ArticlePDF Available

Abstract and Figures

Direct evaporative coolers (DECs) are a low-energy cooling alternative to conventional air conditioning in hot-dry climates. The key component of DEC is the cooling pad, which evaporatively cools the air passing through it. While detailed numerical models of heat and mass transfer have been proposed for the cooling pad, these require many input parameters that are not readily accessible. Alternatively, simplified models lack accuracy and are confined to a common type of cooling pad. To address these limitations, we developed and validated a physics-based model for the evaporative cooling pad that only needs the nominal data to compute the heat and mass transfer with considerable accuracy. The proposed model is implemented in Modelica, an equation-based object-oriented modeling language. For comparison, a basic lumped model from EnergyPlus based on the efficiency curve of the cooling pad is also implemented. The physics-based model exhibits <2% error from the experimental data and the lumped model exhibits a 12.3% error.
Content may be subject to copyright.
Validated Open-source Modelica Model of Direct Evaporative Cooler with
Minimal Inputs
Saranya Anbarasu 1,4, Wangda Zuo 1,2 *, Yangyang Fu 1, Yash Shukla 3, Rajan Rawal 3,4
1 Department of Civil, Environmental and Architectural Engineering, University of Colorado
Boulder, Boulder, CO, USA
2 National Renewable Energy Laboratory, Golden, CO, USA
3 Centre for Advanced Research in Building Science and Energy, CRDF, CEPT University,
Ahmedabad, India
4 Faculty of Technology, CEPT University, Ahmedabad, India
*Corresponding author:
S. Anbarasu, W. Zuo, Y. Fu, Y. Shukla, R. Rawal .2022. "Validated Open-source Modelica Model
of Direct Evaporative Cooler with Minimal Inputs," Journal of Building Performance Simulation, 15
(6), pp. 757–770,
Direct evaporative coolers (DECs) are a low-energy cooling alternative to conventional air
conditioning in hot-dry climates. The key component of DEC is the cooling pad, which
evaporatively cools the air passing through it. While detailed numerical models of heat and mass
transfer have been proposed for the cooling pad, these require many input parameters that are not
readily accessible. Alternatively, simplified models lack accuracy and are confined to a common
type of cooling pad. To address these limitations, we developed and validated a physics-based
model for the evaporative cooling pad that only needs the nominal data to compute the heat and
mass transfer with considerable accuracy. The proposed model is implemented in Modelica, an
equation-based object-oriented modeling language. For comparison, a basic lumped model from
EnergyPlus based on the efficiency curve of the cooling pad is also implemented. The physics-
based model exhibits <2% error from the experimental data and the lumped model exhibits a
12.3% error.
Keywords: Direct evaporative cooler, Modelica, Physics-based model.
1. Introduction
Conventional air conditioners that are based on the vapor compression refrigeration cycle are being
used for cooling in residential and commercial buildings throughout the world. These conventional
systems employ refrigerants like hydrochlorofluorocarbons (such as R-22) and hydrofluorocarbon
(such as R-134a, and R-410A) that have high global warming potential (GWP) (Weubbles, 1994).
Despite these systems having high GWP, they are commercially dominant as conventional systems
are stable in space conditioning (Vakiloroaya et al., 2014). For regions with hot and dry climates,
energy consumption for space cooling is over 60% of the total energy used in buildings
(Boukhanouf et al., 2014). To reduce the energy consumption of conventional systems, alternative
options such as evaporative cooling can be utilized to significantly save energy and reduce CO2
emissions (Dodoo, 2011). The evaporative cooling system is more environmentally friendly as it
uses water instead of refrigerants as the working fluid to cool the air through the process of
evaporation (ASHRAE, 2013). Though direct energy comparisons cannot be made between the
DECs (open systems) and conventional closed systems, DECs can be used to augment the energy
savings of conventional systems. Over the recent years, there has been extensive research on
precooling components that can reduce the peak cooling load of the cooling system. In addition to
DECs, there are (1) indirect evaporative coolers (IEC) that lower the temperature without the
increase in relative humidity; (2) hybrid coolers that have direct, indirect, and DX coils connected
in series; and (3) dew point evaporative coolers that can cool the air below wet bulb temperature
to have proven heat recovery and pre-cooling benefits in hot and dry climates (Sajjad et al., 2021).
One such experimental testing exhibits a 22.9-35.1% peak cooling load reduction by using DEC
and IEC as pre-cooling components for central air-conditioning systems (Chen et al., 2014; Min
et al., 2021).
Figure 1 shows a typical Direct Evaporative Cooler (DEC), with the key component being the
cooling pad that serves as heat and mass exchangers that are built with layers of humidity-
absorbing materials such as cellulose, aspen, paper, etc. The cooling pad is wetted using a pump.
A centrifugal fan blows the air through the wetted cooling pad, which cools the dry air by
increasing the relative humidity. Thermodynamically, the energy required to evaporate the water
is taken from the air in the form of sensible heat and is converted into latent heat. This conversion
of sensible heat to latent heat, without the change in enthalpy, is known as an isenthalpic process.
Figure 2 represents the direct evaporative cooling process on the psychrometric chart as a parallel
line to the wet bulb line and the enthalpy line. Evaporative cooling, therefore, causes a drop in the
air temperature, proportional to the sensible heat drop, and an increase in relative humidity,
proportional to the latent heat gain (Fouda & Melikyan, 2011). The extent of temperature drop
depends on the duration of contact, the surface area of contact, and the mass flow rate of air passing
through the cooling pad. There is also a considerable pressure drop depending on the geometric
configuration of the cooling pad, which must be accounted for by the fan to maintain the specific
outlet mass flow rate.
Figure 1 Diagram of a typical direct evaporative
cooler with a built-in pump and fan
Figure 2 Direct evaporative cooling process on the
psychrometric chart
2. Existing Numerical Heat and Mass Transfer Models of DEC
The passive cooling potential of DECs has been established for various climates in past years
(Venkateswara Rao & Datta, 2020; Saman et al., 2010; ; Jaber & Ajib,
2011), yet research demands still exist due to the limited technical data and numerical models
(Amer et al., 2015). Prior investigations led to diverse approaches to developing numerical models
for evaporative coolers (Table 1); however, they are not open source, not flexible, and require to
be implemented by the user in the preferred computational tool. In 1980, Holman described a
numerical heat and mass transfer method to calculate the performance of evaporative cooling
systems based on the -NTU method of the heat exchanger (Holman, 1980). Maclaine-Cross and
Banks (1981) added a linear function of the air saturation line and a stationary water film. By 1987,
Nu, and Sherwood number Sh, for rigid cellulose media by Dowdy and Karabash (1987). Several
       Nu and Sh correlations for various
evaporative media and configurations using experiments (A. Franco et al., 2010; He et al., 2015).
