was able to distinguish effectively between the selected ensembles as illustr ated in Figure 6, showing very significant differences among Ensembles D and E as compared to Ensembles A, B, and C. Figure 7 indicates Re,T decreases as the metabolic rate in creases. The interaction between ensemble and metabolic rate is clearly seen by observing the differences between metabolic rates within each ensemble in crease corresponding to the reduction to Re,T. There is not much difference in the mean Re,T values between the metabolic rates within Ensemble A indicating good evaporativ e cooling (high permeability). However, progressing through the ensembles, the di fferences between protocols increases indicating the metabolic rate (activity) plays an increasing role in lowering the mean Re,T. Again, this relationship was veri fied by using JMP-IN to test the differences between the protocols for each ensemble. Ensemble A a nd B were not significantly different whereas the others (Ensembles C, D, and E) all resu lted as being very signi ficantly different.
63 As discussed previously, increasing the me tabolic rate results with decreasing the mean Re,T values. This effect is presented di fferently in Figure 13 where the difference among ensembles is distinct for M1, but change s as the metabolic demand increases. For M1, there appears to be a step effect between the ensembles. In M2 and more so in M3, this step effect disappears as the ensemb les appear to reach the lower limit of Re,T. 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 M1M2M3Metabolic RateRe,T A B C D E Figure 13. Effect of Metabolic Rate on Re,T. Havenith et al.  and Pa rsons et al.  explain th e relative decrease in Re,T due to the increased air movement through the clot hing. They use the term pumping action to explain that as an individual moves, air is pumped into and out of their clothing. As the air moves through the clothing, the effective IT and Re,T decreases due to increased convective and evaporative cooling.
64 CONCLUSIONS The primary purpose of this research was to explore the differences among garments based on the total evaporativ e resistance proper ties among different environments and work demands. The secondary purpose was to challenge the relationship of Re,T with respect to changes in P and Emax. Experimental trials were conducted to determine the evaporative resistan ce for five clothing ensembles per phase. The protocols included a fixed metabolic demand under three different environments (levels of relative humidity) for Phase 1, and a fixed relative humidity level with three metabolic demands with for Phase 2. The fundamental step in these studies was being able to distinguish the point just before the transition of compensable heat stress to uncompensable heat stress (Ereq = Emax). Statistical analysis of the data showed that the methodology used was able to distinguish well among the select ed ensembles. Data from Phase 1 found that Ensemble E was different from all others and Ensemb le D was different fr om A and B. More importantly, the data revealed a relations hip with the environment. The mean Re,T values for each ensemble decreases as the hu midity increased. The changes to Re,T due to environment were explored further.
65 The default assumption has been that Re,T remains constant as P changes. This relationship between Re,T and P was challenged and found that Re,T does not stay constant as generally accepted. Envi ronment (relative humidity) effects Re,T as well as P. This relationship needs to be studi ed further before it is fully understood. The Phase 2 analysis resulted with Ense mbles D and E being different from all other ensembles. As expected, with increased activity mean Re,T values decreased. Ensembles D and E had the biggest decreases in Re,T, while Ensembles A, B, and C appeared to reach a lower limit associated with the ensemble permeability properties. The decrease in Re,T from metabolic demand was relate d to the pumping action of air through the ensemble from movement. The null hypothesisÂ’ were rejected for al l three hypothesisÂ’ tested. The data shows (1) there are differences between Re,T values within ensembles, (2) there are differences between Re,T values within ensembles and between the different metabolic rates/demands, and (3) Re,T does not remain constant while P changes.
66REFERENCES 1. Belding, H.S. and T.F. Hatch, Index for evaluating heat stress in terms of resulting physiological strain. Journal of the American Society of Heating and Ventilation Engineers, 1955(Heating, Pi ping and Air Conditioning Section 27): p. 129-135. 2. NIOSH, Occupational Exposure to Hot Environments 1986, U.S. Department of Health and Human Services: Washington DC. 3. Bernard, T.E., F.N. Dukes-Dobos, and J.D. Ramsey, Evaluation and control of hot working environments: Part II The scientific basis (knowledge base) for the guide. International Journal of Industrial Ergonomics, 1994. 14 : p. 129-138. 4. Havenith, G., Heat balance when weari ng protective clothing. Annals of Occupational Hygiene, 1999. 43 (5): p. 289-96. 5. Holmer, I., et al., Clothing convective heat exchange--proposal for improved prediction in standards and models. Annals of Occupational Hygiene, 1999. 43 (5): p. 329-37. 6. Goldman, R.F., Standards for human exposure to heat in Environmental Ergonomics: Sustaining Human Perf ormance in Harsh Environments I.B. Mekjavic, E.B. Banister, and J.B. Mo rrison, Editors. 1988, Taylor & Francis: Philadelphia. p. 99-136. 7. Parsons, K., Human thermal Environments: The effe cts of hot, moderate and cold environments oh human health, comfort and performance 2 ed. 2003, New York: Taylor and Francis. 8. ISO9920, Ergonomics of the thermal environm ent: Estimation of the thermal insulation and evaporative resistance of a clothing ensemble 1995, International Organization for Standardization: Geneva. 9. Bernard, T.E. and F. Matheen, Evaporative resistance and sustainable work under heat stress conditions for tw o cloth anticontamination ensembles. International Journal of Industrial Ergonomics, 1999. 23 : p. 556-564.
67 10. ASTM, F1868-02 Standard Test Method for Thermal and Evaporative Resistance of Clothing Materials Us ing a Sweating Hot Plate 2003, Conshohocken, PA: American Society for Testing and Materials. 11. McCullough, E.A., Thermal Insulation Values for Clothing Ensembles 2003, Institute for Environmental Research Kansas State University: Manhattan. 12. ASTM, Annual Book of ASTM Standards -Part 11.03 2002, Conshohocken, PA: American Society for Testing and Materials. 13. Nilsson, H.O., Comfort Climate Evaluation with Thermal Manikin Methods and Computer Simulation Models ed. D.o.C.a.A. Engineering and S. Royal Institute of Technology. 2004, Stockholm: National Institute for Working Life. 14. Fan, J. and X. Qian, New functions and applications of Walter, the sweating fabric manikin. European Journal of Applied Physiology, 2004. 12 : p. 12. 15. Lind, A.R., A physiological criterion for setti ng thermal environmental limits for everyday work. Journal of Applied Physiology, 1963. 18 : p. 51-6. 16. Belding, H.S. and E. Kamon, Evaporative coefficients fo r the prediction of safe limits in prolonged exposures to work under hot conditions. Federal Proceedings, 1973. 32 (5): p. 1598-1601. 17. Kamon, E. and B. Avellini, Physiologic limits to work in the heat and evaporative coefficient for women. Journal of Applied Physiology, 1976. 41 (1): p. 71-6. 18. Holmer, I. and S. Elnas, Physiological evaluation of the resistance to evaporative heat transfer by clothing. Ergonomics, 1981. 24 (1): p. 63-74. 19. Kenney, W.L., D.E. Hyde, and T.E. Bernard, Physiological Evaluation of LiquidBarrier, Vapor-Permeable Protective Cl othing Ensembles for Work in Hot Environments. American Industrial H ygiene Association, 1993. 54 (7): p. 397-402. 20. Chengalur, S.N., S.H. Rodgers, and T.E. Bernard, Kodak's Ergonomic Design for People at Work 2nd ed. 2004: John Wiley & Sons, Inc. 21. Kenney, W.L., et al., Simultaneous derivation of cl othing-specific heat exchange coefficients. Medical Science of Sports Exercise, 1993. 25 (2): p. 283-9. 22. Dubois, A. and E.F. Dubois, The Measurement of the Surface Area of Man. Archives of Internal Medicine, 1950. 15 : p. 868-881.
68 23. Barker, D.W., S. Kini, and T.E. Bernard, Thermal characteristics of clothing ensembles for use in heat stress analysis. American Industrial Hygiene Association Journal, 1999. 60 (1): p. 32-7. 24. McCullough, E.A., B.W. Jones, and T. Tamura, A Data Base for Determining the Evaporative Resistance of Clothing. ASHRAE Transactions, 1989. 95 (2): p. 316328. 25. Havenith, G., R. Heus, and W.A. Lotens, Resultant clothing insulation: a function of body movement, posture, wind, clot hing fit and ensemble thickness. Ergonomics, 1990. 33 (1): p. 67-84. 26. Nishi, Y., R.R. Gonzales, and A.P. Gagge, Direct measurements of clothing heat transfer properties during sensible and insensible heat exchange with thermal environment. ASHRAE Transactions, 1975. 81 : p. 183-199. 27. Vogt, J.J., et al., Pumping effects on thermal insu lation of clothing worn by human subjects. Ergonomics, 1983. 26 (10): p. 963-74. 28. Nielsen, R., B.W. Olesen, and P.O. Fanger, Effect of physical activity and air velocity on the thermal insulation of clothing. Ergonomics, 1985. 28 (12): p. 161731. 29. Olesen, B.W., et al., Effect of body posture and activ ity on the thermal insulation of clothing: measurements by a moveable thermal manikin. ASHRAE Transactions, 1982. 88 : p. 791-805. 30. Havenith, G., R. Heus, and W.A. Lotens, Clothing ventilation, vapour resistance and permeability index: changes due to posture, movement and wind. Ergonomics, 1990. 33 (8): p. 989-1005. 31. Parsons, K.C., et al., The effects of wind and human movement on the heat and vapour transfer properties of clothing. Annals of Occupational Hygiene, 1999. 43 (5): p. 347-52. 32. Havenith, G., et al., Clothing evaporative heat resist ance--proposal for improved representation in standards and models. Annals of Occupational Hygiene, 1999. 43 (5): p. 339-46. 33. Craig, F.N., Evaporative cooling of men in wet clothing. Journal of Applied Physiology, 1972. 33 (3): p. 331-6.
69 34. Bernard, T.E. Differences in Total Evaporative Resistance Due to Environment and Activity as Observed for Three Clothing Ensembles in International Conference on Environmental Ergonomics 1998. San Diego, CA: International Conference on Environmental Ergonomics. 35. Consolazio, C.R., R.E. Johnson and L.J. Pecora. Physiological Measurements of Metabolic Functions in Man New York: McGraw-Hill (1963). 36. Kuzma, J.W. and S.E. Bohnenblust, Basic Statistics for the Health Sciences 4 ed. 2001, Mountain View, California: Mayfield Publishing Company. 364. 37. Littell, R.C., et al., SAS System for Mixed Models 1996, Cary, NC: SAS Institute Inc.
