Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

SAE TECHNICAL
PAPER SERIES 1999-01-0241

The Vehicle Engine Cooling System Simulation

Part 2 - Model Validation Using Transient Data
Oner Arici, John H. Johnson and Ajey J. Kulkarni
Michigan Technological Univ.

International Congress and Exposition

Detroit, Michigan
March 1-4, 1999

400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

The appearance of this ISSN code at the bottom of this page indicates SAE’s consent that copies of the
paper may be made for personal or internal use of specific clients. This consent is given on the condition,
however, that the copier pay a $7.00 per article copy fee through the Copyright Clearance Center, Inc.
Operations Center, 222 Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sec-
tions 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying such as
copying for general distribution, for advertising or promotional purposes, for creating new collective works,
or for resale.

SAE routinely stocks printed papers for a period of three years following date of publication. Direct your
orders to SAE Customer Sales and Satisfaction Department.

Quantity reprint rates can be obtained from the Customer Sales and Satisfaction Department.

To request permission to reprint a technical paper or permission to use copyrighted SAE publications in
other works, contact the SAE Publications Group.

All SAE papers, standards, and selected

books are abstracted and indexed in the
Global Mobility Database

No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written
permission of the publisher.

ISSN 0148-7191
Copyright 1999 Society of Automotive Engineers, Inc.

Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely
responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in
SAE Transactions. For permission to publish this paper in full or in part, contact the SAE Publications Group.

Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manuscript or a 300
word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE.

Printed in USA
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

1999-01-0241

The Vehicle Engine Cooling System Simulation

Part 2 - Model Validation Using Transient Data
Oner Arici, John H. Johnson and Ajey J. Kulkarni
Michigan Technological Univ.

Copyright © 1999 Society of Automotive Engineers, Inc.

ABSTRACT performance can not be maintained when overcooling

occurs. The performance of the cooling system for diesel
The Vehicle Engine Cooling System Simulation (VECSS) engines is being studied to a much greater extent and
computer code has been developed at the Michigan depth than ever before.
Technological University to simulate the thermal
An optimum system should offer an increase in coolant,
response of a cooling system for an on-highway heavy
oil and engine metal temperatures during cold start or
duty diesel powered truck under steady and transient
low ambient temperatures and light route load conditions.
operation. In Part 1 of this research, the code develop-
It should also offer sufficient cooling of the oil and the
ment and verification has been presented. The revised
engine metal temperatures, under high ambient tempera-
and enhanced VECSS (version 8.1) software is capable
tures and slow vehicle speeds under full load conditions.
of simulating in real-time a Freightliner FLD 120 truck
with a Detroit Diesel Series 60 engine, Behr McCord radi- With increasing complexity of the diesel engine cooling
ator, Allied signal / Garrett Automotive charge air cooler system brought about by various cooling system compo-
and turbocharger, Kysor DST variable speed fan clutch, nent design changes as well as system configuration
DDC oil and coolant thermostat. Other cooling system changes, a need was felt to develop a simulation program
components were run and compared with experimental which would enable the designer to study individual com-
data provided by Kysor Cooling Systems. The experi- ponents and its effect on the overall system. This simula-
mental data were collected using the Detroit Diesel Elec- tion program should also have the following features:
tronic Control’s (DDEC) Electronic Control Module (ECM)
• Have a systematic structure such that individual com-
and the Hewlett Packard (HP) data acquisition system.
ponent models can be added to an existing system or
The engine speed, fuel flow rate and vehicle speed that
modified to conduct parametric studies.
describe the operating condition of the truck are the main
time varying input variables to the code for simulation. • Identify key parameters affecting system perfor-
Ambient temperature and pressure, relative humidity, and mance.
wind velocity are also needed as input. The enhanced • Help to minimize the hardware tests by establishing
VECSS results were validated by comparing the pre- the most important variables to be measured when
dicted top tank coolant temperature, oil sump tempera- experimentally studying the performance of the sys-
ture, and engine power with corresponding experimental tem being modeled for cost benefits and time effi-
data. The results were compared to three stationary and ciency.
three transient on-the-road test runs. In general, the pre- • Define an optimum system design including perfor-
dicted responses match well, within 0.5 to 3.5°C (1 to mance trends and control settings.
6°F) of the experimental data, both qualitatively and
• Simulate a wide range of operating conditions
quantitatively. Intake air temperature generally agreed
(steady state and transient), including the effect of
within 8°C (14°F) between the model and the experimen-
ambient temperature; use various worst case driving
tal data. A parametric study was also conducted with
cycles to design the system.
ambient temperature as the variable in order to study the
performance of the fan and the thermostat. • Be able to simulate various: (a) radiator designs
including area, (b) charge air cooler designs includ-
INTRODUCTION ing area, (c) oil cooler designs including area, (d)
engine power settings, (e) fan design and speeds.
With increased demands for improved fuel economy, Such a program would enable the designer to analyze
increasing engine life, and reducing emissions, while various configurations easily and efficiently for better
maintaining good engine performance, it is necessary to understanding of the operation of modern cooling sys-
have greater control of performance since optimal engine tems. The Vehicle Engine Cooling System Simulation

