Simplified models for the performance evaluation of desiccant wheel dehumidification

INTERNATIONAL JOURNAL OF ENERGY RESEARCHInt. J. Energy Res. 2003; 27:17–29 (DOI: 10.1002/er.856)

Simplified models for the performance evaluation of desiccantwheel dehumidification

M. Beccali*,y, F. Butera, R. Guanella and R.S. Adhikari

Dipartimento di Ricerche Energetiche e Ambientali (DREAM), Universita degli Studi di Palermo, viale delle Scienze bld 9,

90128-Palermo, Italy

SUMMARY

In the present communication, simple models have been presented to evaluate the performance of rotarydesiccant wheels based on different kind of solid desiccants e.g. silica gel and LiCl. The first part of thepaper presents ‘Model 54’ which is developed for silica gel desiccant rotor. The model has been derivedfrom the interpolation of experimental data obtained from the industry and the correlations have beendeveloped for predicting outlet temperature and absolute humidity. The ‘Model 54’ consists of 54coefficients corresponding to each correlation for outlet absolute humidity and temperature and it is foundthat the model predicts very well the performance of silica gel desiccant rotor (Type-I). In the second partof the paper, a psychrometric model has been presented to obtain relatively simple correlations for outlettemperature and absolute humidity. The developed psychometric model is based on the correlationsbetween the relative humidity and enthalpy of supply and regeneration air streams. The model is used topredict the performance of three type of desiccant rotors manufactured by using different kind of soliddesiccants (Type I, II and III). The model is tested corresponding to a wide range of measurement data.The developed psychometric model is simple in nature and able to predict very well the performance ofdifferent kind of desiccant rotors. Copyright # 2002 John Wiley & Sons, Ltd.

KEY WORDS: air conditioning; desiccant systems; rotary wheel; ‘Model 54’; psychrometric model;performance prediction

1. INTRODUCTION

Desiccant cooling (DEC) is an environmental friendly technology and is used to achieve thethermal comfort inside a building. Further, since DEC technology does not use the traditionalrefrigerants, it is getting more importance in respect of decreasing the use of chlorofluor-ocarbons and reduction in Greenhouse gas (GHG) emissions. Desiccant cooling (Nelson et al.,1978; Collier et al., 1982; Lof et al., 1988; Waugaman et al., 1993; Van den Bulck et al., 1986;Lof, 1992) is basically an open cooling cycle consisting of a dehumidifier, sensible heatexchanger and evaporative water spray cooler. In a DEC system, the potential for theapplication of evaporative cooling is increased by means of air dehumidification. The other

Received 15 January 2002Copyright # 2002 John Wiley & Sons, Ltd. Accepted 8 March 2002

Contract/grant sponsor: ICTP Programme for Training and Research.

yE-mail: [email protected]

*Correspondence to: M. Beccali, Dipartimento di Ricerche Energetiche e Ambientali (DREAM), Universita degli Studidi Palermo, viale delle Scienze bld 9, 90128-Palermo, Italy.

interesting feature of this technology is that the low temperature heat (50–808C) required forregeneration can be obtained by integrating it with a solar thermal plant or utilizing waste heatfrom industrial or cogeneration process (Lof, 1992; Henning et al., 2001). Therefore, thedesiccant based air-conditioning reduces the consumption of conventional energy (Babus’Haqet al., 1996; Henning and Erpenbeck, 1996).

In a DEC system, the air dehumidifiers utilize either liquid or solid desiccants. Solid desiccantsuch as silica gel are efficiently utilized for dehumidification process. Silica gel is one of the mostcommonly used desiccant which can absorb up to 40% of its own weight, moreover, itsadsorption characteristics over a wide range of humidity, makes it a first choice solid desiccant.Among other important and popularly used desiccants for dehumidification, the lithiumchloride (LiCl) is one of the most hygroscopic salts available and can absorb water vapour as asolid and then continue to attract moisture even after its absorbed water turns in to liquidsolution.

The solid desiccants are available as rotary wheels as well as fixed bed components. Thesolid desiccant rotary wheels are capable of transferring moisture with the same efficiency.The rotary desiccant wheels have been developed for a range of applications such astextile industries, process and storage industries, indoor swimming pools, clean rooms, etc. Anew attractive application is in the field of air-conditioning for commercial and residentialbuildings.