70 APPENDIX A PARTICIPANT DATA
71 Table A1. Characteristics of Par ticipants in Experimental Trials. Participant Sex Age (years) Height (cm) Weight (kg) Surface Area (m2) Y1S1 M 26 180 95 2.14 Y1S2 F 26 163 52 1.55 Y1S3 M 24 183 86 2.08 Y1S4 M 25 183 77 1.99 Y1S5 F 23 152 63 1.59 Y1S6 F 27 170 91 2.02 Y1S7 M 35 189 101 2.28 Y1S8 F 39 155 46 1.42 Y1S9 M 20 183 130 2.48 Y1S10 M 30 191 110 0.00 Y1S11 M 32 173 71 1.84 Y1S12 M 43 178 112 2.28 Y1S13 M 28 185 95 2.19 Y1S14 F 44 165 65 1.72 Y2S1 F 27 163 52 1.55 Y2S2 M 28 185 95 2.19 Y2S3 F 27 170 91 2.02 Y2S4 M 26 180 95 2.15 Y2S5 M 27 175 98 2.13 Y2S6 M 20 180 83 2.03 Y2S7 M 20 183 72 1.93 Y2S8 M 24 163 64 1.68 Y2S9 M 43 149 75 1.69 Y2S10 M 49 175 86 2.02 Y2S11 F 18 170 56.8 1.66 Y2S12 F 20 157 56.8 1.57 Y2S13 M 21 185 81.8 2.06 Y2S14 M 22 175 66 1.80 Y2S15 M 28 185 86 2.11 Average Std Dev 28.3 8.1 173.9 11.5 81.1 20.1 1.87 0.45
72 APPENDIX B EXPERIMENTAL DATA Â– PHASE 1
Appendix B Experimental Data Â– Phas e 1: Data Dictionary 73 Title Description Code Participant Code Gender Gender of participant Proto Protocol Design: Environment (R2 (20% rh), R5 (50% rh ), R7 (70% rh)) or Metabolic Demand (M1 (80 W/m2), M2 (160 W/m2), M3 (240 W/m2)) Ens Ensemble: (A (work clothes), B (cott on coveralls), C (particle barrier), D (liquid barrier), E (vapor barrier)) Tdb Ambient air temperature (d ry bulb) in degrees Celsius Tpwb Wet bulb air temperature in degrees Celsius Tg Black bulb air temperature in degrees Celsius S(m/s) Speed in meters per second G(%) Grade of treadmill in percentage HR Heart rate Tre Body core temperature (rectal) Tch Skin temperature at the chest Tarm Skin temperature at the upper arm Tth Skin temperature at the thigh Tcalf Skin temperature at the calf Met Calculated metabolic work based on O2 consumption in Watts BSA Body surface area in square meters MSA Met divided by the BSA (W/m2) Tsk Average Skin temperature Psk Partial pressure of the water vapor at the skin Pv Partial pressure of the water vapor in the air Psk-Pv P: Difference between Psk and Pv Tair-Tsk T: Difference between Tdb and Tsk ReT Total evaporative resistance
74 Appendix B (Continued) Experimental Data Â– Phase 1 CodeGenderProtoEnsTdbTpwbTg S(m/s)G (%)HRTreTchTarmTthTcalfTskPskPvBSAMetMSAPsk-PairTair-TskReT S0MR2A47.727.447.71.31010237.734.435.036.036.135.22.214.171.124511643.412.50.015 S0MR5A42.331.342.31.31010338.036.636.136.036.93126.96.36.199.143651712.25.90.011 S0MR7A32.528.432.51.3008937.7188.8.131.524.6184.108.40.206.143111451.9-2.10.014 S0MR2B45.6220.127.116.1109437.636.135.035.237.18.104.22.168.143471623.29.80.015 S0MR5B40.429.940.41.3009637.835.635.236.335.422.214.171.124.143191492.34.90.013 S0MR7B37.431.937.41.30010437.835.536.036.036.235.126.96.36.199411601.51.50.009 S0MR7B33.629.633.61.3008637.834.634.935.033.5188.8.131.52.14208971.6-0.90.017 S0MR7B35.329.935.31.3209637.735.734.236.035.235.184.108.40.206841801.80.10.010 S0MR2C45.825.344.71.3109537.935.436.636.036.836.26.01.82.144692194.29.60.015 S0MR5C38.828.037.11.30011038.036.635.435.835.8220.127.116.11.145322492.82.90.011 S0MR7C35.129.033.61.3009537.436.135.633.834.918.104.22.168.144322022.1-0.10.010 S0MR2D44.624.844.61.30011237.735.536.036.636.436.05.91.82.143561674.18.60.019 S0MR5D39.629.439.61.3309537.735.535.934.835.922.214.171.124.143251522.44.10.014 S0MR5D37.227.937.21.3009337.735.434.235.633.4126.96.36.199.143081442.42.50.015 S0MR7D33.628.032.61.30010238.036.735.936.533.9188.8.131.52.144231982.5-2.30.013 S0MR2E31.316.829.41.3109537.436.635.135.735.7184.108.40.206.143961854.9-4.50.031 S0MR5E29.322.028.01.3109537.436.3220.127.116.1118.104.22.168.144482103.6-6.20.021 S0MR7E30.325.028.41.30010837.636.636.435.134.922.214.171.124.145152413.1-5.60.015 S1FR2A51.628.051.61.34015337.835.136.736.540.4126.96.36.199.541561014.014.70.022 S1FR5A41.030.141.01.33012638.035.235.635.836.3188.8.131.52.541861212.35.30.015 S1FR7A35.230.035.21.34013437.735.836.135.935.9184.108.40.206.542161402.0-0.70.015 S1FR7A36.029.536.01.34015037.935.534.635.734.335.05.63.71.541851201.91.00.016 S1FR2B53.229.553.21.34012738.035.937.637.036.9220.127.116.11.542121373.716.40.016 S1FR5B41.931.041.91.34010138.336.336.936.436.518.104.22.168.541801172.45.40.016 S1FR7B34.830.034.81.43012737.735.335.035.522.214.171.124.91.541681091.8-0.40.017 S1FR2C51.524.049.81.361.512538.036.637.637.337.037.16.31.11.542601685.214.40.021 S1FR5C43.230.542.01.34013937.936.136.636.937.4126.96.36.199.541861212.66.60.017 S1FR7C36.129.836.11.34012737.734.935.235.935.6188.8.131.52.542101361.90.80.014 S1FR2D49.624.549.61.40013738.137.537.039.038.037.184.108.40.2061661085.111.90.030 S1FR5D39.129.939.11.34011438.036.436.636.735.7220.127.116.11.542091352.42.80.016 S1FR7D34.818.104.22.168011438.135.535.735.635.622.214.171.124.54152982.2-0.80.023 S1FR2E37.318.535.41.341.512137.434.532.736.236.4126.96.36.199.542831834.62.60.023 S1FR5E36.226.736.21.43015838.235.737.537.435.8188.8.131.52.541661083.3-0.40.031 S1FR7E32.026.030.01.361.514237.834.935.035.634.635.05.63.01.542761792.7-3.00.016
75 Appendix B (Continued) Experimental Data Â– Phase 1 CodeGenderProtoEnsTdbTpwbTg S(m/s)G (%)HRTreTchTarmTthTcalfTskPskPvBSAMetMSAPsk-PairTair-TskReT S2MR2A51.028.049.31.21011037.936.335.735.9184.108.40.206.22.082461183.914.40.020 S2MR5A40.530.440.21.31011237.536.035.335.735.8220.127.116.11.082811352.24.80.013 S2MR7A39.630.938.71.27010937.736.635.536.134.618.104.22.168.082751322.03.90.013 S2MR2B51.028.851.01.18010038.035.636.837.338.522.214.171.124.083051473.814.10.017 S2MR5B43.530.542.91.19010937.736.135.736.036.336.05.93.52.082631272.47.50.015 S2MR7B36.831.035.61.17010938.036.835.835.335.4126.96.36.199.083541701.80.90.010 S2MR2C45.125.345.11.2009637.6188.8.131.526.136.26.01.92.082621264.18.90.023 S2MR5C41.530.040.71.19011437.836.836.436.539.8184.108.40.206.083061472.94.30.017 S2MR7C36.530.035.41.28010237.635.935.835.835.6220.127.116.11.082941412.10.70.014 S2MR2D49.524.048.51.28011337.936.836.836.336.036.56.11.32.082871384.813.00.023 S2MR5D36.718.104.22.16810922.214.171.1245.6126.96.36.199.72.082521213.11.10.024 S2MR7D38.531.337.21.24011137.936.836.836.338.337.06.34.12.082761332.21.50.015 S2MR2E32.418.532.21.19010237.836.736.536.336.5188.8.131.52.082731314.9-4.10.045 S2MR5E30.922.030.71.20010037.535.535.135.935.8184.108.40.206.083061473.7-4.60.031 S2MR7E220.127.116.11.19012237.937.037.036.737.237.06.33.02.082851373.3-4.80.030 S3MR5A42.030.241.51.33011338.135.736.035.935.818.104.22.168.982851442.46.20.013 S3MR7A36.232.035.11.33012338.336.036.335.635.422.214.171.124.982711371.40.30.010 S3MR2B53.7126.96.36.199011337.935.737.236.3188.8.131.52.71.982621323.417.10.015 S3MR5B43.931.443.51.34011137.9184.108.40.2066.436.36.03.81.982831432.37.60.012 S3MR7B36.531.035.71.32011938.035.836.636.435.836.26.04.11.983291661.90.30.011 S3MR2C45.522.045.51.32010037.934.234.634.836.3220.127.116.11.982671354.510.60.023 S3MR5C40.528.540.01.32010538.235.918.104.22.16822.214.171.124.982831432.74.80.016 S3MR7C36.631.835.71.32011838.035.636.536.536.136.26.04.41.982871451.60.50.011 S3MR2D44.324.743.81.33011538.135.436.234.337.9126.96.36.199.982061044.18.40.027 S3MR5D43.232.542.31.32013737.937.237.437.537.6188.8.131.52.982201112.25.80.016 S3MR5D184.108.40.206.32013238.035.836.935.836.436.36.03.61.982061042.55.00.019 S3MR5D40.429.339.21.320106220.127.116.1118.104.22.168.13.31.982441232.74.00.019 S3MR7D38.131.536.51.32011238.036.536.336.435.636.26.04.21.982821421.81.80.012 S3MR2E33.819.032.21.32010037.535.435.036.236.035.22.214.171.1242791414.6-1.80.035 S3MR5E30.022.029.01.3409337.534.333.835.133.5126.96.36.199.982931483.3-4.10.026 S3MR7E32.326.528.41.33010037.935.435.536.5188.8.131.52.11.982421222.8-3.50.027
