Ijcsi 9 2 2 23 34 PDF

IJCSI International Journal of Computer Science Issues, Vol.
9, Issue 2, No 2, March 2012

ISSN (Online): 1694-0814
www.IJCSI.org 23
Computerized Simulation of Automotive Air-Conditioning

System: Development of Mathematical Model and Its Validation
Haslinda Mohamed Kamar1, Mohd Yusoff Senawi2 and Nazri Kamsah3
1
Department of Thermo-Fluid, Universiti Teknologi Malaysia
81310 UTM Skudai, Johor, Malaysia
2
3
Abstract weight added and the system operation both cause the fuel
A semi-empirical model for simulating thermal and energy consumption to increase. Additional fuel consumption by
performance of an automotive air-conditioning (AAC) system in the AAC system means less fuel economy and more
passenger vehicles has been developed. The model consists of two greenhouse gas emissions. Consequently, augmentation of
sections, namely empirical evaporator correlations and dynamic the AAC efficiency and evaluation of its performance are
load simulation. The correlations used consider sensible and latent
both important. The AAC is often exposed to extremely
heat transfer performance of the evaporator coil. The correlations
were obtained from the experimental data of actual air varying conditions, and substantial efforts are required to
conditioning system for a compact size passenger car. The evaluate its performance. The conditions include: inlet
sensible heat transfer correlation relates the evaporator air off dry- conditions which determine air temperatures entering the
bulb temperature to inlet air dry-bulb temperature, humidity ratio, evaporator and condenser; user choices which determine
evaporator air velocity, condenser inlet air dry-bulb temperature, the evaporator volumetric flow rate and ventilation mode;
condenser air velocity and compressor speed. The latent heat driving patterns which determine the compressor rotational
transfer correlation relates the coil air-off humidity ratio to the speed and the condenser face velocity; and cabin material,
same six independent variables. The dynamic load simulation occupancy, and weather conditions which determine
model was developed based on the z-transfer function method internal and external thermal loads [1].
with a one-minute time step. The cooling load calculations were
performed using heat gain weighting factors. Heat extraction rate
and cabin air dry-bulb temperature calculations were carried out The single largest auxiliary load on an automotive engine is
using air temperature weighting factors. The empirical evaporator caused by the air-conditioner compressor [2]. During peak
sensible and latent heat transfer correlations were embedded in the load it can draw up to 5 to 6 kW power from the vehicle’s
loads calculation program to enable the determination of engine, and this is equivalent to a vehicle being driven
evaporator inlet and outlet air conditions, the cabin air temperature down the road at 56 km/hr [3]. Nationally and globally, the
and relative humidity. Comparisons with road test data indicated additional fuel consumed due to air conditioner usage is
that the program was capable of predicting the performance of the substantial. A study by Rugh et al. [2] indicate air
automotive air-conditioning system with reasonable accuracy. conditioner usage reduces fuel economy by about 20% and
Keywords: Automotive air conditioning (AAC), passenger cabin, increases emissions of nitrogen oxides (NOx) by about 80%
thermal load model.
and carbon dioxide (CO) by about 70%, although the actual
numbers depend on the actual driving conditions.
1. Introduction
Due to the ascending trend in the rate emission of pollutant,
Automotive air-conditioning (AAC) for thermal comfort in the automotive industry has been compelled to accomplish
passenger cabins is now a thing of necessity rather than parallel development of both environmental awareness and
luxury, and cooling is especially needed when travelling in fuel economy by establishing innovative solutions to
summer or throughout the year in countries of hot and improve the quality of the urban environment and to reduce
humid climate. The cost of installation of this air- the rate of fuel consumption to drive the compressor air
conditioning system is generally affordable but the extra conditioner and the climate control system in total. The
Copyright (c) 2012 International Journal of Computer Science Issues. All Rights Reserved.
IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 2, No 2, March 2012
www.IJCSI.org 24
solutions include the use of solar reflective glass [2, 4], 2. Dynamic Loads Analysis Of Passenger
parking cooling systems for non-idling air-conditioning [5], Cabin
demand capacity controlled compressor [6], air inlet
mixture control strategy [7, 8], automatic climate control According to ASHRAE [20] load analysis methods in an
systems [9,10] and reduction in weight and size of air enclosed space can be classified into two groups: steady-
conditioning units [11]. state methods and dynamic methods. Dynamic methods are
more attractive than the steady-state method as they are
The energy crisis can be avoided by using renewable able to model complex phenomena as well as being more
energy or it can be delayed via efficient utilisation of the flexible. The two widely used dynamic methods are the
existing non-renewable energy reserves. The latter heat balance method (HBM) and the weighting factor
approach is probably the only means of avoiding an method (WFM) [20]. The HBM is more accurate than
impending crisis before a practical and economically viable WFM. The simple WFM is chosen over the more rigorous
method of harnessing free energy is found since research HBM for loads calculation in a passenger vehicle for ease
efforts on renewable energy (such as solar and wind energy) of simulation with some small loss in accuracy which
are still on-going. In order to accomplish this goal, an explains why most of the popular building energy
efficient and comprehensive tool for rapid design and simulation programs adopt the weighing factor
prototyping of AAC system for analysis is needed. It also methodology.
must be noted that, increasing the activities of
experimentation and prototyping of any air-conditioning 2.1 The Weighting Factor Method
system, may increase the cost and price tag, especially for
automotive manufacturers of developing countries. In this The WFM is able to accelerate loads calculation by using a
respect, computer simulation can be used for rigorous simplified procedure [21]. It uses heat gain weighting
analysis of the cabin thermal load and air conditioner factors for converting heat gains to cooling loads, followed
interaction. Simulation is economical and not harmful to by the use of another set of weighting factors to determine
the environment, unlike actual testing of prototypes. heat extraction rates and space air temperatures. Heat gain
weighting factors represent parameters in z-transfer
Several integrated models for analyzing AAC performance functions that relate space cooling loads to instantaneous
have appeared in the open literature [2, 12, 13, 14, 15, 16, heat gains. In contrast, air temperature weighting factors
& 17]. Most of them used computational fluid dynamic represent parameters in a z-transfer function that relates
(CFD) and lumped-parameter approaches. Generally, the space air temperature to the net sensible heat removal rate.
CFD scheme is restricted to steady-state cases or very short
simulation periods [18] whereas lumped parameter The heat gain weighting factors are determined by solving
approach involves more analytical efforts and computer the heat balance equations for the space when a unit
time than those dedicated by the original problem [19]. As triangular pulse of space radiant heat gain is applied to the
a result, in this study, a unique approach is adopted in space. The calculation of these factors is subjected to some
developing a computerised simulation tool for analysing initial and boundary conditions. At each time step,
the thermal and energy performance of the AAC system in including the time of introduction, the calculated heat flow
passenger vehicles. The cabin compartment dynamic loads to the space air represents the net cooling load which is also
simulation is made on a minute-by-minute basis using the the heat gain weighting factor. The cooling load for each
z-transfer function methodology. The methodology has type of heat gain is calculated, from [20],
been widely used by building energy simulation programs Qk  v0 qk  v1qk 1  ...  w1Qk 1  w2Qk  2 ... (1)
on an hour-by-hour time interval, but it is rarely been used
in load simulation of passenger vehicles. By using one- where Qk is the actual cooling load at time k, qk is the
minute time interval, it is possible to capture the dynamics
of energy transfer and storage of the vehicles more actual heat gain at time k, and v0 , v1...w1 , w2 ,... are the
accurately. The newly developed tool also directly weighting factors.
considers the performance of the evaporator coil by linking
the thermal loads model and actual coil sensible and latent The air temperature weighting factors can be determined by
heat transfer performance. As a result, this semi-empirical solving the heat balance equations for the space when a
simulation tool is not only comprehensive but also capable unit triangular pulse in space air temperature is introduced
of modelling realistically the actual processes that occur in to the space to generate a sequence of net heat extraction
actual vehicles. rates. The weighting factors are then calculated using z-
transfer function, which relates the net heat extraction rate
and the deviation in room air temperature from a constant
www.IJCSI.org 25
value [21]. The deviation of the air temperature from the fluxes in to the outside surface and out of the inside surface,
reference value at time k, trk is calculated as follows [20], respectively. Generally, car envelopes such as walls, floor
and roof are essentially composite layers of materials with
different thermo-physical properties. For these multi-layer
1  CLk  ERk   p1 (CLk 1  ERk 1 )  
trk    (2) elements, TRFs are obtained using the explicit finite-
g0  p2 (CLk  2  ERk  2 )  ...  g1Tk 1  g 2Tk  2  ... difference method for solving the transient one-
dimensional Fourier heat conduction equation [23].
where CLk is the total cooling load at time k, ERk is the heat
extraction rate of the air-conditioning system at time k, and
 g0 , g1 ,..., p1, p2  are the air temperature weighting factors.
These factors contain information about the space,
particularly the thermal storage capacity of the enclosing
elements. Figure 1 shows the overview of the WFM in
calculating the space air dry-bulb temperature and heat
extraction rate.
Figure 1 Overview of the WFM in calculating the space air dry-bulb