In the literature, many mathematical models have been proposed to assess the performanceof rotary desiccant for a given operating conditions (Maclaine-Cross, 1988; Davangere et al.,1999; Collier and Cohen, 1991; Zheng et al., 1995; Dai et al., 2001). Generally the mathematicalmodels developed for simulating the performance of rotary desiccant wheels are basedon detailed heat and mass transfer phenomena inside the system and are very complex.The research work on solar assisted air-conditioning of buildings is being carried out atPolitecnico di Milano. In this respect, extensive work has been carried out on the performanceevaluation of various kinds of rotary desiccant wheels available in the market. For this, theavailable measurement data on different kind of rotary wheel have been analysed thoroughlyand simplified models have been developed for their performance evaluation (Beccali et al.,2002).

In the present paper, two simple mathematical models have been presented to evaluate theperformance of rotary desiccant wheels in terms of the use of different kind of solid desiccantsviz. silica gel, LiCl, etc. The first part of the paper presents ‘Model 54’ which is developed forsilica gel desiccant rotor. The model has been derived from the interpolation of experimentaldata obtained from the industry and the correlations have been developed for predicting outlettemperature and absolute humidity. The ‘Model 54’ consists of 54 coefficients corresponding toeach correlation for outlet absolute humidity and temperature and it is found that the modelpredicts very well the performance of silica gel desiccant rotor (Type-I). In the second part of thepaper, a psychrometric model has been presented to obtain relatively simpler correlations foroutlet temperature and absolute humidity. The developed psychrometric model is based on thecorrelations between the relative humidity and enthalpy of supply and regeneration air streams.The model is used to predict the performance of three type of desiccant rotors manufactured byusing different kind of solid desiccants. The model is tested corresponding to a wide range ofmeasurement data obtained from the industries. The developed psychrometric model is simpleand able to predict very well the performance of different kind of desiccant rotors (Types I, IIand III).

Copyright # 2002 John Wiley & Sons, Ltd. Int. J. Energy Res. 2003; 27:17–29

M. BECCALI ET AL.18

2. ‘MODEL 54’

2.1. Model Formulation

The measurement data for silica gel solid desiccant wheel has been obtained from themanufacturer. The performance data is available corresponding to a wide range of inlettemperature (20–328C), regeneration temperature (40–808C), inlet absolute humidity (7–16 gKg�1) and regeneration humidity (7–16 gKg�1). The measured data is processed and thevarious equations have been derived from interpolations of measured data related to theperformances of silica gel solid desiccant wheel (Type-I). The equations have been derived tocalculate outlet temperature (Tout) and the absolute humidity (Xout) of the processed air as afunction of the temperature and the humidity of the regeneration air stream (Trig;Xrig) and of theair to be treated (Tin;Xin). The outlet conditions are obtained through a parabolic-shapedfunction linking Xrig to Trig and Trig to Xin: It is interesting to note that the outlet temperature(Tout) and the absolute humidity (Xout) of the processed air is a linear function of Tin and thederived equations can be expressed as follows:

Xout; Tout ¼ ðF1 X 2in þ F2Xin þ F3ÞTin þ ðF4X 2

in þ F5Xin þ F6Þ ð1Þ

where Fi is a second-order polynomial of Trig and can be expressed as follows:

Fi ¼ ðAi T 2rig þ Bi Trig þ CiÞ; i ¼ 1; 6 ð2Þ

where Ai; Bi; Ci are the second order polynomials of Xrig and can be expressed as followinggeneral expression:

Ai; Bi; Ci ¼ aiX 2rig þ bi Xrig þ ci ð3Þ

There are in total 54 coefficients (ai; bi; ci) corresponding to each equation for Xout and Tout: Thecoefficients are summarized in Tables I and II. The model is named as Model ‘54’ since itconsists of 54 coefficients corresponding to each performance parameter. The existence ofquadratic functional relationships among the terms of the equations generates parabolic shapeswith maximum or minimum points.