76 Appendix B (Continued) Experimental Data Â– Phase 1 CodeGenderProtoEnsTdbTpwbTg S(m/s)G (%)HRTreTchTarmTthTcalfTskPskPvBSAMetMSAPsk-PairTair-TskReT S4FR2A46.825.845.41.20011637.536.436.736.7184.108.40.206.91.59142894.210.30.029 S4FR5A220.127.116.11.20013638.035.737.135.535.736.16.03.61.591841162.45.10.017 S4FR7A38.231.538.21.19013037.836.335.736.536.236.26.04.21.592351481.82.10.011 S4FR2B47.725.846.91.19013238.135.037.936.638.036.86.21.91.59134844.310.90.030 S4FR5B18.104.22.168.20012438.036.237.736.336.922.214.171.124.591661042.45.40.018 S4FR7B41.132.039.71.20013538.236.838.036.737.037.26.34.11.59149942.23.90.019 S4FR2C48.326.747.31.20012337.6126.96.36.1996.8188.8.131.52.59135854.111.60.028 S4FR5C46.927.446.71.20012937.635.437.135.936.9184.108.40.206.59118743.710.60.028 S4FR7C220.127.116.11.20012737.735.835.135.834.618.104.22.168.59122772.1-0.30.028 S4FR2D37.331.336.01.20013237.836.036.636.335.836.26.04.21.5994591.81.10.028 S4FR2D46.127.044.21.20014038.037.637.337.337.622.214.171.124.592341474.18.70.021 S4FR5D38.526.537.51.20012737.736.236.936.435.6126.96.36.199.59157993.42.20.031 S4FR2E35.018.433.11.20013437.835.937.437.136.936.86.21.01.591611015.2-1.80.057 S4FR5E35.424.033.71.19013037.734.937.036.836.336.26.02.21.591921213.8-0.80.033 S4FR7 E38.932.437.51.20011737.436.636.4188.8.131.52.14.41.592411511.62.60.010 S5FR2A49.929.048.91.09012137.536.536.936.8184.108.40.206.62.022861423.613.10.017 S5FR5A41.929.740.81.05010834.336.336.436.536.3220.127.116.11.022421202.75.50.018 S5FR7A37.031.236.51.08012337.736.936.936.418.104.22.168.12.023241612.00.30.012 S5FR2B45.527.844.51.05011337.536.537.036.822.214.171.124.62.02182903.78.70.027 S5FR5B41.829.539.31.09011337.336.136.336.036.336.26.03.32.022931452.75.60.015 S5FR7B38.731.536.71.09011437.536.436.336.135.936.26.04.12.023071521.92.50.011 S5FR2C39.123.039.01.04010038.035.736.135.535.3126.96.36.199.022871424.13.40.026 S5FR5C42.230.840.91.09012137.736.837.036.836.6188.8.131.52.022801392.55.40.015 S5FR7C36.830.635.51.05010537.336.335.735.535.635.85.94.02.022351161.91.00.016 S5FR2D47.029.347.21.09011937.336.736.836.6184.108.40.206.92.023061523.310.20.016 S5FR5D40.028.838.51.11011737.536.836.936.536.6220.127.116.11.022741363.03.30.019 S5FR7D36.730.835.31.05011437.336.618.104.22.1686.36.04.02.02177882.00.40.022 S5FR2E32.619.031.01.05010937.636.136.035.935.836.05.91.32.022781384.6-3.40.039 S5FR5E32.823.531.61.09010737.435.936.036.236.036.05.92.32.022751363.7-3.20.031 S5FR7E22.214.171.124.09010137.336.335.5126.96.36.199.82.72.023131553.1-3.50.023
77 Appendix B (Continued) Experimental Data Â– Phase 1 CodeGenderProtoEnsTdbTpwbTg S(m/s)G (%)HRTreTchTarmTthTcalfTskPskPvBSAMetMSAPsk-PairTair-TskReT S6MR2A51.327.350.21.1209237.635.836.836.7188.8.131.52.02.283691624.114.70.017 S6MR5A43.431.341.91.1309637.536.435.936.036.036.16.03.72.282991312.27.30.013 S6MR7A40.933.339.61.11011537.936.736.736.536.7184.108.40.206.283971741.64.30.008 S6MR2B48.925.548.01.1309036.735.035.936.335.8220.127.116.11.284371924.113.20.016 S6MR5B45.233.044.11.1309637.736.336.336.418.104.22.168.22.283101361.98.70.010 S6MR7B42.034.539.31.1101122.214.171.1246.636.936.96.25.02.283961741.35.10.006 S6MR2C49.626.548.41.1309537.6126.96.36.1997.036.36.01.92.283331464.113.30.019 S6MR5C41.429.539.91.1209137.536.3188.8.131.526.26.03.32.283571572.75.20.014 S6MR7C37.631.036.21.1309037.536.036.035.935.836.05.94.12.283631601.91.60.011 S6MR7C41.033.039.21.1609737.4184.108.40.2066.536.36.04.52.283611591.54.70.008 S6MR2D46.325.044.81.13010237.536.636.736.4220.127.116.11.72.284261874.49.60.018 S6MR5D39.127.037.91.1308837.636.435.435.935.518.104.22.168.283581573.13.30.018 S6MR7D36.630.035.51.12010437.636.636.136.336.236.36.03.82.283671612.20.20.014 S6MR2E34.418.032.41.1208437.434.736.036.036.035.65.81.02.283771664.8-1.20.030 S6MR2E33.728.532.41.1309237.535.135.922.214.171.124.93.52.283751652.3-2.10.015 S6MR5E33.925.032.11.1608537.535.736.135.833.9126.96.36.199.284081793.2-1.60.019 S7FR2A54.028.053.01.37213937.836.536.639.539.037.66.52.01.422431714.516.40.017 S7FR5A41.931.340.41.39213637.936.735.636.836.8188.8.131.52.422962092.25.50.009 S7FR7A37.832.036.51.39214438.337.036.637.237.337.06.34.41.422761951.90.80.010 S7FR2B50.925.849.21.42212237.735.536.639.037.036.86.21.61.422411704.614.10.018 S7FR5B45.133.343.71.38214338.037.335.637.036.5184.108.40.206.422842001.88.50.007 S7FR7 B36.430.035.41.36211737.736.535.636.336.336.26.03.81.423322342.20.20.009 S7FR2C49.425.047.91.39211937.736.536.537.843.6220.127.116.11.422261595.211.20.023 S7FR5C42.230.839.11.29211837.836.536.936.636.818.104.22.168.422571812.55.50.012 S7FR7C39.733.038.51.39213038.137.136.537.137.036.96.24.61.422641861.72.80.008 S7FR2D43.922.342.21.40211637.936.736.436.937.622.214.171.124.422671885.07.10.022 S7FR5D43.431.342.21.39213437.937.436.235.7126.96.36.199.71.422751942.56.60.011 S7FR7D37.232.035.81.39213438.037.335.936.934.836.36.04.41.422801981.60.90.008 S7FR2E34.518.533.41.28212337.836.535.836.736.6188.8.131.52.423002125.0-1.90.025 S7FR5E34.025.032.31.39211337.636.537.137.036.43184.108.40.206.422641863.6-2.80.021 S7FR7E34.827.533.01.39211337.736.536.236.636.9220.127.116.11.422922062.9-1.70.015
78 Appendix B (Continued) Experimental Data Â– Phase 1 CodeGenderProtoEnsTdbTpwbTg S(m/s)G (%)HRTreTchTarmTthTcalfTskPskPvBSAMetMSAPsk-PairTair-TskReT S8MR2A52.530.050.81.02012837.735.736.836.338.036.66.12.72.483231303.415.90.016 S8MR5A45.433.544.31.01014637.636.936.636.236.036.56.14.42.484301741.78.90.008 S8MR7A42.335.541.11.01014137.936.938.237.536.918.104.22.168.484311741.14.90.005 S8MR2B52.427.851.11.03012437.435.736.036.237.936.36.02.12.484491813.916.00.015 S8MR5B45.033.544.01.02014637.536.836.836.537.422.214.171.124.484251721.88.20.008 S8MR7B39.030.537.71.00012137.535.435.236.034.53126.96.36.199.484401781.93.70.010 S8MR2C55.029.058.91.03013737.736.936.037.436.036.56.12.32.484661883.918.50.013 S8MR5C41.031.039.91.02010937.135.434.9188.8.131.52.83.82.484221702.05.50.010 S8MR7C38.731.537.51.01013137.537.236.036.236.4184.108.40.206.484271722.02.20.011 S8MR2D48.626.047.01.03013237.836.837.236.936.4220.127.116.11.484831954.411.70.017 S8MR5D43.431.542.11.02012937.636.734.536.436.035.85.93.82.484321742.17.60.010 S8MR7D37.932.036.91.02011837.336.336.036.836.036.36.04.42.484571851.71.70.009 S8MR2E32.017.331.11.03010437.034.834.935.618.104.22.168.02.483971604.7-3.30.033 S8MR5E31.022.529.91.01010337.535.434.935.435.035.25.72.22.484301743.5-4.20.023 S8MR7E34.628.033.51.01012637.434.136.336.234.622.214.171.124.484191692.4-0.70.014 S8MR7E32.325.831.11.03011937.336.135.835.735.6126.96.36.199.484301743.0-3.50.020 S9MR2A51.127.049.91.0308737.336.235.736.737.536.46.12.02.383911654.114.70.017 S9MR5A42.331.041.41.03010837.636.436.436.337.036.56.13.72.383561502.45.80.013 S9MR7A38.331.036.61.0309937.636.736.037.536.636.66.14.02.383771592.11.70.013 S9MR2B37.425.036.31.0308537.036.436.636.737.3188.8.131.52.383751583.80.70.024 S9MR5B41.231.540.01.03010237.436.636.636.136.736.56.14.02.383511482.14.70.012 S9MR2C53.928.052.51.04010137.537.536.837.137.837.36.42.02.383421444.316.60.018 S9MR5C43.832.342.41.04010538.037.237.036.737.037.06.34.12.383761582.26.80.011 S9MR7C37.431.536.11.06010737.738.135.936.536.3184.108.40.206.383831612.00.60.012 S9MR2D43.424.041.61.0308937.337.136.636.136.6220.127.116.11.383651544.56.70.023 S9MR5D38.829.538.51.0409738.236.436.636.636.518.104.22.168.383401432.62.30.017 S9MR7D36.330.535.01.04010137.636.635.936.735.436.26.04.02.383451452.00.10.014 S9MR2E33.118.031.61.0309522.214.171.124.236.236.16.01.12.384111734.9-3.00.032 S9MR5E31.922.830.51.0308937.436.636.336.3126.96.36.199.22.383651543.9-4.50.030 S9MR7E31.625.030.81.0308737.036.335.9188.8.131.52.02.72.383701563.2-4.50.025
