FIELD STUDY OF THE ROTARY DESICCANT SYSTEM USING THE CROMER CYCLE BRONISLAVA VELTCHEVA

FIELD STUDY OF THE ROTARY DESICCANT SYSTEM USING THE CROMER CYCLE By BRONISLAVA VELTCHEVA A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 003

ACKNOWLEDGMENTS First of all I would like to express my deep appreciation to Dr. Yogi Goswami, my advisor and committee cairman, for is guidance, tremendous support and understanding over te last years. I wis to tank my oter committee members, Dr. S. A. Serif and Dr. D. W Han, for teir expert advice and guidance. Special tanks go to Dr. H. A. (Skip) Ingley for is constructive recommendations and elpful suggestions during my researc. I would like also to tank Dr. Carlie Cromer for te discussions on is tecnology. Sincere tanks go to Mr. Carles Garretson for is patience and valuable support in constructing te experimental setup. Very sincere tanks go to my usband, Ivan, witout wose motivation, belief in my abilities and abundant love and support tis degree would never ave been possible. Last, but not least, I wis to express my deep gratitude to my great cousin Ani and er usband Natan for teir measureless support and constant positivism. Tey made tis experience more enjoyable and rewarding. ii

TABLE OF CONTENTS page ACKNOWLEDGMENTS... ii LIST OF TABLES...v LIST OF FIGURES... vi LIST OF ABBREVIATIONS... viii ABSTRACT...x CHAPTER 1 INTRODUCTION...1 1.1 Wy Air Conditioning... 1. Desiccant Cooling Concept...3 CROMER CYCLE...5.1 Parameters Tat Impact Cromer Cycle Performance...8. Literature Review...11.3 Objectives of te Present Study...1 3 EXPERIMENTAL SETUP...1 3.1 Experimental Facility...1 3. Measuring Instrumentation...15 3. Desiccant Weel...18 3.3 Condensate Measuring Equipment... 3. Data Acquisition System... PROTOCOL AND EXPERIMENTAL RESULTS...7.1 Protocol...7. Experimental Results...9.3 Discussion...9 5 CONCLUSIONS...60 iii

APPENDIX A DATA SETS...6 B UNCERTAINTY ANALYSIS...67 C FORTRAN 77 PROGRAM...73 D LATENT COOLING TRENDLINES...77 LIST OF REFERENCES...81 BIOGRAPHICAL SKETCH...83 iv

Table LIST OF TABLES page 1-1 Use of air-conditioning equipment in te USA ouseolds... -1 Performance comparison of a standard air andler to te same equipped wit te Cromer cycle equipment...1-1 Ambient conditions and run times for te data pairs from te baseline set and Cromer set No.1...33 - Ambient conditions and run times for te data pairs from te baseline set and Cromer set No... -3 Ambient conditions and run times for te data pairs from te baseline set and Cromer set No.3...8 - Ambient conditions and run times for two similar days for Cromer cycle configuration but wit different airflows...5 A-1 Summary of all te data sets processed....63 B-1 Uncertainty of experimental measurements...67 B- Average uncertainty...70 B-3 Uncertainty of calculated values...70 B- Calculated values...71 v

Figure LIST OF FIGURES page 3-1. Test ouse...1 3-. Location of te combined temperature and RH sensors...16 3-3. Reference ouse...17 3-. Laminates design....19 3-5. Te experimental weel assembly...0 3-6. Te experimental setup...1 3-7. Te pulley and te driving belt... 3-8. Mecanism for monitoring number of desiccant weel revolutions...3 3-9. Desiccant weel power supply layout...3 3-10. Condensate measuring equipment... 3-11. Condensate measuring vessel and valve...5 3-1. Layout of te data acquisition system...6-1. Indoor temperature and RH profiles from Baseline set....30 -. Indoor temperature and RH profiles for Baseline set and Cromer set No.1....36-3. Comparison of performance caracteristics....37 -. Psycrometric cart for similar days...38-5. Psycrometric cart for similar days...39-6. Indoor temperature and RH profiles for Baseline set and Cromer set No... -7. Comparison of performance caracteristics....5-8. Psycrometric cart for similar days...6-9. Psycrometric cart for similar days...7 vi

-10. Indoor temperature and RH profiles for Baseline set and Cromer set No.3....50-11. Comparison of performance caracteristics....51-1. Psycrometric cart for similar days...5-13. Psycrometric cart for similar days...53-1. Indoor temperature and RH profiles for Cromer cycle wit different flow rates..56-15. Comparison of performance caracteristics......57-16. Psycrometric cart for similar days...58-17. Psycrometric cart for similar days...59 D-1. Latent cooling vs. RH in...78 vii

LIST OF ABBREVIATIONS AC amb. ASHRAE ARI Cond. EER in lat LHR. m NTU Air conditioning Ambient American Society of Heating, Refrigerating and Air-Conditioning Engineers American Refrigeration Institute Condensate Energy efficiency ratio (ratio of cooling in Btu/ to te power input in W) Entalpy [kj/kg] Inside Latent Latent eat ratio Mass flow rate of air [kg/] Number of transfer units. Q Cooling [kj/] rp Ref.H Revolutions per our Reference ouse RH Relative umidity [%] SEECL sens SHR Solar Energy and Energy Conversion Laboratory Sensible Sensible eat ratio viii

T Temperature [ o C] TH UMF U Test ouse Uncertainty magnifying function Uncertainty Subscripts lat sens Latent Sensible ix

Abstract of Tesis Presented to te Graduate Scool of te University of Florida in Partial Fulfillment of te Requirements for te Degree of Master of Science FIELD STUDY OF THE ROTARY DESICCANT SYSTEM USING THE CROMER CYCLE By Bronislava Veltceva May 003 Cair: Yogi Goswami Major Department: Mecanical and Aerospace Engineering Te main objective of tis study is to test (under field conditions) te feasibility and effectiveness of te Cromer cycle. Wen cooling a space to a comfortable condition tere are two types of loads to be removed: te temperature-associated load (sensible load) and te moisture-associated load (latent load). Conventional vapor-compression air-conditioning systems perform well wen te latent load is 5% of te total load or less. In many applications (suc as geograpical locations wit ot and umid climates, restaurants, supermarkets, etc.) te latent load often is iger. In suc cases, te conventional air conditioning system fails to meet te increased latent load. Te Cromer cycle uses a desiccant for enanced deumidification of te air. It is based on installation of a desiccant weel to transfer eat and moisture between te return and supply side of an air andler. Te unique feature of te Cromer cycle is tat regeneration of te desiccant is accomplised by te return air (not by an external eat source). x

To test te Cromer cycle under field conditions, te existing air-conditioning system of a residential ouse was retrofitted to accommodate a desiccant weel. Te weel was alternately switced in and out of te system. Data for te performance of te system were collected and compared for te standard and te Cromer configurations. To screen out any canges due to te ambient conditions only, anoter ouse located close to te Test ouse was also instrumented and monitored. xi

CHAPTER 1 INTRODUCTION Air conditioning (AC) is te process of providing by mecanical means control of temperature, relative umidity, movement and purity of te air. Maintaining a space at a desired indoor condition may be acieved by simple eating (increasing te dry bulb temperature), simple cooling (lowering te dry bulb temperature), umidifying (adding moisture), or deumidifying (removing moisture) te air. Quite often two or more of tese processes are required to bring a space to te desired condition. To maintain te desired comfort conditions an air conditioning system as to andle two loads. Tese are te temperature associated, or sensible load and moisture associated, or latent load. Te sensible load is met simply by canging te dry bulb temperature of te air. To meet te latent load of te space some moisture as to be added or removed from it. Tere are four principal metods [Jones 001] of deumidification: Cooling air to a temperature below its dew point Adsorption Absorption Compression followed by cooling. Te conventional vapor-compression AC system meets te latent load of te space by cooling te air below its dew point and as a result water vapor is condensed from te air. 1

1.1 Wy Air Conditioning Air conditioning as grown rapidly around te world. As sown in Table 1-1 [Energy Information Administration 000] in te last twenty-five years more Americans ave air conditioners in teir omes and use teir AC equipment more often. Furtermore nowadays almost all automobiles are equipped wit AC systems. Table 1-1. Use of air conditioning equipment in te USA ouseolds (percent of ouseolds) Survey year Number of ouseolds (million) Percent wit central air conditioning Percent wit window/wall air conditioning Percent wit no air conditioning National 1978 76.6 3.0 3.8. 1979 77.5.1 30.7 5.1 1980 81.6 7. 30.0.8 1981 83.1 6.9 31.3 1.8 198 83.8 7.9 30. 1.9 198 86.3 9.7 9.9 0. 1987 90.5 33.9 9.8 36. 1990 9.0 38.9 8.8 3.3 1993 96.6 3.5.9 31.6 1997 101.5 7.1 5. 7.5 Sout 1978.6 36.9 37.7 5.5 1979.9 38.5 33.8 7.7 1980 7.0 1. 3.7 6.0 1981 7.7.6 3.0 3. 198 8.1.1 33.7. 198 9.3 7.3 9.8.8 1987 30.9 5.3 9.9 17.9 1990 3.3 59.0 8. 1.9 1993 33.5 65.1.1 10.8 1997 35.9 69.7 3. 07.0 Energy Information Administration (EIA), 000. URL: ttp://www.eia.doe.gov/emeu/consumptionbriefs/recs/actrends/recs_ac_trends.tml. In most commercial and residential applications, te umidity in te space is not directly controlled. Rater it is controlled only indirectly; and increases or decreases as a result of canges in te matc between te sensible and latent capacity of te AC

3 equipment compared to te sensible and latent loads of te space. A conventional AC equipment performs well wen sensible load is 75% of te total cooling load or iger [Kosar et al.1998, p.7]. Tere are many applications, owever, were te latent load is iger tan 5%. Geograpical locations suc as Florida, were te weater is ot and umid for many monts trougout te year; supermarkets wit big display cases; and restaurants are among tose applications. In suc cases, te conventional AC unit often fails to meet te comfort conditions. Tis results in elevated indoor umidity levels, discomfort, and mold and mildew growt. 1. Desiccant Cooling Concept Use of desiccants is one solution to te problem of ig umidity. Desiccants are materials tat ave an affinity for water. Wile conventional AC equipment controls umidity by condensation on a cold surface, desiccant-based systems deumidify by adsorption or absorption in a ydroscopic material. Te process of attracting and olding moisture is described as eiter adsorption or absorption, depending on weter te desiccant undergoes a cemical cange as it takes on moisture. Adsorption does not cange te desiccant, except by te addition of te mass of water vapor; it is similar to a sponge soaking up water. Absorption, on te oter and, canges te desiccant cemically or pysically. Two basic types of desiccants are used: solid desiccants (e.g., silica gels, zeolites and syntetic polymers) and liquid desiccants (e.g., salt solutions and glycols). All desiccants function by te same mecanism transfer of moisture due to a difference between te water vapor pressure at desiccant surface and te surrounding air. Te vapor pressure of water at te desiccant surface depends on te pysical caracteristics of te

desiccant, te temperature of te desiccant, and te amount of water adsorbed in te desiccant. Wen te vapor pressure at te desiccant surface is lower tan tat of te air, te desiccant attracts moisture. Wen te surface vapor pressure is iger tan tat of te surrounding air, te desiccant releases moisture. Equilibrium is reaced wen te vapor pressure in te desiccant is equal to tat in te air. To allow repeated use of te desiccant, it as to be regenerated. Regeneration usually is accomplised by eating te desiccant using an external eat source. Most desiccant cooling systems use a desiccant to andle te latent load before air goes to te cooling device. Te desiccant material picks up moisture from te air before te air is sensibly cooled. Tis way it is not required to cool te air excessively in order to condense moisture from it. Te literature provides extensive overviews of work sowing te effectiveness and energy-saving potentials of desiccants [Pesaran et al. 199, Oberg 1998, Mago and Goswami 001].