2.2. Model validation

To appreciate the accuracy of the developed model, the calculations have been carried out forperformance prediction of a desiccant wheel (Type-I) based on silica gel as desiccant. For this,outlet temperature (Tout) and the absolute humidity (Xout) of the processed air have beencalculated using the developed model for the following range of variation of the parameters:

Xrig ¼ 10216 g Kg�1

Trig ¼ 402808C

Tin ¼ 202348C

Xin ¼ 8215 g Kg�1

The Figures 1 and 2 show the measured and calculated values of DX ðXout � XinÞ and DT ðTout �TinÞ; respectively, corresponding to the measurement parameters for Type-I desiccant wheel. Ithas been observed that the calculated values of DX and DT obtained using ‘Model 54’ are very


DESICCANT WHEEL DEHUMIDIFICATION 19

TableI.

Matrix

ofcoefficients

foroutlet

absolute

humidity(X

out).

Coefficients

Indices

i1

23

45

6

a�0.0000009817708

0.00002486458

�0.000155484

0.00000297891

�0.000770385

0.004898569

Ai

b0.000024125

�0.000610844

0.003809656

�0.000735859

0.01901969

�0.1207189

c�0.0001425729

0.003601979

�0.0224005

0.004374313

�0.1127932

0.7154694

a0.000120938

�0.003064896

0.01911052

�0.003642969

0.09411271

�0.5966672

Bi

b0.002954063

0.0748425

�0.4653281

0.08950781

�2.311069

14.6212

c0.01734688

�0.4383754

2.716457

0.5291188

13.62156

�86.19895

a�0.003395833

0.08619583

�0.5358542

0.10105

�2.610487

16.49429

Ci

b0.0825

�2.093475

12.97393

2.47025

63.78412

�401.8073

c�0.4828167

12.22347

�75.36913

14.5702

�374.4862

2362.187


M. BECCALI ET AL.20

TableII.Matrix

ofcoefficients

foroutlet

temperature

(Tout).

Coefficients

Indices

i1

23

45

6

a0.000004663333

�0.000137699

0.000889096

�0.00015990

0.004608366

�0.02952156

Ai

b�0.000122108

0.003660427

�0.02379164

0.00424914

�0.1237906

0.7970592

c0.00075592

�0.02322678

0.1527826

�0.02691875

0.7979418

�5.187328

a�0.000788528

0.02253221

�0.1440601

0.02616227

�0.7356466

4.675441

Bi

b0.02103875

�0.6067781

3.896302

�0.7045379

19.94981

�127.2275

c�0.1343869

3.932225

�25.44645

4.561903

�130.5749

838.5968

a0.03264739

�0.9065427

5.741306

�1.055388

29.01258

�182.9765

Ci

b�0.8851843

24.70641

�156.8717

28.7686

�794.0862

5018.149

c5.799316

�163.1261

1041.133

�189.9018

5274.441

�33467.1



close to measured values. Therefore, it is quite evident that the developed ‘Model 54’ isable to predict very well the performance evaluation of Type-I desiccant wheel based onsilica gel.

Figure 1. Measured and calculated values of DX ðXout � XinÞ corresponding to the measurementparameters for Type I (silica gel) desiccant wheel.

Figure 2. Measured and calculated values of DT (Tout2Tin) corresponding to the measurement parametersfor Type I (silica gel) desiccant wheel.


M. BECCALI ET AL.22

3. PSYCHROMETRIC MODEL

3.1. Model formulation

The ‘Model 54’ described in previous section is able to predict well the performance of Type-Idesiccant wheel. However, it is noted that for calculating performance parameters for differenttype of wheels (corresponding to measured performance data), one needs 54 new coefficientscorresponding to each equation for Xout and Tout used in the model which is a tedious andcomplex job for the user to calculate performance parameters.

To overcome the complexities of the Model ‘54,’ further simplifications have been made basedon the observation that the measured parameters the UR (relative humidity) and enthalpy (h)can be expressed through a linear correlation of the following type:

DUR ¼ ðURin � URoutÞ ¼ mðURin � URrigÞ þ q ð4Þ

D h ¼ ðhout � hinÞ ¼ m0ðhrig � hinÞ þ q0 ð5Þ

where the enthalpy (h) is a function of absolute humidity and the temperature, can be expressedby following well-established equation:

h ¼ð2501þ 1:805 T ÞX

1000þ 1:006 T ð6Þ

For the calculation of relative humidity, following empirical relation has been developed(Beccali et al., 2002):