79 Appendix B (Continued) Experimental Data Â– Phase 1 CodeGenderProtoEnsTdbTpwbTg S(m/s)G (%)HRTreTchTarmTthTcalfTskPskPvBSAMetMSAPsk-PairTair-TskReT S10MR2A46.925.045.41.34013036.736.736.837.4184.108.40.206.71.833161734.69.80.020 S10MR5A39.729.538.61.32013137.736.836.836.836.8220.127.116.11.833091692.82.90.015 S10MR7A37.431.036.11.31014238.136.837.036.236.9318.104.22.168.833321812.10.60.012 S10MR2B49.525.047.91.32013337.736.437.737.638.422.214.171.124.833301804.912.10.020 S10MR5B39.629.338.11.30013538.136.436.736.936.5126.96.36.199.833261782.83.00.014 S10MR7B32.227.031.51.32012438.536.834.637.036.236.16.03.21.833261782.8-3.90.018 S10MR2C46.423.545.41.34013637.936.637.037.338.4188.8.131.52.833011645.09.20.023 S10MR5C39.528.837.81.31013137.836.536.937.237.5184.108.40.206.832811533.02.60.018 S10MR7C37.330.835.91.31012537.736.336.136.636.736.46.14.01.833231762.10.80.011 S10MR2D43.622.542.31.31013237.736.636.836.3220.127.116.11.31.833211754.96.90.023 S10MR5D37.727.036.51.32012937.836.836.436.737.036.76.22.91.833201753.31.00.018 S10MR7D36.629.035.21.32014638.418.104.22.1687.037.56.43.51.833211752.9-0.80.017 S10MR2E32.316.030.71.31013238.036.936.537.136.722.214.171.124.833411865.5-4.50.034 S10MR5E32.822.531.01.32012837.836.835.236.836.036.26.02.01.833291804.0-3.40.025 S10MR5E33.124.032.51.32014638.336.836.237.636.9126.96.36.199.833371843.8-3.70.023 S10MR7E33.227.031.81.24013338.036.536.136.436.5188.8.131.52.832861562.9-3.20.021 S11MR2A52.228.051.71.03012537.736.336.637.237.4184.108.40.206.284361924.015.40.015 S11MR5A44.032.342.31.03011537.837.037.037.037.037.06.34.02.284271882.27.00.010 S11MR7A40.532.538.51.03013438.237.337.737.136.4220.127.116.11.284471962.03.30.009 S11MR2B52.427.050.91.04010937.735.836.236.036.036.05.91.92.284211854.116.40.015 S11MR5B44.232.543.41.03011037.738.837.036.918.104.22.168.12.284451962.46.60.010 S11MR7B39.631.638.41.03011037.837.236.836.636.522.214.171.124.284171832.12.80.011 S11MR2C52.126.049.91.03010337.636.436.337.137.7126.96.36.199.283901714.615.30.018 S11MR5C42.431.041.41.07011037.736.035.336.336.736.05.93.72.284592022.26.40.009 S11MR7C36.829.836.51.03011837.937.236.336.335.8188.8.131.52.284181842.40.30.013 S11MR2D47.425.546.41.03012837.937.336.937.337.5184.108.40.206.284181844.610.10.019 S11MR5D40.429.338.91.03011237.436.436.735.136.336.26.03.32.283941732.74.20.014 S11MR7D36.129.334.61.03010537.636.537.436.336.4220.127.116.11.283661612.6-0.70.016 S11MR2E36.019.034.71.02011737.836.937.437.318.104.22.168.12.283851695.3-1.20.032 S11MR5E30.923.529.81.02011737.936.936.936.636.722.214.171.124.284151823.8-5.90.026 S11MR7E32.427.031.01.02011937.636.836.836.736.6126.96.36.199.283881703.0-4.40.020
80 Appendix B (Continued) Experimental Data Â– Phase 1 CodeGenderProtoEnsTdbTpwbTg S(m/s)G (%)HRTreTchTarmTthTcalfTskPskPvBSAMetMSAPsk-PairTair-TskReT S12MR2A52.726.051.21.08010437.335.437.037.037.036.188.8.131.52161454.516.20.019 S12MR5A43.131.041.71.07011937.836.737.036.936.6184.108.40.206.183231482.56.30.014 S12MR7A37.731.036.51.09010537.335.435.6220.127.116.11.84.02.183061401.82.10.012 S12MR2B53.728.052.31.08011337.536.536.536.638.036.86.22.12.183021384.216.90.018 S12MR5B42.130.539.91.05012237.535.634.936.334.818.104.22.168.18151692.16.70.020 S12MR5B43.133.042.01.08010337.636.736.836.736.722.214.171.124.183041391.86.40.010 S12MR7B38.631.536.51.07011037.836.336.136.535.336.16.04.22.182961361.82.50.012 S12MR2C53.927.952.81.07011637.736.336.236.9126.96.36.199.02.183331524.217.10.017 S12MR5C42.130.040.21.07011037.536.636.536.636.7188.8.131.52.183101422.75.50.016 S12MR7C39.332.037.61.07010537.736.236.336.736.5184.108.40.206.183061401.82.90.011 S12MR2D45.422.544.01.07011537.436.436.737.036.3220.127.116.11.183061404.98.80.026 S12MR5D38.728.537.51.07010537.836.636.536.836.418.104.22.168.183161452.92.10.019 S12MR5D40.333.538.81.09012137.535.537.237.437.322.214.171.124.183191461.53.60.009 S12MR7D36.029.034.51.06010937.936.436.036.235.236.05.93.52.183331522.40.00.016 S12MR2E33.417.032.21.0809737.335.636.036.536.136.05.90.82.183171455.1-2.60.039 S12MR5E30.523.030.01.0809937.636.235.935.936.236.16.02.32.183251493.6-5.60.031 S12MR7E32.826.532.01.07010337.236.336.235.735.736.06.03.02.183001372.9-3.20.024 S13FR2A56.228.554.51.27013238.336.537.537.537.537.26.32.01.713121824.319.00.015 S13FR5A43.733.042.31.28012238.436.936.938.437.5126.96.36.199.712631542.16.40.011 S13FR7A40.333.838.81.28014338.636.936.736.837.3188.8.131.52.713201871.43.40.007 S13FR2B54.927.053.11.27011837.936.837.037.638.6184.108.40.206.713081804.717.50.017 S13FR5B45.834.044.41.27012838.337.137.337.137.7220.127.116.11.712891691.88.50.009 S13FR7B38.231.036.41.28012638.735.936.036.235.135.85.94.01.713231891.92.40.009 S13FR2C54.028.051.61.27012038.136.437.237.818.104.22.168.01.712641544.316.70.017 S13FR5C41.830.040.41.28012238.237.835.937.138.037.16.33.51.713051782.94.70.014 S13FR7C38.831.537.21.27011738.236.235.036.436.636.05.94.11.712881681.82.80.010 S13FR2D48.526.047.01.28011822.214.171.1246.9126.96.36.199.91.712941724.411.60.019 S13FR5D38.128.536.81.28012338.037.136.73188.8.131.52.13.31.713321942.91.50.014 S13FR7D40.433.038.01.28011838.136.936.436.836.9184.108.40.206.712941721.73.70.009 S13FR2E37.519.535.71.28011238.135.936.836.8220.127.116.11.11.713161845.10.90.027 S13FR2E39.920.538.41.25012738.337.437.738.038.037.76.51.11.713141835.42.20.028 S13FR5E36.025.034.41.28012318.104.22.1686.736.822.214.171.124.713101813.7-0.50.021 S13FR7E32.727.031.41.28011338.236.636.536.236.036.46.13.21.713211872.9-3.70.017
81 APPENDIX C EXPERIMENTAL DATA Â– PHASE 2
Appendix C Experimental Data Â– Phas e 2: Data Dictionary 82 Title Description Code Participant Code Gender Gender of participant Proto Protocol Design: Metabolic Demand (M1 (80 W/m2), M2 (160 W/m2), M3 (240 W/m2)) Ens Ensemble: (A (work clothes), B (cott on coveralls), C (particle barrier), D (liquid barrier), E (vapor barrier)) Tdb Ambient air temperature (d ry bulb) in degrees Celsius Tpwb Wet bulb air temperature in degrees Celsius Tg Black bulb air temperature in degrees Celsius S(m/s) Speed in meters per second G(%) Grade of treadmill in percentage HR Heart rate Tre Body core temperature (rectal) Tch Skin temperature at the chest Tarm Skin temperature at the upper arm Tth Skin temperature at the thigh Tcalf Skin temperature at the calf Met Calculated metabolic work based on O2 consumption in Watts BSA Body surface area in square meters MSA Met divided by the BSA (W/m2) Tsk Average Skin temperature Psk Partial pressure of the water vapor at the skin Pv Partial pressure of the water vapor in the air Psk-Pv P: Difference between Psk and Pv Tair-Tsk T: Difference between Tdb and Tsk ReT Total evaporative resistance
83 Appendix C (Continued) Experimental Data Â– Phase 2 CodeGenderProtoEnsTdbTpwbTg S(m/s)G (%)HRTreTchTarmTthTcalf MetBSA MSA TskPskPv Psk-PairTair-Tsk ReT1FM1A44.9034.0043.200.430.08637.0636.2735.9536.5636.83 1701.55 110 36.346.054.59 1.58.60.009 1FM2A40.1029.2538.601.351.512037.8037.6637.6635.4435.21 2081.55 134 36.736.183.33 2.83.40.018 1FM3A35.8026.7534.651.664.512437.5134.9634.7134.9333.84 3841.55 248 34.665.512.91 2.61.10.010 1FM1B45.4034.0043.100.410.08637.8137.3236.7637.4336.84 981.55 63 37.086.304.55 1.78.30.016 1FM2B42.0031.5040.501.350.013637.8135.0935.6935.7735.62 2501.55 161 35.515.783.92 1.96.50.009 1FM3B36.4026.5034.251.664.515037.7132.9632.8433.8833.49 3091.55 199 33.215.092.80 2.33.20.010 1FM1C44.0030.7543.100.410.010437.2035.7636.1536.5236.41 1281.55 83 36.165.993.54 2.57.80.019 1FM2C40.2029.0038.601.391.511937.5634.5634.5534.9934.79 2571.55 166 34.695.523.25 2.35.50.011 1FM3C33.3023.0031.401.674.515038.1133.4231.9233.4134.70 3761.55 243 33.225.092.12 3.00.10.012 1FM1D43.8031.5041.400.430.08937.0935.8436.0536.0035.84 1211.55 78 35.945.923.80 2.17.90.017 1FM2D37.9028.0036.201.501.511137.2835.8135.8835.0735.53 2971.55 192 35.635.823.11 2.72.30.013 1FM3D35.1026.0032.901.704.511037.7434.7835.6835.1635.21 3841.55 248 35.215.692.75 2.9-0.10.012 