temperature and heat extraction rate [29]. Figure 2 Schematic representation of thermal response factors [30].
2.2 Thermal Response Factors 2.3 Hourly Heat Gains
The application of weighting factor method requires the Hourly heat gain due to conduction through opaque
determination of heat gain weighting factors (v0, v1,) and materials is computed using the transmission matrix
air temperatures weighting factors (g0, g1, g2,…). In the method described by Buffington [24] as follows,
heat balance equations to obtain these factors, transient heat
conduction in solid materials is most easily calculated  t0   A B   ti 
using thermal response factors (TRFs). TRFs are the heat  q   C D   q 
fluxes at the inside and the outside surfaces of a plane slab  0   i (3)
[22] as shown in Figure 2. These heat fluxes result from a where,
unit triangular surface temperature pulse excitation.
Transient one-dimensional heat conduction problems can ti = inside surface temperature of the layer, oC.
be easily solved using the thermal response factor method t0 = outside surface temperature of the layer, oC.
[20]. The major difficulty is in determining the TRFs.
qi = inside surface heat flux, W/m2
Specifically, the Z and Y factors are the heat fluxes in to the q0 = outside surface heat flux, W/m2
inside surface and out of the outside surface of a slab,
respectively, when a unit triangular temperature pulse is This approach assumes that the outdoor air temperature and
applied at the inside surface. When the pulse is applied at solar radiation intensity vary in a periodic manner. Periodic
the outside surface, the X and Y factors represent heat
www.IJCSI.org 26
boundary condition is applied at the outer surface to obtain system in the passenger car. Experimental work was
the analytical solution in terms of a Fourier series conducted to generate the air-side performance data of the
consisting of various harmonics. When allowing for a evaporator coil for the air conditioner of a 1.6L Proton
combined radiative-convective inside and outside surface Wira Aeroback saloon car model.
air film resistances, extra layers of a wall or roof are
included in the calculation [25]. Thus, for a wall or a roof, 3.1 Experimental Setup
t0 represents the sol-air temperature as given by ASHRAE
The experimental setup consisted of three sections: (1)
[20], whereas ti represents the cabin air dry-bulb AAC circuit, (2) a closed air duct work for the evaporator
temperature. section, and (3) an open air duct work for condenser section
as shown in Figure 3.
Solar radiation heat gain through glass windows consists of
transmitted, reflected and absorbed components. The The AAC was a thermal expansion valve system type. The
internal heat gain due to occupants consists of both sensible ductworks were designed and constructed in accordance to
and latent heat components. Both heat gains are computed the British Standard (BS 5141: Part 1: 1975) for of duct-
using the method of ASHRAE [20]. mounted air cooling coils. The AAC system consists of the
original components of Denso air conditioning system used
in a compact size passenger vehicle (1.6L Proton Wira
3. Experimental Work Aeroback) using HFC134a as the refrigerant. The main
components of the AAC system included a swash-plate
Complete simulation of the thermal performance of air- compressor, an evaporator, a condenser, an expansion valve,
conditioning system for a passenger car requires evaporator a receiver drier, a sight glass and connection pipes.
coil performance data. This can be obtained either by
simulation, as shown by Huang [14] or from experimental The air duct work for the evaporator section is a closed
data. The former gives approximate solution while the loop rectangular circuit with an overall length of 5130 mm.
latter will give actual performance data of the evaporator The duct walls between the downstream and upstream of
coil. The latter approach was adopted in this work to the evaporator coil were thermally insulated using 25 mm
acquire a realistic model for simulating the thermal and thick rock wool insulation material. Air was forced through
Figure 3 Schematic diagram of the automotive air-conditioning (AAC) system.
energy performance of an automotive air-conditioning the evaporator coil using an upstream variable speed
www.IJCSI.org 27
centrifugal fan. To improve the air quality, air mixer, honey of velocity readings have been taken using a single air
comb air flow straightener and gauze screens were installed velocity transducer placed successively at each measuring
upstream of the evaporator coil. The internal loads were point as shown in Figure 4 for evaporator ductwork and
simulated using a 2 kW infra red electrical heaters. A steam Figure 5 for condenser ductwork. air dry bulb temperature
humidifier with the steam output of 4 kg/h was distribution across the cross section just upstream and
incorporated in the ducting system to inject water vapor down-stream of the evaporator coil.
into the air stream to provide humidity variation. The
heating and humidification sections were used to adjust the
air temperature and humidity at the evaporator coil entrance.
The air duct work for the condenser section was a 1 2 3 4 5

