Development of a High-Precision Calorimeter for Measuring Power Loss in Electrical Machines

Abstract

This paper is concerned with the development of a high-precision 30-kW (40-hp) calorimeter and is specifically focused on how experimental errors resulting from calorimeter design and operating procedures are eliminated or mitigated. A complete calibration for the calorimetric system using a DC heater is conducted, and two induction motors rated at 5.5 and 30 kW (7 and 40 hp, respectively) are carefully tested both within and outside of the calorimeter. Loss segregation is in accordance with the IEEE 112 method B. Experimental results for the comparison of calorimetric and input-output methods clearly confirm the effectiveness of the calorimeter in terms of accurate power loss measurement. The accuracy of the calorimeter is approximately 5.6 W in the measurement of power loss of up to 4.5 kW with a resolution of 0.1 W.

Full text

Page 1

TeesRep: Teesside University's Research Repository http://tees.openrepository.com/tees/

This full text version, available on TeesRep, is the PDF (final version) of:

Cao, W., Bradley, K. J. and Ferrah, A. et al. (2009) 'Development of a high-precision
calorimeter for measuring power loss in electrical machines', IEEE Transactions on
Instrumentation and Measurement, 58 (3), pp.570-577.

For details regarding the final published version please click on the following DOI link:
http://dx.doi.org/10.1109/TIM.2008.2005083
When citing this source, please use the final published version as above.










Copyright © 2005 IEEE. This material is posted here with permission of the IEEE. Such permission of the IEEE
does not in any way imply IEEE endorsement of any of Teesside University's products or services. Internal or
personal use of this material is permitted. However, permission to reprint/republish this material for
advertising or promotional purposes or for creating new collective works for resale or redistribution must
be obtained from the IEEE by writing to pubs-permissions@ieee.org.

By choosing to view this document, you agree to all provisions of the copyright laws protecting it.









This document was downloaded from http://tees.openrepository.com/tees/handle/10149/94273
Please do not use this version for citation purposes.

All items in TeesRep are protected by copyright, with all rights reserved, unless otherwise indicated.

Page 2

570IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 58, NO. 3, MARCH 2009
Development of a High-Precision Calorimeter for
Measuring Power Loss in Electrical Machines
Wenping Cao, Member, IEEE, Keith J. Bradley, Associate Member, IEEE, and Azzeddine Ferrah, Member, IEEE
Abstract—This paper is concerned with the development of
a high-precision 30-kW (40-hp) calorimeter and is specifically
focused on how experimental errors resulting from calorimeter
design and operating procedures are eliminated or mitigated. A
complete calibration for the calorimetric system using a dc heater
is conducted, and two induction motors rated at 5.5 and 30 kW
(7 and 40 hp, respectively) are carefully tested both within and
outside of the calorimeter. Loss segregation is in accordance with
the IEEE 112 method B. Experimental results for the comparison
of calorimetric and input–output methods clearly confirm the
effectiveness of the calorimeter in terms of accurate power loss
measurement. The accuracy of the calorimeter is approximately
5.6 W in the measurement of power loss of up to 4.5 kW with a
resolution of 0.1 W.
Index Terms—Calorimetry, heat run test, heat transfer, induc-
tion motor, loss measurement, power loss.
I. INTRODUCTION
T
chine drive systems. Advances have been made in theoretical
analysis [1], [2] and numerical simulation [3], [4], but experi-
mental verification has proved to be challenging.
The conventional input–output method [5] for determining
losses and efficiency is effective, but the measurement of total
loss becomes increasingly difficult as machine efficiency or
rating increases. The back-to-back method [6] requires two
identical machines being mechanically coupled together and
forming as a generator–motor set, which may not readily be
available. As a result, the need for high-precision techniques of
measuring the machine losses is long standing.
Inprinciple,calorimetryisgearedtowardtheserequirements.
This technique provides a direct measurement of the total loss
in electrical machines, removing the necessity of subtracting
two similar magnitude input and output power losses. It suf-
fers neither from instrumentation phase shift nor time delay
errors [7]. Consequently, it is suitable for measuring electrical
HE DRIVE for energy efficiency has spurred a growing
number of studies on the efficiency improvement of ma-
Manuscript received August 13, 2007; revised December 12, 2007. First
published September 26, 2008; current version published February 9, 2009.
This work was supported in part by U.K. EPSRC under Grant GR/L05051
and in part by Brook Hansen and Eurotherm Drives Ltd. The Associate Editor
coordinating the review process for this paper was Dr. V. R. Singh.
W.CaoiswiththeSchoolofScienceandTechnology,UniversityofTeesside,
TS1 3BA, Middlesbrough, U.K. (e-mail: w.cao@tees.ac.uk).
K. J. Bradley is with the School of Electrical and Electronic Engineering,
University of Nottingham, NG7 2RD Nottingham, U.K.
A. Ferrah is with the Faculty of Engineering, Sohar University, Sohar 311,
Oman.
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TIM.2008.2005083
machines with any efficiency and power rating under any
excitationandsupplyconditions.Duetotheseinherentfeatures,
calorimetry can potentially achieve a very high level of mea-
surement accuracy and is thus gaining in popularity in testing
electrical machines and power electronic devices [8]–[19].
However, the calorimetric approach is not without experi-
mental problems. Its major difficulties are found in the mea-
surementoffluidproperties[9],preventionofheatleakage[18],
and maintenance of repeatable testing conditions [15]. In this
paper, the development of a 30-kW calorimeter is described,
and these drawbacks are tackled to obtain reliable results over
a relatively long operating time.
II. PRINCIPLE OF CALORIMETER
The underlying principle of calorimetry is to measure the
effect of losses in a stable heat source (e.g., an electrical
machine). This is achieved by placing the test object in an
enclosed thermally insulated container (i.e., calorimeter) and
arranging a cooling fluid to pass through the calorimeter to
exhaust the heat generated by the test object. This force-flowed
coolant fluid can be either gas or liquid.
When thermal equilibrium is attained, the heat loss extracted
from the enclosure is exactly balanced by the power loss
released by the test object. This power loss in steady state can
be expressed by
Ploss= ρ × v × (Cpout× Tout− Cpin× Tin)
where Plossis the power loss (in watts), ρ is the coolant density
(in kilograms per cubic meter), v is the volumetric flow rate (in
cubic meters per second), Cpinand Cpoutare the specific heats
at the entry and exit (in joules per kilogram degree Celsius),
and Tin and Tout are the temperatures at the entry and exit
(in degrees Celsius).
(1)
III. MEASUREMENT ERROR AND ANALYSES
Previously, a balanced-type calorimeter was developed by
the authors and described in [14]. This calorimeter was capable
of measuring power loss of up to 3 kW with an overall accuracy
of approximately 0.5%. Nonetheless, the experimental difficul-
ties encountered were associated with varying fluid properties,
heat leakage, and repeatability of the calorimeter. In addition,
the length of time that the tests took was considerable since the
motor test was followed by a heater run, with each taking at
least 4 h for a load test point. It is very likely that atmospheric
conditions fluctuate over this period of time, giving rise to
measurement error between the two consecutive tests. This is
0018-9456/$25.00 © 2008 IEEE
Authorized licensed use limited to: Teesside University. Downloaded on March 17,2010 at 06:27:14 EDT from IEEE Xplore. Restrictions apply.