1
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

(VECSS) was first developed at Michigan Technological EXPERIMENTAL DATA COLLECTION

University in 1982 to simulate the thermal response of an
on-highway heavy-duty diesel powered truck under The experimental data used for validation of the VECSS
steady state and transient operation. program were collected by Kysor Cooling Systems with
The review and up date of the VECSS model contained the help from Detroit Diesel Corporation. The data were
in version 8.1 of the software is described in the Part I recorded by Kysor using the Hewlett Packard Data Acqui-
companion paper by Arici, et al [1]. sition System (HP-DAS) and Detroit Diesel Corporation
assisted Kysor in accessing the sensor signals in the
The goal of this paper is to present model validation for Detroit Diesel Electronic Controls - Electronic Control
stationary and transient ‘over-the-road’ tests as predicted Module (DDEC - ECM). Both the systems were used in
by the VECSS computer program. The heavy duty truck all the stationary and transient runs, so as to substantiate
configuration used was a Freightliner model the credibility of the data collected by each system.
FLD12064ST. Kysor of Cadillac and Detroit Diesel Cor-
poration collaborated in May-June 1996 to gather exten- The parameters recorded from the DDEC-ECM are; tem-
sive data that would be used in validation of the updated peratures - ambient, top tank coolant, bottom tank cool-
VECSS [2]. Kysor Cooling Systems used a Hewlett ant, sump oil, intake manifold, turbo air in, turbo air out,
Packard (HP) data acquisition system and DDC used the fuel in, exhaust air, fan blast air, air-off-radiator, pressures
- radiator static pressure, drivers side fan total head,
Detroit Diesel Electronic Controls - Electronic Control
intake air manifold pressure, driver side fan static, pas-
Module (DDEC-ECM) to record data on a Freightliner
senger side fan static, engine and fan speeds. In addition
FLD 120 truck with a 12.7 L Series 60 engine, a Behr
six temperatures in several quadrants of the radiator
McCord radiator, an Allied Signal charge-air cooler, and
were recorded.
Kysor Cooling Systems Fan 4035-37494.
The parameters recorded by the HP-DAS were: test cell
ENGINE AND COOLING SYSTEM boost fan speed, radiator top tank coolant temperature,
CONFIGURATION ambient air temperature (located in grill), air temperature
into turbocharger compressor, intake manifold tempera-
Table 1 shows the details of the components used for val- ture, oil sump temperature, thermostat housing coolant
idation in the VECSS system temperature, fuel temperature entering engine, engine
speed, and fan speed.
Shutters, sometimes placed between the radiator and the
fan, were not used for the stationary and transient tests. The road speed, fuel flow rate, and the engine speed
The activation temperatures for the coolant and oil ther- along with ambient temperature and wind speed are the
mostat were: input parameters required for the VECSS program.
These parameters were recorded at two-second intervals
• Coolant thermostat starts to open at 88°C (190°F)
by the DDEC-ECM data acquisition system. Hence, the
and is fully open at 96°C (205°F), with a hysteresis
DDEC-ECM recorded values were used as input parame-
of 1.7°C (3°F) while opening/closing.
ters for the VECSS.
• Oil thermostat starts to open at 104°C (220°F) and is
fully open at 114°C (237°F), with a hysteresis of The VECSS program was validated against four key
1.7°C (3°F) while opening/closing. parameters, the engine power, the coolant radiator top
tank temperature, the oil sump temperature and the
intake manifold air temperature. The data from the HP-
Table 1. Configuration for the VECSS validation
DAS were used extensively for validation.
Component Model No./ type Manufacturer
All of the VECSS model output data are discussed and
Engine Series 60, 12.7 L Detroit Diesel Corp.
analyzed for their significance relative to time varying
characteristics and the magnitudes relative to the compo-
Radiator 405-16366-001 Behr McCord nents being modeled and in turn, relative to the compo-
nent actions, i.e. thermostat operation, fan activation, etc.
Condenser BM 19757 Behr McCord

Charge air cooler 01-23132-00 Allied Signal TEST MATRIX FOR VALIDATION

Fan + Fan drive DST variable speed clutch Kysor of Cadillac Table 2 gives the details of the tests used for validation. A
set of three tests at engine speeds of 1500 rpm, and
Coolant thermostat - Detroit Diesel Corp.
1800 rpm at different percent engine loads, and ambient
Oil Cooler - Harrison Division-GMC temperatures were used for stationary testing. Test no. 1
and 2 were with no airflow from the test cell boost fan
Oil thermostat - Detroit Diesel Corp. while test no. 3 was simulated a 25 m.p.h. RAM air veloc-
ity. For transient testing, three tests at 55%, 75% and
Turbocharger - Garrett Automotive
100% of full load on the vehicle were used over three dif-
ferent vehicle routes.