rectangular open tunnel with an overall length of 663 mm.
Air was forced through the condenser coil using an
upstream constant speed high pressure blower (10 HP). To
provide variation of face air velocity at the condenser, the
blower speed was controlled by using a reducer unit and L1=1.3cm
damper. To improve the air quality in the duct work,
L2=4.9cm
similar arrangements as described for the evaporator were
employed. A 2 kW infra red air heater was used to vary the L3=8.5cm
air dry bulb temperature at the inlet of the condenser coil. L4=12.1cm
3.2 Measuring Instruments L5=15.7cm
L=17cm
The air section of the air-conditioning system was fully
instrumented to record and control the operating parameters.
Figure 4 Position of the measuring points in the evaporator ductwork.
The data acquisition system consisted of a standard laptop
computer with PCMCIA slot, a data acquisition PCMCIA
card and signal conditioning rack. The interface card used
was a NI- PCMCIA-6024E multifunction DAQ card
featuring 16 analogue inputs at 200 kS/s, 12 bits resolution,
1 2 3 4 5 6 7
2 analogue outputs, 8 digital I/O lines and digital triggering. H=18c
The card was interfaced to the laptop computer running
DEWESOFT software. The signal conditioning rack was a
Dewe-Rack-8 type, with 8 slot mainframe for
DEWETRON modules, connection cables for 8 channels
L1=3.5cm
rack to NI- PCMCIA-6024E DAQ card, cables for RS-232
L2=12.8c
and RS-485 interface, and DAQP and PAD modules.
L3=22.1c
L4=31.5c
All DAQP and PAD modules were communicating with
L5=40.8c
DEWETRON system through RS-485 and RS-232
L6=50.1c
interfaces. The DAQP modules were used for velocity and
L7=59.5c
humidity measurement with an accuracy of ±0.05%. The
L=63c
PAD modules were used for RTD temperature sensors with
±0.1% accuracy. In this system, only seven modules were Figure 5 Position of the measuring points in the condenser ductwork.
actively used, specifically, two DAQP modules for velocity
measurement, two DAQP modules for humidity Another RTD temperature sensor was used to measure the
measurement and three PAD modules for RTD temperature air dry bulb temperature distribution across the condenser
sensors. cross section downstream of the condenser coil. All
temperature sensors were pt100 RTD 100 ohm model made
Two RTD temperature sensors were used to measure the of stainless steel with a length and diameter of 6 inches and
Two air velocity windowless transducers with an accuracy 6 mm, respectively. All temperature sensors were
of ±0.5% and probe length of 6 inches were used to connected to PVC air sampling tubes as recommended by
measure the air velocity distribution across the evaporator BS 5141: Part 1:1975 [20]. The RTD temperature sensors
ductwork cross section and the condenser ductwork cross has an accuracy of ±0.3oC has been calibrated by the
section. To obtain average air velocity in the ducts, a series
www.IJCSI.org 28
manufacturer and has been verified using standard