CHAPTER CROMER CYCLE Te Cromer cycle is a desiccant-based tecnology for enanced deumidification of te air. It is based on installing a desiccant weel between te return and supply air streams in an AC system to transfer eat and moisture between te two streams. Unlike oter desiccant-assisted cooling tecnologies, te Cromer cycle does not require external eat source to regenerate te desiccant, but relies on inerent vapor-pressure differential. A general layout of an air andler equipped wit te Cromer cycle is sown in figure -1. A scematic of te weel operation is given in figure -. Te processes tat air undergoes wen passing troug te weel are as follows: cold air wit very ig relative umidity (RH) leaves te cooling coil and passes troug te working side of te weel, cooling te desiccant and transferring moisture to it. At te same time, te warmer air wit lower RH from te conditioned space passes troug te return side of te weel, absorbing moisture and regenerating te desiccant. Te release of moisture into te air returning from te space before it enters te cooling coil increases te latent ratio of te coil, enancing its deumidification abilities. Te RH and temperature difference of te two air streams provides te potential for moisture transfer. Te feature of Cromer cycle tecnology tat distinguises it from oter desiccant-based concepts is tat te return air (rater tan an external eat source) accomplises regeneration of te desiccant. To sow te differences in te cooling process between te standard AC cycle and te Cromer cycle, te corresponding state points of te air are sown on a Psycrometric cart (Figure -3). 5

6 Air Handler Desiccant Weel Supply Air Return Air Weel Motor Figure -1. General layout of te AC andler equipped wit te Cromer cycle Working side Cold and ig RH air from air andler Sligtly warmed and deumidified air to space Cooled and umidified air to air andler Warm and low RH air from space Regeneration side Figure -. Desiccant weel operation

7 Point 1 is te state point of te air tat returns from te conditioned space. For te standard air conditioner configuration, te air at state 1 enters te cooling coil were it is cooled down and deumidified. Point " depicts state of te air as it leaves te coil. Tis point represents te temperature and moisture content of te air tat is supplied to te conditioned space by te standard AC system. Te Cromer cycle is depicted in te same figure wit te solid line passing troug points 1 to. Te desiccant adsorbs 1 s 3 c 5 10 15 0 Temperature [ o C] Figure -3. Psycrometric cart of standard AC cycle and te Cromer cycle. 1 = return air, = air before cooling coil, 3 = air after cooling coil, = supply air (Cromer cycle), s and c = calculation points, = supply air (standard configuration) moisture from te cold and ig RH air leaving te coil. Tis sorption of moisture dries te supply air before it goes to te space, and follows te line between state points 3 and. Te moisture adsorbed by te desiccant is ten re-evaporated into te return air before it reaces te cooling coil. Te process between state points 1 and

8 represents tis process tat also regenerates te desiccant. Te process between points and 3 depicts te work performed by te evaporator cooling coil..1 Parameters Tat Impact Cromer Cycle Performance Te desiccant weel is te eart of te Cromer cycle. Tere are number of parameters tat ave significant impact on te performance of te weel and from tere on te Cromer cycle. Zang and Niu [00] found tree key parameters: Desiccant isoterm sape Maximum desiccant matrix moisture uptake Heat and mass transfer caracteristics of te matrix. Depending on tese parameters but as well as on te operation conditions, te best for te corresponding application geometry of te weel, size of air passage cannels, air flow rate and speed of rotation can be cosen..1. Solid Desiccant Materials Isoterm Sape Adsorption beavior of te solid desiccants depends on: Total surface area Total volume of capillaries Range of capillary diameters. A large surface area gives te adsorbent a larger capacity at low relative umidities. Large capillaries provide a ig capacity for condensed water, wic gives te adsorbent a iger capacity at ig relative umidities. A narrow range of capillary diameters makes te adsorbent more selective in te vapor molecules it can old. Te desiccant isoterm caracterizes ow a desiccant material picks up moisture at different levels of RH. Different desiccant materials exibit different isoterm sapes. Since te

9 adsorption beavior of te solid desiccants depends on te surface caracteristics of te desiccant and te geometry of te internal structure, tey can be engineered and manufactured to produce a variety of isoterm sapes. Figure - [ASHRAE 1997] illustrates tis point using tree silica gels adsorbent materials. Figure -. Adsorption caracteristics of some experimental silica gels (ASHRAE Fundamentals1997, Fig.6, p.1.) Te Cromer cycle application requires te desiccant to adsorb moisture from air coming off te coil tat is cold and close to saturation and desorb moisture to air tat is warmer and at a lower RH. Te desiccant is regenerated by te vapor pressure differential inerent in te RH differences rater tan eat or temperature difference. Terefore, desiccant materials wit isoterms similar to tat of Gel 1 (Figure -) are required.

10.1. Desiccant Matrix Moisture Uptake Desiccant matrix moisture uptake is defined as te moisture adsorbed by a desiccant at 100 % RH per unit mass of desiccant material. Te larger te maximum desiccant moisture uptake, te longer te adsorption and regeneration process times..1.3 Number of Transfer Units Larger number of transfer units (NTU) means more efficient eat and mass transfer witin te desiccant weel. Te optimal performance of a desiccant weel versus te desiccant weel NTU is similar to te maximum desiccant moisture uptake. Te adsorption-side outlet umidity decreases wit NTU. Terefore te performance improves by increasing te NTU. Zeng et al. [1995b] discussed te importance of NTU and ways to modify it..1. Speed of Rotation Te rotational speed of a desiccant weel is te number of rotations tat it undergoes per unit time. Tis speed determines te lengt of time te desiccant stays in te adsorption process as well as te lengt of time it is regenerated. Many autors note tat weels used for air deumidification are more sensitive to te speed of rotation compared to tose for entalpy recovery. Te desiccant weel must be operated at a optimum rotational speed to maximize te deumidification performance and terefore te rotational speed is a critical parameter for optimization [Zeng et al. 1995a]. Depending on te application and te parameters of te weel te correct speed is to be found to provide te optimum eat and mass transfer. Wen a desiccant weel rotates muc faster tan te optimum speed, te adsorption and regeneration processes are too sort wic results in a poor performance. Similarly if te rotary speed is lower tan te optimum ten te adsorption and regeneration processes are too long and more energy is

11 wasted in sensible eating/cooling tan in te sorption process and terefore is less effective.. Literature Review Te main feature of te Cromer cycle tat differentiates it from te conventional air conditioner is tat te dew point (moisture content) of te incoming air is substantially increased by te transfer of moisture to te air before it reaces te cooling coil. An increased average coil temperature results in improved energy efficiency over prior metods of sensible eat transfer for deumidification enancement and in increased deumidification over a conventional air conditioner. [Cromer 1988, p.] Tere are several teoretical analyses of te cycle performance in te literature. Nimmo et al. [1993] developed a simulation model tat calculates te air conditioner Energy efficiency ratio (EER) as a function of te Sensible eat ratio (SHR). Tey use tat model to compare te performance of te Cromer cycle wit tat of eat-pipeaugmented, single-speed air conditioner and an air conditioner wit a variable speed supply air fan. Te simulation results indicate feasibility of te cycle. Wen compared to te oter deumidification alternatives, te Cromer cycle maintains a iger EER over a wide range of SHR values. Rengarajan and Nimmo [1993] carried out a parametric study. First te autors compare te energy use, comfort (defined as te number of ours te space conditions are witin ASHRAE comfort zone) and te total cost (sum of capital costs and operation costs) for single speed air conditioners and variable speed air conditioners eac assessed wit and witout te addition of a desiccant weel and eat pipes. Te results from te parametric study sow tat te AC equipped wit Cromer cycle provides better comfort

1 at low energy use and at a lower total cost. Furtermore te autors evaluate te energy saving potential of te Cromer cycle by comparing it to a ig efficiency air conditioner. Te autors sow tat te ig efficiency AC as iger efficiency and consumes less energy tan te Cromer cycle. Wen, owever, te two are forced to maintain te ASHRAE comfort conditions for applications wit ig latent load for example, Miami te Cromer cycle consumes 10 percent less energy. Results from a study of te Cromer cycle benc test prototype under laboratory ARI test conditions are reported by Cromer [1997] (Table -1). Table -1 Performance of a standard AC andler compared wit an AC andler wit te Cromer cycle Standard AC unit AC unit wit Cromer cycle Improvement % Operational capacity [Btu/r] 53,590 66,38 3.8 Latent cooling [Btu/] 1,017 35,5 15.7 LHR [%] 6. 53.0 103.8 Deumidification [gal/] 1.56 3.93 153. Watts (over test our) 6709 5610 16. EER 7.99 11.8 7.9 (Cromer 1997, Cromer cycle: An energy efficient solution to indoor air quality problems. Engineering Solutions to Indoor Air Quality Problems, p. 9).3 Objectives of te Present Study Te teoretical studies and te laboratory test results publised in te literature sow tat for ig latent load applications te Cromer cycle is significantly superior compared to te standard AC configuration. Publised literature sows tat te Cromer cycle tecnology enables te evaporator coil to meet iger latent loads and it acieves tat at reduced energy consumption.

13 If similar beavior can be verified under field conditions te system could contribute to significant energy savings, wile providing better indoor conditions. Te purpose of tis study was to test te feasibility and effectiveness of te Cromer cycle tecnology under field conditions. To test te performance of te Cromer cycle in field conditions te following was done: Two residential ouses, located close to eac oter were equipped wit te necessary instrumentation. Te existing residential AC system of one of te ouses, called Test ouse, was retrofitted wit te Cromer cycle equipment, wile te AC system of te second ouse, called te Reference ouse, was kept in te standard vapor-compression configuration; Te Cromer equipment was alternately switced in and out of te AC system of te Test ouse. Data for te performance of te AC unit in its standard configuration and wit te Cromer cycle attaced were collected; Indoor conditions maintained in te Test ouse by te AC in its standard configuration and wit te Cromer equipment were compared; Several performance caracteristics as Q total, Q sensible, Q latent, LHR and apparent energy efficiency ratio were calculated and compared for te standard and te Cromer configurations in te Test ouse.

CHAPTER 3 EXPERIMENTAL SETUP For te purposes of te present study an existing residential air conditioning unit, located in te Solar House in te Solar Energy and Energy Conversion Laboratory (SEECL), Gainesville, Florida was cosen. A number of developments tat are now used worldwide originated tere. As recognition of teir important role te SEECL and te Solar House were designated as a Mecanical Engineering Heritage Site in January 003. 3.1 Experimental Facility Te Solar ouse (figure 3-1) as 0.3m ollow concrete block walls, a double wood floor and an aspalt single roof. Te ouse encloses approximately 110 m of Figure 3-1. Test ouse 1

15 living space and is orientated East-West. Tere is no additional insulation added on te walls. Tere is fiberglass bat, R-11 insulation over te conditioned space. Te AC distribution duct system of te ouse is located in te attic, directly above te conditioned space. Te conventional AC system in te Solar ouse is a vertical GrandAire tree ton ig efficiency air conditioner - condenser unit model GS3BA-036KA wit matcing GrandAire air andler GB3BM-036K-A-l0 model. Tis is a direct expansion R- AC system. 3. Measuring Instrumentation For te objective of te study, te Test ouse, tereinafter intercangeably referred as Solar ouse or Test ouse, was instrumented wit measuring devices to monitor and record te desired variables. Temperature and RH measurements were provided by combined sensor transmitters, manufactured by Vaisala Co. HMD60W sensor/transmitter was used to measure te inside RH and temperature. Ambient conditions were monitored by HMD60YO. HMD60Y sensors were used to measure te corresponding temperatures and relative umidities of te return air before te desiccant weel and te AC andler and of te supply air at te exit of te AC andler and after te weel. A scematic of te sensors locations in te duct system is given in figure 3-. All Vaisala sensor/transmitters were installed wit te factory calibration, specified to be ±0.3 o C for te temperature readings and ±% for te RH readings. Periodically, owever, te outside RH sensor was recalibrated because of problems wit condensate formation.