UR ¼ ð18:6715 X þ 1:7976Þe�0:053 T ð7Þ

The above Equations (4)–(7) are valid for each kind of desiccant wheel and m; q; m0; q0 are theonly parameters to be calculated. In the present work, the coefficients m; m0; q; q0 have beencalculated by analysing the field data obtained from the industries for three different kind ofdesiccant wheels. We named them Types I, II and III, and among them two are based on silicagel solid desiccant and another one is based of LiCl as solid desiccant. The range of availablefield data used to develop the model corresponds to Xrig ¼ 10216 gKg�1, Trig ¼ 402808C,Tin ¼ 202348C, Xin ¼ 8215 gKg�1.

3.2. Performance parameters

System of equations (4)–(7) has been solved for calculating outlet absolute humidity (Xout) andtemperature (Tout) and the following solution is obtained.

Xout ¼½e0:053 Tout ð0:9428URrig þ 0:0572URinÞ � 1:7976�

18:671ð8Þ

ðUR e0:053 Tout � 1:7976Þ18671:0

¼ðhout � 1:006 ToutÞð2501� 1:805 ToutÞ

ð9Þ

It is to be noted that for calculating the outlet temperature (Tout), Equation (9) has to be solvedby iteration.

On analysing field data for three different kind of desiccant wheel, it has been found thatEquation (4) for relative humidity difference (UR) can be expressed by following equation which



is valid for all kind of desiccant wheel:

URout ¼ ð0:9428URrig þ 0:0572URinÞ ð10Þ

However, for enthalpy calculations, there is a need to find out coefficients m0 and q0

corresponding to different type of desiccant wheels. We have developed following equations forenthalpy calculations corresponding to various types of desiccant wheels, which can beexpressed as follows:

Desiccant wheel Type-I:

hout ¼ ð0:1312 hrig þ 0:8688 hinÞ ð11Þ

Desiccant wheel Type-II:

hout ¼ ð0:1861 hrig þ 0:8139 hinÞ ð12Þ

Figure 3. Calculated values of DUR along with experimental values for Type I (silica gel) desiccant wheel.

Figure 4. Calculated values of Dh along with experimental values for Type I (silica gel) desiccant wheel.


M. BECCALI ET AL.24

Desiccant wheel Type-III:

hout ¼ ð0:1148 hrig þ 0:8852 hinÞ � 0:9474 ð13Þ

The calculated values of DUR and Dh along with experimental values for different kind ofdesiccant wheels are shown in Figures 3–8. It can be observed that the developed correlationscalculate reasonably well the DUR and Dh for all kind of desiccant rotors (the highest errorfound in the calculations of Dh corresponds to R2 � 0:88).

4. RESULTS AND DISCUSSION

To appreciate the developed psychrometric model, numerical calculations have been carried outfor performance prediction of three different kinds of desiccant wheels (Types I, II and III). The

Figure 5. Calculated values of DUR along with experimental values for Type II (LiCl) desiccant wheel.

Figure 6. Calculated values of Dh along with experimental values for Type II (LiCl) desiccant wheel.



results are presented in the form of absolute humidity and temperature difference between outletand inlet of process air (DX and DT ). Figures 9–14 show the calculated values of DX and DTversus the experimental values corresponding to Types I, II and III desiccant wheels. It can beseen from the figures that the developed model predicts quite well the performance of all kind ofdesiccant wheels.

The developed mathematical models serve as a useful tool for designing a desiccanttechnology for building climatization. It is further envisaged that these models can be includedin the simulation tools, which would allow for comparing desiccant technology vis-"aa-vis otherconventional air-conditioning technologies. Moreover, it also allows selecting the appropriateheating source available for desiccant technology.

Figure 7. Calculated values of DUR along with experimental values for Type III (silica gel) desiccant wheel.

Figure 8. Calculated values of Dh along with experimental values for Type III (silica gel) desiccant wheel.


M. BECCALI ET AL.26

Figure 10. Measured and calculated values of DT (Tout2Tin) corresponding to the measurement parametersfor Type I (silica gel) desiccant wheel.

Figure 9. Measured and calculated values of DX (Xout2Xin) corresponding to the measurement parametersfor Type I (silica gel) desiccant wheel.

