View PDF

Access through your institution

Purchase PDF

Journal of Building
Engineering
Volume 32, November 2020, 101813

Experimental based multi-objective optimisation

for structured packed bed liquid desiccant
dehumidification systems
Mrinal Bhowmik a, P. Muthukumar a, b , R. Anandalakshmi c

Show more

Share Cite

https://doi.org/10.1016/j.jobe.2020.101813 Get rights and content

Abstract
In last few decades, a substantial advancement has been observed in dehumidifier
modules for their energy-conservative and environment-friendly nature; however,
still uncertainty in decision-making is of concern in selection of optimum input
parameters to attain an effective performance. In this perspective, an experimental
study is performed in an adiabatic packed bed, counter-flow liquid desiccant
dehumidifier fuelled with the blend of lithium bromide and calcium chloride
solution at various operating conditions with different control settings. It is
observed that air mass flux rate (Fa), solution concentration (ζ) and air specific
humidity (ωa) have significant impact on condensation rate (CR), whereas, air mass
flux rate (Fa) and solution temperature (Ts) are the most effective parameters for
moisture effectiveness and latent heat factor, respectively. The experimentally
obtained results for multiple response performance characteristics are
represented based on inlet process parameters using gene expression
programming (GEP). The proficiency of developed GEP meta-models is provided
by statistical parameters, Taylor diagram and Theil uncertainty. Subsequently,
GEP meta-model based fuzzy logic is developed for optimising dehumidifier inlet
process parameters in terms of multi-responsive performance characteristics
using Genetic Algorithm. The optimisation results showed that dehumidifier
performance characteristics have a tendency to become optimal at
Fa = 0.766 kg/m2-s, Ta = 30.745 °C, ωa = 0.023 kgwv/kgda, Fs = 1.812 kg/m2-s,
Ts = 24.01 °C, and ζ = 48.1%. Finally, experiments were conducted at optimum
conditions and observed proximity between the predicted and experimental
outcomes, which further validates the current optimisation approach.

Graphical abstract

Download : Download high-res image (301KB) Download : Download full-size image

Introduction
Global warming potential (GWP) and ozone depletion potential (ODP) derived
from the leakage of refrigerant from Heating Ventilation Air Conditioning (HVAC)
systems are the prominent contributors for the degradation of air standard
quality, which play a significant role in global warming. In HVAC, the vapour
compression air conditioning system (VCAs) is the best option for heavy-duty
applications to maintain indoor air quality. However, the VCAs consume more
electricity and the recent stringent environment norms imposed on the
conventional refrigerants made to think towards the alternative systems. The
primary reason for using such environmentally harmful and energy-intensive
chemical refrigerant systems is lack of alternative well-developed advanced air
conditioning systems. Further, minimising the usage of these stringent
refrigerants is a prime concern of HVAC industries. In this perspective, active
research works have been focused on minimising the use of these refrigerants [1]
and explore the usage of the liquid or solid based dehumidifier system for HVAC
applications. Dehumidifiers are also conducive to energy saving as the humidity
level can be monitored without cooling the air below its due point temperature.
Generally, the liquid desiccants are most effective to remove airborne biological
contamination and can kill the bacteria and viruses, which are present in the
ambient air [2]. Further, the separate controlling capability of humidity and
temperature with greater economy and less noise compared to the solid desiccants
makes liquid desiccants a better option over solid desiccant systems [3]. Elsarrag et
al. [4] performed theoretical and experimental research on structured packed
dehumidifier using triethylene glycol (TEG) as a liquid desiccant. Their research
sets a design guideline for the structured packed based desiccant centred air
dehumidifier. Further, the experimental outcomes were compared with the
theoretical model for the confirmation of experimental results. Subsequently, it
was observed that the effectiveness of such dehumidifier ranging from 0.72 to
0.91 at different operating conditions. Wen et al. [5] carried out an investigation on
liquid desiccant dehumidification using potassium formate (KCOOH) desiccant
solution. The vapour pressure of KCOOH solution was measured at different
concentrations and temperatures. It was found that the vapour pressure of
KCOOH solution with concentration of 64.3%–73.3% was similar to that of
lithium chloride (LiCl) solution with concentration of 33%–38% when the solution
temperature ranged from 45 °C to 65 °C. Lychnos and Davies [6] conducted
experiments and theoretical modelling to investigate the performance of a liquid
desiccant system using magnesium chloride (MgCl2) desiccant. It was shown that
the moisture absorption rate was ranged from 154 to 194 g h−1 m−3 at different
operating conditions. Due to these advantages, the liquid dehumidifier is widely
adopted in many developed countries. However, the dehumidifier suffers from a
disadvantage in providing effective performance compared to VCAs due to poor
thermo-physical properties of desiccant materials. This shortcoming in the
desiccant systems can be overcome by introducing an effective desiccant mixture.
Recently, many research efforts have been dedicated towards the exploration of
desiccant mixtures, which will provide better dehumidification performance.