1FM1E39.4029.0037.400.280.010837.6536.5437.2536.8237.40 1161.55 75 36.986.273.31 3.02.40.033 1FM2E33.1019.0031.101.401.514138.1734.8935.4436.7134.56 2791.55 180 35.355.731.25 4.5-2.30.027 1FM3E29.5021.0027.101.704.517238.3533.6833.0534.3533.69 3851.55 248 33.635.211.92 3.3-4.10.015 2MM1A46.1034.0043.950.380.012337.3636.2736.7436.4637.31 2222.19 101 36.666.164.51 1.69.40.010 2MM2A41.6030.0038.801.140.012037.8036.0736.1336.3036.08 3962.19 181 36.145.983.46 2.55.50.012 2MM3A42.4031.0040.201.591.013138.3335.9636.5736.4336.28 5062.19 231 36.306.043.73 2.36.10.008 2MM2B40.4031.0039.201.150.011337.4635.6436.4332.2836.13 3852.19 176 35.305.713.86 1.95.10.009 2MM1C42.8032.0041.150.370.011137.6336.5736.8337.5836.47 1912.19 87 36.836.214.03 2.26.00.018 2MM2C44.4033.0042.151.070.011737.7436.8736.6136.4937.26 3992.19 182 36.796.204.26 1.97.60.008 2MM3C39.1029.0038.401.551.012238.2436.4135.5436.0934.33 4592.19 210 35.675.833.33 2.53.40.011 2MM1D42.1032.5041.000.270.012737.5136.6037.0536.5836.99 2352.19 107 36.816.214.25 2.05.30.014 2MM2D39.5028.0038.001.100.012038.0035.7536.2036.3536.28 3352.19 153 36.115.973.01 3.03.40.017 2MM3D38.8029.0037.851.630.012738.2935.3435.0836.5335.69 4742.19 216 35.575.803.35 2.53.20.010 2MM2E35.5025.5033.801.090.010137.3636.5436.5036.3736.56 3102.19 142 36.506.102.59 3.5-1.00.026 2MM3E31.6021.7529.801.630.012638.0535.7936.3236.3735.87 4782.19 218 36.085.961.94 4.0-4.50.022
91 APPENDIX D SAS CODE AND ANALYSIS Â– PHASE 1
Appendix D SAS Code Â– Phase 1 92 options nodate nonumber; libname Vc 'F:\USF\NIOSH Studies\evap res Yr1\' ; SAS Code for Analyzing Re,T for Phase 1; %macro mean1 (var1, var2, var3, var4); Proc Means data=Vc.ret n mean var std stddev; title "SAS Analysis of Phase 1 Data" ; Class &var2 &var3 &var4; var &var1; Run; %mend ; % mean1 (ReT, ensemble); % mean1 (ReT, ensemble, proto); % mean1 (ReT, proto); %macro anov1 (var1, var2, var3, var4); Proc glm data=vc.ret; title "Three way ANOVA using Proc GLM for &var1 Data" ; Class &var2 &var3 &var4; Model &var1 = &var2 &var3 &var4; lsmeans &var2 &var3 &var4 /pdiff adjust=Tukey alpha= 0.05 ; run; %mend ; % anov1 (ReT, ensemble, proto, subj); %macro anov2 (var1, var2, var3, var4); Proc glm data=vc.ret; title "Three-way ANOVA of &var1 data set: Testing Interaction of &var2 x &var3" ; Class &var2 &var3 &var4; Model &var1 = &var2 | &var3 &var4; *lsmeans &var2 | &var3 /pdiff adjust=Tukey alpha=0.05; run; %mend ; % anov2 (ReT, ensemble, proto, subj); %macro mixed1 (var1, var2, var3, var4); Proc mixed data=vc.ret; title "Analysis of $var1 using the Mixed Model" ; Class &var2 &var3 &var4; Model &var1 = &var2 &var3; Random &var4; LSmeans &var2 &var3 /adjust=tukey alpha= .05 ; run; %mend ; % mixed1 (ReT, ensemble, proto, subj);
93 Appendix D (Continued) SAS Analysis Â– Phase 1 SAS Analysis of Phase 1 Data The MEANS Procedure Analysis Variable : ReT ReT N Ensemble Obs N Mean Variance Std Dev ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ A 42 42 0.0139524 0.000019656 0.0044335 B 44 44 0.0143409 0.000026044 0.0051033 C 43 43 0.0158140 0.000031060 0.0055731 D 46 46 0.0178696 0.000030649 0.0055362 E 45 45 0.0265333 0.000076118 0.0087246 ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ
94 Appendix D (Continued) SAS Analysis Â– Phase 1 SAS Analysis of Phase 1 Data The MEANS Procedure Analysis Variable : ReT ReT Ensemble Proto Obs N Mean Variance Std Dev ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ A R2 13 13 0.0183846 0.000014923 0.0038630 R5 14 14 0.0128571 8.2857143E-6 0.0028785 R7 15 15 0.0111333 9.1238095E-6 0.0030206 B R2 14 14 0.0187857 0.000023258 0.0048227 R5 15 15 0.0126000 0.000013971 0.0037378 R7 15 15 0.0119333 0.000015210 0.0038999 C R2 14 14 0.0202857 0.000018220 0.0042685 R5 14 14 0.0148571 0.000022440 0.0047370 R7 15 15 0.0125333 0.000022981 0.0047938 D R2 15 15 0.0220667 0.000017781 0.0042167 R5 18 18 0.0168333 0.000026147 0.0051134 R7 13 13 0.0144615 0.000020936 0.0045756 E R2 16 16 0.0328125 0.000090696 0.0095234 R5 15 15 0.0261333 0.000021981 0.0046884 R7 14 14 0.0197857 0.000031566 0.0056184 ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ
95 Appendix D (Continued) SAS Analysis Â– Phase 1 SAS Analysis of Phase 1 Data The MEANS Procedure Analysis Variable : ReT ReT N Proto Obs N Mean Variance Std Dev ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ R2 72 72 0.0228056 0.000063483 0.0079676 R5 76 76 0.0167368 0.000042516 0.0065205 R7 72 72 0.0138750 0.000028364 0.0053258 ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ
96 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Class Level Information Class Levels Values Ensemble 5 A B C D E Proto 3 R2 R5 R7 Subj 14 S0 S1 S10 S11 S12 S13 S2 S3 S4 S5 S6 S7 S8 S9 Number of observations 220
97 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Dependent Variable: ReT ReT Sum of Source DF Squares Mean Square F Value Pr > F Model 19 0.00962095 0.00050637 32.80 <.0001 Error 200 0.00308801 0.00001544 Corrected Total 219 0.01270896 R-Square Coeff Var Root MSE ReT Mean 0.757021 22.09211 0.003929 0.017786 Source DF Type I SS Mean Square F Value Pr > F Ensemble 4 0.00475024 0.00118756 76.91 <.0001 Proto 2 0.00273468 0.00136734 88.56 <.0001 Subj 13 0.00213603 0.00016431 10.64 <.0001 Source DF Type III SS Mean Square F Value Pr > F Ensemble 4 0.00455585 0.00113896 73.77 <.0001 Proto 2 0.00261372 0.00130686 84.64 <.0001 Subj 13 0.00213603 0.00016431 10.64 <.0001
98 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN Ensemble ReT LSMEAN Number A 0.01412073 1 B 0.01452913 2 C 0.01597198 3 D 0.01778303 4 E 0.02648958 5 Least Squares Means for effect Ensemble Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 4 5 1 0.9891 0.1958 0.0002 <.0001 2 0.9891 0.4312 0.0012 <.0001 3 0.1958 0.4312 0.1972 <.0001 4 0.0002 0.0012 0.1972 <.0001 5 <.0001 <.0001 <.0001 <.0001
99 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN Proto ReT LSMEAN Number R2 0.02250895 1 R5 0.01663240 2 R7 0.01419533 3 Least Squares Means for effect Proto Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 1 <.0001 <.0001 2 <.0001 0.0007 3 <.0001 0.0007
100 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN Subj ReT LSMEAN Number S0 0.01571161 1 S1 0.01970261 2 S10 0.01902724 3 S11 0.01580000 4 S12 0.01873757 5 S13 0.01440995 6 S2 0.02066667 7 S3 0.01845979 8 S4 0.02531243 9 S5 0.02046667 10 S6 0.01456730 11 S7 0.01433333 12 S8 0.01355455 13 S9 0.01815476 14 Least Squares Means for effect Subj Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 4 5 6 7 1 0.1657 0.4479 1.0000 0.5757 0.9995 0.0263 2 0.1657 1.0000 0.2526 1.0000 0.0133 1.0000 3 0.4479 1.0000 0.5685 1.0000 0.0620 0.9966 4 1.0000 0.2526 0.5685 0.6946 0.9994 0.0506 5 0.5757 1.0000 1.0000 0.6946 0.0985 0.9827
101 Appendix D (Continued) SAS Analysis Â– Phase 1 6 0.9995 0.0133 0.0620 0.9994 0.0985 0.0013 7 0.0263 1.0000 0.9966 0.0506 0.9827 0.0013 8 0.7469 0.9998 1.0000 0.8356 1.0000 0.1847 0.9544 9 <.0001 0.0075 0.0012 <.0001 0.0004 <.0001 0.0792 10 0.0415 1.0000 0.9991 0.0757 0.9936 0.0022 1.0000 11 0.9999 0.0195 0.0859 0.9998 0.1327 1.0000 0.0020 12 0.9992 0.0135 0.0620 0.9991 0.0972 1.0000 0.0013 13 0.9457 0.0013 0.0083 0.9471 0.0145 1.0000 <.0001 14 0.8992 0.9984 1.0000 0.9412 1.0000 0.3456 0.9068
102 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer Least Squares Means for effect Subj Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 8 9 10 11 12 13 14 1 0.7469 <.0001 0.0415 0.9999 0.9992 0.9457 0.8992 2 0.9998 0.0075 1.0000 0.0195 0.0135 0.0013 0.9984 3 1.0000 0.0012 0.9991 0.0859 0.0620 0.0083 1.0000 4 0.8356 <.0001 0.0757 0.9998 0.9991 0.9471 0.9412 5 1.0000 0.0004 0.9936 0.1327 0.0972 0.0145 1.0000 6 0.1847 <.0001 0.0022 1.0000 1.0000 1.0000 0.3456 7 0.9544 0.0792 1.0000 0.0020 0.0013 <.0001 0.9068 8 0.0002 0.9789 0.2375 0.1804 0.0342 1.0000 9 0.0002 0.0530 <.0001 <.0001 <.0001 0.0002 10 0.9789 0.0530 0.0034 0.0023 0.0002 0.9488 11 0.2375 <.0001 0.0034 1.0000 1.0000 0.4181 12 0.1804 <.0001 0.0023 1.0000 1.0000 0.3373 13 0.0342 <.0001 0.0002 1.0000 1.0000 0.0886 14 1.0000 0.0002 0.9488 0.4181 0.3373 0.0886