 
 
 
 
 
Dai and Sumathy (2002) proposed an elaborate model with governing equations of liquid film and
gas phases, as well as the interface conditions between the media. Kachhwaha and Prabhakar
(2010) incorporated the impact of elevated water temperature on the cooling efficiency, by adding
dimensionless cor   
(2017) proposed an energy and mass conservation model of humid air and water in a one-
dimensional geometry by applying correlations for heat and mass transfer coefficients. Most of
these methods can predict the performance of DECs with 85-98% accuracy but require few
measured values from experiments to supplement the mathematical model. Using assumptions to
parameters such as the temperature of water at the media interface, the number of segments in the
cooling pad, enthalpy correction factor, Nu, and Sh, etc., can result in significant variations in the
model prediction. Thus, there is a need for a new DEC model, which only requires the basic and
easily accessible input information yet provides accurate predictions.
Table 1 Existing research on heat and mass transfer of DECs
(Holman, 1988)
& Banks, 1981)
(Kettleborough &
Hsieh, 1983)
(Dowdy et al.,
Dowdy &
Karabash, 1987)
(Halasz, 1998)
Theoretical paper
(Camargo &
Ebinuma, 2003)
(Dai & Sumathy,
Outlet temperature
prediction ±
(Wu et al., 2009)
Outlet temperature
prediction ±
(Kachhwaha &
Prabhakar, 2010)
(Fouda &
Melikyan, 2011)
Outlet temperature
prediction ± 0.7°C
(Sodha &
(Crawley et al.,
Sourbron, 2017)
Correspondingly, in the building energy modeling and simulation industry, there are limitations in
the availability of validated DEC models in the existing simulation tools. DOE-2 (Winkelmann et
al., 1993) and IES (IESVE, 2011) have validated single and two-stage DEC models, where the
outlet conditions are based on efficiency (user input value). EnergyPlus contains component
models for direct and indirect evaporative coolers based on an industrial standard CelDek cooling
pad (Crawley et al., 2001). EnergyPlus also gives a research special component that calculates
operation efficiency using an efficiency modifying curve and part load fraction of a static
efficiency input. Thus, the EnergyPlus models are only limited to cellulose cooling pads. With
advancements in interactive buildings, there arrives a need to dynamically test the performance of
systems and components that are non-linear and complex  , 2009). Thus, a flexible
simulation method that can satisfy the above needs is desired.
To overcome these issues, this research develops and validates a physics-based DEC model
which is (1) capable of accurate heat and mass transfer predictions using the easily available
catalog data; (2) flexible enough to be modeled and simulated for various needed, such as an
entire DEC system model, cooling pad model, individual blocks that can facilitate alternative
heat and mass transfer equations testing, etc.; (3) an open-source model contribution to support
the growing needs of the modeling community (freely available at:
To cater to these needs, we have implemented the new DEC model in Modelica, which is an
equation-based object-oriented modeling approach, capable of testing complex, dynamic, and non-
linear systems (Elmqvist & Mattsson, 1997). Modelica facilitates component-based modeling
which is useful to build the heat and mass transfer equations as individual blocks and integrate
them as cooling pad components and then as a system. There are many open-source Modelica
libraries for building systems such as Modelica Buildings library (MBL), IDEAS, AixLib,
Building Systems, etc. that are under constant development (Wetter et al., 2014; Jorissen et al.,
2018; Mehrfeld et al., 2016; and Plessis et al., 2014). Models similar to the direct evaporative
cooler are not available in the commonly used open-source Modelica libraries, which also
substantiates the need for developing this open-source model in Modelica. The rest of this paper
is organized as follows: Section 3 introduces two mathematical models for the evaporative cooling
pad with a varying degree of input parameters, (i) a lumped model, and (ii) a detailed physics-
based model. Section 4, describes the Modelica implementation of the evaporative cooling pad
models and the integrated DEC system model. Section 5, describes the evaluation of the cooling
pad with the DEC system by comparing the performance to the experimental data from the
literature. At last, simulation results are summarized and concluding remarks of this paper are
3. Mathematical Model Description
3.1 Lumped Cooling Pad Model
The lumped cooling pad model is based on the model implemented in the well-known simulation
program EnergyPlus (Crawley et al., 2001). The outlet conditions of the lumped model depend
primarily on the efficiency of the cooling pad . The function used in EnergyPlus is derived
            and pad
thicknesses. The least-squares routine produced an eleven-term multi-variate fit using a third-order
quadratic. This equation is limited to the commonly used   cooling pad. The
efficiency equation needs modifications to support various cooling pad media and can be
determined only through experiments.
where is the thickness󰇛󰇜 and is the velocity at the face of the cooling pad󰇛󰇜. Using
the calculated from eq.(1), the dry bulb temperature of the outlet air can be estimated using the
efficiency relationships of evaporative systems:
 
 
where and are the inlet and outlet dry bulb temperature 󰇛󰇜 and  is the inlet
wet bulb temperature 󰇛󰇜. As evaporative cooling is an isenthalpic process, the inlet wet bulb
temperature  is equal to the outlet wet bulb temperature ,
 
Based on this assumption, the resulting humidity ratio of the outlet , is calculated using the
psychrometric properties of air described in Appendix. The volume flow rate of water evaporated
󰇗󰇛󰇜, which is added to the airside is determined using,
󰇗 󰇗󰇛 󰇜
where 󰇗, is the mass flow rate of air󰇛󰇜;  and  are the outlet and inlet humidity
ratios 󰇛
󰇜; and is the density of water 󰇛󰇜 (standard density of water at 25°C is
used). Once the properties of the outlet air are determined, the total volume flow rate of water
consumption can be determined using the,
󰇗 󰇗 󰇗 󰇗
󰇗 󰇗
󰇗 󰇗
 󰇗
where 󰇗 is the total volume of water consumed󰇛󰇜; 󰇗 is the volume flow rate of water
leaving as droplets on the supply side 󰇛󰇜; 󰇗 is the volume flow rate of water drained from
the sump to counter the build-up of solids in the water that would otherwise occur because of
evaporation󰇛󰇜;  is the drift factor ( if the system has no losses); and  is
the ratio of solids in blowdown water compared to freshwater.