16 3 1 Figure 3-. Location of te combined temperature and RH sensors Energy consumption of te AC unit was measured using watt-our transducer Model WL0R-05, manufactured by Oio Semitronics Inc, Oio. Energy consumption of te weel was measured separately using a watt-our transducer Model WL0R-09, manufactured by te same company. Bot transducers were installed wit te factory calibration wic is specified to be ±0.5% of full scale. Te pulse pick up of te wattour transducers provides a pulse for every 10W consumed. Air flow measurements were provided by using a ot wire anemometer, manufactured by Comark Ltd. Its accuracy is specified to be ±3% of te reading. A data acquisition system based on LABTECH software was setup to acquire te signals from te corresponding measuring equipment. A program was designed to scan eac probe every 5 seconds, ten te readings were averaged for eac 0-second-period and recorded in a file. Anoter ouse, named te Reference ouse for tis study located next to te Test ouse, was monitored to be used as a control. Toug te reference building (figure 3-3), is located very close to te Test ouse, it is not te same. It is larger-

17 approximately 180 m (as compared to 110m test ouse), oriented Nort-Sout, as different glass area and uses 3 ton Trane condenser and air andler. Figure 3-3. Reference ouse Despite te differences, owever, wen te indoor conditions of te two ouses were compared it was establised tat te umidity levels maintained by te corresponding conventional AC systems were quite similar. Terefore it was decided tat tis Reference ouse could serve as a control for te purposes of te study. Since te ambient conditions vary, collecting data from tis Reference ouse would provide an additional control comparison. Monitoring simultaneously te Reference ouse, were no canges are made and te Test ouse would make it possible to estimate to wat extent te different comfort conditions maintained in te Test ouse wit te Cromer tecnology, are due to enancing te AC unit wit te Cromer equipment as opposed to te canges in te ambient conditions. If wit te Cromer cycle considerable decrease in

18 te RH levels is observed but te same trend is observed in te Reference ouse also ten te reason could be favorable ambient conditions rater tan te Cromer equipment. If owever, in te Test ouse considerable canges in te indoor conditions are observed between te standard configuration and te Cromer cycle wile in te Reference ouse te indoor conditions are maintained te same, ten it could be concluded tat te canges were due to te Cromer cycle. Te Reference ouse was equipped wit HMD60W sensor/transmitters for monitoring te inside temperature and RH and wit a WL0R-05 watt-our transducer for te AC energy consumption measurement. Te monitoring system of te Reference ouse was connected to te data acquisition system of te Test ouse. 3. Desiccant Weel Te desiccant weel, used in te present study, was manufactured by AirXcange Company, Rockland, MA. It was 0.075m wide, 0.9m diameter and consisted of 6 removable segments. 3..1 Desiccant Material In te study two types of desiccants, called Desiccant A and Desiccant B, were tested. Te first test was conducted wit Desiccant A material - te typical entalpy weel desiccant tat AirXcange company uses for teir entalpy weel products. Te desiccant weel ad flat laminate segments structured in an ideal parallel plate geometry. Te laminates were arranged continuously wit one flat and one structured layer (Figure 3- (A)). Te structured layer ad small conical internal dimples to separate te laminates and define te geometry of te matrix. For te second and te tird test te segments were replaced wit laminates tat were not flat but waved. Te waved

19 Figure 3-. Laminates design. (A) Flat laminates. (B) Waved laminates. segments ad axial ridges (Figure 3- (B)) to determine te geometry wile providing an obstruction for te air carryover from one side of te weel to te oter. Te new segments laminates were coated wit different silica gel, called desiccant type B. Te second desiccant type was suggested by Dr. Cromer. Under laboratory conditions te inventor ad tested several types of desiccants and found type B one to ave superior performance for Cromer cycle applications. 3.. Speed of Rotation In te current study te Cromer cycle performance was tested wit te desiccant weel rotating at two different speeds: 10 revolutions per our-a speed of rotation found to be te optimal in teoretical simulations [Nimmo et al. 1993]. It is to be admitted, toug, tat tis rotational speed was found optimal for a different desiccant type; revolutions per our-determined by Dr. Cromer, based on laboratory tests of a desiccant weel wit segments, coated wit type B desiccant.

0 3..3 Weel Accommodation Since desiccant weel transfers moisture between te return and supply air streams, te important moment in retrofitting an existing AC system wit te Cromer cycle is to reorient te air in order to direct te air flow troug te weel (Figure 3-5). Figure 3-5. Te experimental weel assembly Te original setup of te air andler in te Solar ouse in te standard AC configuration was tat te air andler was taking return air directly from te space being

1 conditioned via a sort straigt duct connected to a return air grille. Anoter sort and straigt duct connected te air andler supply side to a plenum box from wic a spider type distribution system brougt te air to eac room of te ouse. In its standard Figure 3-6. Te experimental setup

configuration te existing vertical air-andling unit in te Solar ouse fitted in 1x1xm (lengt by widt by eigt) space. In order to connect te weel witout modifying te distribution system, but at te same time to avoid any sarp turns, te retrofitted configuration took a 3.5x1.5x3m space (Figure 3-6). Tere was information tat in previous tests certain slipping ad been observed (te belt tat drives te weel skided on its surface, tus canging its speed of rotation). In order to avoid tat, in tis study te flat belt tat originally came wit te weel was replaced by a grooved belt (Figure 3-7). Figure 3-7. Te pulley and te driving belt A mecanism (Figure 3-8) was designed for monitoring te number of rotations of te desiccant weel. Tests sowed tat wit te new belt te problem of te belt slipping on te weel s surface was eliminated.

3 Figure 3-8. Mecanism for monitoring number of desiccant weel revolutions 3.. Power Supply of te Desiccant Weel Te desiccant weel is supposed to rotate only wen air is blowing in te AC system. Tis requirement was acieved by connecting te motor of te weel to te power supply via a relay as sown in te Figure 3-9. From Control Relay Relay FUSE Line 10V/AC Neutral Blower Motor 0V/AC Weel s Electric Motor 10V/AC Figure 3-9. Desiccant weel power supply layout

3.3 Condensate Measuring Equipment During te field study it was recognized tat it would be useful to measure automatically te water condensed by te cooling coil. For te purposes of te present test, a system was designed to measure te volume of te condensed water. Te scematic of te system and a picture of te condensate collecting part are given in figures 3-10 and 3-11 respectively. Te equipment was connected to te data acquisition system and was designed to send a pulse for eac 110ml of water removed by te cooling coil. Water Level Level Sensor N S Magnet Timer Relay FUSE 10V/AC Neutral Solenoid Valve 10V/AC Figure 3-10. Condensate measuring equipment 3. Data Acquisition System A data acquisition system (Figure 3-1) based on LABTECH software was setup to acquire te signals from te corresponding measuring equipment. Te program was designed to scan eac probe every 5 seconds, ten te readings were averaged for eac 0-second-period and recorded in a file. Since all variables were recorded at eac 0-

5 second-period, data collected are very detailed and made it possible to obtain information about beginning and end of eac cycle, its duration, state points of te air before and after te weel bot on its working and regeneration side, condensate removal, and te exact energy consumption. Figure 3-11. Condensate measuring vessel and valve

+17V Probe -0mA In.1 Analog Signals I I= -0 ma Us= 1-5V.1. In.8 M U X Gain Amp. ADC Data C o m p. +17V ma / 0mA Current to TTL Converter ma = Log.0 0mA = Log.1 TTL Level In1 In3 Counter Control Logic Address B U S 6 Data Acquisition System Digital Signals +17V DC +17V AC Figure 3-1. Layout of te data acquisition system

CHAPTER PROTOCOL AND EXPERIMENTAL RESULTS.1 Protocol An experimental protocol was developed to study operation of te AC system wen retrofitted wit te Cromer cycle equipment. It as to be noted tat no attempt was made to acieve operation under ARI conditions. Te goal was to test te performance of te Cromer cycle in field conditions. All data, collected for te periods of operation of te air andler as a conventional AC unit, were collected in a baseline set. Te data for te periods, wen te AC system was enanced wit te Cromer cycle, were collected in te Cromer set. Cromer set includes several data sets because Cromer tecnology was tested wit different desiccant types and at different speeds of rotation of te desiccant weel. Te protocol involved alternate switc of te Cromer equipment in and out of te AC system of te Solar ouse. Since ambient conditions cange alternated switc made it possible to obtain data bot for te baseline set and for te Cromer set under variety of ambient conditions. As already mentioned in Capter 3, during te first test te weel was rotated at 10 rp, a speed found to give optimum performance for te Cromer cycle under computer simulations [Nimmo et al. 1993]. Te second test was conducted wit te weel rotating at rp, a speed tat te inventor found to be te optimum under laboratory test of a weel, covered wit type B desiccant. For te tird test conducted wit te second weel and type B desiccant, te speed of rotation was reverted back to 10 rp in order to ceck 7

8 ow it would influence te overall performance and if tis would reduce te eat recovery from wat was observed in te preceding test. Before te beginning of eac field test, te system was run in its standard vaporcompression configuration. For te first test te termostat was set at o C and te blower was set at low speed. Te baseline test was run for a week. After tat te Cromer cycle equipment was installed. Te Cromer tecnology was tested wit te desiccant weel, consisting of flat laminate segments, coated wit desiccant type A and wit a weel rotation speed of 10 rp. Te tests were conducted for a two-week period and under te same settings as te preceding baseline test. Te data were collected in te Cromer set and stored as set No.1. For te next test te Cromer configuration was disconnected and data collected for te operation of te AC system in its standard configuration again. Te test was conducted for two weeks, te termostat set at o C and te blower speed at medium. After tis two-week period, te Cromer cycle was connected again. Tis time, owever, te performance was tested wit waved laminate segments, coated wit desiccant type B and at a speed of rotation of rp. Under te settings of o C and medium fan speed te test was run for a two-week period and te data collected in te Cromer set as set No.. Te next step was to switc te Cromer configuration out of te system for 10 days so as to collect more data for te baseline set. After tis 10-day-period te Cromer configuration was connected in te system again. Te new test was conducted for weeks. During tis test te settings, as well as te desiccant weel, were te same as in

9 set No. tests. Te only difference was tat te rotational speed of te weel was reverted back to 10 rp. Data from tis test were stored in te Cromer set as set No.3. Appendix A contains tables for all data sets collected. Eac table presents summarization of te ambient conditions, duration of operation, condensate removal and energy consumption of te corresponding AC systems in te Test ouse and in te Reference building. Te detailed data sets include extended tables for every day. Tese extended tables give information about te ambient conditions, te indoor conditions, temperatures and relative umidities at different places of te return and supply air ducts, number of cycles of operation, duration of eac cycle, condensate removed by te coil and energy consumption of te AC units bot in te Solar ouse and in te Reference building. Tese detailed tables are available on a compact disk.. Experimental Results..1 Baseline Set - AC Unit in Standard Configuration Data, collected during te periods wen te AC system in te Solar ouse was operating in a conventional cycle, are presented in te baseline data set and used for comparison. Plots of te typical indoor RH levels maintained in te Solar ouse and in te Reference building during te periods wen te AC systems in bot ouses were operating in te standard vapor-compression configuration are illustrated in figure -1. As indicated, despite te differences between te size of te two ouses and te different AC equipment, te inside RH levels maintained in te Test ouse were found to be similar to te inside RH levels maintained in te Reference ouse, and somewere in te range of 5-50%.