Page 3

CAO et al.: HIGH-PRECISION CALORIMETER FOR MEASURING POWER LOSS IN ELECTRICAL MACHINES 571
an inherent flaw in the balanced calorimeter, which assumes
that the two conditions remain unchanged [9], [14].
The lessons for developing the 30-kW calorimeter have been
learned, and these difficulties have carefully been dealt with in
the improved design.
A. Measurement of Fluid Properties
Both calorimeters use air as the coolant because of its
convenience and low cost. There is apparently no need for a
heat exchanger inside the calorimeter to boost heat transfer,
as in a liquid-cooled calorimeter. In the event of fluid leakage,
air causes neither environmental hazards nor damage to the
laboratory equipment. Using a conditioning system, air can
readily be cooled down to a controlled level prior to its supply
to the calorimeter. This way, the machine is not operated
at excessively high temperatures, as is the case in some
calorimeters [16].
To reduce the duration of calorimetric tests, the new design
is of direct-type calorimeter [20]. As such a consequence, an
accurate determination of fluid properties becomes vital to
power calculation. However, this is far from easy to achieve,
particularly in the case of using gas (air) as the coolant fluid. In
practice, most flow-measuring devices for air measure volume
flow rather than mass flow. The density of air changes with
temperature and pressure. Atmospheric air also has a varying
water vapor content that significantly impacts the specific heat
of the mixture. Without a doubt, the use of high-precision
instruments can reduce the measurement errors in humidity,
pressure, and flow rate to a satisfactory level.
Clearly, the balanced calorimeter or the series calorimeter
[15] can remove the need for measuring fluid properties and
associated measurement errors. They need to employ a resistive
heater to replicate the load test and to operate along with the
machine in the sequence of time (balanced) or space (series).
Again, there is no guarantee that consecutive tests would be
conducted under exactly the same conditions over a relatively
long period of time.
B. Temperature Measurement
However, it is always a problem to measure the mean air
temperature with sufficient precision. Ideally, the temperature
of the fluid measured at the inlet and outlet ports should be
single valued. In the real world, there would be temperature
gradients present across the inlet and outlet ducts as well as
mass flow and velocity, as shown in Fig. 1. In a steady fully
developed forced-ventilated flow, the air temperature distribu-
tion across a circular duct exhibits a parabolic shape. Previous
studies have shown that the temperature gradient between the
duct center and the circumference could be up to 2◦C in a
150-mm diameter duct [14]. It is obvious that using a single
temperature sensor to measure the bulk temperature will un-
avoidably lead to measurement uncertainty that can easily
exceed the accuracy required. This error is often overlooked.
In practice, there are two approaches available for minimizing
the error. One is adopting multiple temperature sensors in a
grid to find a mean temperature [15], and the other is using a
Fig. 1.Temperature equalizer in a force-ventilated duct.
temperature equalizer to average the fluid temperature at the
temperature measurement point [21]. The latter is used in this
design and is illustrated in detail in Fig. 1.
The temperature equalizer used in this paper comprises an
array of 25-mm diameter cooper tubes welded together to form
a 200-mm diameter honeycomb. This is then connected to a sin-
gle centrally located PT 100 temperature sensor. The purpose
of this equalizer is twofold. First, it divides the main air flow
into many subflows so that the difference in temperature (and
flow rate) across the cross section of the duct is considerably
reduced. Second, by employing high-conductivity copper tubes
and platinum wires, the temperature distribution is equalized.
It is only after the measurement errors in the bulk air tem-
peratures at the inlet and outlet ducts have been significantly
reduced that it becomes possible for the coolant temperatures
to be controlled to 0.01◦C of their desired values by both sets
of heaters.
C. Heat Flux Leakage Paths
Three modes of heat transfer, i.e., convection, conduction,
and radiation, are involved in the thermal-dynamic performance
of a calorimeter, as shown in Fig. 2. In general, convection aids
intransmittingthepowerlossasheattothecoolantmediumand
does not cause any heat leakage as long as the box is perfectly
airtight. Conduction is undesired in this case since it creates
heat flux leakage paths through the box boundaries. Radiation
contributes little to heat leakage. However, the use of aluminum