Wen et al. [7] carried out an experimental study on the liquid desiccant system by
adding polyvinyl pyrrolidone (PVP) surfactant. It was observed that the mass
transfer performance was improved and the dehumidification rate was increased
with an average of 26.1% compared to the pure LiCl operation. Bhowmik et al. [8]
carried out a numerical investigation to pattern the impact of LiCl + CaCl2 blends
on dehumidifier's outlet parameters at various L/G ratios. It was found that blends
contained 5% CaCl2 and 35% LiCl was more suitable combination for air
dehumidification application among the other prepared blends at L/G ratio of 2.
Qu et al. [9] tested 1-Ethyl-3-methylimidazolium acetate [EMIM][OAc] for
adsorption and desorption process. The test results indicated that [EMIM][OAc]
solution has comparable vapour pressure with LiBr or LiCl solution at a higher
mass fraction of [EMIM][OAc]. Luo et al. [10] investigated the counter-flow
dehumidifier to study the performance of 1-Ethyl-3-methylimidazolium
Tetrafluoroborate ([EMIM]BF4) aqueous solution. The results showed that the
dehumidification performances using [EMIM]BF4 was a little lower than that of
LiBr and subsequently, mentioned that this performance difference could be
compensated by increasing the mass concentration of [EMIM]BF4. Ren et al. [11]
introduced a new working solution of phase change materials (PCM) into aqueous
LiCl desiccant to analyse the dehumidification intensity. The results showed that
the dehumidification efficiency was improved. Ameel et al. [12] conducted
experiments by mixing zinc chloride and lithium chloride desiccants with a
proportion of 1:1.63. It was observed that dehumidification efficiency was
improved at the expense of high cost. The LiCl based desiccant mixtures are more
corrosive, expensive and crystallise at low temperature. Alternatively, Donate et al.
[13] studied the mixture of organic salts (sodium formate/methanol) with LiBr
desiccant and observed that the moisture absorption capability of such mixtures
was high compared to pure LiBr. Lucas et al. [14] investigated the water vapour
absorption capacity of LiBr and sodium formates (CHO2Na) based desiccants with
n-octanol. The results showed that such mixtures increase the water vapour
absorption capability in the presence of n-octanol. The LiBr + CHO2Na at 45% by
mass with 1000 ppm of n-octanol solution showed a better water vapour
absorption capacity than LiBr solution. However, LiBr possesses a demerit of
solubility and is responsible for uneven performance profile throughout the
operation, which can be minimised by introducing calcium chloride (CaCl2), an
inorganic salt (typically synthesised from calcium carbonate (CaCO3) [15].
Moreover, CaCl2 is the most easily available and cheapest desiccant for
dehumidification, but its dehumidification performance greatly depends on
operating conditions [16,17]. Therefore, it is essential to study the effect of CaCl2
on LiBr system at optimal operating conditions to get the maximum performance
from the system. In this paper, an effort is made to fulfil this considerable
research gap in the desiccant performance of a dehumidification system. Table 1
shows other relevant research literature on mixed desiccant performance studies.

However, the studies on desiccant mixture based dehumidifier operation have

established that proper blend ratio is essential for maximum dehumidifier
performance. The key challenge is to identify desiccant materials of higher
performance and to predict the performance parameters effectively to minimise
the time consuming experimental runs. The following relevant literatures
reported the studies, which are majorly involved in the development of a model
using regression techniques such as Non-Linear Regression [24], Artificial Neural
Network [25], Response Surface Method [26], and Least Square Support Vector
Machine [27,28]. However, such techniques lack accuracy due to limited functional
operator availability to establish the relationship between independent and
dependent parameters. In this perspective, Koronaki et al. [28] also developed
some empirical correlation between the input and output variables. It is observed
that their model could reach up to maximum R2 value of 86%. This limitation can
be improved using Gene Expression Programming (GEP), which can forecast the
model with more accuracy due to the availability of various functions and advance
adaptive learning algorithms.