103 Appendix D (Continued) SAS Analysis Â– Phase 1 Three-way ANOVA of ReT data set: Testing Interaction of ensemble x proto The GLM Procedure Dependent Variable: ReT ReT Sum of Source DF Squares Mean Square F Value Pr > F Model 27 0.00989843 0.00036661 25.04 <.0001 Error 192 0.00281053 0.00001464 Corrected Total 219 0.01270896 R-Square Coeff Var Root MSE ReT Mean 0.778854 21.51080 0.003826 0.017786 Source DF Type I SS Mean Square F Value Pr > F Ensemble 4 0.00475024 0.00118756 81.13 <.0001 Proto 2 0.00273468 0.00136734 93.41 <.0001 Ensemble*Proto 8 0.00022513 0.00002814 1.92 0.0587 Subj 13 0.00218838 0.00016834 11.50 <.0001 Source DF Type III SS Mean Square F Value Pr > F Ensemble 4 0.00446782 0.00111696 76.30 <.0001 Proto 2 0.00258380 0.00129190 88.26 <.0001 Ensemble*Proto 8 0.00027748 0.00003468 2.37 0.0187 Subj 13 0.00218838 0.00016834 11.50 <.0001
104 Appendix D (Continued) SAS Analysis Â– Phase 1 Analysis of ReT using the Mixed Model The Mixed Procedure Model Information Data Set VC.RET Dependent Variable ReT Covariance Structure Variance Components Estimation Method REML Residual Variance Method Profile Fixed Effects SE Method Model-Based Degrees of Freedom Method Containment Class Level Information Class Levels Values Ensemble 5 A B C D E Proto 3 R2 R5 R7 Subj 14 S0 S1 S10 S11 S12 S13 S2 S3 S4 S5 S6 S7 S8 S9 Dimensions Covariance Parameters 2 Columns in X 9 Columns in Z 14 Subjects 1 Max Obs Per Subject 220 Observations Used 220 Observations Not Used 0 Total Observations 220
105 Appendix D (Continued) SAS Analysis Â– Phase 1 Analysis of ReT using the Mixed Model The Mixed Procedure Covariance Parameter Estimates Cov Parm Estimate Subj 9.648E-6 Residual 0.000015 Fit Statistics -2 Res Log Likelihood -1697.9 AIC (smaller is better) -1693.9 AICC (smaller is better) -1693.9 BIC (smaller is better) -1692.7 Type 3 Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Ensemble 4 200 73.65 <.0001 Proto 2 200 85.00 <.0001
106 Appendix D (Continued) SAS Analysis Â– Phase 1 Analysis of ReT using the Mixed Model The Mixed Procedure Least Squares Means Standard Effect Ensemble Proto Estimate Error DF t Value Pr > |t| Alpha Proto R5 0.01664 0.000945 200 17.59 <.0001 0.05 Proto R7 0.01418 0.000951 200 14.90 <.0001 0.05 Least Squares Means Effect Ensemble Proto Lower Upper Proto R5 0.01477 0.01850 Proto R7 0.01230 0.01605 Differences of Least Squares Means Effect Ensemble Proto Ensemble Proto Adjustment Adj P Alpha Lower Upper Ensemble A B Tukey-Kramer 0.9900 0.05 -0.00207 0.001277 Ensemble A C Tukey-Kramer 0.2002 0.05 -0.00352 -0.00016 Ensemble A D Tukey-Kramer 0.0002 0.05 -0.00532 -0.00200 Ensemble A E Tukey-Kramer <.0001 0.05 -0.01402 -0.01069 Ensemble B C Tukey-Kramer 0.4309 0.05 -0.00311 0.000222
107 Appendix D (Continued) SAS Analysis Â– Phase 1 Analysis of $var1 using the Mixed Model The Mixed Procedure Differences of Least Squares Means Standard Effect Ensemble Proto Ensemble Proto Estimate Error DF t Value Pr > |t| Ensemble B D -0.00326 0.000831 200 -3.92 0.0001 Ensemble B E -0.01195 0.000836 200 -14.30 <.0001 Ensemble C D -0.00182 0.000836 200 -2.17 0.0310 Ensemble C E -0.01051 0.000839 200 -12.52 <.0001 Ensemble D E -0.00869 0.000827 200 -10.52 <.0001 Proto R2 R5 0.005875 0.000649 200 9.06 <.0001 Proto R2 R7 0.008333 0.000657 200 12.67 <.0001 Proto R5 R7 0.002458 0.000649 200 3.79 0.0002 Differences of Least Squares Means Effect Ensemble Proto Ensemble Proto Adjustment Adj P Alpha Lower Upper Ensemble B D Tukey-Kramer 0.0011 0.05 -0.00490 -0.00162 Ensemble B E Tukey-Kramer <.0001 0.05 -0.01360 -0.01031 Ensemble C D Tukey-Kramer 0.1943 0.05 -0.00347 -0.00017 Ensemble C E Tukey-Kramer <.0001 0.05 -0.01217 -0.00886 Ensemble D E Tukey-Kramer <.0001 0.05 -0.01032 -0.00706 Proto R2 R5 Tukey-Kramer <.0001 0.05 0.004596 0.007154 Proto R2 R7 Tukey-Kramer <.0001 0.05 0.007037 0.009630 Proto R5 R7 Tukey-Kramer 0.0006 0.05 0.001179 0.003738
108 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data without Ensemble E The GLM Procedure Class Level Information Class Levels Values Ensemble 4 A B C D Proto 3 R2 R5 R7 Subj 14 S0 S1 S10 S11 S12 S13 S2 S3 S4 S5 S6 S7 S8 S9 Number of observations 175
109 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data without Ensemble E The GLM Procedure Dependent Variable: ReT ReT Sum of Source DF Squares Mean Square F Value Pr > F Model 18 0.00369821 0.00020546 24.04 <.0001 Error 156 0.00133330 0.00000855 Corrected Total 174 0.00503151 R-Square Coeff Var Root MSE ReT Mean 0.735010 18.81614 0.002923 0.015537 Source DF Type I SS Mean Square F Value Pr > F Ensemble 3 0.00042199 0.00014066 16.46 <.0001 Proto 2 0.00167469 0.00083735 97.97 <.0001 Subj 13 0.00160153 0.00012319 14.41 <.0001 Source DF Type III SS Mean Square F Value Pr > F Ensemble 3 0.00038764 0.00012921 15.12 <.0001 Proto 2 0.00153307 0.00076653 89.69 <.0001 Subj 13 0.00160153 0.00012319 14.41 <.0001
110 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data without Ensemble E The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN Ensemble ReT LSMEAN Number A 0.01405075 1 B 0.01450732 2 C 0.01593123 3 D 0.01785464 4 Least Squares Means for effect Ensemble Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 4 1 0.8884 0.0184 <.0001 2 0.8884 0.1109 <.0001 3 0.0184 0.1109 0.0126 4 <.0001 <.0001 0.0126
111 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data without Ensemble E The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN Proto ReT LSMEAN Number R2 0.01974149 1 R5 0.01431330 2 R7 0.01270317 3 Least Squares Means for effect Proto Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 1 <.0001 <.0001 2 <.0001 0.0092 3 <.0001 0.0092
112 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data without Ensemble E The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN Subj ReT LSMEAN Number S0 0.01386180 1 S1 0.01864754 2 S10 0.01741667 3 S11 0.01325000 4 S12 0.01573967 5 S13 0.01258333 6 S2 0.01700000 7 S3 0.01574064 8 S4 0.02341347 9 S5 0.01783333 10 S6 0.01342597 11 S7 0.01283333 12 S8 0.01100000 13 S9 0.01545805 14 Least Squares Means for effect Subj Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 4 5 6 7 1 0.0023 0.1066 1.0000 0.9036 0.9974 0.2522 2 0.0023 0.9987 0.0007 0.3645 <.0001 0.9798 3 0.1066 0.9987 0.0389 0.9731 0.0060 1.0000 4 1.0000 0.0007 0.0389 0.6569 1.0000 0.1047 5 0.9036 0.3645 0.9731 0.6569 0.2661 0.9981
113 Appendix D (Continued) SAS Analysis Â– Phase 1 6 0.9974 <.0001 0.0060 1.0000 0.2661 0.0200 7 0.2522 0.9798 1.0000 0.1047 0.9981 0.0200 8 0.9165 0.4003 0.9773 0.6858 1.0000 0.2956 0.9984 9 <.0001 0.0057 0.0001 <.0001 <.0001 <.0001 <.0001 10 0.0376 1.0000 1.0000 0.0126 0.8645 0.0016 1.0000 11 1.0000 0.0009 0.0499 1.0000 0.7337 1.0000 0.1316 12 0.9997 0.0002 0.0126 1.0000 0.3985 1.0000 0.0389 13 0.3989 <.0001 <.0001 0.8314 0.0047 0.9879 0.0001 14 0.9840 0.3128 0.9427 0.8693 1.0000 0.5182 0.9923
114 Appendix D (Continued) SAS Analysis Â– Phase 1 Three way ANOVA using Proc GLM for ReT Data without Ensemble E The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer Least Squares Means for effect Subj Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 8 9 10 11 12 13 14 1 0.9165 <.0001 0.0376 1.0000 0.9997 0.3989 0.9840 2 0.4003 0.0057 1.0000 0.0009 0.0002 <.0001 0.3128 3 0.9773 0.0001 1.0000 0.0499 0.0126 <.0001 0.9427 4 0.6858 <.0001 0.0126 1.0000 1.0000 0.8314 0.8693 5 1.0000 <.0001 0.8645 0.7337 0.3985 0.0047 1.0000 6 0.2956 <.0001 0.0016 1.0000 1.0000 0.9879 0.5182 7 0.9984 <.0001 1.0000 0.1316 0.0389 0.0001 0.9923 8 <.0001 0.8808 0.7592 0.4316 0.0063 1.0000 9 <.0001 0.0005 <.0001 <.0001 <.0001 <.0001 10 0.8808 0.0005 0.0163 0.0036 <.0001 0.7982 11 0.7592 <.0001 0.0163 1.0000 0.7191 0.9150 12 0.4316 <.0001 0.0036 1.0000 0.9588 0.6660 13 0.0063 <.0001 <.0001 0.7191 0.9588 0.0236 14 1.0000 <.0001 0.7982 0.9150 0.6660 0.0236