thermometer having an accuracy of ±0.2oC. Table 1 The ranges of the test variables applied in the study.
Two air velocity windowless transducers with an accuracy

of ±0.5% and probe length of 6 inches were used to
measure the air velocity distribution across the evaporator
ductwork cross section and the condenser ductwork cross
section. To obtain average air velocity in the ducts, a series
of velocity readings have been taken using a single air
velocity transducer placed successively at each measuring
point as shown in Figure 4 for evaporator ductwork and
Figure 5 for condenser ductwork.
Point sampling for each set of experimental work is time 3.4 Development of Empirical Correlations
consuming. To overcome this problem a correlation is
obtained between averaged air velocity and the centre point A multi-linear regression of the 273 experimental data has
air velocity. Least squares regression yields the correlations been made to fit two power equations. These power
for evaporator ductwork with a coefficient of determination, equations were obtained using the standard regression
r2 = 0.997 and for condenser ductwork with r2 = 0.977. The method described by Chapra and Canale [26]. The
subsequent experiments were conducted by taking one independent variables are the compressor speed, Nc (rpm),
reading at the centre of the ductwork only and the average the coil (evaporator) inlet air dry-bulb temperature, t1 (oC),
air velocities are obtained from the correlations. Two the coil air face velocity, Ve (m/s), the condenser air face
Honeywell humidity sensors model 4129R with measuring velocity, Vc (m/s), the condenser inlet air dry bulb
range of 0 to 100% RH with an output of 4-20mA and temperature, tci (oC), and the coil inlet air specific humidity,
accuracy of ±1% were used to measure the air relative w1 (kgv/kgda ).
humidity just upstream and downstream of the evaporator
coils.
3.3 Description of the Test Procedures

The experimental work was conducted at steady-state
conditions. The tests were run at different compressor
speeds with varying heat load at the coil. The compressor
speed was set at 1500, 2000, 2500 and 3000 rpm by
controlling the motor speed using a frequency inverter. At
each compressor speed, the air flowing through the
condenser was adjusted between 2.2 and 4.8 m/s using a
reducer unit and damper. At each air velocity through the
condenser, the air velocity of the evaporator coil was
maintained at blower speed 1, blower speed 2, blower
speed 3 and blower speed 4. At each blower speed, the air
dry-bulb temperature at the inlet of the evaporator was kept
at 25, 30 and 35oC by varying the energy input to the Figure 6 Predicted and measured coil air off dry-bulb temperatures.
electric heater upstream of the evaporator. At each
evaporator inlet air temperature, the air dry-bulb The first correlation relates the coil air off dry bulb
temperature at the inlet of the condenser was set at 32, 35 temperature, t2 (oC ) to the independent variables,
and 40oC, by varying the energy input to the electric heater
upstream of the condenser. To stabilize the air-conditioning t2  2.67104( N c ) 0.2155 (tei )0.4472 (Ve )0.4319 (Vc ) 0.0882
system, the air conditioner was run 30 minutes prior to each
test. Table 1 shows the ranges of the test variables applied ...(tci )0.4799 ( w1 )0.4623
(4)
in the study.
The second correlation relates the coil air off specific
humidity, w2 (kg/kg) to the independent variables and as
follows,
www.IJCSI.org 29
w2  0.00206( N c ) 0.1590 (tei ) 0.4289 (Ve ) 0.5103 ...

(Vc ) 0.0977 (tci )0.4438 ( w1 )0.4784
(5)
Figure 6 shows the predicted coil air off dry bulb

temperature of equation (4) compared to the measured
values. The degree of agreement is very good with the
coefficient of determination, r2 of 0.94. A plot of predicted
coil air off specific humidity of equation (5) against
measured data also indicates good agreement with r2 = 0.92,
as shown in Figure 7.
Figure 7 Predicted and measured coil air off specific humidity.