Further, an effort is also needed to be given on the study of optimisation

approaches for liquid desiccant cooling systems (LDCs) to achieve better
performance with an applicable strategy in building HVAC systems. In this
perspective, Ge et al. [29] established an optimal control approach for air-chilled
ceiling system (liquid desiccant based dedicated outdoor) using genetic algorithm
(GA) to obtain maximum performance from the system. Similarly, Qi et al. [30]
optimised control parameters for LDCs using GA in a prospective to achieve the
maximum energy savings from the system for buildings with different operating
conditions. Wang et al. [31] implemented particle swarm optimisation (PSO)
algorithm in liquid desiccant regenerator to optimise the energy consumption of
the system in such a way that the regeneration rate keeps remained at an
acceptable level. Wang et al. [32] used Self-Adaptive Differential Evolutionary
(SADE) algorithm as global optimised operation strategy in LDCs for energy
savings perspective and found that optimisation strategy results could save the
energy unto 8.5%. Literature reveals that the study related to optimisation on
LDCs is limited and predominantly focused on the system design, however, the
control parameters setting optimisation to obtain the enhanced performance from
the system is still significantly lacking, which can be considered as a significant
research gap in real-time dehumidifier performance optimisation. In general, GA
is a stochastic method in nature, and its search operators mostly depend on
arbitrary statistics [33]. Fuzzy logic is applied to provide the intermediate results
within certain constraints and is successfully emulated the results using genetic
algorithm [34]. Therefore, this paper is aimed to use a novel optimisation
approach comprises of GEP-Fuzzy-GA to find out the optimum inlet parameters
to obtain optimum dehumidifier performance.

It is well known that maintaining the desired humidity level in health centres,
research laboratories, pharmaceutical and food processing industries is very
important. It is noted that the desiccant mixture concertation/blending ratio is
one of the critical factors to maintain the desired humidity level. In order to
ensure this, a systematic and robust approach is necessary in terms of energy
conservation and energy efficiency perspective, which is given less attention in the
literature. In the past two decades, although significant energy models were
studied based on heat and mass transfer models in the liquid desiccant
dehumidifier, prediction of dehumidifier performance parameters with higher
accuracy and a methodology to optimise the independent parameters to improve
its energy efficiency are not yet explored to reduce the energy cost of the system.

Therefore, this paper discusses an energy-efficient methodology applied to an

HVAC system like air dehumidification system for utilising optimal energy
resources rather than any modifications to existing HVAC systems. Firstly, an
effort has been made to explore the potential of LiBr + CaCl2 as an aqueous
desiccant mixture as working fluid in liquid desiccant dehumidifier through
rigorous experiments at different operational settings. Secondly, an effective meta-
model is developed to predict the dehumidifier performance parameters
accurately, and an empirical correlation is patterned for the performance
parameters as a function of inlet process parameters. Lastly, a novel paradigm
using GEP meta-model based fuzzy logic combined with genetic algorithm is
developed for tackling multiple responsive optimisation inferences to transform
multiple responses into a single performance criterion through Multiple
Performance Criteria Index (MPCI) for controlling input parameters to get
optimum outputs for real-time implementation strategy in view of energy-saving
perspective. The proposed optimisation approach in this paper is further validated
with experimental results. The methodology of the current work is represented in
Fig. 1 and can be extended to optimise the input process parameters of a
dehumidifier with various other desiccant mixture to achieve enhanced
performance.

Section snippets

Performance indices
The dehumidification process is predominantly a mass transfer dominated
process. Generally, in the mass transfer process sorption heat and latent heat of
condensation/evaporation are released. Subsequently, the energy balance equation
cannot be expressed by only sensible heat factor (i.e., not only as temperature
function). However, Gandhidasan [35] showed that temperature effectiveness
might be negative/positive/even meaningless in the dehumidification process.
Therefore, in the present study, …

Test desiccants and their properties

Experimental investigation was carried out to explore the dehumidifier
performance using LiBr and CaCl2 desiccant mixture. In order to ensure the
mixture constraints of input parameters, a laboratory-based test was performed.
The absorption rate of pure LiBr and CaCl2, and LiBr + CaCl2 mixtures at different
time-steps were represented by the same research group in Bhowmik et al. [36]. It
was found that the overall moisture absorption rate of 85:15 (LiBr: CaCl2) mixture
was higher compared to…

Experimental setup
The dehumidifier setup mainly consists of dehumidification chamber, solution
tank, pump, and a blower (Fig. 4a). Dehumidification chamber is made of stainless
steel material, and an anti-corrosive paint is used to double ensure the corrosion
resistance. Dehumidification chamber is the key part of the setup, and that is
insulated by two layers of nitrile foam rubber (each layer thickness of 6 mm). The
dimensions of the dehumidifier packing chamber are determined based on the
design guidelines…

Energy balance between air and desiccant solution

In order to validate adiabatic condition of the experiments, the energy balance
analysis was conducted. The overall energy exchange (Qe) between the air and
desiccant solution for the dehumidifier with counter-flow configuration can be
estimated as

Air enthalpy:

Change in solution enthalpy: where ha, hs, , ,Cpm

and Cps are enthalpy of air, enthalpy of solution, mass flow rates of air, mass flow
rates of solution, specific heat of moist air…