T [ o C] & RH [%] 70 60 50 0 30 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day T [oc] & RH [%] 70 60 50 0 30 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day T [ o C] & RH [%] 70 60 50 0 30 T [ o C] & RH [%] 70 60 50 0 30 30 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day Figure -1. Indoor temperature and RH profiles from Baseline set. (A) 09/18/001. (B) 09/0/001. (C) 06/10/00. (D) 08/05/00.

31.. Set No.1 Cromer Cycle Configuration wit te Desiccant Weel Using Flat Laminate Segments and a Rotational Speed of 10 rp Te immediate observation, after conducting tis first test of te retrofitted AC system, was tat installation of te desiccant weel introduced a considerable increase of te pressure drop in te AC system. Tis resulted in 35% decrease in te airflow rate (te mass flow rate dropped from 1680 kg/ down to 1080 kg/). Anoter observation was tat wit te weel in place tere was an increase in te average operational time of te air conditioning system. Since in Cromer configuration te desiccant weel transfers moisture and eat between te ig pressure side and te low pressure side of te weel, wit te radial and flat design of te laminates a significant re-circulation of air from supply to te return side of te weel was observed. Pysically it was easy to feel tis re-circulation but it was not quantified matematically. Calculations were conducted to quantify te amount, but because of te considerable uncertainty involved, no precise value could be given. Wit regard to te ability of te retrofitted AC system to improve te indoor conditions, tis data set is not very persuasive. Most of te time te Cromer cycle system was maintaining indoor relative umidities in te range of 5-50% as did te conventional arrangement. Tere were days toug, wen for a limited time during te day te inside RH levels were below 0%. Review of te data sows tat tese were te days wen te AC system was operating for more tan 5 ours a day. It was felt tat te reason for te inconclusiveness of any enancement of performance by te Cromer cycle was because te tests were conducted in te mont of October wen te ambient temperatures were lower. Te termostat setting was satisfied for longer periods and under tese circumstances run-time fraction of te air conditioner was low.

3 For more detailed evaluation of te AC performance, pairs of days wit similar ambient conditions from te baseline set and te Cromer set were cosen and compared. Days in eac pair were cosen based on meeting te following criteria: Ambient conditions witin eac pair were very similar Average ambient temperatures were ig trougout te day. Te following parameters were calculated: Total cooling capacity Q ( ) total Sensible cooling capacity ( ) Sensible eat ratio Q sensible = m& 1, [kj/] (-1) = m& ', [kj/] (-) Q sensible SHR = (-3) Q total Latent cooling capacity ( ' ) Q latent = m& 1, [kj/] (-) Latent eat ratio Q latent LHR = (-5) Q total Apparent energy efficiency ratio 1 Qtotal Btu / EERapp =, PowerInput W (-6) were 1, and ' (Figure -3) are te entalpy of te return air, entalpy of te supply air and entalpy of a condition wen te air is at te temperature of te return air but wit te umidity ratio of te supply air. Te symbol m& is used for te mass flow rate of te air. As an illustration, te results for two suc pairs are presented. Te ambient conditions during tose pairs of days, te corresponding run times of te systems in te 1 Te energy efficiency ratio is defined as Total cooling/power input under standard ARI conditions. Here EER was estimated under field conditions, terefore it was called EERapp.

33 two ouses monitored and te condensate removed by te air andler in te Test ouse are sown in table -1. Table -1. Ambient conditions and run times for te data pairs from te baseline set and Cromer set No.1 Date A B C D E F T amb >5 o C T amb >8 o C RH TH Ref.H Conden amb. duration duration sate ours o C ours o C % ours ours liters Baseline 09/17/01 9:50 8. 6:15 8.7 53.8 6:0:0 9:55:0 1.9 Cromer 10/05/01 10:10 8.79 6:55 9.85 55.09 7::38 8:36:19 5. Baseline 09/19/01 1:00 8.6 7:10 9.05 65.59 9:13:0 11:5:0.6 Cromer 10/06/01 11:35 8.8 6:35 9. 71.10 10:6:0 10:0:3 30.3 A = Time wen ambient temperature was above 5 o C. Left sub column denotes number of ours ambient temperature was above 5 o C. Rigt sub column denotes te average ambient temperature during tat time. B = Case were ambient temperature was above 8 o C. Left sub column denotes number of ours ambient temperature was above 8 o C. Rigt sub column denotes te average ambient temperature during tat time. C = Average ambient RH during te time of day wen te ambient temperature was above 5 o C. D = Duration of operation of te AC unit in te Test ouse. E = Duration of operation of te AC unit in te Reference ouse. F = Amount of water removed from te air by te AC coil in te Test ouse. Cromer cycle resulted in more water removal from te air. Te average operational time, owever, was increased. Data sow tat, wen te systems in te Solar ouse and in te Reference ouse were bot in standard configuration, on average it took 30% longer time for te air conditioner in te Reference ouse to meet te termostat setting. Wen, owever, te Cromer equipment was installed in te Solar ouse, tis

3 difference in te operational times was reduced. Furtermore, days were observed wen te operational time of te AC system in te Test ouse was longer tan te corresponding time in te Reference ouse. Comparison of te temperatures and relative umidities, maintained in te Solar ouse and in te Reference ouse are given in figure -. Te calculated performance parameters for te AC system in te Solar ouse are presented in a grapical form in figure -3. Eac point on te plots represents te average for one run period of te AC unit. It as to be noted tat since te sensor used to measure te temperature and RH after te coil was located after te fan, te measurements include te additional eat generated by te fan motor. Terefore, te cooling calculated ere is less tan te actual cooling performed by te cooling coil. To facilitate te calculations a Fortran program for calculation of te performance caracteristics was developed. Te print of te program developed is given in Appendix C. Te program used a link to te software package PROPATH (PROgram PAckage for THermopysical properties of fluids), courteously given for use to te SEECL by te PROPATH Group [Propat Group 001]. Te states and te corresponding cooling processes of te air in te Solar ouse are plotted on Psycrometric carts in figures - and -5. For plotting te carts an ASHRAE Psycrometric Cart software was used [Hands Down Software group 199]. In tese figures te states of te air wen te system was in its standard vapor compression configuration are depicted by te symbol S, were S1 is te return air state and S is te supply air state. Te symbol C depicts te states of te air wen te AC system was modified wit te Cromer equipment. Te state points are also depicted by

35 numbers 1 to, were 1 depicts return air state, -state of te air after te desiccant weel on its regeneration side, 3 -state of te air as it leaves te AC andler, and - state of te supply air after te desiccant weel, as depicted in figure -3. As illustrated in figures - and -3, te following observations are made for te Cromer configuration as compared to te standard configuration: Observed up to 50% increase in te latent eat ratio (latent cooling to total cooling); Observed approximately 5% increase in te latent cooling; Observed approximately 15% decrease in te total cooling performed; Observed approximately 30% decrease in te sensible cooling; Observed 15% reduction in te apparent energy efficiency ratio; and RH levels maintained in te space are sligtly lower. Increase in te LHR and te latent cooling observed in te Cromer configuration test, can be explained from te fact tat te desiccant weel desorbs moisture into te return air before it reaces te coil. Terefore te air is wetter and closer to its dew point, wic switces te total cooling toward more latent cooling. Te observed decrease in te sensible cooling could be explained from te reduced inlet air temperature and te reduced air flow due to te desiccant weel. Te desiccant weel reduces te temperature of te air before it reaces te cooling coil because of evaporative cooling. However, wen te corresponding initial and final states of te air in te standard configuration are compared wit te Cromer cycle on psycrometric cart, it is observed tat te entalpy difference between te inlet and te

70 70 T [ o C] & RH [%] 60 50 0 30 T [ o C] & RH [%] 60 50 0 30 0 1 3 5 7 9 11 13 15 17 19 1 Time of day 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day T [ o C] & RH [%] 70 60 50 0 30 T [ o C] & RH [%] 70 60 50 0 30 36 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day Figure -. Indoor temperature and RH profiles for Baseline set and Cromer set No.1. (A) Baseline set-09/17/01. (B) Cromer set-10/05/01. (C) Baseline set-09/19/01. (D) Cromer set-10/06/01.

Q total [kj/] 30000 8000 6000 000 000 EER [Btu/ / W] 10.0 9.5 9.0 8.5 8.0 7.5 0000 0.50 17 0 5 7 30 3 35 T amb [oc] 7.0 1000 17 0 5 7 30 3 35 T amb [oc] LHR 0.0 0.30 0.0 Q lat [kj/] 10500 9000 7500 6000 500 37 0.10 30 35 0 5 50 55 60 RH in [%] 3000 30 35 0 5 50 55 60 RH in [%] Figur e -3. Comparison of performance caracteristics. (A) Q total vs T amb. (B) EER vs T amb. (C) LHR vs RH in. (D) Q lat.vs RH in.

Figure -. Psycrometric cart for similar days (Baseline set, 09/17/01 Cromer set No.1, 10/05/01) 38

Figure -5. Psycrometric cart for similar days (Baseline set, 09/19/01 Cromer set No.1, 10/06/01) 39

0 outlet condition of te air is almost te same. Te increased moisture of te air results in a wetter coil tat facilitates te eat transfer tus offsetting te negative impact of te reduced air temperature on te eat transfer. Terefore te observed decrease in te sensible cooling is mostly due to te dramatic decrease in te airflow rate, caused by te introduction of te desiccant weel. Anyway, for Cromer cycle applications it would be better if te motor of te blower is located before te cooling coil. Tis way te eat input of te motor would offset te pre-cooling of te air resulting from te air passing troug te regeneration side of te desiccant weel. Decrease in te sensible cooling exceeds te corresponding increase in te latent cooling resulting in a decrease in te total cooling. Decrease in te apparent energy efficiency ratio observed can be explained wit te reduction of te total cooling performed by te coil. Decrease in te sensible cooling, performed by te evaporator coil, inevitably increases te operational time of te AC system. Since only te termostat setting controls te air andler, for one and te same setting te Cromer configuration will take longer time before it is able to satisfy te termostat. Terefore te system equipped wit te Cromer cycle will operate longer and will ave iger overall energy consumption. Wit te Cromer cycle in place te RH of te air in te supply ductwork is considerably lower. Tis way te requirement of te ASHRAE Standard 6 tat calls for maintaining te umidity in te ducts below 70% is accomplised, someting tat is really very difficult to satisfy wit te conventional AC vapor compression configuration. Drier ducts prevent fungus and bacteria from growing so te space conditioned is at a lower risk from suc contamination.