2
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

STATIONARY EXPERIMENTAL AND VECSS when the oil sump temperature first reaches 104°C (220
PREDICTED RESULTS °F). The oil thermostat and the oil sump temperature take
about 1800 seconds to stabilize. The predicted airflow
For all the stationary tests, the vehicle was stationary. A across the radiator depends on the fan speed which in
boost fan was used in the test cell to pressurize air over turn is a function of the engine speed and the radiator top
the vehicle to simulate airflow to the engine compartment tank temperature. As the top tank coolant temperature
from either wind speed or the vehicle speed. Load was reaches 88°C (190°F), the fan speed starts increasing
applied by means of a dynamometer. The coolant used and later reaches a steady speed of 1140 rpm. Table 4
was water. The air-conditioning was off and hence there gives the summary comparison of the VECSS predicted
was no energy release from the condenser. The cab parameters with the experimentally measured values
heater was also non-functional in all the stationary tests. when the system is at steady state condition. All the sim-
Initially, all the temperatures in the VECSS program were ulation runs were conducted on the PC with a 300 MHz
at ambient temperature. The “warming up” of the VECSS Pentium-II processor.
was done by using the DDEC-ECM measured engine
speed and fuel flow rate as input data to accurately repre- Table 2. Test Matrix for VECSS Validation
sent this transient operation of the engine / cooling sys-
tem from ambient temperature to steady state TEST TYPE No. TEST DESCRIPTION

temperature. The ambient temperature for the VECSS STATIONARY:

program was maintained constant at the average mea- Tests run with: 1. 1500 RPM / 60% of full engine load* / No RAM Air
sured ambient temperature, even though the ambient Ambient:23 deg. C ( 73 deg. F)
* 100% water
temperature for the experimental tests varied 1.7°C 2. 1800 RPM / 80% of full engine load* / No RAM Air
* Kysor DST variable
(3°F). Table 3 gives the input parameter values to the speed clutch Ambient: 21 deg. C (70 deg.F)
VECSS program at steady state. * Without shutters 3. 1800 RPM / 75% of full engine load* / 25mph RAM
Air
A sample time traces of the results obtained for test no.1 Ambient:17 deg. C ( 62 deg.F)
are shown in Figure 1,2. ON THE ROAD:

In Figure 1 the first two time traces show the time traces Tests run with: 4. Cadillac to Spring Lake, Ambient: 18.9 deg.C (66
deg.F)
for the input data to the VECSS program. All the other * 50/50 Ethylene gly- 13609 kg (30,000 lbs) (55% of full load on vehicle)
time traces in that figure give the VECSS predicted col & water
5. Spring Lake to Plymouth, Ambient:1 deg.C ( 34
results. * Kysor DST variable deg.F)
speed clutch 18712 kg (41,250 lbs) (75% of full load on vehicle)
During the first 750 seconds, the fuel flow rate slowly
* Without shutters 6. Brighton to Spring Lake, Ambient: 11 deg.C (52
increases. This corresponds to the warm-up period for deg.F)
the engine test and the VECSS program. After 750 sec- 24949 kg (55,000 lbs) (100% of full load on vehi-
cle)
onds, the engine speed, and the fuel flow rate reach a
steady state value. The predicted engine power follows *% full load was calculated as the ratio of measured fuel flow to fuel flow from the
full load power curve for this engine
more or less the fuel flow rate curve. This predicted
engine power is the power obtained right at the engine
output shaft into the transmission and includes the power
used for the coolant pump, oil pump, and the fan. A repre- Table 3. Steady state input parameters for VECSS
sentative parameter for the engine operating conditions, program
i.e., the piston surface temperature, stabilizes after 1200
seconds. Time traces for intake air temperature and Input parameter Test No.1 Test No.2 Test No.3
mass flow rate are also shown in Figure 1. The top tank Engine speed (rpm) 1500 1800 1800
coolant temperature rises from the initial ambient temper-
Fuel flow rate (kg/min) 0.590 0.726 0.671
ature to a steady state value of 93°C (198°F). The time
trace for the percent open coolant thermostat shows that % Full load 60 80 75
it is first activated when the top tank coolant temperature
RAM air velocity (mph) 0 0 25
reaches 88°C (190°F). This results in flow of the coolant
through the radiator and hence a drop in the coolant tem- Average ambient temp. (deg.C) 23 21 17
perature. The ripples in top tank temperature time trace Total simulation real time (sec) 2800 2800 3750
are due to the opening and closing of the thermostat until
it reaches a steady state opening position. The oil ther-
mostat time trace involved modeling it with a 5% leak-
age, starting at time 0. The thermostat begins to activate