3.2 Physics-based Cooling Pad Model
The physics-based cooling pad model is built based on the governing equations of the heat and
moisture transfer between water and air. The rate of sensible , and latent heat transfer ,
along a small thickness of the cooling pad  is defined as,
󰇛 󰇜
where is the surface area 󰇛󰇜 represented as the product of breadth and height (󰇜 of
the cooling pad, is the dry bulb temperature󰇛󰇜, is the temperature of water film󰇛󰇜,
is the convective heat transfer coefficient󰇛󰇜, is the mass transfer coefficient󰇛󰇜,
is the latent heat of vaporization󰇛󰇜, is the saturated humidity ratio󰇛
󰇜, and
is the humidity ratio󰇛
󰇜. Considering that the rate of sensible heat removed from the
air is equal to the latent heat gain rate from the evaporation of water we get,
By assuming the inlet boundary conditions to be   , integrating eq.(9) we can
obtain the change of temperature with thickness as (Wu et al., 2009),
  󰇧
󰇗 󰇨
where  is the wet-bulb temperature 󰇛󰇜, 󰇗 is the mass flow rate of air 󰇛󰇜, and  is
the specific heat of the air 󰇛󰇜. Through analysis of various numerical methods and
empirical correlations of heat and mass transfer of DECs (He et al., 2015), we identified that the
area of heat transferred used in the heat transfer equation of the cooling pad was the most
significant one to predict the outlet conditions. Wu et al., (2009) presented the use of  which is
the pore surface coefficient per unit volume 󰇛󰇜, which represents the total area that is in
contact with the air. The is specific for various cooling pad material and configuration (e.g.,
cellulose of 45°by 45° flutes with 147 sheets,  ). This easily available is the
key to achieve model accuracy. Thus, the total area of heat transfer is defined as,
By assuming water film temperature is approximately equal to  and 󰇗 the
cooling efficiency is derived by solving eq.(2) and eq.(11),
󰇧 
The derived is a function of , , ,and . The specific to the cooling pad can be
determined by using the empirical correlations of   , for convective heat
transfer for flow across banks of tubed (Incropera et al., 1996), which is similar to that of the
evaporative cooling pad media, with additional non-dimensional geometric parameter
where   is the Reynolds number,  is the Prandtl number, is the
characteristic length of the cooling pad󰇛󰇜, which is calculated by dividing the volume V󰇛󰇜, of
the cooling pad by the total wetted surface area , and coefficients C, a, and b are determined
based on experimental testing of the cooling pad.  and  are determined by,
 
where, is the dynamic viscosity of air󰇛󰇜; is the density of air 󰇛󰇜; is the thermal
conductivity of air 󰇛󰇜. For our modeling, we have identified the  correlations for the
commonly used cooling pad media described in Table 2.
Table 2 Nu correlations for different cooling pad media
pad media
Pore surface
coefficient per
unit volume
Nusselt’s number correlations
(Dowdy &
Karabash, 1987)
coated rigid
(Rawangkul et
al., 2008)
Coconut coir
(Liao et al.,
937<Re< 1390
(Maurya et al.,
Using the calculated from eq.(13), we can determine the using eq.(2). The corresponding
is determined based on the assumption  , using the psychrometric calculations
described in Appendix. The water consumed is calculated using eq.(4)-󰇛7󰇜, similar to the lumped
model. Finally, the sensible heat transfer and latent heat transfer are calculated using the
determined outlet properties of the air,
󰇗 
󰇗󰇛 󰇜
Pressure drop introduced by the pad media is critically important as it can impact the mass flow
rate of air through it, as well as the power consumed by the fan to cater to the pressure drop. A
universal pressure drop expression specific to evaporative cooling pads proposed by Franco et al.
(2014) is used to determine 󰇛󰇜,
 󰇡
The  accounts for the impact of the mass flow rate of water 󰇗, , . The coefficients a, b,
and c can be calibrated for different cooling pads using values from the literature. The pump and
fan models used in this work are from the Modelica buildings library and the underlying numerical
equation used can be referred from Wetter (2013).
4. Model Implementation in Modelica
In this section, we first describe the Modelica implementation of the two variants of the cooling
pad introduced in Section 3. Then we present the implementation of the DEC system model (Figure
1), using the developed cooling pad models and the existing fan and pump model from the
Modelica buildings library (Wetter et al., 2014). Figure 3 shows the hierarchical structure of the
DEC cooling pad model, which consists of three functions: heat transfer, mass transfer, and flow
resistance. These functions are realized by using Modelica blocks which as commonly called
components, implemented at different levels. Three top-level components include efficiency, heat
transfer, and water consumption; and two bottom-level components include static conservation
equation and flow resistance. The combination of efficiency, heat transfer, and static conservation
equation components realize the function of heat transfer. While the water consumption
component with static conservation equation realizes the function of mass transfer. Finally, a flow
resistance component with a fixed flow coefficient maintains the pressure drop. Both the variants
of the cooling pad have the same internal hierarchical structure (Figure 4), yet the difference
between the two cooling pad models is the underlying mathematical equation in the internal
modules. This provides flexibility to interchange the variants for different modeling purposes. The
pad models are built using the standard four-port interface of MBL with mixing volume and flow
resistances connected on both the air and waterside. The ports are assigned with predefined media
of MBL, the Buildings.Medium.Air and Buildings.Medium.Water. The mixing volume
component represents the air and water flowing through the cooling pad, to which the heat and
mass are exchanged. These mixing volumes have the static conservation equation for energy and
mass balance implemented and hence can calculate the outlet conditions of the medium based on
the values from the input connections. The mixing volume on the airside (volAir) has a moisture
port to add the evaporated water mass flow rate 󰇗 to the medium. Conversely, the total mass
flow rate of water 󰇗 required for the process of evaporative cooling is removed from the fluid
port of the mixing volume on the waterside (volWat). The flow resistance creates a pressure drop
between the inlet and outlet ports for different mass flow rates 󰇗, using the nominal pressure drop
value , either from calculations, catalog or from engineering standards.
Figure 3 Hierarchical structure of the cooling pad model.
