A hybrid dehumidifier model for real-time performance monitoring, control and optimization in liquid desiccant dehumidification system

Applied Energy 111 (2013) 449–455

Contents lists available at SciVerse ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

A hybrid dehumidifier model for real-time performance monitoring, control and optimization in liquid desiccant dehumidification system Xinli Wang a,b, Wenjian Cai b,⇑, Jiangang Lu a, Youxian Sun a, Xudong Ding b a b

State Key Laboratory of Industrial Control Technology, Department of Control Science and Engineering, Zhejiang University, Hangzhou 310027, China EXQUISITUS, Centre for E-City, School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore

h i g h l i g h t s  Study the heat and mass transfer process in dehumidifier for LDDS.  A simplified yet accurate model for real-time performance optimization is developed.  The model requires no iterative computations and is easy for engineering application.

a r t i c l e

i n f o

Article history: Received 23 September 2012 Received in revised form 8 May 2013 Accepted 10 May 2013 Available online 4 June 2013 Keywords: Liquid desiccant Dehumidifier Heat transfer and mass transfer Hybrid model Levenberg–Marquardt method Performance assessment and optimization

a b s t r a c t In this paper, a simplified, yet accurate hybrid model to predict the heat and mass transfer processes in a packed column liquid desiccant dehumidifier is developed. Starting from energy and mass balance principles, and by lumping the geometric parameters and fluids’ thermodynamic coefficients as constants, the derived model only requires two equations together with total seven parameters for predicting the heat and mass transfer status in the dehumidifier. Commissioning information together with Levenberg–Marquardt method can be used to identify these parameters. Compared with the existing liquid desiccant dehumidification system dehumidifier models, the proposed model is very simple, accurate and does not require iterative computations. Experimental results demonstrate their effectiveness in predicting heat and transfer performances over a wide operating range. The model is expected to be applied in operational optimization, performance assessment, fault detection and diagnosis in liquid desiccant dehumidification system. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction In recent years, Liquid Desiccant Dehumidification System (LDDS) for air dehumidification has emerged as a viable alternative to conventional mechanical based dehumidification schemes where air is cooled below the dew point. The main advantages of the LDDS include: (1) possible energy savings by shifting the energy use away from electricity towards renewable or low grade energy, such as solar energy, geothermal energy, and waste energy from industrial processes [1]; (2) flexibility in operation to achieve independent temperature and humidity control; and (3) employ environment-friendly hygroscopic salt solutions as working fluids which do not contribute to ozone depletion [2]. The research on air dehumidification by LDDS may trace back to 1955 when Lof [3] first designed an open-cycle air-conditioning system using triethylene glycol as liquid desiccant. Since then, hundreds of research results have been published to refine the ⇑ Corresponding author. Tel.: +65 6790 6862; fax: +65 6793 3318. E-mail address: [email protected] (W. Cai). 0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apenergy.2013.05.026

system in the areas of system design [4–6], experimental investigation [7,8] and performance analysis [9–11], where the fundamental heat and mass transfer processes were intensively investigated using finite difference, effectiveness NTU or empirical models [12]. Among the three modeling approaches, finite difference model has most frequently been used in investigation of LDDS performances for its accuracy. Factor and Grossman [13] and Gandhidasan et al. [14] developed theoretical models to test the columns to predict the performance of air dehumidification and solution regeneration under various operating conditions, and experiment results showed very good agreement with the theoretical model. Oberg and Goswami [15] presented a finite difference model and carried out detailed experimental investigations of heat and mass transfer inside the dehumidifier. Fumo and Goawami [16], Yin et al. [17] employed a modified Oberg and Goswami’s model to study the LDDS with aqueous lithium chloride solution and random packing. Linear approximation was introduced to obtain analytical solution to the finite difference model and more accurate outlet conditions can be predicted by this method [18]. Mesquita et al. [19] developed a finite difference model for internally