sheets covering the inner surfaces of the box may promote heat
exchange in a slightly favorable way. Heat loss by radiation is
estimated to be negligible and appears in the linearity error of
the calorimeter.
Overall, heat leakage primarily occurs at conductive con-
nections, e.g., the walls, supporting legs, and machine shaft
linkage. The heat leakage through the walls can be minimized
by using high insulation materials and active wall temperature
control [11]. Heat leakage through connection ports can be
overcome by active port temperature control in parallel with
further insulation arrangements [14].
Authorized licensed use limited to: Teesside University. Downloaded on March 17,2010 at 06:27:14 EDT from IEEE Xplore. Restrictions apply.

Page 4

572IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 58, NO. 3, MARCH 2009
Fig. 2.Heat transfer in the calorimeter.
D. Repeatability
In addition to accuracy, repeatability is also a key perfor-
mance indicator of any measuring instrument. It is self-evident
to maintain a repeatable testing environment for calorimetric
tests to be successful. In absolute terms, repeatability involves
dismantling and reassembling the calorimeter, putting the same
motor back into the calorimeter, and repeating the same test.
This not only measures the calorimeter repeatability but also
that of the supply system, of the loading system, and of the test
motor loss. The problem is, therefore, split into parts.
Repeatable supply and load systems can reasonably be re-
alized, but repeatability of the power delivered by the test
machine itself is not guaranteed. Oil seals, the grease content
in the bearings, and the changes between tests to the residual
thrust between the two ball bearings that are commonly used to
support the rotors of small commercial machines all give rise to
variation of friction loss.
The ability to identify the power loss in the calorimeter is
largely dictated by the measurement noise. Although averaging
over long time periods can minimize the noise level on the
measurement, it would also lose instant information during a
test and would not remove the long-term drifts arising from
variations of ambient temperature, humidity, and pressure.
IV. DESIGN CONSIDERATIONS
The 30-kW calorimeter is designed to have a maximum
temperature rise of 20◦C at a constant mass flow rate of
0.2 kg/s of dry air. The purpose of fixing the mass flow rate
is to ensure that the temperature rise between inlet and outlet
is independent of the air’s pressure and humidity. This capacity
corresponds to a maximum power loss of 4.5 kW, representing
the power loss in a 30-kW motor with the lowest efficiency
of 85%. The calorimeter has internal dimensions of 1.3 m ×
0.8 m × 0.8 m (L × W × H), which are suitable for machine
frame sizes of up to motor 250 S. The loading system works
best for machines rated at between 0.75 and 30 kW.
The calorimeter has a low thermal conductivity across the
walls.Asandwichformof100-mm-thickexpandedpolystyrene
between two layers of 6-mm-thick aluminum sheet is employed
Fig. 3.
walls removed). (b) Internal view.
Experimental setup of the calorimeter. (a) External view (end and side
in the structure. Polystyrene is a reliable and cheap material for
thermal insulation. Aluminum can conveniently be cut, shaped,
and welded. It also has a high mechanical strength and thermal
conductivity. The use of aluminum sheets covering all the
surfaces inside and outside of the enclosure helps to achieve an
isothermal condition on the walls. An additional advantage of
selecting polystyrene and aluminum to construct the enclosure
is their light weight, which potentially reduces the thermal time
constant of the calorimeter.
Fig. 3 shows the experimental setup, and Fig. 4 illustrates
the schematic of the whole calorimetric system. Air is initially
cooled and then fed to a chamber containing the main circula-
tion fan. This feeds through the main heater to a plenum cham-
ber supplying a 4-m-long, 100-mm-diameter tube containing
the high-pressure drop, multiple pipe flow straighteners, and
volume-measuring turbine flowmeter. After flow measurement,
a second plenum chamber collects the air that is then fed
through a fine-tuned heater to the calorimeter.
The test machine is mounted on a steel bedplate supported
by four 50-mm-diameter, 2-mm-thick steel tubes. The tubes are
filled with insulation to prevent convection currents. Tempera-
ture sensors are installed in the tubes just inside and outside of
the calorimeter. A heating arrangement using external heaters
is designed to maintain no more than a 0.1◦C temperature
Authorized licensed use limited to: Teesside University. Downloaded on March 17,2010 at 06:27:14 EDT from IEEE Xplore. Restrictions apply.