115 Appendix D (Continued) SAS Analysis Â– Phase 1 Three-way ANOVA of ReT data set: Testing Interaction of ensemble x proto without Ensemble E The GLM Procedure Dependent Variable: ReT ReT Sum of Source DF Squares Mean Square F Value Pr > F Model 24 0.00371890 0.00015495 17.71 <.0001 Error 150 0.00131261 0.00000875 Corrected Total 174 0.00503151 R-Square Coeff Var Root MSE ReT Mean 0.739122 19.03931 0.002958 0.015537 Source DF Type I SS Mean Square F Value Pr > F Ensemble 3 0.00042199 0.00014066 16.07 <.0001 Proto 2 0.00167469 0.00083735 95.69 <.0001 Ensemble*Proto 6 0.00001444 0.00000241 0.28 0.9479 Subj 13 0.00160777 0.00012367 14.13 <.0001 Source DF Type III SS Mean Square F Value Pr > F Ensemble 3 0.00038033 0.00012678 14.49 <.0001 Proto 2 0.00153345 0.00076673 87.62 <.0001 Ensemble*Proto 6 0.00002069 0.00000345 0.39 0.8820 Subj 13 0.00160777 0.00012367 14.13 <.0001
116 APPENDIX E SAS CODE AND ANALYSIS Â– PHASE 2
Appendix E SAS Code Â– Phase 2 117 options nodate nonumber; libname Vc 'F:\USF\NIOSH Studies\evap res Yr2\' ; SAS Code for Analyzing Re,T for Phase 2; %macro mean1 (var1, var2, var3, var4); Proc Means data=Vc.ret n mean var std stddev; title "SAS Analysis of Pase 2 Data" ; Class &var2 &var3 &var4; var &var1; Run; %mend ; % mean1 (ReT, ensemble); % mean1 (ReT, ensemble, M); % mean1 (ReT, M); %macro anov1 (var1, var2, var3, var4); Proc glm data=vc.ret; title "Three way ANOVA using Proc GLM for &var1 Data" ; Class &var2 &var3 &var4; Model &var1 = &var2 &var3 &var4; lsmeans &var2 &var3 &var4 /pdiff adjust=Tukey alpha= 0.05 ; run; %mend ; % anov1 (ReT, ensemble, M, subj); %macro anov2 (var1, var2, var3, var4); Proc glm data=vc.ret; title "Three-way ANOVA of &var1 data set: Testing Interaction of &var2 x &var3" ; Class &var2 &var3 &var4; Model &var1 = &var2 | &var3 &var4; lsmeans &var2 | &var3 /pdiff adjust=Tukey alpha= 0.05 ; run; %mend ; % anov2 (ReT, ensemble, M, subj); %macro mixed1 (var1, var2, var3, var4); Proc mixed data=vc.ret; title "Analysis of $var1 using the Mixed Model" ; Class &var2 &var3 &var4; Model &var1 = &var2 &var3; Random &var4; LSmeans &var2 &var3 /adjust=tukey alpha= .05 ; run; %mend ; % mixed1 (ReT, ensemble, M, subj);
118 Appendix E (Continued) SAS Analysis Â– Phase 2 SAS Analysis of Phase 2 Data The MEANS Procedure Analysis Variable : ReT ReT N Ensemble Obs N Mean Variance Std Dev ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ A 44 44 0.0114318 5.5533827E-6 0.0023566 B 42 42 0.0121667 8.0934959E-6 0.0028449 C 46 46 0.0126304 9.2603865E-6 0.0030431 D 45 45 0.0152889 0.000016846 0.0041044 E 48 48 0.0235833 0.000031525 0.0056147 ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ
119 Appendix E (Continued) SAS Analysis Â– Phase 2 SAS Analysis of Phase 2 Data The MEANS Procedure Analysis Variable : ReT ReT Ensemble M Obs N Mean Variance Std Dev ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ A M1 14 14 0.0110714 4.0714286E-6 0.0020178 M2 15 15 0.0125333 8.9809524E-6 0.0029968 M3 15 15 0.0106667 2.2380952E-6 0.0014960 B M1 14 14 0.0135714 6.4175824E-6 0.0025333 M2 14 14 0.0117857 3.8736264E-6 0.0019682 M3 14 14 0.0111429 0.000011824 0.0034386 C M1 16 16 0.0149375 0.000013663 0.0036963 M2 15 15 0.0119333 3.352381E-6 0.0018310 M3 15 15 0.0108667 1.8380952E-6 0.0013558 D M1 15 15 0.0183333 0.000024238 0.0049232 M2 15 15 0.0152000 4.6E-6 0.0021448 M3 15 15 0.0123333 4.8095238E-6 0.0021931 E M1 16 16 0.0282500 0.000026200 0.0051186 M2 15 15 0.0239333 0.000013781 0.0037123 M3 17 17 0.0188824 0.000010610 0.0032573 ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ
120 Appendix E (Continued) SAS Analysis Â– Phase 2 SAS Analysis of Phase 2 Data The MEANS Procedure Analysis Variable : ReT ReT N M Obs N Mean Variance Std Dev ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ M1 75 75 0.0174800 0.000051794 0.0071968 M2 74 74 0.0151216 0.000028136 0.0053043 M3 76 76 0.0129605 0.000016545 0.0040676 ÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒÂƒ
121 Appendix E (Continued) SAS Analysis Â– Phase 2 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Class Level Information Class Levels Values Ensemble 5 A B C D E M 3 M1 M2 M3 Subj 15 1 2 3 4 5 6 7 8 9 10 11 12 13 15 16 Number of observations 225
122 Appendix E (Continued) SAS Analysis Â– Phase 2 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Dependent Variable: ReT ReT Sum of Source DF Squares Mean Square F Value Pr > F Model 20 0.00603568 0.00030178 33.04 <.0001 Error 204 0.00186321 0.00000913 Corrected Total 224 0.00789889 R-Square Coeff Var Root MSE ReT Mean 0.764117 19.91167 0.003022 0.015178 Source DF Type I SS Mean Square F Value Pr > F Ensemble 4 0.00468863 0.00117216 128.34 <.0001 M 2 0.00080444 0.00040222 44.04 <.0001 Subj 14 0.00054261 0.00003876 4.24 <.0001 Source DF Type III SS Mean Square F Value Pr > F Ensemble 4 0.00465386 0.00116347 127.39 <.0001 M 2 0.00080189 0.00040094 43.90 <.0001 Subj 14 0.00054261 0.00003876 4.24 <.0001
123 Appendix E (Continued) SAS Analysis Â– Phase 2 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN Ensemble ReT LSMEAN Number A 0.01151359 1 B 0.01204646 2 C 0.01261354 3 D 0.01528889 4 E 0.02358058 5 Least Squares Means for effect Ensemble Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 4 5 1 0.9257 0.4212 <.0001 <.0001 2 0.9257 0.9050 <.0001 <.0001 3 0.4212 0.9050 0.0004 <.0001 4 <.0001 <.0001 0.0004 <.0001 5 <.0001 <.0001 <.0001 <.0001
124 Appendix E (Continued) SAS Analysis Â– Phase 2 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN M ReT LSMEAN Number M1 0.01728715 1 M2 0.01506492 2 M3 0.01267376 3 Least Squares Means for effect M Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 1 <.0001 <.0001 2 <.0001 <.0001 3 <.0001 <.0001
125 Appendix E (Continued) SAS Analysis Â– Phase 2 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer LSMEAN Subj ReT LSMEAN Number 1 0.01540000 1 2 0.01415582 2 3 0.01540000 3 4 0.01393333 4 5 0.01553333 5 6 0.01466667 6 7 0.01350728 7 8 0.01888148 8 9 0.01166546 9 10 0.01493333 10 11 0.01686667 11 12 0.01632821 12 13 0.01493333 13 15 0.01440000 14 16 0.01452426 15 Least Squares Means for effect Subj Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 1 2 3 4 5 6 7 8 1 0.9992 1.0000 0.9918 1.0000 1.0000 0.9173 0.1104 2 0.9992 0.9992 1.0000 0.9977 1.0000 1.0000 0.0067 3 1.0000 0.9992 0.9918 1.0000 1.0000 0.9173 0.1104 4 0.9918 1.0000 0.9918 0.9815 1.0000 1.0000 0.0012
126 Appendix E (Continued) SAS Analysis Â– Phase 2 5 1.0000 0.9977 1.0000 0.9815 1.0000 0.8676 0.1502 6 1.0000 1.0000 1.0000 1.0000 1.0000 0.9992 0.0144 7 0.9173 1.0000 0.9173 1.0000 0.8676 0.9992 0.0002 8 0.1104 0.0067 0.1104 0.0012 0.1502 0.0144 0.0002 9 0.0691 0.7382 0.0691 0.7852 0.0487 0.3305 0.9414 <.0001 10 1.0000 1.0000 1.0000 0.9999 1.0000 1.0000 0.9927 0.0321 11 0.9918 0.5839 0.9918 0.3385 0.9968 0.7998 0.1295 0.8853 12 0.9999 0.8667 0.9999 0.6662 1.0000 0.9712 0.3555 0.5638 13 1.0000 1.0000 1.0000 0.9999 1.0000 1.0000 0.9927 0.0321 14 0.9999 1.0000 0.9999 1.0000 0.9995 1.0000 1.0000 0.0060 15 1.0000 1.0000 1.0000 1.0000 0.9998 1.0000 0.9997 0.0060