450

X.L. Wang et al. / Applied Energy 111 (2013) 449–455

Nomenclature Aa As b1–b4 c1–c3 c4–c7 cp C D Da Dcalc Dp Dreal Ds e f g h ha hs H k ka ks KG _ m _a m _s m N

convection heat transfer area of process air (m2) convection heat transfer area of desiccant solution (m2) constant (dimensionless) parameters in heat transfer (dimensionless) parameters in mass transfer (dimensionless) specific heat of desiccant solution (J/(kg °C)) constant (dimensionless) diameter of the structured packing (m) diffusivity of process air (m2/s) calculated data (dimensionless) nominal size of packing material (m) experimental data (dimensionless) diffusivity of desiccant solution (m2/s) constant (dimensionless) constant (dimensionless) gravitational constant (m/s2) heat transfer coefficient (W/(m2 °C)) heat transfer coefficient of process air convection (W/ (m2 °C)) heat transfer coefficient of desiccant solution convection (W/(m2 °C)) Henry’s law constant (Pa) thermal conductivity (W/(m2 °C)) mass transfer coefficient of gas phase convection in the dehumidifier (kg/(m2 s Pa)) mass transfer coefficient of liquid phase convection in the dehumidifier (kg/(m2 s Pa)) overall mass transfer coefficient in the dehumidifier (kg/ (m2 s Pa)) mass flow rate of fluid (kg/s) mass flow rate of process air (kg/s) mass flow rate of desiccant solution (kg/s) mass transfer rate in the dehumidifier (kg/m2 s)

cooled parallel plate liquid desiccant dehumidifiers. However, the process of developing and solving finite difference models are quite complex and iterative computation is absolutely necessary as the outlet conditions of desiccant solution are generally unknown, which makes the modeling approach unsuitable for real-time optimization and the performance estimation. For the other two methods, Chen et al. [20] proposed NTU models in for a packed-type LDDS with two different flow configurations: parallel flow and counter flow. Compared with other models in literature, better accuracy can be obtained by using the proposed NTU model. Liu et al. [21] conducted a simulation of heat and mass transfer process with the corresponding data collected in a cross-flow dehumidifier and regenerator and a theoretical model with NTU input parameter was developed. Further, the authors showed that the analytical solutions of the available NTU model could be utilized in the optimization of the LDDS [22]. Khan and Ball [23,24] conducted sensitivity analysis for heat and mass transfer process of a packed-type liquid desiccant system to identify the performance of dehumidifier and regenerator through empirical method. Although the empirical methods are simple, some key parameters involving the performance of LDDS should be known in advance which may become very complicated in applications. Moreover, accuracy will decrease if the NTU or empirical model is expanded over a wider region. In this paper, a simplified, yet accurate dehumidifier model to predict the heat and mass transfer processes for real-time performance monitoring, control and optimization in LDDS is developed based on the hybrid modeling approach [25–27]. Starting from energy and material balance principles, two simple non-linear

pa,in ps;in Pa,sat Q R Ra Rh Rs Ta,in Ts,in V

at ax l la ls xs /a

qa qs

water vapor pressure of inlet process air in the dehumidifier (Pa) equilibrium water vapor pressure of inlet desiccant solution in the dehumidifier (Pa) saturated water vapor pressure (Pa) heat transfer rate in the dehumidifier (W) ideal gas constant (J/(mol °C)) thermal resistance of process air convection (°C/W) overall thermal resistance in the dehumidifier (°C/W) thermal resistance of desiccant solution convection (°C/ W) temperature of inlet process air in the dehumidifier (°C) temperature of inlet desiccant solution in the dehumidifier (°C) volume flow rate of fluid (m3/s) specific surface area (m2/m3) wetted specific surface area (m2/m3) viscosity of fluid (Pa s) viscosity of process air (Pa s) viscosity of desiccant solution (Pa s) concentration of desiccant solution (%) relative humidity of process air (%) density of process air (kg/m3) density of desiccant solution (kg/m3)