Figure 8 Flow chart for the CARSIM program.
4. Car Simulation (CARSIM) Program

Figure 8 shows the flow chart for the program. The input
data are submitted in three files, each with different types
of data. The first file contains the thermo-physical data of
the car, the second file contains the custom z-transfer
function weighting factors for the car, and the last file
contains the twenty-four hourly weather data.
A separate system program named WFAC was used to

specifically generate two sets of weighting factors: heat
gain weighting factors (v0, v1, …) and air temperature
weighting factors (g0, g1, ..) that are required by the Figure 9 Simplified schematic diagram of car air conditioning system.
CARSIM program. The weighting factors were calculated
for a one minute simulation time interval. Heat gain weighting factors are then used to convert the
heat gains to cooling loads using the heat gain weighting
The weather data submitted to CARSIM consisted of factors. Solar heat gains are calculated using the method
typical days of actual weather data for a twenty-four hour described by ASHRAE [19], followed by conversion to
period. The outdoor air data are the dry bulb temperature, cooling loads using the weighting factors obtained from
specific humidity, and direct and diffuse solar radiation WFAC. The total cooling load profile is then used to
intensity on the horizontal plane. Transient heat conduction calculate the cabin air dry bulb temperature and specific
through roof, walls and floor are calculated using the humidity when the air conditioner is functioning. When the
transmission matrix method. air conditioner is not active, only cabin air dry bulb
temperature is calculated.
www.IJCSI.org 30
for cabin air temperature is obtained by simply setting

Figure 9 shows the simplified schematic diagram of the air ERk in equation (6) to zero to give
conditioning system of a passenger vehicle. During the
CLk  TRC   j 1 g j (TRC  t3,k  j )
period the air conditioner is active, the seven governing 
equations with seven unknowns are [27]: t3,k 

 g0 (13)
ERk  CLk   g j (TRC  t3,k  j ) W (6)
j 0 Constant room temperature (TRC) is the one-day averaged
 cabin air temperature. Initially, a value is assumed for TRC,
ERk  1.23V (t3, k  t2, k ) W (7)
and it is updated after each one-day cycle of calculation.
Convergence is achieved when
t2,k  a0 N ca1t1,ak2Vea 3Vca 4t0,a 5k w1,ak6 o
C (8)
w2,k  b0 N cb1t1,b k2Veb 3Vcb 4t0,b 5k w1,b 6k kg/kg TRCold  TRCnew

(9)   0.001
TRCold (14)
QL , k  3010 V ( w3, k  w2, k ) W (10)
For a typical simulation, convergence is attained after about
t1,k  ( XOA)t0,k  (1  XOA)t3,k o
C (11) five to six cycles of calculation, with 1440 minutes
calculation per cycle. With air conditioning turned on, the
w1,k  ( XOA) w0, k  (1  XOA) w3,k kg/kg (12) coil sensible load, QC,S and the coil latent load, QC,L are
calculated as follows:
where,
QC , S  1.23V (t1,k  t2,k ) W (15)
t0,k = ambient air dry bulb temperature at time k, oC
t1,k = coil inlet air dry bulb temperature at time k, C o QC , L  3010 V ( w1, k  w2,k ) W (16)
t2,k = coil outlet air dry bulb temperature at time k, oC 4.1 Validation of CARSIM Program
o
t3,k = cabin air dry bulb temperature at time k, C The test vehicle used was a four-door 1.6L Proton Wira
w0,k = ambient air specific humidity at time k, kg/kg Aeroback. The age of the vehicle was about four years. The
vehicle and its air conditioning system have gone through
w1,k = coil inlet air specific humidity at time k, kg/kg standard maintenance routines, and were in good operating
w2,k = coil outlet air specific humidity at time k, kg/kg condition.
w3,k = cabin air specific humidity at time k, kg/kg 4.1.2 Test Setup
ERk = heat extraction rate of cabin at time k, W The experimental road test was carried out from October to
CLk = total sensible cooling load of cabin at time k, W November 2008. During each test, there were two to three
occupants in the vehicle cabin, including the driver. Each
QL ,k = cabin latent load (from occupant) at time k, W trip began in Skudai, Johor on city streets, progressed north
on PLUS highway to Pagoh, Johor, regressed to south on
XOA = fraction of outside air the same highway and ended in Skudai, Johor. Total
Nc = compressor speed, rpm/1000 mileage for each test was approximately 300 kilometres.
Ve = air velocity at coil outlet, m/s The objective of the test was to measure the cabin air dry-
bulb temperature of the test vehicle. The tests were
Vc = condenser air face velocity, m/s
performed at two different vehicle speeds, 90 km/h and
a0, a1, a2, a3, a4, a5, a6 = coil sensible heat correlation 105 km/h respectively. These speeds were chosen so that
coefficients (see Eq. (4)) the compressor speeds of the air conditioning system were
similar to the speeds during the laboratory test. For each
b0, b1, b2, b3, b4, b5, b6 = coil latent heat correlation
vehicle speed, the air blower was regulated at blower speed
coefficients (see Eq. (5))
1 and blower speed 4. For each condition, the temperature
The seven unknowns are t1,k , t2,k , t3,k , w1,k , w2,k , w3,k was recorded at every two to five minutes for 10 to 20
minutes duration depending on the weather and traffic
and ERk . Equations (6) through (12) are solved by condition during the test.
iteration. When air conditioning is not active, the solution
www.IJCSI.org 31
A total of eight type-T thermocouples were used to measure model of the test car. The air-conditioned floor area and
cabin temperature distribution as shown in Figure 10. All passenger cabin volume were 2.33 m2 and 2.54 m3,
thermocouples were connected to a data acquisition system respectively. For zoning purposes, it was modelled as one
through extension wire to record thermocouples readings as thermal zone.
shown in Figure 11.
Figure 10 Locations for data measurement (top view).
The data acquisition system consisted of a standard