Figure 4 Icon of cooling
pad model
4.1 Lumped Model for DEC Cooling Pad
Figure 5 represents the Modelica implementation for the lumped cooling pad model. The air/water
inlet and outlet ports enable the connection of the cooling pad to a DEC system or an AHU. The
three top-level functions in Figure 3 are implemented as Modelica components. These components
use the user input values for the geometric and thermal characteristics of the cooling pad to
determine the outputs. In addition, there are sensors connected to the air and waterside to measure
the inlet properties of the medium. The component EffLum calculates the , using the inlet air
velocity and pad thickness and outputs the value. The component HeaTraLum, uses the output
of EffLum as its input and determines the outlet dry bulb temperature . The output is used by
the component WatCon to computes 󰇗, which is added to the mixing volume on the airside
(volAir) and 󰇗that is removed from the mixing volume on the waterside (volWat). The nominal
pressure drops across the cooling pad and pipe connecting the pump and the top of the cooling pad
is also a user input (determined from manufacturers' catalogs).
Figure 5 Modelica model of the lumped cooling pad
4.2 Physics-based Model for Cooling Pad
Figure 6, represents the Modelica implementation of the physics-based cooling pad. The
hierarchical structure is similar to the lumped cooling pad model but differs in the detailed
equations implemented (section 3.2). The block Effphy uses the geometric characteristics of the
cooling pad and the psychrometric properties of inlet air to calculate the heat transfer coefficient
, outlet dry bulb temperature  and outlet mass fraction . These outputs are used by both
the HeaTraPhy and WatCon components. The HeaTraPhy, calculates and outputs the sensible and
latent heat transfer, . The component WatCon uses  and to calculate the
mass flow rate of water evaporated 󰇗 and the total mass flow rate of water consumed 󰇗
that is added to the air side (volAir) and removed from the water side (volWat) respectively. For
the flow resistance function, the pressure drops , across the cooling pad for the inlet mass flow
rate 󰇗 is calculated using eq.(20) and is plugged into the dpPad. The waterside  is calculated
based on the 󰇗 and  from the catalog specification of the DEC system in the
Figure 6 Modelica model of the physics-based cooling pad
4.3 Direct Evaporative Cooler Model
The DEC system model is implemented using the developed cooling pad models and the pump
and fan component model from MBL. Figure 7 is the icon model of the DEC system and Figure 8
shows its internal components. The DEC system model is reduced to two ports representing only
the airside, as the water is assumed to be recirculating within the reservoir. The cooling pad can
be interchanged based on the modeling requirements. The fan model is implemented as an rpm
input fan, as most commercially available DEC systems run on multiple speeds. The fan curves
specific to the system evaluated can be used as the model input. The DECs typically have constant
mass flow pumps; hence the pump model is implemented as mass flow input, requiring 󰇗
and  as inputs. This model outputs the total mass flow rate of water consumed as
watCon; pump and fan power as pumP and fanP respectively.
5. Model Validation
For validation, we implemented the DEC with a physics-based cooling pad connected to a source
and a sink representing the boundary conditions of the inlet and outlet air (Figure 9). A duct
resistance, dry bulb, and wet bulb sensors are added to the outlet of the DEC. Various experimental
data from the literature are used for evaluating different parameters of the model represented in
Table 3. The DEC with lumped cooling pad is also tested for comparison. As the lumped model is
less sophisticated, it only calculates a few outputs, and thus has limited evaluation. The root mean
squared error (RMSE) and normalized mean bias error (NMBE) metrics are used to evaluate the
performance of the developed models. RMSE can indicate  predict the overall
load shape that is reflected in the data, and positive and negative values of NMBE can determine
if the model over or underpredicts the data points.
Figure 7 Modelica icon of the DEC
Figure 8 Modelica model of the DEC system model with built-in
fan and pump
Table 3 Model outputs used for comparing the performance.
Model outputs
Reference results for comparison
(Davis & Elberling, 2007)
Pressure drops,
(Franco et al. 2014)
Power, P
(Davis & Elberling, 2007)
Pad media
(Jain & Hindoliya, 2011; Rawangkul et
al., 2008; Maurya et al., 2014)
Water consumption, 󰇗
(Davis & Elberling, 2007)
5.1 Evaluation of the Efficiency Performance
The measured data from the performance test report of the Breezier evaporative cooler (Icon170)
by PG&E is used for the validation (Davis & Elberling, 2007). This performance test follows the
ASHRAE 133 standard for testing the DEC. The Modelica simulations are run with the same
boundary conditions as that of the experiment (=31.8 - 42.8 °C,  = 18.8-19.5 °C, =
0 -2.5 Kg/s, at 5 different rpm) with the catalog specifications of the Breezier DEC system (Icon
170) and CelDek cooling pad (Munters 440). The manufacturer's performance and system curves
are used for calibrating the fan model (Seeley International, 2015). The outlet conditions
Figure 9 Modelica model of the DEC system with boundary conditions used for validation
 from the simulation results are used to calculate the efficiency  using the
relationship from eq.(2). The cooling efficiency of the DEC system for various mass flow rates is
presented in Figure 10. The of the physics-based model closely follows the experimental data,
whereas the lumped model follows a similar trend with an offset. The clustered points in
experimental data represent a slight change in at each speed with varied inlet conditions,
however, this is not distinctly reflected in the physics-based model. Figure 11 presents the at
varied inlet air dry bulb temperature and wet bulb depression (  ). The physics-
based model closely overlaps the experimental at  between 14-19 °C, exhibiting a minimum
error for = 40.8 °C cases. From Table 4, the lumped model underpredicts the by 11.5%
with an RMSE=9.9, whereas the physics-based model overpredicts by 0.6% with an RMSE=1.2.
As the lumped model follows the experimental data with a constant offset, if calibrated using a
correction coefficient for the curve fit eq.(1), it can result in a lesser error. But obtaining a
correction factor for each type of cooling pad can be difficult in practice. Hence, the physics-based
model has the advantage compared to the lumped model due to its accurate prediction without any
calibration of the cooling pad and system parameters.
Figure 10 Efficiency of the cooling pad for various inlet mass flow rates of air.
Figure 11 Efficiency of the cooling pad for various wet-bulb depressions.
5.2 Evaluation of Water Consumption
The volume of water consumption is an important aspect of assessing the performance of DEC;
therefore, it is essential to have an accurate prediction. The lumped and the physics-based cooling
pad models use the same equation for calculating the water consumption, yet the difference in
󰇗, occurs due to the variation in , which is a function of . A similar trend as that of the
efficiency performance can be observed for water consumption values represented in Figure 12.
The vertical lines of data points represent the water consumption at each speed and the physics-
based model closely follows the experimental water consumption data. While observing the water
consumed at varied inlet air dry bulb temperature and wet bulb depression (Figure 13), there is a
minimum error for WBD between 14-22 °C for =40.8 to 42.8 °C. From Table 4, the lumped
model underpredicts the water consumed by 13.7% with an RMSE=0.001, whereas the physics-
based model overpredicts by 0.9% with an RMSE=0.0004.