Subscripts a process air G gas phase in inlet out outlet s desiccant solution sat saturated

equations can be obtained to present the heat and mass transfer process in dehumidifier. Through determining the process input– output variables and lumping dimensional parameters into seven characteristic parameters, Levenberg–Marquardt method can be applied to carry out the identification to the seven parameters by real-time experimental data. The proposed model need no iterative computation and is simple and accurate compared with the existing models. Experimental results show that the prediction errors are mostly fall within 10%, which validated the effectiveness of the modeling approach. Therefore, the model is expected to find its applications in real-time performance monitoring, control and optimization. 2. The operating principle of LDDS A basic schematic diagram of Liquid Desiccant Dehumidification System (LDDS) operating with lithium chloride as desiccant is shown in Fig. 1, which is composed of two packing columns: dehumidifier and regenerator. The dehumidifier is to remove the moisture in the process air and the regenerator is to concentrate the diluted solutions from the dehumidifier to an acceptable concentration. The change of desiccant water vapor pressure in LDDS is illustrated in Fig. 2 and its working principle is briefly described below:  In the dehumidifier, the strong desiccant solution is firstly cooled by the cooler until the state A is reached, and sprayed on top of the dehumidifier, directly contacting with process air, drawn from the bottom of the column by a fan, in a

451

X.L. Wang et al. / Applied Energy 111 (2013) 449–455

Supply Air A Outlet Air Cooler

C

Dehumidifier

Heater Regenerator

HEX D

Regeneration Fan

B Regeneration Pump

Dehumidification Fan

Dehumidification Pump

Desiccant Surface Vapor Pressure

Fig. 1. Schematic diagram of a basic liquid desiccant dehumidification system.

Dehumidification

A

B

B

C

Heating

C

D

Regeneration

D

A

Cooling

C

B

D A

Desiccant Concentration Fig. 3. Schematic of a typical heat and mass transfer processes in dehumidifier. Fig. 2. Change of desiccant water vapor pressure during the cycle.

counter-flow configuration. The water vapor pressure of high concentration cooled desiccant solution is lower than that of process air and the surplus moisture is forced to migrate from process air towards to the solution. Water vapor pressure difference between desiccant solution and process air acts as the driving force for the mass transfer. This process is represented by line A–B in the figure.  To recover the diluted solution back to strong affiliation to moisture, desiccant solution is heated (B–C) by a heater before it is pumped into the regenerator to contact with regenerating air with a higher water vapor pressure.  Since the desiccant solution has higher water vapor pressure than that of regenerating air, mass transfer takes place in the opposite direction to that which occurs in the dehumidifier and the moisture absorbed in dehumidification progress can be transferred from the desiccant solution to regenerating air as indicated by line C–D.  Even desiccant solution is concentrated; it still has high surface vapor pressure due to its high temperature. Therefore, cooling (D–A) is needed to cool down the solution so that the desiccant can reach to state A to complete the cycle. In LDDS, the humidity and temperature of supply air are determined by the heat and mass transfer process in the dehumidifier. Meanwhile, a large amount of cooling energy will be consumed to balance the latent and sensible load in process air. To manipulate the pump and fan so that both demanded air humidity and

temperature with a maximal thermal efficiency can be achieved in the dehumidifier. For the purpose of real-time performance monitoring, control and optimization, a simplified and efficient model is necessary. The model can be derived from physical principle of heat and mass transfer processes in the dehumidifier which is shown schematically as in Fig. 3. To simplify the derivation process in the subsequent sections, the following assumptions are made: 1. 2. 3. 4.

The packed column is adiabatic. Vaporization of desiccant is ignored. The heat and mass transfer is in steady state. Mass variations in liquid desiccant solution and process air are neglected during dehumidification. 5. Constant desiccant properties under temperature variations. 3. Heat transfer model In the dehumidifier, sensible heat transfer occurs between the cooled falling film solution and the rising process air through the channels of packing materials. There are three routes for heat transfer: process air convection, the interface conduction and cooled solution convection. According to the energy balance and heat transfer theory, the heat transfer rate due to the temperature differences in liquid and gas phases can be expressed as:

Q¼

T a;in  T s;in Rh

ð1Þ

452

X.L. Wang et al. / Applied Energy 111 (2013) 449–455

where Q, Ta,in, Ts,in and Rh are the heat transfer rate, the inlet process air temperature, the inlet desiccant solution temperature and the thermal resistance to overall heat transfer, respectively. In theory, the overall thermal resistance Rh can be divided into three parts, namely thermal resistance of air side convection, interface thermal conduction resistance and solution convection thermal resistance. However, the interface conduction thermal resistance can be neglected as the interface between the two fluids is very thin and with very small thermal resistance. The overall thermal resistance therefore can be represented as:

Rh ¼ Ra þ Rs

ð2Þ

where Ra, Rs are the thermal resistances of process air convection and of solution convection, respectively. The heat transfer rate between the interface and the fluid moving over it depends on the characteristics of both the interface and the moving fluid, such as the interface geometry, the fluid’s velocity as well as the temperature differences. In the dehumidifier, the desiccant solution and process air are driven by the pump and fan, respectively. It is therefore possible to consider the heat transfer between process air and desiccant solution as forced convection heat transfer. For the forced convection heat transfer in dehumidifier, the heat transfer coefficient h, is influenced by the packing diameter D, the fluid density q, the fluid velocity v, and also affected by the mean film temperature through fluid’s specific heat cp, viscosity l, and thermal conductivity k, respectively. Through dimensional analysis, the following equation has been proposed [28]:

 e hD Dqm cp lf ¼C k l k

ð3Þ

where the values of the unknown parameters, C, e and f, need to be exactly determined. For steady flow, assumption can be made that both the fluid density q and the volume flow rate V remain constant during the dehumidification. Accordingly, the product of qV (the mass flow _ is unchanged. Furthermore, l and k are approximately conrate m) stants if there is only a small change in fluid’s temperature. So Eq. (4) can be rewritten as:

h¼C



 _ e cp lf k 4m _e ¼ bm D plD k

ð4Þ

where b ¼ Cð4=plDÞe ðcp l=kÞf k=D. The thermal resistance to overall heat transfer in the dehumidifier can be written as:

_ es þ b2 Aa m _ ea 1 1 b 1 As m Rh ¼ þ ¼ _ es b2 Aa m _ ea hs As ha Aa b1 As m

Q¼

1

ðT a;in  T s;in Þ

  N ¼ K G pa;in  ps;in

ð7Þ

where N, KG, pa,in and ps;in are the mass transfer rate, overall mass transfer coefficient in the dehumidifier, water vapor pressure of inlet process air and the equilibrium water vapor pressure of desiccant solution, respectively, and based on the two film theory of mass transfer and Henry’s law [29], KG can be calculated by,

KG ¼

1 ka

1 þ Hk1 s

ð8Þ

where ka, ks and H are the convective mass transfer coefficient of gas phase, the convective mass transfer coefficient of liquid phase and Henry’s law constant, respectively. For the mass transfer between process air and desiccant solution in packed column, Onda proposed the respective correlations on the mass transfer coefficients in gas phase and liquid phase [30]

ka ¼ a1

ks ¼ a2

!e1 

_a 4m

la qa Da

2

at la pD _s 4m

!e 2 

ax ls pD2

ls qs Ds

f1

f2

ðat Dp Þg1

ðat Dp Þg2

at Da

ð9Þ

RT a 

qs ls g

j2 ð10Þ

where at is the specific surface area, ax is the wetted specific surface area, Dp is the nominal size of packing material, which depend on the geometry of packing and are unchanged for a certain kind of packing, Da, qa, la and Ds, qs, ls are the diffusivity, density, viscosity of process air and desiccant solution, respectively. la, qa, Da ls qs Ds can assumed to be constants because the temperature differences of desiccant solution and process air are not too big (less than 10 °C for desiccant solution and 20 °C for process air). Therefore, the form can be simplified as:

_ a Þ e1 ka ¼ b3 ðm ð5Þ

1 T a;in

ð11Þ

and

_ a, m _ s , ha, hs, Aa and As are the process air flow rate, desiccant where m solution flow rate, heat transfer coefficient of process air convection, heat transfer coefficient of desiccant solution convection, the convection heat transfer area of process air and the convection heat transfer area of desiccant solution, respectively. Combining Eqs. (1) and (5), the heat transfer in dehumidifier can be described by:

_ cs 3 c1 m _ s c3 m þ c2 ðm _aÞ

at the interface between the two fluids; and (3) the resistance within desiccant solution. Since a high solution flow rate is often applied in the dehumidifier, mass transfer at the interface between process air and desiccant solutions is very fast when compared to the process within either process air or desiccant solution. The resistance at the interface is considered to be very small and therefore can be neglected. Furthermore, it can be assumed that an equilibrium state in terms of mass transfer exists at the interface. Consequently, the following relationship to describe the mass transfer with driving force in terms of gas phase water vapor pressure can be obtained:

ð6Þ

where c1 = b1As, c2 = b1As/b2Aa, c3 = e.

_ s Þ e2 ks ¼ b4 ðm where

b3 ¼ a1 ð4=at la pD2 Þe1 ðla =qa Da Þf1 ðat Dp Þg1 at Da =R b4 ¼ a2 ð4=ax ls pD2 Þe2 ðls =qs Ds Þf2 ðat Dp Þg2 ðqs =ls gÞj2 Substituting Eqs. (11) and (12) into (8),

KG ¼

4. Mass transfer model

KG ¼ In dehumidification process, transferring water vapor from process air to strong desiccant solution has to overcome three resistances: (1) the resistance in process air it itself; (2) the resistance

ð12Þ

1 T a;in _ a Þe1 b3 ðm

þ

1 _ s Þe2 Hb4 ðm

¼

_ s Þ e2 Hb4 ðm b4 _ s Þe2 ðm _ a Þe1 1 þ T a;in H b ðm

_ s Þc6 c 4 ðm _ s Þc 6 ðm _ a Þc7 1 þ c5 T a;in ðm

ð13Þ

3

ð14Þ

where c4 = Hb4, c5 = Hb4/b3, c6 = e2, c7 = e1. Hence, the mass transfer process can be finally described as:

453

X.L. Wang et al. / Applied Energy 111 (2013) 449–455

N¼

  _ s Þc6 c 4 ðm  c6 c7 pa;in  ps;in _ s Þ ðm _ aÞ 1 þ c5 T a;in ðm

ð15Þ

where the water vapor pressure of inlet process air, pa,in, can be calculated based on the definition of relative humidity:

pa;in ¼ ua;in Pa;sat

ð16Þ

where /a,in is the relative humidity of inlet process air, and the saturation water vapor pressure, Pa,sat, which is affected only by the temperature of water vapor, can be expressed as a function of temperature according to empirical data available [31]:

Pa;sat ¼ a0 T 2a;in þ a1 T a;in þ a2

ð17Þ

For temperature between 15 °C to 35 °C, the constants are:

a0 = 4.9569, a1 = 54.75, a3 = 1437.4. Since desiccant solution is ranging from 15 to 25 °C in temperature and from 27% to 40% in concentration in dehumidifier. Based on the thermodynamic properties of aqueous solutions of the chlorides lithium [32], we obtain through algebraic fitting:

ps;in ¼ b0 þ b1 T s;in þ b2 xs þ b3 T 2s;in þ b4 x2s þ b5 T s;in xs

ð18Þ

where xs is the concentration of desiccant solution and the constants are: b0 = 1.1491997, b1 = 0.11242725, b2 = 0.06183172, b3 = 0.0014521543, b4 = 0.0011080673, b5 = 0.0035782429. Thus, Eqs. (6) and (15) describe the adiabatic simultaneous heat and mass transfer process in the dehumidifier with only seven parameters. Compared with the existing models, the present model is characterized by fewer characteristic parameters, which can be determined by real-time experimental data using Levenberg– Marquardt method [26]. 5. Validation and discussion To validate the proposed dehumidifier model, an experimental rig operating with lithium chloride aqueous as desiccant, as shown in Figs. 4 and 5, is used. The schematic diagram is also provided in Fig. 6. The system capacity is:  Maximum air flow rate 500 m3/h,  Cooling capacity 1.7 kW,  Dehumidification 5 kg/h. which can meet the sensible and latent demand for fresh air supply in a room with area of 100 m2 in Singapore. The column shell, made of polypropylene, 1100 mm in height, is filled with structured packing with face dimensions 400 mm  400 mm  600 mm. A liquid distributor at the top of column provides a good liquid