notebook with USB TC-08 furnished with a PicoLog
software. The thermocouples were calibrated using
standard thermometer of accuracy 0.2 C . The cabin
o
supply airflow velocity was measured at frontage of the air

vents by using a Testo 450 vane anemometer. The vane Figure 12 Typical simplified model of the test car.
anemometer measuring ranges were between 4 to 40 m/s,
with an accuracy of 0.5%. These test data were used as a The CARSIM program used the typical weather data for
basis for comparison with the predictions by CARSIM Singapore for the November– December months. The
program, for the same vehicle type. weather data was a ten-year averaged actual weather data
developed by Moller and Wooldridge [28] for the computer
program BUNYIP.
The CARSIM program predictions of the air dry-bulb

temperature variation inside the car cabin are shown in
Figures 13 through 16. Figure 13 shows the averaged air
dry-bulb temperature variation for the vehicle speed of 105
km/h, air-conditioning (A/C) fan blower at level 4 and
three occupants.
Figure 14 shows the averaged air dry-bulb temperature

variation for the vehicle speed of 105 km/h, A/C fan blower
at level 1 and three occupants. Figure 15 shows the
averaged air dry-bulb temperature variation for the vehicle
speed of 90 km/h, A/C fan blower at level 4 and three
occupants. Finally, Figure 16 shows the averaged air dry-
bulb temperature variation for the vehicle speed of 90 km/h,
A/C fan blower at level 1 with three occupants.
Figure 11 Experiment apparatus: (A) USB TC-08 is connected to a laptop; (B)
Temperature reading displayed on line; (C) USB TC-08; (D) Thermocouples
arrangement.
4.2 Comparison Between Simulation and Test Data

The CARSIM program requires thermo-physical data of the
test car as input data. Figure 12 shows a typical simplified
www.IJCSI.org 32
Figure 13 Variation of car cabin air temperature for vehicle speed of 105
km/h, blower 4 and three occupants. Figure 15 Variation of car cabin air temperature, t3 for vehicle speed of 90
km/h, blower 4 and three occupants.
Figure 14 Variation of car cabin air temperature, t3 for vehicle speed of