3
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

Figure 1. Time traces of select parameters for stationary Figure 2. Time traces of select parameters for stationary
Test No. 1 Test No. 1 (Continued)

4
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

As can be noted from Table 4 for the stationary tests, the TRANSIENT EXPERIMENTAL AND VECSS
oil sump temperature was predicted 0.5 to 1°C (1 to 2°F) PREDICTED RESULTS
lower than the experimental values. This can be attrib-
uted to the following reason. The oil sump temperature Both of the data systems were setup on the truck and the
was measured using only one thermocouple to represent truck was run over various routes in Michigan (refer Table
the entire oil sump, and might not be representative of 2). The tests were initiated in the starting cities and con-
the mass average temperature calculated by the VECSS ducted for the specified amount of time towards the desti-
since it could be close to an oil return from any of the nation city. Load on the truck was varied by changing the
auxiliaries (piston cooling, bearing, etc.). Furthermore, weight of load it carried. The cab-heater was off, just as
the differences between the actual and the predicted are in the case of the stationary tests. The coolant used was
only 0.5 to 1°C (1 to 2°F) low, which is well within the 50% water and 50% ethylene glycol (The VECSS pro-
experimental accuracy of the thermocouple used to mea- gram has been modeled to simulate either 50% water /
sure the oil sump temperature. 50% ethylene glycol or 100% water as the coolant [3]).
The predicted radiator top tank coolant temperature for Table 5 gives the input parameters for the transient tests.
each of the three tests agreed with the actual data to Most of these parameters do not remain constant with
within 0.5°C (1°F), which was within the experimental time, in contrast to the steady state tests, so either a
accuracy of the thermocouples used to measure the range or a time averaged value is given in this table. The
data. The predicted intake air temperature was within 8°C time averaged value is the average of about 2000 contin-
(14°F) and the power was within 4% of the experimen- uous data points. Figures 3,4 show sample time traces
tally recorded data, which also were within the accuracy for test no. 4 (Cadillac to Spring Lake) which are the time
of the experimental measurements. traces for key input and output variables for the VECSS
program.
Comparison between test no.1 and test no.2 from Tables
3,4 showed that with the ambient temperature being
Table 4. Comparison of predicted and experimental
close to the same values, the engine surface tempera-
stationary test parameters
tures for liner, head and piston increased with increase in
speed and percent full load of the engine. This also Test No.1 Test No.2 Test No.3
1500 RPM / 1800 RPM / 1800 RPM /
resulted in higher temperature rise for the coolant across 60% full 80% full 75% full
the engine. Due to this, the radiator top tank temperature Parameters
engine load / engine load engine load /
no RAM air / / no RAM air 25 m.p.h
was higher. The calculated and measured oil sump tem- 23 deg. C / 21 deg. C RAM air /
ambient ambient 17 deg. C
perature predictions for test no. 1 and 2 were also higher. ambient
The intake air temperature going into the engine was Pred. Expt. Pred. Expt. Pred. Expt.
higher at 80% load and 1800 rpm (test no. 2) as com- Engine Power (kW) 176 184 218 224 199 205
Exhaust temperature ( deg.C) 562 - 528 - 500 -
pared to 60% load and 1500 rpm (test no. 1).
Piston surface (deg.C) 313 - 328 - 312 -
The effects of change in ambient temperature and the Head surface ( deg.C) 253 - 261 - 249 -
Liner surface (deg.C) 155 - 161 - 155 -
RAM air can be studied by comparing test no. 3 and test
Coolant temperature rise across 3 - 3.5 - 3 -
no. 2 (Table 3,4). At the same engine speed and sightly engine (deg.C)
lower load, the engine liner, head and piston surface tem- Radiator top tank (deg. C) 93 93 94 94 91.5 91
peratures were lower at lower ambient temperatures. Radiator bottom tank (deg.C) 56 - 57 - 51 -
Thermostat flow through radiator (%) 9.9 - 10.5 - 9 -
This is because, for lower ambient temperatures (17°C Total time for which radiator thermostat 2163 - 2148 - 2827 -
(62°F) for test no. 3 and 21°C (70°F) for test no. 2) the was active (sec)
intake air temperature is lower for the same cooling air- Intake Manifold (deg. C) 38 46 50 55 44 49
Boost pressure (k Pa) 176 - 196 - 197 -
flow. The lower engine surface temperatures gave
Intake air mass (kg/min) 18.6 - 26 - 24 -
decreased coolant temperature rise across the engine A/F ratio 31.5 - 35 - 36 -
because of lower heat transfer to the coolant. Lower cool- Oil Sump (deg.C) 107 108 110 111 109 109
ant temperature results in a lower percentage opening of Oil thermostat factor open (%) 13.8 - 14.4 - 15 -
the thermostat (9 and 10.5%) and hence less coolant Actual air flow (ambient temp., m3/min) 125 - 152 - 141 -
Cooling air mass (kg/min) 102 - 126 - 120 -
flow through the radiator. Also, fan speed which is a func-
Fan speed (rpm) 1140 - 1480 - 1170 -
tion of the coolant temperature is lower for test no. 3. Time for which fan was fully engaged 0 - 0 - 0 -
Hence, although test no. 3 has RAM air of 25 m.p.h and (sec)
test no. 2 had no airflow, the cooling air mass remains Simulation real time (sec) - 2720 - 2640 - 3530
CPU time required (sec) 7800 - 7780 - 10,500 -
nearly the same. The oil sump temperature was also
slightly lower (109 and 110°C (228 and 230°F) from the
model) due to lower surface temperatures in the engine In Figure 3 the first three time traces show the engine
and higher heat rejection from the oil sump to the lower speed, vehicle velocity and the fuel flow rate which are
ambient air temperature. For the intake air temperature, input data to the VECSS program.
the lower ambient temperature results in lower intake air
temperature (44 and 50°C (111°F and 122°F)).