1..3 Set No. Cromer Cycle Configuration wit te Desiccant Weel Using Waved Laminate Segments and a Rotational Speed of rp In te Cromer set No. performance of te cycle was tested wit new desiccant segments for te weel. Te laminates of te new segments were coated wit silica gel, desiccant B type. In order to reduce te pressure drop te distance between te laminates was sligtly increased. Te larger air cannels, owever, would increase te recirculation of te air from te ig pressure to te low pressure side of te weel. To avoid tat te new laminates were not flat as in te set No.1 test but waved. After replacing te segments wit te new ones and converting two of te flexible turns in te duct into rigid ones and setting te blower speed to medium small improvement in te pressure drop in te system was observed. Te volume flow rate increased from 1080 kg/ to 1180 kg/. Tis, owever, was still below te specifications, provided by te manufacturer of te air andler. Te test sowed tat under ot ambient conditions, te retrofitted AC unit maintains considerably lower indoor RH levels. Wen compared to te umidity levels wit te standard configuration, it is seen tat te Cromer cycle enables te conventional AC system, in ot days, to acieve and maintain approximately 0% lower indoor RH levels. Logically te more ours te unit works te more uniform te profile of te umidity maintained. Tis test unambiguously verified tat te Cromer cycle is able to enance te deumidification, performed by te cooling coil. Again pairs of days wit similar ambient conditions from te baseline set and te present set were cosen for more detailed comparison. Ambient conditions and duration of operation of te AC system in te Test ouse in its standard configuration and wit te Cromer cycle for two suc pairs of days are sown in table -.

Table -. Ambient conditions and run times for te data pairs from te baseline set and Cromer set No. o o Date T amb >5 C T amb >8 C RH TH Ref.H Conden amb. duration duration sate ours o C ours o C % ours ours liters Baseline 06/07/0 10:05 6.93 3:05 30.3 76.16 5:53: 6:5:1 18.6 Cromer 06/6/0 10:30 7.3 3:55 8.67 7.30 7:0:59 6:3:00 30.03 Baseline 06/09/0 11:0 7.39 5:15 8.30 66.86 :57:18 6:3:0 19.7 Cromer 06/7/0 11:05 7.88 6:00 9.06 71.11 7:3:38 6:56:1 7.17 Cromer configuration resulted in increased water removal. Te average operational time of te unit, owever, was also increased. Comparison between te data collected in Cromer No. set and te data from te Baseline set for bot ouses sow tat wit te Cromer cycle te AC system in te Test ouse started operating at considerably longer run cycles. Furtermore, wen compared to te time of operation of te AC system in te Reference ouse, it is seen tat wile under similar ambient conditions te unit in te Reference ouse keeps its time of operation more or less te same, te retrofitted unit in te Test ouse increases its time of operation. Plots of te indoor conditions maintained are sown in figure -6. Figure -7 gives plots of te calculated performance caracteristics of te AC system for te standard configuration and for te Cromer cycle. State points of te air and te corresponding cooling processes are plotted on psycrometric carts as sown in figures -8 and -9. As wit te Cromer set No.1 test, ere also similar positive and negative consequences are observed. Only te magnitudes are different. Following observations are made for te Cromer cycle:

3 35% increase in te latent eat ratio 5 % reduction in te s ensible cooling 30% reduction in total cooling 5% lower energy efficiency ratio. In tis test if look at te latent cooling by taking te average of te data it would appear tat Q lat is at te same order of magnitude for te standard and for te Cromer configuratio n. Tis could lead to misleading conclusions. Te reas on for tat is tat te Cromer cycle maintains muc lower indoor RH tan te standard configuration. If tests were run wen te inside RH were te same for te two configurations tan te Cromer cycle would provide considerably iger latent cooling as sown by te trend lines in te corresponding Q = f InsideRH plots, given in Appendix D. An important observation tat came from tis series of tests was tat in field conditions te speed of rotation of te weel is very important for te eat transfer ability. At a rotational speed of rp te desiccant weel recovered more eat. Wile in Cromer set No.1 test te supply air was injected into te space at a temperature o lat ( ) approximately C lower tan in te standard configuration, in Cromer set No. te opposite was observed. Tis increased eat recovery is clearly illustrated on te corresponding Psycrometric carts on figures -8 and -9. Obviously te excess eat recovery is undesirable for AC applications. It reduces te sensible cooling tus increasing te time te system operates to satisfy te termostat setting and from tere te corresponding overall energy consumption. During tis test, it was observed tat wen te AC came on after idle periods longer tan 6-7 ours, an unpleasant odor was introduced in te space in te first to 3

70 70 60 60 T [ o C] & RH [%] 50 0 30 T [ o C] & RH [%] 50 0 30 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day 70 70 T [ o C] & RH [%] 60 50 0 30 T [ o C] & RH [%] 60 50 0 30 0 1 3 5 7 9 11 13 15 17 19 1 3 T ime of day 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day Figure - 6. Indoor temp erature and RH profiles for Baseli ne set and Crom er set No.. (A) Baseline set- 06/07/ 0. (B) Cromer set- 06/6/0. (C) Baseline set-06/09/0. (D) Cromer set-06/7/0.

33000 11 Q total [kj/] 30000 7000 000 1000 EER [Btu/ / W] 10 9 8 7 18000 0.6 0 6 8 30 3 3 T amb [ o C] 6 1500 0 6 8 30 3 3 T amb [ o C] 0.5 13000 5 LHR 0. 0.3 Q lat [kj/] 11500 10000 0. 8500 0.1 30 33 36 39 5 8 51 5 RH in [%] 7000 30 33 36 39 5 8 51 5 RH in [%] Figure -7. Comparison of performance caracteristics.. (A) Q total vs T amb. (B) EER vs T amb. (C) LHR vs RH in. (D) Q lat.vs RH in.

Figure -8. Psycrometric cart for similar days (Baseline set, 06/07/0 Cromer set No., 06/6/0) 6

Figure -9. Psycrometric cart for similar days (Baseline set, 06/09/0 Cromer set No., 06/7/0) 7

8 minutes of its operation. Tis could be explained wit te desiccant material picking not only moisture but oter gases as well and evaporating tem back... Set No.3 Cromer Cycle Configuration Using Desiccant Weel wit Waved Laminate Segments and a Rotational Speed of 10 rp In tis test te desiccant weel ad te same segments as in te set No.. Te only difference was tat te speed of te desiccant weel was reduced to 10 rp. Tis was done in order to ceck ow te performance would cange and to see if te reduced rotational speed would reduce te eat recovery observed in te preceding test. Ambient conditions and duration of operation of te AC system in te Test ouse in its standard configuration and wit te Cromer cycle for two similar pairs of days are given in table -3. Table -3. Ambient conditions and run times for te day pairs from te baseline set and Cromer set No.3 Date T amb >5 o C T amb >8 o C RH TH Ref.H Conden amb. duration duration sate ours o C ours o C % ours ours liters Baseline 08/06/0 16:00 7.96 1:30 31.99 75.31 11:1:18 13:58:1 3. Cromer 08//0 16:15 7.68 1:5 31.0 70.90 1:56:18 13:58:39 38. 7 Baseline 08/01/0 1:30 7.06 5:00 30.13 76.80 7:05:15 8:0:0. 9 Cromer 08/6/0 1:15 6.5 5:15 9.9 79.3 11:9:19 10:7:1 37. Comparison of te indoor RH maintained and te AC system performance caracteristics for te two pairs of similar days are sown grapically in figures -10 and -11. Te corresponding states of te air and te cooling processes are plotted on psycrometric carts and given in figures -1 and -13

9 Wen performance of te retrofitted air andler in te Cromer set No.3 tests is compared to te standard configuration under similar ambient conditions, te following is observed: Increased deumidification of te space -15% to 0% lower indoor RH levels are maintained; 0% increase in te latent eat ratio; 0% decrease in te sensible cooling; 0% decrease in te total cooling; 15% decrease in te energy efficiency ratio. During tis test again unpleasant odor was introduced in te conditioned space during te first few minutes of te AC system operation in te Cromer configuration after longer idle periods..3 Discussion As it was mentioned earlier, te main problem wit retrofitting an A C system w it te Cromer cycle equipment is t e enormous increase in te pressure drop, wic reduces te air flow rate and fro m tere many performance caracteristics deteriorate. An inter esting question tat was not answe red in tis researc is ow would te performance cange if te same airflow rat e could be provid ed wit t e Cromer equipment in place as witout it. An attempt was made to address tis question. Te air intake section connected to te Cromer unit was removed in order to allow te return air to enter te weel directly, some flexible duct connections were replaced wit rigid ones and te sections after te AC andler and after te weel were extended and widened. All tese canges increased te air flow rate to about 10% below te flow rate in te standard configuration.

70 70 T [ o C] & RH [%] 60 50 0 30 T [ o C] & RH [%] 60 50 0 30 0 1 3 5 7 9 11 13 15 17 19 1 3 Ti me of day 70 0 1 3 5 7 9 11 13 15 17 19 1 3 Tim e of d ay 70 T [ o C] & RH [%] 60 50 0 30 T [ o C] & RH [%] 60 50 0 30 50 0 1 3 5 7 9 11 13 15 17 19 1 3 Ti me of day 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day Figure -10. Indoor temperature and RH profiles for Baseline set and Cromer set No. 3. (A) Baseline set-08/06/0. (B) Cromer set- 08//0. (C) Baseline set-08/01/ 0. (D) Cromer set-08/6/0.

Q total [kj/] 3000 30000 8000 6000 000 000 EER [Btu/ / W] 11 10 9 8 7 0000 0. 0 6 8 30 3 3 T amb [ o C] 6 10000 0 6 8 30 3 3 T amb [ o C] 0.35 9000 51 LHR 0.3 0.5 0. Q lat [kj/] 8000 7000 6000 0.15 5 8 30 33 35 38 0 3 5 8 RH in[%] 5000 5 8 30 33 35 38 0 3 5 8 RH in [%] Figure -11. Comparison of performance caracteristics. A) Q total vs T amb. (B) EER vs T amb. (C) LHR vs RH in. (D) Q lat.vs RH in.

Figure -1. Psycrometric cart for similar days (Baseline set, 08/06/0 Cromer set No.3, 08//0) 5

Figure -13. Psycrometric cart for similar days (Baseline set, 08/01/0 Cromer set No.3, 08/6/0) 53

5 A new test, called te trial test, was run. Unfortunately weater deteriorated soon after te beginning of tose test. As a result very limited data were obtained. Ambient conditions, operational times and water removal for selected similar pairs of days from Cromer set No.3 and te trial test are given in table -. Wit te increased air flow in te trial test decrease in te duration of operation and in te condensate removal is observed. Indoor temperature and RH profiles are sown in figure -1. Comparison of te calculated performance caracteristics of te AC system in Cromer configuration wit different air flows is sown in figure -15. Table -. Ambient conditions and run times for two similar days for Cromer cycle configuration but wit different airflows Date T amb >5 o C T amb >8 o C RH TH Ref.H Conden amb. duration duration sate ours o C ours o C % ours ours Liters Cromer 08/06/0 11:55 30.18 5:15 30.9 67.93 9:16:1 8:5:0 8. 9 Trial test 08//0 1:35 8.05 7:15 9.3 71.3 8:18:19 8:13:1 7. 06 Cromer 08/01/0 7:35 7.58 :30 30. 75.97 8:33:38 8:57:59 8. 8 Trial test 08/6/0 7:15 7.9 3:10 8.80 73.95 7:05:0 5:09:0. 09 Te data, altoug quite limited to draw any general conclusions, confirmed tat te reduction in sensible cooling capacity and in te energy efficiency, observed in te previous tests, was indeed mostly due to te reduced airflow. As it can be seen from te trial test, wit te increased airflow te Cromer cycle still was able to maintain lower RH levels. It maintained iger LHRs tat imply greater latent fraction and better deumidification. Te increase in te airflow rate troug te system, toug, improved bot te total cooling and te energy efficiency. Te psycrometric carts (Figures -16 and -17) sow tat te states of te air in te Cromer cycle are very similar regardless of

55 te air flow rate. Since, owever, te sensible cooling performed by te coil increases wit te increased air flow, it is expected tat te unit will on for sorter periods to satisfy te termostat. Tis is expected to result in lower overall energy consumption. Te sorter run periods, owever, will make te indoor RH fluctuate more compared to te case wit te lower flow rate.