Fig. 5. Photograph of dehumidifier: 1 – dehumidifier (packed column); 2 – cooler; 3 – pump; 4 – airflow meter; 5 – solution flow meter; 6 – humidity/temperature transmitter.

distribution and a stainless wire mesh is installed to remove desiccant droplets carried by air with high velocities. The air and solution flow rate can be regulated by the Variable Speed Drive (VSD) equipped in fans and pumps during the operation process to meet different cooling and dehumidification demands under different outdoor air conditions. The regenerator on the left is installed to supply desiccant solution with a constant concentration (e.g. 37%). For a certain demand, steady state can be obtained by regulating the cooler and the dehumidification pump. Some sensors are installed in the rig to measure air flow rate, air temperature and relative humidity, solution flow rate, solution temperature and solution density. Table 1 lists the main features of the different sensors. With these sensors, heat transfer rate and mass transfer rate can be determined as the real-time operating data collected by the data acquisition system installed in the dehumidifier. In order to evaluate the effectiveness of the model, two error indexes, Relative Error (RE) and Rooting-Mean-Square of Relative Error (RMSRE), are used to show the accuracy of the present model:

RE ¼

jDreal  Dcalc j  100% Dreal

RMSRE ¼

Fig. 4. Experimental rig of LDDS.

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uP   u M Dreal Dcalc 2 t i¼1 Dreal M

ð19Þ

ð20Þ

Operating data of the dehumidifier is collected and then model parameters can be identified by the acquired data through nonlinear least squares method, the parameters are evaluated as c1 = 127.3482, c2 = 69.6502, c3 = 1.2300, c4 = 0.5468, c5 = 0.6804, c6 = 0.5773 and c7 = 0.8773. To validate the model in the application performance prediction, 270 points are tested in wide working ranges (10–90% of the designed capacity) with heat transfer rate from 0.2 kW to 1.5 kW and mass transfer rate from 0.002 kg/m2 s to 0.010 kg/m2s, respectively. Figs. 7 and 8 show the comparison results of heat transfer rate between the predicted values and the

454

X.L. Wang et al. / Applied Energy 111 (2013) 449–455

Fig. 6. Schematic diagram of the experimental rig.

Table 1 Specification of the sensors. Sensors

Type

Accuracy

Range

Solution temperature sensor Solution flow meter

PT100, 3-Wire Magnetic flow meter Glass hydrometer Probe

0.15 °C ±0.5%

0–100 °C 0–25 L/min

1 kg/m3

1.10– 1.30 kg/m3 0–100%, 0– 60 °C 0–600 m3/h

Density meter Humidity/temperature transmitter Airflow meter

Blade

0.5%, 0.1 °C ±0.5%

Relative error [%]

15

10

5

0

0

50

100

+10% error

150

200

250

300

Fitting points

1 -10% error

Fig. 8. Relative error for heat transfer.

-3

11

0.5

0

0

0.5

1

1.5

Catalog heat transfer rate [kw] Fig. 7. Model prediction for heat transfer.

experimental data and the RE, respectively, using the universal correlations. 95.9% of the 270 data points are within the RE of ±10% with the average RE of 4.79% and RMSRE of 0.0572. While the comparison results of the predicted values of mass transfer rate in the dehumidifier by the present model with the corresponding experimental data and the RE are shown in Figs. 9 and 10, respectively. 97.3% of the 270 data points are falling within the RE of ±10% with the average RE of 4.26% and RMSRE of 0.0521. The experimental results show that the proposed model has very good agreements with the experimental testing (RE < 10% [26]). It is accurate enough for real-time applications in performance monitoring, control and optimization for dehumidifier in LDDS. A qualitative comparison is also made between the existing dehumidifier models and the