5
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

Figure 3. Time traces of select parameters for transient

Test No. 4 Figure 4. Time traces of select parameters for transient
Test No. 4 (Continued)

6
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

The predicted engine power follows the fuel flow rate The predicted top tank and oil sump temperatures agree
curve trend. One representative parameter for the engine closely with the experimental data both qualitatively and
operating conditions, the piston surface temperature, is quantitatively within 1°C (2°F) and 3.5°C (6°F) respec-
shown in Figure 3. Time traces for intake air temperature tively. Intake air temperature is predicted within 2.5°C
and mass flow rate are also shown. The ripples similar to (4.5°F).
the ones seen in the stationary test time traces (Figure 3)
Figures 6 gives the comparison for the oil sump tempera-
in the top tank temperature time trace are due to the
ture and Figure 7 gives the intake manifold air tempera-
opening and closing of the thermostat. The oil sump tem-
ture for the predicted and experimental data. The oil
perature slowly increases from the ambient temperature
temperature is initially at ambient temperature. The leak-
as the test time progresses. When it reaches 99°C
age through the oil thermostat is assumed at 5%. As the
(210°F), the oil thermostat begins to activate. The work-
oil temperature increases to 99°C (210°F), the oil thermo-
ing of the thermostat maintains the oil sump temperature
stat begins to open. The working of the oil thermostat
between 98 and 101°C (208 and 214°F). The predicted
maintains the average oil sump temperature at 97°C
airflow across the radiator follows the same trend as the
(207°F). The average temperature for the experimentally
vehicle velocity, since the fan speed which is a function of
measured data is 100°C (212°F). The trend in general
the coolant temperature and the engine speed, runs at a
lies within 3°C (5°F) qualitatively. The time averaged
nearly constant speed of 624 rpm as shown in Figure 4.
intake air temperature predicted by the VECSS program
is 27°C (81°F). The comparison of the trends for pre-
Table 5. Input parameters for the VECSS program for dicted and experimental data is shown in Figure 7. The
transient tests time averaged experimental data for intake air tempera-
Input parameter Test No.4 Test No.5 Test No.6
ture is 25.5°C (78°F), which is 1.5°C (3°F) less than the
predicted value.
Cadillac to Spring Lake Brighton to
Spring Lake to Plymouth Spring Lake
Comparison Plots for Test No 5 – The experimental
Engine speed (rpm) 1380 to 1450 1450 to 1500 1600 to 1650
test was run at an ambient temperature of 1°C (34°F)
Fuel flow rate range (kg/min) 0 to 1.2 0 to 1.2 0 to 1.2 with a load of 18712 kg (41,250 lbs) (75% of full load on
Average fuel flow (kg/min) 0.48 0.54 0.57
vehicle). The VECSS program took 11,400 secs of CPU
time to simulate 4200 secs of real time. For the top tank,
Cruising vehicle speed (m.p.h.) 57-60 60-65 60-65
oil sump, and the intake air temperatures, the time aver-
Ambient temperature (deg. C) 19 1 15.5 aged predicted and experimental data are 87 and 88, 97
% load on vehicle (24949.2 kg full 55 75 100
and 99.5, 11 and 13.5 deg. C(190 and 190, 207 and 211,
load) 51 and 56 deg. F) respectively.
Total simulation real time (sec) 5000 4000 4000
Comparison Plots for Test No 6 – A load of 24949 kg
(55,000 lbs) (100% full load on the vehicle) was carried
Comparison Plots for Test No 4 – For this test, the out at an ambient temperature of 15.5°C (60°F) for this
truck carried a load of 13609 kg (30,000 lbs), which was experimental test run. The VECSS program took 11,500
equal to 55% of the full load (24949 kg (55,000 lbs)) it secs of CPU time to simulate 4500 secs of real time. For
can carry on the vehicle. The VECSS program took the top tank, oil sump and the intake air temperatures,
14,250 secs of CPU time to simulate 5000 secs. of real the time averaged predicted and the experimental data
time. Figure 5 shows the comparison between the are 87 and 87, 97 and 100.5, and 22 and 24°C (189 and
VECSS predicted radiator top tank coolant temperature 189, 207 and 213, and 72 and 75°F) respectively.
and the experimental data. The time trace at the bottom
shows the percent opening of the coolant thermostat. As Table 6 gives the summary of the results obtained for the
the coolant temperature reaches 88°C (190°F), the ther- transient tests. This table presents predicted and experi-
mostat begins to open. The ripples in the coolant temper- mental data either in a range or as a time averaged value.
ature are due to initial opening and closing of the coolant With increase in the load on the vehicle, the average fuel
thermostat. The coolant temperature reaches an average flow rate increases, increasing the average power for test
temperature of 87°C (189°F). The small fluctuations in no. 4, test no. 5, and test no. 6 from 133 to 150 to 160
the temperature in the time range 1000-5000 seconds kW(181 to 205 to 218 hp). This conclusion can be made
show the continuous working of the thermostat. A similar because the load due to terrain effects (up and down hill
trend is seen in the experimental data. The average tem- effects) is probably averaged out at about the same value
perature for the experimental data is 88°C (190°F). for all the three tests although some of the differences in
power could be from terrain differences for the three runs.