Figure 12 Water consumed for various inlet mass flow rates of air.
Figure 13 Water consumed for various wet-bulb depression and inlet dry bulb temperature.
5.3 Evaluation of Power Consumption
In Modelica models, the predicted power consumption of the fan is based on the pressure drop
curve and the system curve used as input. To minimize errors, the system curves derived from the
experimental result are used for calibration (Davis & Elberling, 2007), instead of the system design
curves available from the catalog (Seeley International, 2015). Figure 14 presents the total energy
consumed by both pump and the fan. The pumps consume a constant 30W throughout the operation
and the variation in power is contributed by the fan. Both the lumped and physics-based cooling
pad models have a similar trend and closely followed the experimental data. The experiment data
showed slight variations of the power consumption at a particular speed/rpm, which can be
identified as the clustered data points; however, this is not observed in both lumped and the
physics-based model. From Table 4, the lumped model underpredicts the power consumed by
11.7% with an RMSE=75.1, and the physics-based model underpredicts by 3.6 % with an
RMSE=28.7. Although the lumped and physics-based models use the same fan models, the latter
is more accurate due to the precise calculation of the pressure drop across the cooling pad, which
impacts the fan power.
Figure 14 Power consumed at various mass flow rates of air
5.4 Evaluation of Pressure Drop
The experiment results from Franco et al. (2014) are used for validating the pressure drop equation
implemented in the DEC model. The pressure drops values from the simulation, for two different
pad thicknesses ( = 50mm and 100mm, cellulose pads with 45° by 45° flutes) under several inlet
air velocities , are compared with the experimental results. The coefficients of the generic
pressure drop eq.(20) has been calibrated for this experimental case:
󰇢󰇛󰇜 (21)
Since the lumped cooling pad model uses a nominal pressure drop user input value for calculations,
only the physics-based cooling pad model is considered for validation. The fan model is calibrated
 From Figure
15, the pressure drop curve closely aligns with the experimental data for velocities less than 2.5m/s,
whereas for higher velocities there is a larger error observed. From Table 4, the physics-based
model underpredicts the pressure drop by 4.7% for a 50 mm thick cooling pad and 10.1% for a
100 mm thick cooling pad. Moreover, the velocity prediction of the model also is under-predicted
between 2-6.8%. Although the accuracy of pressure drop prediction for thicker cooling pads
reduces considerably, it is within 10% acceptable limit. Thus, the model can compute the
pressure drop for different thicknesses of the cooling pad without the need for nominal pressure
drop values as input.
Figure 15 Pressure drops across the different thicknesses of the cooling pad (d=50mm, 100mm)
and air velocity.
5.5 Evaluation with Different Pad Materials
Traditional evaporative cooling media is made of aspen or wooden cooling pads. With
advancements in research and testing, the traditional media is replaced by engineered, plastic, or
ceramic coated cellulose and rigid cellulose cooling pads. The model is simulated for four different
pad media, varying in pore surface coefficient per unit volume , and  correlation coefficients.
Figure 16 presents the simulated (S) and measured (M) for four different pad media for different
. The model can accurately predict the of GLASdek and CELdek media with -0.37% and 1.8%
errors, respectively. Whereas the error in predicting the for aspen and coir media is 4.8% and -
9.11% respectively. This difference in error percentages is due to the lack of availability of the
values for the coir and aspen pads. There pads are organically made, in contrast to the factory-
made CELdek or GLASdek pads whose are precisely determined. Calibrating the of aspen and
coir pads can result in reduced error.
Figure 16 Saturation efficiency for different face velocities for various evaporative
cooling media
Summarizing the results in Table 4, it is evident that the physics-based cooling pad model has an
accurate prediction with NMBE between 0.6% to 3% for parameters related to heat and mass
transfer, and NMBE between 2% to 10% for parameters related to the pressure drop. Whereas the
lumped model has the NMBE in the range of 11-13%, with lesser outputs and hence cannot be
used for detailed simulations. The physics-based model can predict the overall load shape that is
reflected in the data based on the RMSE values. Apart from the lesser RMSE and NMBE, the
physics-based cooling pad model is also found to be capable of simulating the performance of
different evaporative cooling media. Currently, we have implemented the Nu correlations for four
commonly used cooling pad media, which can be extended as required.
Table 4 Summary of errors
Parameters evaluated
Lumped model
Physics-based model
NMBE (%)
NMBE (%)
Water consumption
Power consumption
Pressure drop
Various pad media (efficiency)
6. Conclusion
In this paper, we systematically formulated, implemented, and validated a direct evaporative cooler
model in Modelica, comprising two variants of evaporative cooling pads. (i) a physics-based model
for the cooling pad that only needs the nominal catalog data to compute the heat and mass transfer
with less than 1.9 % error (average NMBE). (ii) a simplified model of the cooling pad (i.e., the
lumped model) from EnergyPlus exhibiting a 12.3% error (average NMBE) compared to the
experimental data. Both the models require the dimensions of the cooling pad (L, B, H) and
nominal pump and fan curves as inputs. The key difference in the accuracy is due to the use of
pore coefficient area per unit volume in the physics-based model, which accounts for the porosity
and the area of heat and mass transfer . On the other hand, the lumped model from EnergyPlus
uses a multivariate curve fit equation specific to the cooling pad used, which is not easily available
in manufacturers' catalogs.
Other advanced models discussed in the literature section require additional input parameters such
as the temperature of water at the media interface, parameters specific to the configuration of the
cooling pad, heat and mass transfer correlation coefficients, correlation coefficients for elevated
water temperatures, etc. for accurate model prediction; and few of the inputs are determined via
experiments. Therefore, considering the balance between the availability of inputs and model
accuracy, the physics-based model developed in this paper is well capable of performing detailed
energy simulation with the easily accessible catalog data. With the component-based Modelica
implementation, the developed model can be re-used at various scales such as (1) DEC system-
level simulations combined with a room thermal model; (2) Simulating the cooling pad as a pre-
cooling component for a central cooling system; (3) Combining the cooling pad model with the
indirect evaporative cooling component and Dx coil to test the performance of a hybrid cooling
system; (4) Humidity based control of the cooling pad, etc.
This research can be further extended to develop a package of evaporative coolers (direct, indirect,
hybrid) which can aid the design and development of low energy cooling systems, pre-cooling
peak load savings, develop new evaporative media, and establish comfort and energy-efficiency
control algorithms for hot and dry climate zones.