Calculated mass transfer rate [kg/(m2.s)]

Calculated heat transfer rate [kw]

1.5

x 10

10 9 8

+10% error

7 -10% error

6 5 4 3 2 2

3

4

5

6

7

8

Catalog mass transfer rate [kg/(m2.s)]

9

10 -3

x 10

Fig. 9. Model prediction for mass transfer.

proposed model, as shown in Table 2 showing that the proposed model can be expanded over a wider operating range with satisfactory accuracy compared with the NTU or empirical model.

X.L. Wang et al. / Applied Energy 111 (2013) 449–455

15

455

(973 Program: 2012CB720500). Their support is gratefully acknowledged.

Relative error [%]

References 10

5

0

0

50

100

150

200

250

300

Fitting points Fig. 10. Relative error for mass transfer.

Table 2 Comparison of different dehumidifier models. Model

Geometric Iterative computation

Model technique

Model application

Finite difference [13–15] NTU [20–22]

Yes

Yes

Physical

Yes

Yes

Physical

Empirical model [23,24] Present model

No

No

Empirical

No

No

Hybrid

Design and simulation Design and control Design and control Control and optimization

6. Conclusions In this paper, a simple model used for performance monitoring, control and optimization of operating liquid desiccant dehumidifiers was proposed. By analyzing the heat and mass transfer, the hybrid model with only seven characteristic parameters was developed to evaluate the performance of heat and mass transfer rates in the liquid desiccant dehumidifier. Unlike other exiting models, this model focuses only on input–output related variables with the packed column so that iterative computation process can be avoided. Moreover, complex parameters such as packing geometric specifications, fluids’ thermodynamic properties are treated as lumped parameter and can be captured and determined through identification. The present model which contains only seven parameters is simple, flexible, with acceptable accuracy and easy to be applied in engineering compared with the existing models of dehumidifier. According to the results of validation, the present dehumidifier model is accurate and possible to be used for performance monitoring, control and optimization (RMSER < 0.1) over a wide operating range. It should be pointed out that, however, for long period of operating time, model parameters may vary greatly and should be reevaluated periodically to ensure its accuracy in performance prediction. The performance optimization using the developed model is currently under study and the results will be reported later. Acknowledgements This work was supported by National Research Foundation of Singapore under the Grant NRF2011 NRF-CRP001-090, the National Natural Science Foundation of China (NSFC) (No. 21076179), and the National Basic Research Program of China