70 70 60 60 T [ o C] & RH [%] 50 0 T [ o C] & RH [%] 50 0 30 30 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day 70 70 60 60 56 T [ o C] & RH [%] 50 0 T [ o C] & RH [%] 50 0 30 30 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day 0 1 3 5 7 9 11 13 15 17 19 1 3 Time of day Figure -1. Indoor temperature and RH profiles for Cromer cycle wit different flow rates. (A) Cromer set-no.3-09/03/0. (B) Trial test-10/1/0. (C) Cromer set-no.3-08/19/0. (D) Trial test-10/13/0.

35000 11 Q total [kj/] 30000 5000 0000 EER [Btu/ / W] 10 9 8 7 15000 0.5 1 3 5 7 9 31 33 35 T amb [ o C] 6 1000 1 3 5 7 9 31 33 35 T amb [ o C] 0.5 1500 57 LHR 0. 0.35 0.3 Q lat [kj/] 11000 9500 8000 0.5 6500 0. 6 8 30 3 3 36 38 RH in [%] 5000 6 8 30 3 3 36 38 RH in [%] Figure -15. Comparison of performance caracteristics.. (A) Q total vs T amb. (B) EER vs T amb. (C) LHR vs RH in. (D) Q lat.vs RH in.

Figure -16. Psycrometric cart for similar days (Cromer set No.3, 09/03/0 Trial test, 10/1/0) 58

Figure -17. Psycrometric cart for similar days (Cromer set No.3, 08/19/0 Trial test, 10/13/0) 59

CHAPTER 5 CONCLUSIONS Te objective of te present study was to test te performance of te Cromer cycle in field conditions. Based on wat was observed in te field test conducted a few conclusions can be drawn. Te field tests confirmed te feasibility of te Cromer cycle. In ot and umid locations te tecnology enances te deumidification potential of te standard AC system. Te field test, owever, did not confirm te predicted energy savings. In field conditions lower indoor RH levels were acieved wit a corresponding increase in te overall energy consumption. In summary, te field study of te Cromer cycle tecnology confirmed tat under one and te same termostat setting, an AC system retrofitted wit te Cromer cycle maintains lower indoor RH levels in te space. Te retrofitted system, owever, accomplises te AC at a iger overall energy consumption. operation: Te following additional observations were made in regard to te Cromer cycle Introduction of te Cromer cycle equipment increases te pressure drop in te system tus reducing te airflow rate. Terefore wen an existing AC syste m is to be retrofitted wit te Cromer cycle larger air andling unit as to be inst alled to overcome te additional pressure drop; Te Cromer cycle enables a conventional AC system to maintain lower indoor umidity levels; Te Cromer cycle increases te latent eat ratio of te AC system; 60

61 Te Cromer cycle reduces te sensible cooling performed by te cooling coil, wic results in longer run time to satisfy te termostat setting. Terefore under one and te same termostat setting te Cromer cycle as a iger overall energy requirement compared to te standard configuration; Te air in te supply ductwork is muc drier, tat elps prevent fungus and bacteria from growing in te duct linings, so te building is at lower risk from ealt problems caused by suc contamination; For a sort period ( to 3 minutes) of operation after a longer idle period, an unpleasant odor is introduced into te space. Te equipment, at least in te configuration tested in te present study, requires considerable additional space tat not many omeowners can spare and may be willing to dedicate. In te tests conducted on ot days te Cromer cycle AC system maintained indoor relative umidities around 30%. For residential ouses suc low umidity levels toug are neiter required nor recommended. Terefore during days wit ig ambient temperatures, te termostat for te Cromer cycle AC system could be set at a iger temperature. Because of te enanced deumidification abilities, even wit te iger indoor temperature te Cromer cycle would be able to maintain te indoor conditions witin te ASHRAE comfort zone. Te iger termostat setting, owever, would result in sorter operational time and reduction in te overall energy consumption. Furtermore for applications, were maintaining low umidity levels is a must, te tecnology is feasible and may be considered as a possible solution. For te same termostat setting te AC system in Cromer cycle configuration consumes more energy tat te conventional ig-efficiency AC configuration. Te Cromer cycle, owever, increas es te moisture removal capabilities of te cooling coil and AC system is able to provide and maintain muc lower indoor RH levels tan te standard configuration.

APPENDIX A DATA SETS Tis Appendix contains summarized information for all te data sets processed for te purposes of te Cromer cycle field study.

Table A-1. Summary of all te data sets processed Solar House Reference House Tamb>5 o C Tamb>8 o C Tamb>3 o C RH (5) Date Duration AC Weel Energy Cond. Duration Energy Duration Temp Duration Temp. Durat Temp. 63 Hours W W W Liters Hours W ours o C ours o C ours o C % BASELINE SET 17-Sep-001 6:0:0 17730-17730 1.9 9:55:0 3000 9:50 8. 6:15 8.7 - - 53.8 18-Sep-001 6:39:18 18660-18660 0.8 9:10:18 9790 8:0 6.67 :00 8.88 - - 71.6 19-Sep-001 9:13:0 700-700.6 11:5:0 39790 1:00 8.6 7:10 9.05 - - 65.59 0-Sep-001 9:58:3 9770-9770 6.5 1:09:38 80 1:00 8.9 8:30 30.3 - - 6.81 1-Sep-001 11:35:59 3110-3110 8. 13:6:59 7010 13:35 8.1 9:10 30.85 - - 70.5 -Sep-001 9::59 710-710 30.9 11:5:0 10 9:00 30.68 8:15 31.07 - - 63.39 CROMER SET No.1 3-Oct-001 1:57:0 530 80 50 8.5 3:5:00 1380 8:30 7.39 3:5 8.6 - -.01 -Oct-001 :53:38 1370 190 13660 9.5 6:: 3710 9:0 8.35 5:50 8.93 - - 50.06 5-Oct-001 7::38 1080 360 10 5. 8:36:19 8530 10:10 8.79 6:55 9.85 - - 55.09 6-Oct-001 10:6:0 9970 30 3000 30.3 10:0:3 33500 11:35 8.8 6:35 9. - - 71.10 7-Oct-001 :59:0 790 10 8060 16.9 3:07:19 9680 - - - - - - 8-Oct-001 1:5:0 850 70 90 5.6 3:39:59 11810 5:55 5.93 - - - - 59.80 10-Oct-001 :18:0 1160 160 11800 13. 5:38:00 18650 7:0 6.8 1:0 8.33 - - 6.93 11-Oct-001 6:3:1 17670 300 17970 1 7:0:3 310 8:55 7.89 :0 8.66 - - 53.13 1-Oct-001 6:57:0 18890 30 1910 19.8 6:50:0 910 8:0 6.06 0:17 8.15 - - 67.9 13-Oct-001 8::00 800 300 3100 6.5 10:10:0 3310 10:10 7.71 :35 9.1 - - 61.01 1-Oct-001 8:35:0 3390 370 3760 8.3 6:19:58 0650 6:5 6.03 - - - - 8.31 15-Oct-001 3:3:0 9990 170 10160 - :0:38 15170 6:0 7. 1:15 8.8 - - 38.36 16-Oct-001 :17:38 11670 180 11850 - :3:18 160 7:5 7.0 0:0 7.80 - - 9.30 BASELINE SET 6-Jun-00* 8:16:0 560-560 1.7 9:10:0 31800 10:5 9.03 7:5 9.68 - - 6.5 7-Jun-00 5:53: 1770-1770 18.3 6:5:1 1650 10:05 6.93 3:05 30.3 - - 76.16

Table A-1. Continued Solar House Reference House Tamb>5 o C Tamb>8 o C Tamb>3 o C RH (5) Date Duration AC Weel Energy Cond. Duration Energy Duration Temp Duration Temp. Durat Temp. Hours W W W Liters Hours W ours o C ours o C ours o C % 8-Jun-00 3:0:0 11060-11060 1.0 :13:0 1190 7:05 6.79 1:0 8.9 - - 7.51 9-Jun-00 :57:18 15110-15110 19.5 6:3:0 50 11:0 7.39 5:15 8.30 - - 66.86 10-Jun-00 5:5:0 16600-16600 0.7 8:06:0 780 1:30 7.8 6:15 8.5 - - 66.05 11-Jun-00* 3:6:58 1150-1150 13.9 :36:01 1580 5:50 6.87 0:50 8.50 - - 75.6 1-Jun-00 7:07:39 060-060.9 8:5:00 30590 1:55 8.7 7:00 9.6 0:55 3.5 70.31 13-Jun-00 8:5:0 600-600 6. 11:6:00 1550 10:50 30.68 9:0 30.9 :55 33.15 61.91 1-Jun-00 10:0:0 3180-3180 30.5 1:36:00 50 15:5 9.9 11:15 31.37 :5 3.85 63.9 15-Jun-00 9:5:00 30950-30950 5.6 11:8:0 1010 15:00 7.5 11:05 31.17 3:55 3.70 68.53 16-Jun-00 8:00:0 870-870 0.7 9:9:0 33190 13:5 9.9 10:55 9.9 - - 5.03 17-Jun-00 :3:0 7050-7050 9. :50:00 930 1:10 5.18 - - - - 76.5 18-Jun-00 rainy and cold - - - 0:36:0 1970 - - - - - - 19-Jun-00* 0::00 50 50.5 - - :55 6.9 0:0 8.53 - - 7.31 CROMER SET No. 19-Jun-00* :0:0 170 190 1910 17.7 3:18:39 1170 3:15 8.51 :50 9.9 - - 59.90 0-Jun-00 5:0:1 1560 30 15850 3.8 5:35: 1880 9:0 6.96 :15 9.0 - - 70.18 1-Jun-00 :1:0 600 100 6500 11.3 :0:19 8730 0:0 5 - - - - 7.50 -Jun-00 0:1:00 1960 30 1990 3.9 - - - - - - - - - 3-Jun-00 :0:00 11910 160 1070 19.7 :57:59 1680 9:05 6.17 :30 8.6 - - 80.79 -Jun-00 6:07:1 1790 70 18190 5. 5:53:39 0090 9:05 8.08 6:5 8.6 - - 67.89 5-Jun-00 3:6:00 10850 170 1100 16.8 5:00:19 17550 5:05 7.8 :0 8.89 - - 71.81 6-Jun-00 7:0:59 370 30 690 30.0 6:3:00 3380 10:30 7.3 3:55 8.67 - - 7.30 7-Jun-00 7:3:38 1560 30 1880 7. 6:56:1 70 11:05 7.88 6:00 9.06 - - 71.11 8-Jun-00 10:1:0 30080 50 30530 3.5 8:8:0 3090 11:0 9.63 8:35 30.7 1:50 3.30 7.5 9-Jun-00 7:36:38 1980 30 300 7.0 6:8: 180 6:0 9.33 :0 30.8 - - 80.0 30-Jun-00 5:06:0 1800 0 1500 0.7 :58:00 17000 6:5 7.1 :35 9.6 - - 87.5 6