Model Availability
The DEC model along with the underlying components and base classes are freely available in our
open-source GitHub location: The
model package also includes the example and validation cases discussed in this paper. Any future
additions and revisions to the models can also be found in the same location.
The first author was initially funded by IUSSTF (Indo-US Science and Technology Forum) via
the BHAVAN (Building Energy Efficiency Higher & Advanced Network) student internship
program at the University of Colorado Boulder. This research was supported by the National
Science Foundation under Awards No. IIS-1802017. BIGDATA: Collaborative Research: IA: Big
Data Analytics for Optimized Planning of Smart, Sustainable, and Connected Communities. This
work also emerged from the IBPSA Project 1, an international collaborative project conducted
under the umbrella of the International Building Performance Simulation Association (IBPSA).
Project 1 aims to develop and demonstrate a BIM/GIS and Modelica Framework for building and
community energy system design and operation.
This work was also supported by the Department of Science and Technology (DST), Government
-energy-cooling Technologies: Assessment to
Tech-transfer (DELTA-      on order number
TMD/CERI/BEE/2016/056(C) and TMD/CERI/BEE/2016/056(G).
A. Franco, D. L. Valera, A. Madueño, A. Peña, Franco, A., Valera, D. L., Madueño, A., & Peña,
A. (2010). Influence of water and air flow on the performance of cellulose evaporative
cooling pads used in mediterranean greenhouses. Transactions of the ASABE, 53(2), 565
Amer, O., Boukhanouf, R., & Ibrahim, H. G. (2015). A Review of Evaporative Cooling
Technologies. International Journal of Environmental Science and Development, 6(2), 111
ASHRAE. (2013). 2013 ASHRAE HandbookFundamentals (IP).
Boukhanouf, R., Ibrahim, H. G., Alharbi, A., & Kanzari, M. (2014). Investigation of an
Evaporative Cooler for Buildings in Hot and Dry Climates. Journal of Clean Energy
Technologies, 2(3), 221225.
Camargo, J., & Ebinuma, C. (2003). A Mathematical Model for Direct and Indirect Evaporative
Cooling Air conditioning Systems. Revista Da Engenharia Térmica, 2(2), 3034.
Chen, Y., Luo, Y., & Yang, H. (2014). Fresh Air Pre-cooling and Energy Recovery by Using
Indirect Evaporative Cooling in Hot and Humid Region - A Case Study in Hong Kong.
Energy Procedia, 61, 126130.
Crawley, D. B., Lawrie, L. K., Winkelmann, F. C., Buhl, W. F., Huang, Y. J., Pedersen, C. O.,
Strand, R. K., Liesen, R. J., Fisher, D. E., & Witte, M. J. (2001). EnergyPlus: Creating a
New-generation Building Energy Simulation Program. Energy and Buildings, 33(4), 319
Dai, Y. J., & Sumathy, K. (2002). Theoretical Study on a Cross-flow Direct Evaporative Cooler
Using Honeycomb Paper as Packing Material. Applied Thermal Engineering, 22(13), 1417
Davis, R. A., & Elberling, L. (2007). Laboratory Evaluation of the Breezair Icon Direct
Evaporative Cooler. In Pacific Gas and Electric Company.
Dodoo, A. (2011). Life Cycle Primary Energy Use and Carbon Emission of Residential
Buildings. Mid Sweden University, 115, 82.
Dowdy, J. A., & Karabash, N. S. (1987). Experimental Determination of Heat and Mass Transfer
Coefficients in Rigid Impregnated Cellulose Evaporative Media. ASHRAE Transactions,
93(2), 382395.
Dowdy, J. A., Reid, R. L., & Handy, E. T. (1986). Experimental Determination of Heat-and
Mass-transfer Coefficients in Aspen Pads. ASHRAE Transactions, 92(2), 6070.
Elmqvist, H., & Mattsson, S. E. (1997). Modelica the Next Generation Modeling Language
an International Design Effort. Proceedings of the 1st World Congress on System
Simulation, 15.
Fouda, A., & Melikyan, Z. (2011). A Simplified Model for Analysis of Heat and Mass Transfer
in a Direct Evaporative Cooler. Applied Thermal Engineering, 31(5), 932936.
Franco, A., Valera, D. L., & Peña, A. (2014). Energy Efficiency in Greenhouse Evaporative
Cooling Techniques: Cooling Boxes Versus Cellulose Pads. Energies, 7(3), 14271447.
Halasz, B. (1998). A General Mathematical Model of Evaporative Cooling Devices. Revue
Generale de Thermique, 37(4), 245255.
He, S., Gurgenci, H., Guan, Z., Huang, X., & Lucas, M. (2015). A review of wetted media with
potential application in the pre-cooling of natural draft dry cooling towers. Renewable and
Sustainable Energy Reviews, 44, 407422.
Holman, J. P. (1988). Thermodynamics. McGraw-Hill, New York.
IESVE. (2011). Version 6.2, Integrated Environmental Solutions Virtual Environment.
Incropera, F. P., DeWitt, D. P., Bergman, T. L., & Lavine, A. S. (1996). Fundamentals of heat
and mass transfer (Vol. 6). Wiley New York.
Jaber, S., & Ajib, S. (2011). Evaporative Cooling as an Efficient System in Mediterranean
Region. Applied Thermal Engineering, 31(1415), 25902596.
Jain, J. K., & Hindoliya, D. A. (2011). Experimental Performance of New Evaporative Cooling
Pad Materials. Sustainable Cities and Society, 1(4), 252256.
Jorissen, F., Boydens, W., & Helsen, L. (2018). Validated Air Handling Unit Model Using
Indirect Evaporative Cooling. Journal of Building Performance Simulation, 11(1), 4864.
Kachhwaha, S. S., & Prabhakar, S. (2010). Heat and Mass Transfer Study in a Direct
Evaporative Cooler. Journal of Scientific and Industrial Research, 69(9), 705710.
Kettleborough, C. F., & Hsieh, C. S. (1983). The Thermal Performance of the Wet Surface
Plastic Plate Heat Exchanger used as an Indirect Evaporative Cooler. Journal of Heat
Transfer, 105(2), 366373.
Applied Thermal Engineering, 113, 819.
Kowalski, P., & K
Industrial Building in Poland. Journal of Building Engineering, 32, 101555.