[1] Xiong ZQ, Dai YJ, Wang RZ. Development of a novel two-stage liquid desiccant dehumidification system assisted by CaCl2 solution using exergy analysis method. Appl Energy 2010;87(5):1495–504. [2] Tu M et al. Simulation and analysis of a novel liquid desiccant air-conditioning system. Appl Therm Eng 2009;29(11–12):2417–25. [3] Lof GOG. Cooling with solar energy. In: Proceedings of the world symposium on applied solar energy. Stanford Research Institute; 1955. [4] Xiong Z, Dai Y, Wang R. Development of a novel two-stage liquid desiccant dehumidification system assisted by CaCl2 solution using exergy analysis method. Appl Energy 2010;87(5):1495–504. [5] La D et al. An experimental investigation on the integration of two-stage dehumidification and regenerative evaporative cooling. Appl Energy 2012. [6] Xiao F, Ge G, Niu X. Control performance of a dedicated outdoor air system adopting liquid desiccant dehumidification. Appl Energy 2011;88(1):143–9. [7] Sultan G, Hamed AM, Sultan A. The effect of inlet parameters on the performance of packed tower-regenerator. Renew Energy 2002;26(2):271–83. [8] Audah N, Ghaddar N, Ghali K. Optimized solar-powered liquid desiccant system to supply building fresh water and cooling needs. Appl Energy 2011;88(11):3726–36. [9] Liu XH, Jiang Y, Yi XQ. Effect of regeneration mode on the performance of liquid desiccant packed bed regenerator. Renew Energy 2009;34(1):209–16. [10] Xiao F, Ge GM, Niu XF. Control performance of a dedicated outdoor air system adopting liquid desiccant dehumidification. Appl Energy 2011;88(1):143–9. [11] Ge GM, Xiao F, Xu XH. Model-based optimal control of a dedicated outdoor airchilled ceiling system using liquid desiccant and membrane-based total heat recovery. Appl Energy 2011;88(11):4180–90. [12] Mei L, Dai Y. A technical review on use of liquid-desiccant dehumidification for air-conditioning application. Renew Sustain Energy Rev 2008;12(3):662–89. [13] Factor HM, Grossman G. A packed bed dehumidifier/regenerator for solar air conditioning with liquid desiccants. Sol Energy 1980;24(6):541–50. [14] Gandhidasan P, Ullah MR, Kettleborough C. Analysis of heat and mass transfer between a desiccant-air system in a packed tower. J Sol Energy Eng 1987;109:89. [15] Öberg V, Goswami DY. Experimental study of the heat and mass transfer in a packed bed liquid desiccant air dehumidifier. Trans – Am Soc Mech Eng J Sol Energy Eng 1998;120:289–97. [16] Fumo N, Goswami DY. Study of an aqueous lithium chloride desiccant system: air dehumidification and desiccant regeneration. Sol Energy 2002;72(4):351–61. [17] Yin Y, Zhang X, Chen Z. Experimental study on dehumidifier and regenerator of liquid desiccant cooling air conditioning system. Build Environ 2007;42(7):2505–11. [18] Ren CQ, Jiang Y, Zhang YP. Simplified analysis of coupled heat and mass transfer processes in packed bed liquid desiccant-air contact system. Sol Energy 2006;80(1):121–31. [19] Mesquita LCS, Harrison SJ, Thomey D. Modeling of heat and mass transfer in parallel plate liquid-desiccant dehumidifiers. Sol Energy 2006;80(11):1475–82. [20] Chen XY, Jiang LZ, Qu KY. Analytical solution of adiabatic heat and mass transfer process in packed-type liquid desiccant equipment and its application. Sol Energy 2006;80(11):1509–16. [21] Liu XH, Jiang Y, Qu KY. Heat and mass transfer model of cross flow liquid desiccant air dehumidifier/regenerator. Energy Convers Manage 2007;48(2):546–54. [22] Liu X et al. Analytical solutions of coupled heat and mass transfer processes in liquid desiccant air dehumidifier/regenerator. Energy Convers Manage 2007;48(7):2221–32. [23] Khan AY. Sensitivity analysis and component modelling of a packed-type liquid desiccant system at partial load operating conditions. Int J Energy Res 1994;18(7):643–55. [24] Khan A, Ball H. Development of a generalized model for performance evaluation of packed-type liquid sorbent dehumidifiers and regenerators. ASHRAE Transactions. 1992;98:525–33. [25] Wang YW et al. A simplified modeling of cooling coils for control and optimization of HVAC systems. Energy Convers Manage 2004;45(18– 19):2915–30. [26] Jin GY et al. A simplified modeling of mechanical cooling tower for control and optimization of HVAC systems. Energy Convers Manage 2007;48(2):355–65. [27] Ding XD et al. Evaporator modeling – a hybrid approach. Appl Energy 2009;86(1):81–8. [28] Waldram JR. The theory of thermodynamics. Cambridge: Cambridge University Press; 1985. [29] Benitez J. Principles and modern applications of mass transfer operations. Wiley-Interscience; 2009. [30] Onda K, Takeuchi H, Okumoto Y. Mass transfer coefficients between gas and liquid phases in packed columns. J Chem Eng Jpn 1968;1(1):56–62. [31] Dean JA. Lange’s handbook of chemistry. 15th ed. McGraw-Hill; 1999. [32] Conde MR. Properties of aqueous solutions of lithium and calcium chlorides: formulations for use in air conditioning equipment design. Int J Therm Sci 2004;43(4):367–82.