Table A-1. Continued Solar House Reference House Tamb>5 o C Tamb>8 o C Tamb>3 o C RH (5) Date Duration AC Weel Energy Cond. Duration Energy Duration Temp Duration Temp. Durat Temp. Hours W W W Liters ours o C ours o C ours o C % 1-Jul-00 7:06: 0930 80 110 9.6 8:57:01 3100 11:15 8.07 6:10 9. - - 78.86 -Jul-00 10:05:0 950 60 9910 35.1 7:53:0 730 10:5 9.0 8:35 9.76 - - 75.83 3-Jul-00 8:05:0 3500 390 3890 8. 7:00:0 60 6:55 9.77 5:30 30.63 0:15 3.17 77.1 -Jul-00 7:00:0 0370 300 0670 5.3 6:03:0 1550 9:05 7.59 3:0 9.67 - - 83.01 BASELINE SET 1-Aug-00 7:05:15 180-180.9 8:0:0 8880 1:30 7.06 5:00 30.13 - - 76.8 -Aug-00 5:5:00 17780-17780 18. 8:05:0 760 7:30 7.1 :5 9.7 - - 75.81 3-Aug-00 5:8:00 17360-17360 17.5 7:3:00 5700 7:10 8. :0 9.15 - - 69.65 -Aug-00 :37:0 13710-13710 17.7 5:17:0 19770 :55 7.5 :15 8.66 - - 78.81 5-Aug-00 8:51:53 700-700 7.0 10::0 37350 15:10 9.0 9:5 30.8 0:30 3.31 70.7 6-Aug-00 11:1:18 35850-35850 3. 13:58:1 9310 16:00 7.96 1:30 31.99 7:00 33.5 75.31 7-Aug-00 8:3:3 670-670 9.9 9:0:00 3800 0:50 6.8 :5 8.61 - - 79.7 8-Aug-00 :8: 1350-1350 1.1 7:16:38 660 10:50 8.8 7:50 9.18 - - 7.5 9-Aug-00 :1:01 160-160 15.0 6:36:01 060 6:5 6.59 3:50 8.5 - - 65. 10-Aug-00 6:3:00 19960-19960 1.8 9:0:0 3070 1:05 8.9 7:55 9. - - 59.13 CROMER SET No.3 1-Aug-00 :51:57 13930 10 110 19.6 :07:59 13690 3:5 6.98 1:00 8.8 - - 77.9 13-Aug-00 7:7: 0 30 50 8.9 6:0:01 30 8:35 8.1 5:0 9.6 - - 68.87 1-Aug-00 8:39:3 6670 390 7060 30.0 9:11:00 30 10:00 7.6 7:0 30.39 - - 76.80 15-Aug-00 6:15:00 18810 90 19100. 6::00 370 5:35 9.03 :30 9.7 - - 7.06 16-Aug-00 10::59 3360 80 3370 35. 1:11:0 390 15:50 8.5 9:05 9.75 1: 3.9 73.30 17-Aug-00 11::0 35190 510 35700 0.5 1:0:0 950 11:5 7.5 8:05 30.9 :50 3.79 78.8 18-Aug-00 9:33:0 9130 0 9550 3.7 9:17:0 3730 7:50 9.75 6:05 30.7 0:35 3. 66.80 19-Aug-00 8:33:38 5180 380 5560 8.8 8:57:59 3100 7:35 7.58 :30 30. 0:0 3.15 75.97 65

Table A-1. Continued Solar House Reference House Tamb>5 o C Tamb>8 o C Tamb>3 o C RH (5) Date Duration AC Weel Energy Cond. Duration Energy Duration Temp Duration Temp. Durat Temp. Hours W W W Liters Hours W ours o C ours o C ours o C % 0-Aug-00 6:0:36 1780 60 1750 1.3 :37:0 150 3:0 7.6 1:05 8.96 - - 76.33 1-Aug-00 9:5:0 790 0 830 9.5 10:56:57 3880 1:10 9.0 9:10 30.1 - - 66.71 -Aug-00 10:19:18 3050 60 31000 30.7 11:5:59 0000 1:0 9.19 9:10 30.98 :5 3.9 6.1 3-Aug-00 11::39 370 500 350 33.1 10:5:00 37850 15:50 9.3 10:50 30.91 1:50 3.37 6.03 -Aug-00 1:56:18 3860 570 3910 38.7 13:58:39 980 16:15 7.68 1:5 31. 6:5 3.91 70.90 5-Aug-00 1:3:01 360 650 3910.0 13:10:1 6380 18:15 7.61 9:55 30. 5:05 33.0 7.56 6-Aug-00 11:9:19 350 530 35070 37. 10:7:1 36170 1:15 6.5 5:15 9.9 - - 79.3 7-Aug-00 6:38: 1980 90 19570 5.3 6:11:01 1360 6:55 7.17 1:30 8.1 - - 78.7 8-Aug-00 8::39 0 00 80 7.7 10:09:3 35050 9:0 7.96 5:50 8.86 - - 73.60 9-Aug-00 6:37:59 1910 310 1950.6 7:3:01 530 8:05 8.08 5:10 8.9 - - 71.55 30-Aug-00 :35:1 1310 10 13330 16.7 :1:38 1100 :5 7.5 1:30 9.06 - - 78.9 31-Aug-00 6:03: 17670 70 1790.0 6:0:19 060 10:10 6.9 0:0 8.19 - - 78.0 1-Sep-00 8:56:18 6390 0 6810 9.6 8:51:0 31180 13:05 8.3 5:50 30.1 - - 73.56 -Sep-00 10:07: 970 30 30150 31. 9:3:59 330 15:00 8.31 7:35 9.99 - - 71.51 3-Sep-00 9:16:1 7090 0 7510 8.5 8:5:0 30300 11:55 30.18 5:15 30.9 - - 67.93 TRIAL TEST 11-Oct-00* 8:8:18 700 380 70 3. 8:3:00 9090 10:0 9.3 7:30 30.5 - - 6.71 1-Oct-00 8:18:19 5010 380 5390 7.1 8:13:1 870 1:35 8.05 7:15 9.3 - - 71.3 13-Oct-00 7:05:0 190 300 1590.1 5:09:0 1860 7:15 7.9 3:10 8.8 - - 73.95 1-Oct-00 5:15: 1580 30 16070 19.7 5:3:0 0530 6:55 8.11 :30 9.06 - - 67.5 66

APPENDIX B UNCERTAINTY ANALYSIS Tis appendix presents te uncertainty analysis carried out to evaluate te unc ertainty of te experimental measurements as well as of te quantities calculated using tese measured variables. Devices used to measure te relative umidity, temperature, air velocity, etc. ave certain accuracy tat impacts te uncertainty of te corresponding experimentally measured quantities. Table B-1 presents te uncertainty of te experimental measurements. Table B-1. Uncertainty of experimental measurements Quantity Instrument Uncertainty Re lative umidity HMD60U/YO ± % Temperature HMD60U/YO ± 0.3 o C Air veloc ity Hot Wire Anemometer ± 3% reading Circumference Ruler ± 0.5 cm Watt-ours Transducer Model WL0R-05 ± 0.5% F.S. Te uncertainty analysis of te calculated values, t at use experimentally measured quantities, was done following te metod, described by Coleman [1999]. Briefly te metod involves te following procedure: Let certain experimental result R, is a function of N number of measured variables X i : R = R (X 1, X, X 3,..., X n ) (B-1) Ten te uncertainty in te result is given by U R R + R U X U X... U 1 X n X + R = + (B-) X 1 X n 67

68 were te U are te uncertainties in te measured variables. X i By dividing eac term in te equation by X i R and by multiplying eac term on te rigt-and side by ( X i X i ), we obtain te uncertainty equation in a nondimensionalised form: U R R X = R 1 R X 1 U 1... X X R U X X n R U X n + X + + 1 R X X R X n X n (B-3) were U R R is te relative uncertainty of te result. Te factors U X i X i are te relative uncertainties for eac variable. Te factors in te parentesis tat multiply te relative uncertainties of te variables are called uncertainty magnification factors (UMFs). Tey indicate te influence of te uncertainty of te corresponding variable on te uncertainty in te result. If a UMF is greater tan 1 tis indicates tat te uncertainty in te variable is magnified as it propagates troug te data reduction equation into te result. If UMF value is less tan 1 ten te uncertainty in te variable is diminised as it propagates troug te data reduction equation into te result. For instance, tis metod applied to our study to determine te uncertainty in te Latent cooling ( ' ) Q lat = m 1 gives te following: Qlat + Q lat = Qlat U + 1 Q lat U U U (B-) ' m 1 ' m Te partial derivatives of Qlat in respect to te tree variables, used for its calculation: Q lat = m 1 Q ; lat = m ' Qlat and = ( ' ) m 1 (B-5)

69 By substituting in te equation and nondimensionalising we obtain U Qlat U ' U 1 1 ' U m Q lat = 1 ' 1 + 1 ' ' + m (B-6) All te calculated quantities, te functions of tose quantities, te partial derivatives of tese functions wit respect to te variables tey depend on and te corresponding non-dimensional form of te uncertainty equation, are listed in table B-. In te current study uncertainties of te quantities calculated ave been determined by picking several days in eac set of data, averaging te daily values for eac day, finding te corresponding uncertainty and after tat finding te average uncertainty for eac data set. Tis process is illustrated in table B-. Te iger uncertainties observed in respect to Q lat and LHR are as a result of te comparatively small moisture removal. As it can be seen from te corresponding uncertainty equations, tis small moisture removal makes te denominator (1-') small and consequently te UMFs are ig. Terefore in teses cases te uncertainties of te measured quantities magnifies as tey propagate troug te data reduction equations into te final results for Q lat and LHR.

70 Table B-. Average uncertainty Date U latent U sens U LHR U total U EER U 1 U U ' % % % % % % % % No weel 6 June, 00 31.16 7.9 9.6 6.3 6.5.30.17.7 8 June, 00 5.91 10.68.3 7.88 7.90.0.07.73 9 June, 00.76 11.38 1.1 8.10 8.11.17.05.68 1 June, 00 6.3 8..17 6.58 6.60.5.09.66 1 June, 00 9.35 8.05 7. 6.7 6.8.7.11.80 16 June, 00 33.07 9.8 8.58 8.09 8.11.3.13.73 AVERAGE 8.1 9.31 5.8 7.3 7..5.10.7 Weel ( rpm) 7 June, 00 1.1 11.63 18. 7.59 7.60.56.60.1 8 June, 00 3.89 11.8 0.97 7.68 7.70.57.67.60 3 July, 00.5 11.19.55 7.5 7.6.56.66.67 AVERAGE.51 11.37 1.5 7.57 7.59.56.6.9 Weel (11 rpm) 13 Aug., 00 19.3 8.8 16.93 6.13 6.15.3.55 3.86 8 Aug., 00. 8.30 1.36 6.59 6.61.5.67 3.81 30 Aug., 00.16 8.70 19.6 6.53 6.55.5.60.3 31 Aug., 00 0.53 8.73 17.79 6.53 6.55..58 3.76 1 Sep., 00.76 8.39.01 6.6 6.6.53.6.0 Sep., 00 5.6 8.7.51 6.66 6.68.57.69 3.7 3 Sep., 00 3.89 7.6 1.98 5.83 5.85.03.69 3.9 AVERAGE.89 8.8 0.3 6.1 6.3.0.63 3.91 Table B- 3. Uncertainty of calculated values Quantity No weel Weel ( rpm) Weel (11 rpm) 1. Entalpy Point 1: ±.5 % Point : ±.10 % Point ': ±.7 % Point 1: ±.56 % Point : ±.6 % Point ': ±.9 % Point 1: ±.0 % Point : ±.63 % Point ': ± 3.91 %. Mass flow rate ± 3.1 % ± 3.1 % ± 3.1 % 3. Sensible cooling ± 9.31 % ±11.37 % ± 8.8 %. Latent cooling ± 8. % ±.51 % ±.89 % 5. Total cooling ± 7.3 % ± 7.57 % ± 6.1 % 6. LHR ± 5.8 % ±1.5 % ±0.3 % 7. EER ± 7. % ± 7.59 % ± 6.3 %