Liao, C. M., Singh, S., & Wang, T. Sen. (1998). Characterizing the Performance of Alternative
Evaporative Cooling Pad Media in Thermal Environmental Control Applications. Journal
of Environmental Science and Health, 33(7), 13911417.
Maclaine-Cross, I. L., & Banks, P. J. (1981). A General Theory of Wet Surface Heat Exchangers
and its Application to Regenerative Evaporative Cooling. Journal of Heat Transfer, 103(3),
Maurya, R., Shrivastava, N., & Shrivastava, V. (2014). Performance evaluation of alternative
evaporative cooling media. International Journal of Scientific & Engineering Research,
5(10), 676684.
Mehrfeld, P., Lauster, M., Huchtemann, K., & Müller, D. (2016). Multi-mode Model of an Air
Handling Unit for Thermal Demand Calculations in Modelica. Building Simulation and
Optimization, Third IBPSA - England Conference.
Min, Y., Chen, Y., Shi, W., & Yang, H. (2021). Applicability of Indirect Evaporative Cooler for
Energy Recovery in Hot and Humid Areas: Comparison with Heat Recovery Wheel.
Applied Energy, 287, 116607.
Plessis, G., Kaemmerlen, A., & Lindsay, A. (2014). BuildSysPro: a Modelica Library for
Modelling Buildings and Energy Systems. Proceedings of the 10th International Modelica
Conference, 96, 11611169.
Rawangkul, R., Khedari, J., Hirunlabh, J., & Zeghmati, B. (2008). Performance Analysis of a
New Sustainable Evaporative Cooling Pad made from Coconut Coir. International Journal
of Sustainable Engineering, 1(2), 117131.
Sajjad, U., Abbas, N., Hamid, K., Abbas, S., Hussain, I., Ammar, S. M., Sultan, M., Ali, H. M.,
Hussain, M., Rehman, T. ur, & Wang, C. C. (2021). A Review of Recent Advances in
Indirect Evaporative Cooling Technology. International Communications in Heat and Mass
Transfer, 122, 105140.
Saman, W., Bruno, F., & Tay, S. (2010). Technical Background Research on Evaporative Air-
conditioners and Feasibility of Rating their Energy Performance. In Institute for sustainable
systems and technologies.
Seeley International. (2015). The BreezairIcon Exq Series.
Sodha, M. S., & Somwanshi, A. (2012). Variation of Water Temperature along the Direction of
Flow: Effect on Performance of an Evaporative Cooler. Journal of Fundamentals of
Renewable Energy and Applications, 2, 16.
 & Wetter, M. (2009). Co-simulation of Innovative Integrated
HVAC Systems in Buildings. Journal of Building Performance Simulation, 2(3), 209230.
Vakiloroaya, V., Samali, B., Fakhar, A., & Pishghadam, K. (2014). A Review of Different
Strategies for HVAC Energy Saving. Energy Conversion and Management, 77, 738754.
Venkateswara Rao, V., & Datta, S. P. (2020). A Feasibility Assessment of Single to
Multi/Hybrid Evaporative Coolers for Building Air-conditioning Across Diverse Climates
in India. Applied Thermal Engineering, 168, 114813.
Wetter, M. (2013). Fan and Pump Model that has a Unique Solution for any Pressure Boundary
Condition and Control Signal. Proceedings of 13th Conference of International Building
Performance Simulation Association, 35053512.
Wetter, M., Zuo, W., Nouidui, T. S., & Pang, X. (2014). Modelica Buildings library. Journal of
Building Performance Simulation, 7(4), 253270.
Weubbles, D. J. (1994). The Role of Refrigerants in Climate Change. International Journal of
Refrigeration, 17(1), 717.
Winkelmann, F. C., Birdsall, B. E., Buhl, W. F., Ellington, K. L., Erdem, A. E., Hirsch, J. J., &
Gates, S. (1993). DOE-2 Supplement: Version 2.1 E. Lawrence Berkeley Lab,CA (United
Wu, J. M., Huang, X., & Zhang, H. (2009). Numerical Investigation on the Heat and Mass
Transfer in a Direct Evaporative Cooler. Applied Thermal Engineering, 29(1), 195201.
This section describes the commonly used method for calculating the humidity ratio  from dry
 and wet bulb temperature  of the process air (inlet). The  of the air is calculated
by assuming  . The relative humidity of the outlet air is calculated using:
󰇧 
󰇧 
󰇣 󰇡
where is the saturation vapor pressure at dry bulb and is the saturation vapor pressure at the
wet bulb,  is the actual vapor pressure and  is the atmospheric pressure. The dew point
temperature  is calculated by:
 
The partial vapor pressure  at  is calculated as,
 
Using and , the  is calculated using:
 
ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
This paper describes a simulation model to calculate thermal energy demand of air handling units (AHU) centrally installed in buildings with focus on laboratories. The model’s design gradually supports energy demand calculations of multiple buildings, e.g. on district level. The AHU is modelled in the open source, object-oriented modelling language Modelica®. The model uses particular operation modes while neglecting dynamic transitions as this reduces computational effort to allow simulations on district level. A comparison of simulation results to experimental results, gained with a test bed at the Institute for Energy Efficient Buildings and Indoor Climate, RWTH Aachen University gives insights into the model’s accuracy. The results justify using the model in district simulations, where reduced calculation efforts outweigh acceptable deviations. Moreover, this paper presents thermal demand simulations of a research site with 195 buildings and a comparison to monitoring data. The computed hourly heating power is satisfying compared to this measured data, e.g. in terms of the coefficient of determination with a value of R²=0.939.
Full-text available
In this paper different cooling pads have been studied. Three types of cooling pad made of a cellulose, aspen fiber, and coconut coir were comparatively studied. This study is performed in summer and based on weather conditions of Bhopal, India; maximum dry bulb temperature of air 41.2 ˚C and 26.1 ˚C wet bulb temperature. The relative humidity of 31.1 % is carried out from online psychrometric calculator. The primary air velocity considered varies between 0.5 m/s to 3.0 m/s and the performance of the cooling pads are analyzed based on the saturation efficiency, leaving air temperature, specific humidity, relative humidity, cooling capacity and water consumption. Graphs are plotted for variation in Saturation efficiency and cooling capacity for different materials of pad with air mass flow rate. Saturation efficiency of the cooling pads made of cellulose was in the ranges of 64.55 to 55.29 %, for aspen it was 80.99 to 68.86%, and for coconut coir 68.15 % to 50.79 % was observed. It is seen that the saturation efficiency decreases with increase in velocity of air and the cooling capacity increases with air velocity. It is also observed that the Aspen fiber has higher efficiency while coconut coir has lower and the water consumption rate increases with air mass flow rate.