Page 5

CAO et al.: HIGH-PRECISION CALORIMETER FOR MEASURING POWER LOSS IN ELECTRICAL MACHINES 573
Fig. 4. Schematic of the 30-kW calorimetric system.
gradient along the supporting legs inside and outside. The
drive shaft is totally enclosed, and the bearing outside the
calorimeter is insulated from the calorimeter wall. Similarly,
it is oil cooled/heated to keep the temperature gradient in the
shaft exit to be within 0.1◦C.
The outer walls of the calorimeter are fitted with a further
layer of insulation and temperature-controlled heaters to ensure
as close as possible an isothermal condition. The latter is
achieved by a simple “bang-bang”-type control with ±0.1◦C
hysteresis. The temperature over the heated wall is not uniform,
differing by as much as 0.5◦C from the mean, but a weighted
integration of the temperature with respect to area over each
wall gives a mean temperature difference of approximately
0.1◦C. The outer insulation reduces the power required by
the heaters and, therefore, the temperature gradients in the
interinsulation walls.
The thermal time constant of the calorimeter is estimated to
be 40 min. If the power loss is to be measured to an accuracy
of 0.2%, the operational time for a single test point should be
at least 6.2 times the thermal time constant, which is approxi-
mately 4 h. When a machine is installed inside the calorimeter
for testing, the thermal constant will be altered and the test
duration will be elongated, depending on the machine size.
V. SYSTEM CONTROL
Energy flow control and power calculation are realized on a
control computer via a LabView interface. A screen shot of the
control program is given in Fig. 5. The sensors situated in the
entry port of the calorimeter box measure the air’s humidity,
pressure, temperature, and volumetric flow rate and send these
to the control computer via RS 232 and IEEE 488 interfaces.
The LabView Virtual Instrument is employed to control a
14-bit data acquisition card (DAC) feeding the field controller
of a dc generator, which is driven by an induction motor. PID
control is implemented with digital filtering to enable the input
power to the calorimeter to be defined in the increment of 1 W.
Voltage attenuators and current shunts are calibrated to 0.01%
and 0.02%, respectively.
For a calorimetric system to work best, it is critical to
maintain a steady and repeatable testing environment over
the long test period. This condition includes stable inlet and
ambient temperatures, effective mass flow rate, supply and
loading systems, and measurement instrumentation.
A. Inlet Air Temperature
The inlet air temperature is controlled through two heaters.
The air coolant is initially cooled down to around 12◦C using
two refrigeration units and subsequently heated up to approxi-
mately 14.5◦C using a main heater (crude) in the first chamber.
Prior to flowing into the calorimeter box, the coolant is further
heated up to 15◦C using a fine-tuned heater in the second
chamber. The coolant temperature at the point of measurement
iscontrolledtobeconstantwithin0.01◦Cofthesettemperature
point (15◦C).
B. Air Flow Rate
The air intake fan used to maintain the air flow through the
calorimeter box is driven by a three-phase induction machine
fed by an inverter. The air volumetric flow rate is dynamically
Authorized licensed use limited to: Teesside University. Downloaded on March 17,2010 at 06:27:14 EDT from IEEE Xplore. Restrictions apply.