7
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

In Table 6 although the load on the vehicle for test no. 5 is ing test no. 4 and 6 substantiate the above reasoning.
more than that of test no. 4, the piston, head, and the Test no. 6 is at 100% load on the vehicle with an ambient
liner temperatures are lower for test no. 5. This can be temperature of 15.5°C (60°F) for which the predicted
explained as follows. The ambient temperature is lower intake air temperature is 22°C (72°F). This is slightly
for test no. 5 resulting in lower intake air temperature. lower than the intake air temperature for test no. 4 and
Lower intake air temperatures even with increased power hence the engine surface temperatures are also slightly
resulted in lower engine surface temperatures. Compar- lower.

Table 6. Predicted and experimental parameters for transient tests

Test No.4 Test No.5 Test No.6

55% load 75% load 100% load

Parameters (13609 kg) on vehicle / (18712 kg) on vehicle / (24949 kg) on vehicle /
19 deg. C ambient 1 deg. C ambient 15.5 deg. C ambient

Pred. Expt. Pred. Expt. Pred. Expt.

Average power (kW) 133 - 150 - 160 -

Piston surface range (deg. C) 274-300 - 263-278 - 270-300 -

Head surface range (deg. C) 258-268 - 238-247 - 249-261 -

Liner surface range (deg. C) 150-166 - 146-155 - 153-168 -

Averaged coolant temperature rise across engine (deg. C) 3.2 - 2.7 - 3.3 -

Time averaged radiator top tank (deg. C) 87 88 87 88 87 87

Time averaged radiator bottom tank (deg. C) 51 - 31 - 57 -

Coolant thermostat flow (%) 8-10 - 3-5 - 8-10 -

Total time for which radiator thermostat was active (sec) 4570 - 3600 - 4190 -

Time averaged intake manifold (deg. C) 27 25.5 11 13.5 22 24

Intake air mass (kg/min) 14-17 - 19-22 - 16-20 -

Average A/F ratio 34 - 36.5 - 38 -

Average fuel economy (miles/gal) 6.3 - 6.1 - 6 -

Averaged oil sump (deg. C) 97 100 97 99.5 97 100.5

Actual cooling airflow range (at ambient temp. m3/min) 100-118 - 130-144 - 131-142 -

Fan speed (rpm) 624 - 660 - 659 -

Fan fully engaged time (sec) 0 - 0 - 0 -

Airflow mass range (kg/min) 83-97 - 127-141 - 105-114 -

Total real time (sec) - 5000 - 4200 - 4500

CPU time required (sec) 14,250 - 11,400 - 11,500 -

8
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

Figure 5. Comparison of top tank temperature and thermostat % open for Test No. 4