The indirect evaporative cooler (IEC), used as a novel energy recovery component for central air-conditioning (AC) systems, can cool and dehumidify the fresh air by capturing the waste thermal energy of exhaust air. To facilitate its implementation in hot and humid areas, the applicability of the hybrid AC system integrated with IEC needs to be addressed. This study quantitatively evaluated the cooling and energy-saving potentials of an IEC for energy recovery and compared it to a traditional hybrid AC system with a heat recovery wheel (HRW). On-site performance measurements were conducted in a wet market located in Hong Kong, where two air-handling units were integrated with a newly-designed IEC prototype and a commercial HRW respectively. Simulation models of the two hybrid AC systems were established based on TRNSYS platform by incorporating the numerical model of IEC and HRW respectively. The field-measurement data was used to validate the component models, and further calibrate the system models. To compare the regional adaptability of the two systems, annual simulations were conducted among 8 selected cities in southern China. Results showed that the total cooling capacities of IEC and HRW are closely related to local ambient relative humidity. Compared with the baseline AC system, the AC + IEC provides an annual energy saving intensity of 64.2–73.4 MJ/m² for cities with hot and moderate humid climates, which is more competitive than the AC + HRW (45.5–51.8 MJ/m²). The annual energy saving ratios of the two types of hybrid systems range from 14.4% to 26.4%.
Indirect evaporative cooler with dew point cooling has great potential to replace the conventional mechanical vapor compression system in air conditioning industry. In this paper, a detailed review has been conducted on recent developments in the indirect evaporative cooling (IEC) systems as well as their associated design parameters and operating conditions for higher cooling effectiveness and cooling capacity. The current review also consolidates the design and performance of various indirect evaporative cooling systems (such as classical indirect evaporative cooler, regenerative, dew point cooler, and Maisotsenko cycle based cooler). Furthermore, integration of these indirect evaporative cooling systems with other cycles is elaborated. In addition, the thermal management potential of the indirect evaporative coolers in various applications is highlighted. It is found that major design parameters include system configuration, inlet airflow conditions, channel geometry, and evaporative material. It is also found that counter-flow arrangement with higher inlet air temperature, lower inlet air humidity, smaller channel height, smaller inlet air velocity, higher channel length, and higher working to intake air ratio, triangular shape of channel with fabric evaporative material yields higher cooling effectiveness and efficiency. This review indicates that the IEC systems could be a potential replacement for the conventional vapor compression cycle based cooling systems in buildings and other applications.
The use of evaporative cooling in industrial buildings is becoming increasingly popular nowadays. However, there is a lack of detailed and sufficient evaluation in the literature of simple evaporative cooling systems used in this type of buildings, particularly in the climate of Central and Eastern Europe. Therefore, this article evaluates evaporative cooling (EC) systems for an assumed industrial building located in Poland using the TRNSYS 17 software. The analysis of six systems located in four Polish cities (Koszalin, Lublin, Suwałki and Wrocław) was carried out. Simple EC systems using both direct and indirect evaporative cooling were assessed. The systems were evaluated in terms of the number of hours that each system does not reach the room air temperature setpoint (TTF), electricity and water consumption and the seasonal coefficient of performance (SCOP). The results have shown that the largest number of the TTF hours is for systems located in Lublin and the smallest in Koszalin. Depending on the analysed system, the number of the TTF hours is for Koszalin and Lublin from 178 h to 611 h and from 429 h to 1194 h, respectively. By generalizing the results, EC systems have the potential to save energy in the climate of Poland. Out of the analysed systems, the S5 system, which uses both direct and indirect evaporative cooling and only outside air for humidifying, has the greatest potential for use. It was also found that systems using outside air for humidifying instead of exhaust air reach higher energy efficiency and lower indoor air temperature more.
Air conditioning systems are responsible for increasing men's workefficiency as well for his comfort, mainly in the warm periods of the year.Currently, the most used system is the mechanical vapor compressionsystem. However, in many cases, evaporative cooling system can be aneconomical alternative to replace the conventional system, under severalconditions, or as a pre-cooler in the conventional systems. It leads to areduction in the operational cost, comparing with systems using onlymechanical refrigeration. Evaporative cooling operates using inducedprocesses of heat and mass transfer, where water and air are the workingfluids. It consists in water evaporation, induced by the passage of an airflow, thus decreasing the air temperature. This paper presents the basicprinciples of the evaporative cooling process for human thermal comfort,the principles of operation for the direct evaporative cooling system andthe mathematical development of the equations of thermal exchanges,allowing the determination of the effectiveness of saturation. It alsopresents some results of experimental tests in a direct evaporative coolerthat take place in the Air Conditioning Laboratory at the University ofTaubaté Mechanical Engineering Department, and the experimental resultsare used to determinate the convective heat transfer coefficient and tocompare with the mathematical model.
This paper presents the validated Modelica model of a Menerga™ Adsolair™ type 58 Air Handling Unit. The unit consists of two fans, four dampers, three air filters and further allows heat recovery (including free cooling), indirect evaporative cooling and active cooling. The first principles model uses manufacturer data sheet information and optionally also measurement data to calibrate some parameters for increased accuracy. Model validation is based on measurement data and results in an average normalized absolute error for the electrical power consumption of approximately 1%. A custom developed control strategy is compared with the measured performance of a proprietary controller. The model is available open source as part of the IDEAS Modelica library.
This paper presents a numerical model for a compact direct-contact cross-flow air/water heat exchanger where evaporating water cools down an air stream, and where an innovatively designed metallic direct evaporative pad enhances air-water interaction. The numerical model implements energy and mass conservation equations of humid air and water in a one-dimensional geometry by applying correlations for heat and mass transfer coefficients. The system of ordinary differential equations is solved by central-finite discretisation using Matlab. The effective hydraulic diameter is isolated as the only unknown model parameter, and is determined by a parameter estimation using experimental data available from a producer of such a direct evaporative pad. The numerical model is able to predict the air outlet temperature, with an maximal error of 1.33 % compared to experimental data for different inlet temperature and humidity values. Humid air properties inside and at the outlet of the direct evaporative pad, the pad effectiveness and the water consumption can be evaluated by the presented model. The use of the numerical model is demonstrated with examples analysing the impact on heat exchanger effectiveness of a changed geometry (design analysis) and of varying air inlet conditions for a given geometry (operational analysis).