Table B. Calculated values Quantity Functio n Partial derivatives Latent cooling Q lat Sensi ble cooling Q Total coo ling Q total Q ( '. lat lat = m 1 1 Q lat. sens m( ) sens ' ' Q lat m Q sens ' Q = = Q m. Q Q total = m( ) total 1 = ). Q = m = ( ' ) 1. m Q sens = ' 1 Q m. m Q total = ( 1. = m. sens = m (. total = m ; ; ; ) ) ; U Q U Q U = 1 Uncertainty equation U Q lat 1 ' lat Q sens sens = 1 ' 1 1 ' ' ' 1 ' U U ' ' U + ' + ' U U Q total 1 Q total = + U m + m U + m m U m + m 1 1 1 71

Table B. Continued 7 Latent eat ratio R lat = 1 R lat ' ( ) ' R lat = ) ' lat = ( ) 1 ' R lat = ( ) ( )( ) 1 1 1 1 ' ' ' ' 1 ' + + = U U U R U lat R lat = 1 ' R sens ( ) 1 1 ' R sens = ( ) 1 1 ' R sens = ( ) 1 1 R sens = ( ) ( )( ) 1 1 1 1 1 1 ' ' ' ' ' ' + + = U U U R U sens R sens in total W Q EER = Q total W in EER 1 = W W EER = + = in W total Q EER W U Q U EER U in total 1 1 1 ( 1 R 1 1 1 1 ' ' EER in Q total Sensible eat ratio R sens Energy efficiency ratio

APPENDIX C FORTRAN 77 PROGRAM FOR CALCULATION OF PERFORMANCE CHARACTERISTICS OF THE COOLING CYCLE Te following program was written to calculate te parameters used to evaluate te performance of te cooling cycle. Tese parameters are Qtotal, Qlat, Qsens, LHR, SHR and apparent energy efficiency ratio. Te program also obtains suc termopysical properties of te moist air as entalpy and umidity ratio. To accomplis te latter te program uses a link to te software package PROPATH (PROgram PAckage for THermopysical properties of fluids). PROGRAM CROMER.FOR C ***************************************************************** C * THIS PROGRAM CALCULATES THE FOLLOWING PERFORMANCE * C * CHARACTERISTICS : * C * LATENT AND SENSIBLE COOLING * C * LATENT HEAT RATIO AND SENSIBLE HEAT RATIO * C *TOTAL COOLING PERFORMED BY AN AC HANDLER * C *APPARENT ENERGY EFFICIENCY RATIO * C * AND BASED ON TEMPERATURE AND RELATIVE HUMIDITIES * C *FINDS ENTHALPY AND ABSOLUTE HUMIDITY * C **************************************************************** C C C C C C C C C C DECLARATIONS OF TABLE COLUMNS REAL TEMPO(5), ENTH(5,10),HUMR(5,10),TIME(5),HDIF(5) DECLARATION OF THE MAIN ARRAYS REAL INPUT (5,13),RESULT (5,9),PROPERTY(5,11), POWER (5) INTEGER AIRFLOW IROW - NUMBER OF ROWS I3 - NUMBER OF COLUMNS (1 FOR AC WITH NO WHEEL) CHARACTER*10 FILESOURCE, FILERESULT, PROPERTIES, 1 ENTEXT, HRTEXT INFORMATION REQUIR ED FOR THE LINK TO THE PROPERTY PROGRAM COMMON/UNIT/ KPA, MESS 73

7 KPA=1 MESS=1 P=1.0135 C C DATA INPUT - COMMUNICATION WITH THE USER C ENTEXT=' H' HRTEXT=' W' PRINT*, 'ENTER THE NAME OF THE SOURCE FILE + EXTENSION' READ*, FILESOURCE PRINT*, 'ENTER A NAME FOR THE RESULT FILE +EXTENSION' READ*, FILERESULT PRINT*,'PLEASE ENTER THE AIR FLOW [kg/]' READ*, AIRFLOW PRINT*,'PLEASE ENTER NUMBER OF ROWS' READ*, IROW C C DATA INPUT - PROGRAM READS FROM THE USER'S FILE C OPEN (0, file=filesource,access='direct', 1 FORM='FORMATTED',RECL=10) DO 10 I1=1,IROW READ(0, FMT=100) (TEMPO(I), I=1,10),TIME(I1),POWER(I1) DO 11 I3=1,10 INPUT(I1,I3)=TEMPO(I3) 11 CONTINUE INPUT(I1,I3)=TIME(I1) INPUT(I1,I3+1)=POWER(I1) 10 CONTINUE 100 FORMAT (1X, 10(F6.), F6., 1X,F7.1) CLOSE(0) C C LIN K TO PROPATH PROPERTY PROGRAM FOR DETERMINATION OF C ENTHALPY AND ABSOLUTE HUMIDITY C DO 13 I1=1,IROW DO 1 I3=1,9, ENTH(I1,I3)=HC(P,INPUT(I1,I3),(INPUT(I1,(I3+1)))/ 100) HUMR(I1,I3)=XC(P,INPUT(I1,I3),(INPUT(I1,(I3+1)))/100) PROPERTY(I1,I3)= ENTH(I1,I3)*0.001 PROPERTY(I1,I3+1) = HUMR(I1,I3)*1000 1 CONTINUE 13 CONTINUE C C CALCULATION OF POINT Hprime=f(Wsup,Tret) C C Wsup = PROPERTY(I1,10), Tret = INPUT(I1,3) DO 15 I1=1,IROW TEMPO(I1)=HD(P,INPUT(I1,3),(PROPERTY(I1,10)*0.001)) PROPERTY(I1,11)=TEMPO(I1)*0.001 PRINT*,'HPR=',PROPERTY(I1,11) 15 CONTINUE C C C C CALCULATION OF SHR, LHR, COOLING CAPACITY AND EER

75 C CALCULATION OF SENSIBLE HEAT RATIO (SHR) C AND LATENT HEAT RATIO (LHR) C DO 17 I1=1,IROW HDIF(I1)=PROPERTY(I1,3)-PROPERTY(I1,9) RESULT(I1,1)=(PROPERTY(I1,11)-PROPERTY(I1,9))/HDIF(I1) RESULT(I1,)=(PROPERTY(I1,3)-PROPERTY(I1,11))/HDIF(I1) 17 CONTINUE C C CALCULATION OF TOTAL COOLING CAPACITY C Q total = (1-) x (mass flow rate) C DO 18 I1= 1,IROW RESULT(I1,3)=AIRFLOW RESULT(I1,)=HDIF(I1)*AIRFLOW RESULT(I1,5) =RESULT(I1,)*0.978 PRINT*,'COOLING=', RESULT(I1,) 18 CONTINUE C C CALCULATION OF EER = (TOTAL COOLING)/(POWER INPUT) C DO 19 I1=1, IROW RESULT(I1,6)=INPUT(I1,1) RESULT(I1,7)=RESULT(I1,5)*INPUT(I1,11)/RESULT(I1,6) 19 CONTINUE C C CALCULATION OF SENSIBLE AND LATENT COOLING C DO I1= 1,IROW RESULT(I1,8)=(PROPERTY(I1,11)-PROPERTY(I1,9))*RESULT(I1,3) RESULT(I1,9)=(PROPERTY(I1,3)-PROPERTY(I1,11))*RESULT(I1,3) CONTINUE C C RESULTS RECORDED IN A FILE C OPEN (1, file=fileresult) WRITE(1, FMT=107) WRITE(1, FMT=108) 'DATE IS:', FILESOURCE WRITE(1, FMT=107) WRITE(1, FMT=106) 'Tt','RHt','Tret','RHret','Tbef','RHbef', 1 'Taft','RHaft','Tsup','RHsup',' Time' WRITE(1, FMT= 106) 'oc ',' % ',' oc ',' % ',' oc ',' % ', 1 ' oc ',' % ',' oc ',' % ',' ours' DO 0 I1=1, IROW WRITE(1, FMT=103) I1,(INPUT(I1,I3), I3=1,10),INPUT(I1,11) 0 CONTINUE WRITE(1,101) (ENTEXT,I1,HRTEXT,I1, I1=1,5), Entext,'Pr' DO 16 I1=1,IROW WRITE(1, FMT=10),I1, (PROPERTY(I1,I3), I3=1,11) 16 CONTINUE 101 FORMAT (/5(5X, A, I1, 5X, A,I1),6x, A,a3) 10 FORMAT (I,1X,5(F8.3, F8.3),1X,F8.3) WRITE(1, FMT=107) WRITE(1, FMT=10) 'SHR','LHR','FLOW','COOLING','COOLING', 1 'POWER','EER','SENS COOL','LAT COOL' WRITE(1, FMT=10) ' - ',' - ','kg/',' kj/ ',' Btu/',

76 1 ' W ',' - ','kj/','kj/' DO 1 I1=1,IROW WRITE(1, FMT=105) I1,(RESULT(I1,I3),I3=1,9) 1 CONTINUE C C 103 FORMAT (1X, I, 10(F8.), 1X,F8.) 10 FORMAT (5X, A, X,A, 1X,A6, (X,A8), X,A5, X,A, 1 3X,A9, 3X,A9) 105 FORMAT (I,1X,(1X,F5.), X,F6.0, (F10.), X,F6.0, 1 X,F5., 1X,F10., X,F10.) 106 FORMAT (5X,A5, X,A5, 3X,A5, X,A6, X,A5, X,A7, (1X,A7), 1 3X,A6) 107 FORMAT (//) 108 FORMAT ((X,A10)) CLOSE(1) END

APPENDIX D LATENT COOLING TRENDLINES

1000 1000 1500 1500 Q lat [kj/] 11000 9500 Q lat [kj/] 11000 9500 8000 8000 6500 8 31 3 37 0 3 6 9 5. RH in [%] 6500 6 9 3 35 38 1 7 50 RH in [%] 78 Cromer cycle y = 8.59x + 69. R = 0.9693 Standard configuration y = 7.01x - 11711 R = 0.9051 Cromer cycle y = 73.6x + 770.5 R = 0.7958 Standard configuration y = 585.7x - 17076 R = 0. 8916 Figure D-1. Latent cooling vs. RH in. (A) Baseline 06/07/0, 06/09/0; Cromer 06/6/0, 06/7/0. (B) Baseline 06/10/0, 06/13/0; Cromer 06/6/0, 06/7/0. (C) Baseline 08/01/0, 08/06/0; Cromer 08//0, 08/6/0. (D) Baseline 08/0/0, 08/05/0; Cromer 08/16/0, 08/19/0. (E) Baseline 08/03/0, 08/06/0; Cromer 08/19/0, 08//0. (F) Cromer set No.3-08/19/0, 09/03/0; Trial test- 10/1/0, 10/13/0.

10000 11000 9000 10000 Q lat [kj/] 8000 7000 Q lat [kj/] 9000 8000 6000 7000 5000 5 30 35 0 5 50 Cromer configuration y = 38.61x - 615.7 R = 0.653 RH in [%] Standard configuration y = 03.5x - 980.5 R = 0.71 6000 8 31 3 37 0 3 6 9 Cromer configuration y = 333.3x - 38 R = 0.9 RH in [%] Standard configuration y = 66.38x - 1367 R = 0.793 79 Figure D-1. Continued

10500 1500 9500 11000 Q lat [kj/] 8500 7500 Q lat [kj/] 9500 8000 6500 6500 5500 5 8 31 3 37 0 3 6 9 Cromer cycle y = 36.13x - 11.9 R = 0.8813 RH in [%] Standard configuration y = 505.1x - 1006 R = 0.698 5000 6 8 30 3 3 36 38 Cromer test No.3 y = 358.5x - 305 R = 0.93 RH in [%] Trial test y = 335.63x - 358.81 R = 0.56 80 Figure D-1. Continued