Figure 6. Comparison of sump temperature and thermostat % open for Test No. 4

9
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

Figure 7. Comparison of intake air temperaure for Test No. 4

Lower surface temperatures in the engine give lower tem- a high ambient temperature of 43°C (110°F) and at a low
perature rise in the coolant flowing across the engine ambient temperature of -29°C (-20°F). The driving cycle
because of lower heat transfer from the engine (refer fuel rate and engine speed used for these tests was
Table 6). At 75% load on the vehicle and lower ambient based on stationary test no. 2 which is at 21°C (70°F).
temperature (test no. 5), the radiator top tank remains the Test no. 2 used 0 m.p.h. ram air speed while these runs
same and the radiator bottom tank temperature is much used 58 m.p.h. to simulate the vehicle travel at this
lower. speed.
The oil sump temperature remains almost constant for all All the three tests were run with 50/50 ethylene glycol
the three tests. More than the load on the vehicle or heat and water as the coolant. The air-conditioning and the
rejection to the ambient temperature from the sump, the cab heater were in the off condition. No shutters were
oil thermostat plays a dominant role to maintain the tem- used for the tests. The fan used was variable speed
peratures constant. clutch.
For the tests number 4, 5, and 6, the fan operates at an Table 7 gives the input parameter values to the VECSS
average speed of 624, 660, and 659 rpm. respectively. program for the three runs.
Hence, the cooling airflow varies as the vehicle speed
Table 8 gives the summary of the results obtained from
changes. The vehicle speed and fan speed for test no. 4
the three tests.
are less than that for test no. 5, which results in lower air-
flow for test no. 4. But, for test no. 4 & 5, the average
vehicle speed and fan speed are the same. Since the Table 7. Input parameters for ambient temperature tests
ambient temperature is lower for test no. 5, the airflow is
Low Moderate High
higher as the air is more dense. The system resistance, Input parameter
Ambient Ambient Ambient
which is proportional to the air density decreases giving
higher volume flow rate. Engine speed (rpm) 1800 1800 1800

Fuel flow rate (kg/min) 0.725* 0.725 0.725

EFFECT OF AMBIENT TEMPERATURE
RAM air velocity (mph) 58 58 58
A parametric study of the cooling system under station-
ary conditions was conducted with ambient temperature Ambient temperature (deg. C) -29 21 43

as the variable. These computations help to evaluate the

Total simulation real time (sec) 2800 2800 2800
performance of the thermal control devices like the fan
and thermostat and to determine the effect of this impor- *80% of full load fuel flow
tant variable on the engine cooling system performance.
The tests were carried out using the VECSS program at

10
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

Discussion of Ambient Test Results – Observed trends

Table 8. Results of parametric study, the effect of from ambient tests are shown in Figures 8,9. At higher
ambient temperature ambient temperature, the intake air is less dense as com-
pared to the intake air at lower ambient temperature.
Low Moderate High
ambient ambient ambient Hence, for the same boost pressure, more mass of intake
air flows into the engine as ambient temperature
Parameters
-29 21 43 decreases. As the fuel flow rate is the same for all the
deg. C deg. C deg. C
ambient ambient ambient three tests, more intake air mass into the cylinder helps in
efficient combustion of all the fuel. This results in higher
Pred. Pred. Pred.
energy release and higher expansion work. Hence the
Engine power (kW) 225 214 211
power increases. The surface temperatures for the piston,
head and the liner inside the engine increase with
Exhaust temperature (deg. C) 424 525 554 increase in ambient temperature. This is because the
engine runs at a higher temperature as the intake air tem-
Piston surface (deg. C) 270 331 348
perature coming into the engine increases as ambient air
Head surface (deg. C) 239 291 304
temperature increases. Higher surface temperatures
results in higher heat transfer rate to the coolant which
Liner surface (deg. C) 159 173 184 gives higher rise in temperature for the coolant across the
engine. As the engine speed is the same for all the three
Coolant temperature rise across engine
(deg. C)
1.5 2.5 3.2 simulation runs, the coolant flow rate across the engine is
the same. But, since the surface temperatures are higher,
Radiator top tank coolant (deg. C) 88 91 92 more energy transfer takes place between the engine and
the coolant.
Radiator bottom tank coolant (deg. C) 0.2 62 80
The rate of rise in coolant temperature from ambient tem-
Coolant thermostat flow factor (%) 3.5* 12.3 31.2 perature to the thermostat activation temperature, 88°C
Average heat rejection out of the radia-
(190°F) is more in the case of 43°C (110°F) ambient tem-
61.7 70 74
tor (kW) perature. The radiator cannot reject enough heat to the air
flowing over it at the higher ambient temperature and
Total time for which radiator thermostat
was active (sec)
1945 2210 2360 hence the coolant temperature for the top tank radiator
rises above the fan activation temperature. This increases
Intake air temperature (deg. C) 2.8 49 64 the fan speed, in order to flow more air over the radiator.
Intake air mass flow rate (kg/min) 28.2 25.8 24.6 For the low ambient temperature test, as the coolant tem-
perature in the system slowly rises to the thermostat acti-
Boost pressure (kPa) 207 208 208 vation temperature, the thermostat activates. This results
A/F ratio 39 35 34
in coolant flow through the radiator. Since, the ambient
temperature is low, the coolant temperature drop in the
Oil sump temperature (deg. C) 107 107 108 radiator is high. The dropping coolant temperature corre-
sponds to closing of the thermostat. Hence, the thermo-
Oil thermostat percent open (%) 15.4 18.4 20.4 stat for low ambient test continuously opens and closes,
Fan speed (rpm) 824 1023 1363
with the top tank coolant temperature fluctuating between
86 to 89°C (187 to 192°F).
Fan engagement time (sec) 2800 2800 2800
Another observation is that the heat transfer to the air
Fan fully engaged time (sec) 0 0 0 flowing across the radiator increases with increasing
ambient temperature. The increased temperature of the
Actual airflow at ambient temp (m3/ 141 155 183 air at the exit of the radiator, causes the air volume flow
min)
rate to increase. This can be explained as follows. The
Cooling airflow mass (kg/min) 168 186 204 increase in temperature at the radiator outlet for higher
ambient temperature causes the density of the air to
* The coolant thermostat factor fluctuates between 0 and 7 percent. decrease. The system resistance, which is proportional to
the air density also decreases giving higher volumetric
flow rate.