127 Appendix E (Continued) SAS Analysis Â– Phase 2 Three way ANOVA using Proc GLM for ReT Data The GLM Procedure Least Squares Means Adjustment for Multiple Comparisons: Tukey-Kramer Least Squares Means for effect Subj Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: ReT i/j 9 10 11 12 13 14 15 1 0.0691 1.0000 0.9918 0.9999 1.0000 0.9999 1.0000 2 0.7382 1.0000 0.5839 0.8667 1.0000 1.0000 1.0000 3 0.0691 1.0000 0.9918 0.9999 1.0000 0.9999 1.0000 4 0.7852 0.9999 0.3385 0.6662 0.9999 1.0000 1.0000 5 0.0487 1.0000 0.9968 1.0000 1.0000 0.9995 0.9998 6 0.3305 1.0000 0.7998 0.9712 1.0000 1.0000 1.0000 7 0.9414 0.9927 0.1295 0.3555 0.9927 1.0000 0.9997 8 <.0001 0.0321 0.8853 0.5638 0.0321 0.0060 0.0060 9 0.2016 0.0006 0.0036 0.2016 0.4931 0.3648 10 0.2016 0.9138 0.9943 1.0000 1.0000 1.0000 11 0.0006 0.9138 1.0000 0.9138 0.6392 0.6747 12 0.0036 0.9943 1.0000 0.9943 0.9075 0.9274 13 0.2016 1.0000 0.9138 0.9943 1.0000 1.0000 14 0.4931 1.0000 0.6392 0.9075 1.0000 1.0000 15 0.3648 1.0000 0.6747 0.9274 1.0000 1.0000
128 Appendix E (Continued) SAS Analysis Â– Phase 2 Three-way ANOVA of ReT data set: Testing Interaction of ensemble x M The GLM Procedure Dependent Variable: ReT ReT Sum of Source DF Squares Mean Square F Value Pr > F Model 28 0.00642128 0.00022933 30.42 <.0001 Error 196 0.00147761 0.00000754 Corrected Total 224 0.00789889 R-Square Coeff Var Root MSE ReT Mean 0.812934 18.09023 0.002746 0.015178 Source DF Type I SS Mean Square F Value Pr > F Ensemble 4 0.00468863 0.00117216 155.48 <.0001 M 2 0.00080444 0.00040222 53.35 <.0001 Ensemble*M 8 0.00040395 0.00005049 6.70 <.0001 Subj 14 0.00052425 0.00003745 4.97 <.0001 Source DF Type III SS Mean Square F Value Pr > F Ensemble 4 0.00470708 0.00117677 156.09 <.0001 M 2 0.00074264 0.00037132 49.25 <.0001 Ensemble*M 8 0.00038560 0.00004820 6.39 <.0001 Subj 14 0.00052425 0.00003745 4.97 <.0001
129 Appendix E (Continued) SAS Analysis Â– Phase 2 Analysis of $var1 using the Mixed Model The Mixed Procedure Model Information Data Set VC.RET Dependent Variable ReT Covariance Structure Variance Components Estimation Method REML Residual Variance Method Profile Fixed Effects SE Method Model-Based Degrees of Freedom Method Containment Class Level Information Class Levels Values Ensemble 5 A B C D E M 3 M1 M2 M3 Subj 15 1 2 3 4 5 6 7 8 9 10 11 12 13 15 16 Dimensions Covariance Parameters 2 Columns in X 9 Columns in Z 15 Subjects 1 Max Obs Per Subject 225 Observations Used 225 Observations Not Used 0 Total Observations 225 Iteration History
130 Appendix E (Continued) SAS Analysis Â– Phase 2 Analysis of ReT using the Mixed Model The Mixed Procedure Covariance Parameter Estimates Cov Parm Estimate Subj 2.001E-6 Residual 9.135E-6 Fit Statistics -2 Res Log Likelihood -1864.0 AIC (smaller is better) -1860.0 AICC (smaller is better) -1859.9 BIC (smaller is better) -1858.6 Type 3 Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Ensemble 4 204 127.77 <.0001 M 2 204 43.94 <.0001
131 Appendix E (Continued) SAS Analysis Â– Phase 2 Standard Effect Ensemble M Ensemble M Estimate Error DF t Value Pr > |t| Adjustment Ensemble A B -0.00057 0.000653 204 -0.87 0.3837 Tukey-Kramer Ensemble A C -0.00110 0.000638 204 -1.72 0.0862 Tukey-Kramer Ensemble A D -0.00378 0.000641 204 -5.90 <.0001 Tukey-Kramer Ensemble A E -0.01209 0.000633 204 -19.10 <.0001 Tukey-Kramer Ensemble B C -0.00053 0.000646 204 -0.82 0.4138 Tukey-Kramer Ensemble B D -0.00321 0.000650 204 -4.94 <.0001 Tukey-Kramer Ensemble B E -0.01152 0.000640 204 -18.00 <.0001 Tukey-Kramer Ensemble C D -0.00268 0.000634 204 -4.23 <.0001 Tukey-Kramer Ensemble C E -0.01099 0.000625 204 -17.57 <.0001 Tukey-Kramer Ensemble D E -0.00830 0.000629 204 -13.21 <.0001 Tukey-Kramer M M1 M2 0.002225 0.000498 204 4.47 <.0001 Tukey-Kramer M M1 M3 0.004615 0.000492 204 9.37 <.0001 Tukey-Kramer M M2 M3 0.002390 0.000495 204 4.83 <.0001 Tukey-Kramer Differences of Least Squares Means Adj Adj Effect Ensemble M Ensemble M Adj P Alpha Lower Upper Lower Upper Ensemble A B 0.9065 0.05 -0.00186 0.000718 . Ensemble A C 0.4215 0.05 -0.00236 0.000158 . Ensemble A D <.0001 0.05 -0.00505 -0.00252 . Ensemble A E <.0001 0.05 -0.01334 -0.01084 . Ensemble B C 0.9246 0.05 -0.00180 0.000745 . Ensemble B D <.0001 0.05 -0.00449 -0.00193 . Ensemble B E <.0001 0.05 -0.01278 -0.01025 . Ensemble C D 0.0003 0.05 -0.00393 -0.00143 . Ensemble C E <.0001 0.05 -0.01222 -0.00975 . Ensemble D E <.0001 0.05 -0.00954 -0.00707 . M M1 M2 <.0001 0.05 0.001244 0.003206 . M M1 M3 <.0001 0.05 0.003644 0.005586 . M M2 M3 <.0001 0.05 0.001415 0.003365 .
132 APPENDIX F JMP IN DATA ANALYSIS Â– PROTOCOLS
Appendix F 133 Re,T Response to Protocol (Environment) by Ensemble
Appendix F (Continued) 134 Re,T Response to Protocol (Environment) by Ensemble
Appendix F (Continued) 135 Re,T Response to Protocol (Environment) by Ensemble
Appendix F (Continued) 136 Re,T Response to Protocol (Environment) by Ensemble
Appendix F (Continued) 137 Re,T Response to Protocol (Environment) by Ensemble
Appendix F (Continued) 138 Re,T Response to Protocol (Metab olic Demand) by Ensemble
Appendix F (Continued) 139 Re,T Response to Protocol (Metab olic Demand) by Ensemble
Appendix F (Continued) 140 Re,T Response to Protocol (Metab olic Demand) by Ensemble
Appendix F (Continued) 141 Re,T Response to Protocol (Metab olic Demand) by Ensemble
Appendix F (Continued) 142 Re,T Response to Protocol (Metab olic Demand) by Ensemble
143 APPENDIX G JMP IN DATA ANALYSIS Â– M2R5
Appendix G 144 Re,T Response to Ensemble by Phase (M2R5 Dataset)
Appendix G (Continued) 145 Re,T Response to Ensemble by Phase (M2R5 Dataset)
Appendix G (Continued) 146 Re,T Response to Ensemble by Phase (M2R5 Dataset)
Appendix G (Continued) 147 Re,T Response to Ensemble by Phase (M2R5 Dataset)
Appendix G (Continued) 148 Re,T Response to Ensemble by Phase (M2R5 Dataset)
149 APPENDIX H GRAPHS OF RE,T VERSUS P BY ENSEMBLE
Appendix H Graphs of Re,T versus P by Ensemble 150 Phase 1 Ensemble A0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 11.522.533.544.55 PRe,T Phase 2 Ensemble A0.006 0.008 0.010 0.012 0.014 0.016 0.018 0.020 1.522.533.5 PRe,T
Appendix H (Continued) Graphs of Re,T versus P by Ensemble 151 Phase 1 Ensemble B0.0000 0.0050 0.0100 0.0150 0.0200 0.0250 0.0300 0.0350 11.522.533.544.55 PRe,T Phase 2 Ensemble B0.000 0.005 0.010 0.015 0.020 0.025 1.522.533.5 PRe,T
Appendix H (Continued) Graphs of Re,T versus P by Ensemble 152 Phase 1 Ensemble C0.000 0.005 0.010 0.015 0.020 0.025 0.030 11.522.533.544.555.5 PRe,T Phase 2 Ensemble C0.000 0.005 0.010 0.015 0.020 0.025 0.030 126.96.36.199.188.8.131.52.93.13.3 PRe,T
Appendix H (Continued) Graphs of Re,T versus P by Ensemble 153 Phase 1 Ensemble D0.0000 0.0050 0.0100 0.0150 0.0200 0.0250 0.0300 0.0350 11.522.533.544.555.5 PRe,T Phase 2 Ensemble D0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 22.533.54 PRe,T
Appendix H (Continued) Graphs of Re,T versus P by Ensemble 154 Phase 1 Ensemble E0 0.01 0.02 0.03 0.04 0.05 0.06 123456 PRe,T Phase 2 Ensemble E0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 184.108.40.206.24.7 PRe,T
ABOUT THE AUTHOR Major Victor Caravello received a Bachel orÂ’s Degree in Industrial Technology from Binghamton University in 1989. He was commissioned a 2nd Lieutenant in the United States Air Force in 1990 and bega n his career in the military as a bioenvironmental engineer. He complete d a M.S. in Toxicology from Texas A&M University in 1998. He has published technical reports dealing with human health risk assessments. He entered the Ph.D. program in occupational and environmental health at the University of South Florida in 2001. While at the University of South Florida, Major Caravello did re search in the Heat Stress Laboratory and served in various leadership positions within the student chapter of the Human Factors and Ergonomics Society. He has coauthored two publications in heat stress and presented three research papers at national level meetings Â– the American Industrial Hygiene Conferen ce and Exposition and the Amer ican College of Sports Medicine Annual Meeting.