Figure 11. Measured and calculated values of DX (Xout2Xin) corresponding to the measurementparameters for Type II (LiCl) desiccant wheel.



Figure 12. Measured and calculated values of DT (Tout2Tin) corresponding to the measurement parametersfor Type II (LiCl) desiccant wheel.

Figure 13. Measured and calculated values of DX (Xout2Xin) corresponding to the measurementparameters for Type III (silica gel) desiccant wheel.

Figure 14. Measured and calculated values of DT (Tout2Tin) corresponding to the measurement parametersfor Type III (silica gel) desiccant wheel.


M. BECCALI ET AL.28

NOMENCLATURE

ACKNOWLEDGEMENTS

One of the authors (R.S. Adhikari) undertook this work with the support of the ‘‘ICTP Programme forTraining and Research in Italian Laboratories. Trieste, Italy’’. The authors gratefully acknowledge theconcerned industries for providing the measurement data on rotary desiccant wheels. Some of theinformation provided by various scientists on desiccant cooling under IEA Task (Task 25: Solar AssistedAir Conditioning of Buildings) is also duly acknowledged.

REFERENCES

Babus’Haq RF, Olsen H, Probert SD. 1996. Feasibility of using an integrated small-scale CHP unit plus desiccant wheelin a leisure complex. Applied Energy. 53:179–192.

Beccali M, Butera F, Guanella R, Adhikari RS. 2002. Performance evaluation of rotary desiccant wheels, Proceedings ofWorld Renewable Energy Congress (WREC)-VI, Cologne, Germany, July 2002.

Collier RK, Barlow RS, Arnold FH. 1982. An overview of open-cycle desiccant cooling systems and materials. ASMEJournal of Transfer 104:28–34.

Collier RK and Cohen BM. 1991. An analytic investigation of methods for improving the performance of desiccantcooling system. ASME Journal of Solar Energy Science and Engineering 113:157–163.

Davanagere BS, Sherif SA and Goswami DY. 1999. A feasibility study of a solar desiccant air-conditioning system,Part}I: psychrometrics and analysis of the conditioned zone. International Journal of Energy Research 23:7–21.

Dai YJ, Wang RZ and Zhang HF. 2001. Parameter analysis to improve rotary desiccant dehumidification usingmathematical model. International Journal of Thermal Science 40:400–408.

Henning HM and Erpenbeck T. 1996. Integration of solar assisted open cooling cycles into Building ClimatizationSystems. Proceedings of EuroSun’96, Freiburg, Germany, September 16–19.

Henning HM, Erpenbeck T, Hindenburg C, Santamaria IS. 2001. The potential of solar energy use in desiccant coolingcycles. International Journal of Refrigeration 24:220–229.

Lof GOG. 1992. Desiccant systems. In Solar Air Conditioning and Refrigeration, Sayigh AAM, McVeigh JC (eds).Oxford: UK, Pergamon Press.

Lof GOG, Cler G, Brisbane T. 1998. Performance of a solar desiccant cooling system. ASME Journal of Solar EnergyScience and Engineering 110:165–171.

Maclaine-Cross IL. 1988. Proposal for a desiccant air conditioning system. ASHRAE Transaction 94:1997–2009.Nelson JS, Beckman WA, Mitchell JW, Close DJ. 1978. Simulation of the performance of open cycle desiccant systems

using solar energy. Solar energy 21:273–278.Van den Bulck E, Mitchell JW, Klein SA. 1986. The use of dehumidifiers in desiccant cooling and dehumidification

systems. ASME Journal of Heat Transfer 108:684–692.Waugaman DG, Kini A, Kettleborough C. 1993. A review of desiccant cooling systems. ASME Journal of Energy

Resources Technology 115:1–8.Zheng W, Worek WM, Novosel V. 1995. Performance optimisation of rotary dehumidifiers. ASME Journal of Solar

Energy Science and Engineering 117:40–44.

Symbols

h ¼ enthalpy (kJ kg�1)T ¼ Temperature (8C)UR ¼ relative humidity (%)X ¼ absolute humidity (g kg�1)

Subscript

in ¼ inlet (in the process side of the wheel)out ¼ outlet (in the process side of the wheel)rig ¼ in the regeneration side of the wheel



Documents

Simplified models for the performance evaluation of desiccant wheel dehumidification