105 km/h, blower 1 and three occupants.
Figure 16 Variation of car cabin air temperature, t3 for vehicle speed of 90
As seen from Figure 13, the predicted temperatures are in a km/h, blower 1 and two occupants.
good agreement with the test data with a deviation of 0.5ºC
in average. Overall, the CARSIM program estimates of
zone temperature are slightly lower than those obtained
5. Conclusion
from the test data. The predicted data is about 0.25% to A semi-empirical computerized simulation program
4.05% lower than the experimental data. Figure 14 shows (CARSIM) for simulating the thermal and energy
that the deviations between predicted and measured values performance of an automotive air-conditioning system has
are between 0.86% and 3.32%. Figure 15 shows that the been developed based on the z-transfer function method.
agreement between predicted and measured temperatures is The empirical evaporator sensible and latent heat transfer
quite good, with a deviation of 0.57ºC on average. The correlations embedded in the loads calculation program
largest and lowest deviations are 4% and 0.31% enable the determination of evaporator inlet and outlet air
respectively. Overall, the predicted values of zone conditions and the cabin air conditions. The program was
temperature are slightly higher than those obtained from validated using data obtained from actual road test for
the test data. various operating conditions. The results obtained from the
CARSIM program were found to agree very well with
Figure 16 shows that the agreement between the predicted those obtained from actual road test, with a discrepancy of
and measured temperatures is quite good, with a deviation 2 to 4%.
of 0.29ºC on average with maximum and minimum errors
of about 2.69% and 0.2% respectively. It is evident that the
agreement between predicted and measured car cabin Acknowledgement
temperatures is good.
The authors would like to acknowledge the support from
Universiti Teknologi Malaysia and funding provided by
Ministry of Science, Technology and Innovation, Malaysia
www.IJCSI.org 33
and Research Management Centre, Universiti Teknologi [11] Jin, J., Chen, J. and Chen, Z., Development and
Malaysia through IRPA grant 74062. Validation of a Microchannel Evaporator Model for a
CO2 Air-conditioning System. Applied Thermal
Engineering. Development of Compact Heat
References Exchangers for CO2 Air-conditioning Systems.
Applied Thermal Engineering (2010), doi: 10.1016/
[1] Amr Gado, Yunho Hwang and Reinhard Radermacher, j.applthermaleng.2010.06.019.
Dynamic Behavior of Mobile Air-Conditioning [12] Kohler, J., Kuhn, B. and Beer, H., Numerical
Systems. ProQuest Science Journals; HVAC&R Calculation of The Distribution of Temperature And
Research 14 (2) (2008) 307 – 312. Heat Flux In Buses Under The Influence of The
[2] Rugh, J. P. and Hendricks, T. J., Effect of Solar Vehicle Air-Conditioning System. ASHRAE
Reflective Glazing on Ford Explorer Climate Control, Transactions. 96 (Part 1) (1990) 432 – 446.
Fuel Economy, and Emissions. Society of Automotive [13] Currle, J., Harper, M., Ross and F., Heid, T.,
Engineers (SAE) Technical Paper Series. Paper No. Evaluation of HVAC System of Passenger Cars and
2001-01-3077. Prediction of The Microclimate in The Passenger
[3] Johnson, H. V., Fuel Used for Vehicle Air Compartment by Application of Numerical Flow
Conditioning: A State-by-State Thermal Comfort- Analysis. Society of Automotive Engineers (SAE)
Based Approach. Society of Automotive Engineers Technical Paper Series. (1997) Paper No. 971788.
(SAE) Technical Paper Series. Paper No. 2002-01- [14] Huang Chi-Chang Daniel, A Dynamic Computer
1957. Simulation Model For Automobile Passenger
[4] Farrington, R. B., Rugh, J. P. and Barber, G. D., Effect Compartment Climate Control and Evaluation. PhD
of Solar Reflective Glazing On Fuel Economy, Thesis Michigan Technological University (1998).
Tailpipe Emissions and Thermal Comfort. Society of [15] Khamsi, Y. and Petitjean, C., Dynamic Simulation of
Automotive Engineers (SAE) Technical Paper Series. Automotive Passenger Compartment And It’s Air
Paper No. 2000-01-2694. Conditioning System. Society of Automotive
[5] Kampf, H. and Schmadl, D., Parking Cooling Systems Engineers (SAE) Technical Paper Series. Paper No.
for Truck Cabins. Society of Automotive Engineers 2000-01-0982.
(SAE) Technical Paper Series. Paper No. 2001-01- [16] Ali Heydari and Saeed Jani, Entropy-Minimized
1728. Optimization of an Automotive Air-conditioning and
[6] Watanabe, Y. and Sekita, M., Saving Power by HVAC System. Society of Automotive Engineers
Demand Capacity Controlled Compressor. Society of (SAE) Technical Paper Series. Paper No. 2001-01-
Automotive Engineers (SAE) Technical Paper Series. 0592.
Paper No. 2002-01-0232. [17] Rugh, J., Integrated Numerical Modeling Process for
[7] Forrest, W. O. and Bhatti, M. S., Energy Efficient Evaluating Automobile Climate Control Systems.
Automotive Air Conditioning System. Society of Society of Automotive Engineers (SAE) Technical
Automotive Engineers (SAE) Technical Paper Series. Paper Series. (2002) Paper No. 02FCC-70.
Paper No. 2002-01-0229. [18] Hensen, J., Bartak, M. and Drkal, F., Modeling and
[8] M. Khamis Mansour, Md Nor Musa, Mat Nawi Wan Simulation of a Double-Skin Facade System.
Hassan and Khalid M. Saqr, Development of Novel ASHRAE Transactions. 108 (2) (2002) 1251-1259.
Control Strategy for Multiple Circuit, Roof Top Bus [19] Ingersoll, J. G., Thomas, G. K., Maxwell, L. M. and
Air Conditioning System in Hot Humid Country. Niemec, R. J., Automobile Passenger Compartment
Energy Conversion and Management. 44 (2008) 1455– Thermal Comfort Model – Part 1: Compartment Cool-
1468. Down / Warm-Up Calculation. Society of Automotive
[9] Michalek, D., Gehsat, C., Trapp, R. and Bertram, T., Engineers (SAE) Technical Paper Series. Paper
Hardware-in-the-loop-Simulation of a Vehicle Climate No. 920265 232 – 242.
Controller with a Combined HVAC and Passenger [20] ASHRAE 1993. Handbook of Fundamentals. Atlanta,
Compartment Model. Advanced Intelligent GA: American Society of Heating, Refrigerating and
Mechatronics; Proceedings, 2005 IEEE/ASME Air-Conditioning Engineers, Inc.
International Conference on Digital Object Identifier. [21] Mitalas, G. P., Comments on the Z-Transfer Function
(2005) 1065-1070. Method for Calculating Heat Transfer in Building.
[10] Farzaneh, Y. and Tootoonchi, A. A., Controlling ASHRAE Transactions. 84 (1) (1978) 667-674.
Automobile Thermal Comfort Using Optimized [22] Mitalas, G. P. ,Calculation of Transient Heat Flow
Fuzzy Controller. Applied Thermal Engineering. 28 Through Walls and Roofs. ASHRAE Transactions. 72
(2008) 1906–1917 (2) (1968) 182 – 188.
www.IJCSI.org 34
[23] Mohd Yusoff Senawi, Development of A Building