11
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

Figure 8. Trends observed from ambient tests Figure 9. Trends observed from ambient tests
(continued)

12
 Downloaded from SAE International by Stony Brook Univ, Monday, August 13, 2018

The oil sump temperature is only slightly higher (108 REFERENCES

compared to 107°C) at higher ambient temperatures,
because of higher temperatures in the engine and hence 1. Arici, O., Johnson, J.H., Kulkarni, A., “The Vehicle Engine
more heat rejection to the engine and in turn to the oil Cooling System Simulation. Part 1 - Model Development”,
Society of Automotive Engineers, SAE Paper 1999-01-
(Table 8). The oil temperature is controlled by the thermo- 0240, 1999.
stat which is open 20.4% at 43°C ambient temperature 2. Ursini, V.J., “Ram Air as a Function of Wind Speed”, Kysor
as compared to 15.4% at -29°C. Engineering Test Report 85010, April 11, 1987.
3. Ursini, V.J., “A Computer Simulation Program for Evalua-
CONCLUSIONS tion of the Cooling System Performance of a Diesel Pow-
ered Truck”, Master’s Thesis, Michigan Technological
University, 1982.
The main objective of this paper was to use the VECSS
(version 8.1) software program with the addition of cool-
ing airflow model in order to validate the calculated
parameters as compared to the experimental data for
stationary and transient runs and to conduct a parametric
study of variables like the ambient temperature.
The conclusions are:
1.The stationary data predictions for the radiator top tank
temperature, and oil sump temperature were modeled to
be within 0.5 to 1°C (1 to 2°F), and the intake air temper-
ature to within 8°C (14°F) of the experimental data. The
engine power was predicted within 4% of the experimen-
tal data.
2.For the transient tests, the time averaged coolant, oil
sump, and the intake air temperatures were within 1°C
(2°F), 3.5°C (6°F), and 2.5°C (4.5°F) of the experimental
data respectively.
3.The parametric studies showed that at high ambient
temperature of 43°C (110°F), the thermostat is 31.2%
open and the fan airflow is the main variable affecting the
cooling of the coolant. At a low ambient temperature of
-29°C (-20°F), the radiator top tank coolant temperature
is at 88°C (190°F), therefore, the fan is at a speed of 824
rpm and the thermostat activity (continuously opens and
closes, with a maximum opening of 7%) is the predomi-
nant system component for maintaining coolant tempera-
tures within the design value of 88 to 96°C (190 to
205°F).

ACKNOWLEDGEMENTS

The authors acknowledge the continued financial and

technical support by Kysor Cooling Systems of the
Schwitzer Group. In particular, we want to express our
appreciation to Vince Ursini from Kysor for his efforts in
collecting the vehicle test data used for the validation,
and valuable comments on the thermostat and fan mod-
els. Detroit Diesel Corporation is gratefully acknowledged
for providing us with extensive test data on their Series
60 12.7 L engine and also other engine related informa-
tion including help in modeling the oil circuit. John Fisher,
Chris Prichard, and Dr. Dan Ducu of DDC gave us signif-
icant help and feedback on our modeling and validation
effort.

13