Summary and conclusions

In this article, an experimental study on mixed desiccant system tested with
optimised dehumidifier inlet parameters to obtain the maximum performance
characteristics is presented. The GEP meta-model is established based on the
experimental results. The GEP meta-model is found to possess higher prediction
capability as experimental results. The fuzzy logic is made to get the intermediate
results within certain constraints which successfully emulated using genetic
algorithm. A novel approach is…

CRediT authorship contribution statement

Mrinal Bhowmik: Conceptualization, Methodology, Software, Investigation,
Writing - original draft. P. Muthukumar: Validation, Supervision, Writing - review
& editing. R. Anandalakshmi: Data curation, Investigation, Writing - review &
editing.…

Declaration of competing interest

The authors declare that they have no known competing financial interests or
personal relationships that could have appeared to influence the work reported in
this paper.…

Acknowledgements
Authors are ostensibly thankful to the Engineering Section, Indian Institute of
Technology Guwahati, Assam, India, for their financial support to develop the
experimental setup (Project No.: IITG/ENGG/AEE/EL81).…

References (60)

S.D. Atalay et al.

Performance analyses of statistical approaches for modeling electricity consumption of a
commercial building in France
Energy Build. (2019)

A. Zendehboudi et al.
Performance investigation of the cross-flow closed-type heat-source tower using
experiments and an adaptive neuro-fuzzy inference system model
Energy Build. (2019)

M. Bhowmik et al.
Experimental based multilayer perceptron approach for prediction of evacuated solar
collector performance in humid subtropical regions
Renew. Energy (2019)

O. Ünal et al.
Fuzzy logic approach to predict stress-strain curves of steel fiber-reinforced concretes in
compression
Build. Environ. (2007)

H. Pombeiro et al.
Comparative assessment of low-complexity models to predict electricity consumption in
an institutional building: linear regression vs. fuzzy modeling vs. neural networks
Energy Build. (2017)

C.S. Kang et al.

Fuzzy logic-based advanced on – off control for thermal comfort in residential buildings
Appl. Energy (2015)

N. Li et al.
Multi-objective optimisation for designing of high-speed train cabin ventilation system
using particle swarm optimisation and multi-fidelity Kriging
Build. Environ. (2019)

Y. Murad
Joint shear strength models for exterior RC beam-column connections exposed to
biaxial and uniaxial cyclic loading
J. Build. Eng. (2020)

T. Ayata et al.
An investigation for predicting the effect of green roof utilisation on temperature
decreasing over the roof surface with Gene Expression Programming
Energy Build. (2017)

H.A. Milukow et al.

Estimation of the Darcy–weisbach friction factor for ungauged streams using gene
expression programming and extreme learning machines
J. Hydrol. (2019)

View more references

Cited by (2)

Experimental study of coupled heat and mass transfer phenomena between air
and desiccant in a solar assisted thermal liquid desiccant system
2021, International Journal of Thermal Sciences

Show abstract

Experimental investigation on structured packed bed liquid desiccant

dehumidifier: An optimal mixture design of experiments strategy
2021, International Journal of Refrigeration

Show abstract

Recommended articles (6)

Research article

The problem of the equivalent time to start the internal frost resistance test on
self-compacting concrete with supplementary cementitious materials
Journal of Building Engineering, Volume 32, 2020, Article 101503

Show abstract

Research article

Suitability evaluation of cold sources in green buildings based on a field

study
Journal of Building Engineering, Volume 32, 2020, Article 101795

Show abstract

Research article

Investigation of parameters affecting direct shear behavior of concrete

strengthened with CFRP laminate
Journal of Building Engineering, Volume 32, 2020, Article 101480

Show abstract

Research article

Service life of GGBFS concrete under carbonation through probabilistic method

considering cold joint and tensile stress
Journal of Building Engineering, Volume 32, 2020, Article 101826

Show abstract

Research article

In vitro deterioration study of concrete and marble by Aspergillus

tamarii
Journal of Building Engineering, Volume 32, 2020, Article 101774

Show abstract

Research article

Application of multiple step change in shear rate model for determination of

thixotropic behaviour of cement pastes
Journal of Building Engineering, Volume 32, 2020, Article 101494

Show abstract

View full text

© 2020 Elsevier Ltd. All rights reserved.

About ScienceDirect Remote access Shopping cart Advertise Contact and support
Terms and conditions Privacy policy

We use cookies to help provide and enhance our service and tailor content and ads. By continuing you agree to the
use of cookies.
Copyright © 2021 Elsevier B.V. or its licensors or contributors. ScienceDirect ® is a registered trademark of Elsevier
B.V.
ScienceDirect ® is a registered trademark of Elsevier B.V.