Energy Analysis Package (BEAP) And Its Application
To The Analysis of Cool Thermal Energy Storage
Systems. Ph.D Thesis, University Technology of
Malaysia (2000).
[24] Buffington, D. E., Heat Gain by conduction Through
Exterior Walls and Roofs – Transmission Matrix
Method. ASHRAE Transactions. 81 (2) (1975) 89 –
101.
[25] Mohd Yusoff Senawi, Software Development For
Building Energy Analysis. Master Thesis, University
Technology of Malaysia (1992).
[26] Chapra, S.C. and Canale, R.P., Numerical Methods for
Engineers, Mc Graw Hill, New York, USA 2006.
[27] Haslinda Mohamed Kamar, Computerised Simulation
of Automotive Air Conditioning System. Ph.D Thesis,
University Technology of Malaysia (2009).
[28] Moller, S. K. and Wooldridge, M. J., User’s Guide for
the Computer Program BUNYIP: Building Energy
Investigation Package (Version 2.0), Highett, Victoria,
Australia 1985.
[29] McQuiston, F. C. and Parker, J. D. (1994). Heating,
Ventilating, and Air Conditioning Analysis and Design.
4th ed. United States : John Wiley & Sons, Inc.
[30] Kimura, K., Scientific Basis of Air Conditioning.
London, Applied Science Publishers Ltd. (1977).
Haslinda Mohamed Kamar, is a member of ASHRAE. She

received her Bachelor’s Degree in Mechanical Engineering from
University of Glasgow, Scotland in 1993, Master and PhD in
Mechanical Engineering from University Teknologi Malaysia in 1997
and 2009, respectively. She is now a senior lecturer in the Faculty
of Mechanical Engineering, Universiti Teknologi Malaysia. Her
areas of interest are automotive air-conditioning system, thermal
comfort & energy efficiency in hot climates, indoor air quality (IAQ),
natural ventilation as passive cooling strategy in buildings and
computational fluid dynamics (CFD) modeling and simulation.
Mohd Yusoff Senawi, graduated with Bachelor of Mechanical

Engineering from The University of New South Wales in 1987.
Received Post Graduate Diploma in Computer Science, Master and
PhD in Mechanical Engineering from Universiti Teknologi Malaysia
(UTM) in 1989, 1993 and 2000, respectively. Currently a Senior
Lecturer at UTM. Research interests include cooling loads
calculation and energy analysis of air conditioning systems.
Nazri Kamsah is a member of ASHRAE. He received his

Bachelor’s Degree in Mechanical Engineering from University of
Sunderland, United Kingdom in 1983, Masters of Engineering
(Mechanical) from Universiti Teknologi Malaysia in 1988, and PhD
in Mechanical Engineering from University of New Hampshire,
Durham, USA in 2001. He is currently a senior lecturer in the
Faculty of Mechanical Engineering, Universiti Teknologi Malaysia.
His areas of interest include computational solid mechanics, finite
element modeling and simulation, thermal management in
microelectronics, thermal comfort and energy efficiency in buildings,
natural ventilation as passive cooling strategy for buildings, indoor
air quality (IAQ) and computational fluid dynamics (CFD) modeling
and simulation.

Ijcsi 9 2 2 23 34 PDF

Uploaded by

Copyright:

Available Formats

Ijcsi 9 2 2 23 34 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ijcsi 9 2 2 23 34 PDF

Uploaded by

Copyright:

Available Formats

IJCSI International Journal of Computer Science Issues, Vol.

9, Issue 2, No 2, March 2012

Computerized Simulation of Automotive Air-Conditioning

Figure 1 Overview of the WFM in calculating the space air dry-bulb

2.2 Thermal Response Factors 2.3 Hourly Heat Gains

Figure 3 Schematic diagram of the automotive air-conditioning (AAC) system.

The air duct work for the condenser section was a 1 2 3 4 5

3.2 Measuring Instruments L5=15.7cm

manufacturer and has been verified using standard

Two air velocity windowless transducers with an accuracy

3.3 Description of the Test Procedures

w2  0.00206( N c ) 0.1590 (tei ) 0.4289 (Ve ) 0.5103 ...

Figure 6 shows the predicted coil air off dry bulb

Figure 7 Predicted and measured coil air off specific humidity.

4. Car Simulation (CARSIM) Program

A separate system program named WFAC was used to

for cabin air temperature is obtained by simply setting

equations with seven unknowns are [27]: t3,k 

w2,k  b0 N cb1t1,b k2Veb 3Vcb 4t0,b 5k w1,b 6k kg/kg TRCold  TRCnew

Figure 10 Locations for data measurement (top view).

The data acquisition system consisted of a standard

supply airflow velocity was measured at frontage of the air

The CARSIM program predictions of the air dry-bulb

Figure 14 shows the averaged air dry-bulb temperature

4.2 Comparison Between Simulation and Test Data

Figure 14 Variation of car cabin air temperature, t3 for vehicle speed of

[23] Mohd Yusoff Senawi, Development of A Building

Haslinda Mohamed Kamar, is a member of ASHRAE. She

Mohd Yusoff Senawi, graduated with Bachelor of Mechanical

Nazri Kamsah is a member of ASHRAE. He received his

You might also like