Homogeneidade Sirilanca.pdf

7/27/2019 Homogeneidade Sirilanca.pdf

1/6

Elektrotehniki vestnik 72(4): 189-194, 2005Electrotechnical Review: Ljubljana, Slovenija

Measurement Device and Procedure for ThermocoupleInhomogeneity Detection

Miha Hiti, Jovan Bojkovski, Valentin Batagelj and Janko DrnovekUniversity of Ljubljana, Faculty of Electrical Engineering,

Laboratory of Metrology and Quality, Traka 25, 1000 Ljubljana

E-pota: [email protected]

Abstract. The paper presents an automatic system for thermocouple inhomogeneity detection achieved by moving

the thermocouple through temperature gradients and measuring any Seebeck voltage variation. It is a flexible

portable system enabling accurate and reproducible measurements even for a great number of repetitions and long

measurement periods. Also, the workload on the user performing such measurements is reduced. The vital part ofthe system is a microcontroller controling the stepper motor for carriage movement. It is also used for

communication with a PC computer, where a custom-written program in LabVIEW serves as a user interface. The

mounting of the system is developed for vertical as well as horizontal thermometer movement. This allows the

system to take advantage of the existing laboratory equipment since it can easily be mounted on different baths and

furnaces. A brief description of procedures for thermocouple inhomogeneity detection is given. Measurement results

of thermal emf for inhomogeneity detection are presented for different heat sources such as water bath, oil bath andheat-pipe furnace.

Key words: thermocouples, uncertainty, temperature measurement

Naprava in postopek za ugotavljanje nehomogenosti termo*lenov

Povzetek. Nehomogenost termo*lenov je pri merjenju

temperature vir negotovosti, ki lahko mo*no vpliva na

merilni rezultat. V prispevku je predstavljen

avtomatiziran sistem za dolo*anje poloaja in velikostinehomogenosti s pomikanjem termo*lena skozi

temperaturno polje. Sistem omogo*a to*ne, ponovljive

meritve, tudi kadar potrebujemo veliko ponovitev in

dolge *ase meritev. Osnova sistema je mikrokrmilnik za

krmiljenje kora*nega motorja in za komunikacijo zosebnim ra*unalnikom. Uporabniki vmesnik in

komunikacijski vmesnik na osebnem ra*unalniku sta

bila razvita s programskim paketom LabVIEW. Sistem

je primeren tako za vodoravne kot za navpi*ne pomikein omogo*a uporabo na obstoje*ih pe*eh in kopelih. V

prispevku so opisane razli*ne metode za ugotavljanjenehomogenosti termo*lenov in predstavljeni rezultati

merjenja termonapetosti za ugotavljanje nehomogenosti

pri vodni kopeli, oljni kopeli in pe*i s toplotno cevjo.

Klju*ne besede: termo*leni, negotovost, merjenje

temperature

1 Introduction

Thermocouple sensors are most widely used

temperature sensors in industrial applications. They

allow robust, simple and low-cost measurement of

temperature in a wide range, from below -200 C toover +2000 C. However, a serious error can be present

in thermocouple measurements and remain unnoticed

even when they are calibrated. During use,

thermoelectric properties of thermocouples can degrade,

especially at higher temperatures or as a result ofmechanical strain. Such degradation can be perceived as

drift, although it usually does not affect the wholethermocouple, but mainly alters the parts of

thermocouple wires exposed to heat. The magnitude of

the error as a result of thermocouple degradation

depends on immersion depths and heat exposure duringuse.

For accurate measurements with thermocouple

thermometers it is necessary to determine any possible

inhomogeneities in thermocouple wires. Detection of

inhomogeneity has two main objectives: first to detect

distribution of any inhomogeneity of the thermocouple,and second to determine the minimum uncertainty that

can be expected during use. Uncertainty during use can

not be smaller than uncertainty caused by

inhomogeneities, since the immersion depth of thePrejet 18. January, 2005

Odobren 1. April, ,2005


2/6

190 Hiti, Bojkovski, Batagelj, Drnovek

thermocouple can be such that inhomogeneities are

found in the maximum temperature gradient.

In this paper a measurement system for detecting

thermocouple inhomogeneities is presented. In

connection with a PC computer and existing

measurement equipment it enables fully automatic

detection of thermocouple inhomogeneities by

measuring thermocouple output at different immersion

depths.

2 Methods For Thermoelectric

Inhomogeneity Testing

A thermocouple consists of two dissimilar conductors,

i.e. thermocouple wires, joined to form a circuit. It

produces a current in a closed circuit when one junctionis at a different temperature from the other. The current

is produced by electromotive force (emf), also known asSeebeck voltage. The Seebeck voltage E(T) is

calculated by Eq. (1):

=2

1

)()(

T

T

dTTsTE (1)

where s(T) is the Seebeck coefficient of a homogeneous

thermocouple segment and T1 and T2 represent local

temperatures at each end of the thermocouple segment.It can be seen from Eq. (1) that the emf is not produced

at the thermocouple junctions but is generated along the

whole thermocouple length.Thermoelectric properties of thermocouple wires

may vary as a result of physical and chemical variations.

These variations can lead to changes of the localSeebeck coefficient along thermocouple wires during

use. In general, the Seebeck coefficient is a function not

only of temperature but also of ambient pressure, elastic

strain, magnetic field and other variables. Since the

thermal emf is generated at temperature gradients along

the whole wire length, local changes in Seebeckcoefficient can cause an error in the measured voltage at

a given temperature.

If inhomogeneous thermocouples are used at

different immersion depths, accurate temperaturemeasurements are impossible. Even if the immersion

depth of thermocouples is not changed during use, theyare in most cases removed for calibration, which is

usually performed at one immersion depth only. At

calibration the temperature gradient along the

thermocouple wires differs from the temperature

gradient at use, thus the inhomogeneity introducedduring use can not be detected and corrected. Since

inhomogeneity of the Seebeck coefficient is one of

major sources of uncertainty in thermocouple

temperature measurement, a method for inhomogeneity

testing has to be employed and uncertainty must be

evaluated accordingly.Several different methods and systems exist for

detection of inhomogeneities, however they usually

require dedicated equipment, [1, 2]. Such systems are

based on applying a local temperature gradient to the

thermocouple wire and moving it along the wire.

Various heat sources can be used to assure the required

temperature gradient such as hot-air blowers, small

flames and different furnaces and baths. Hot-air

blowers, small flames and other concentrated heat

sources cause a temperature rise in a narrow region on

the thermocouple wire with two temperature gradients,

while keeping both ends of the thermocouple at the

same temperature. In furnaces and baths the temperature

gradient is present only in one region along the

thermocouple wire with one end of the thermocouple

totally immersed in the heat source and one end at a

room temperature. Local heating of the thermocouple

wire is suitable for testing local inhomogeneities but is

not appropriate for detecting inhomogeneities along an

extensive length of the wire, [3].

With respect to motion control, different procedures

can be used when measuring thermocouple

inhomogeneities, [4]. The simplest method is manually

advancing the thermocouple into the heat source or

moving the heat source along the thermocouple and

measuring the emf. Since it is difficult to maintain a

constant rate of movement manually, a motorised

movement control is more appropriate. For mostaccurate inhomogeneity testing, an intermittent motion

is recommended, where a thermocouple is slowly

moved in steps of few millimetres. Such motion can berealized quite easily by using a computer or a

microcontroller and an appropriate system for linear

movement. The rate of movement can thus be definedand accurate positioning can be assured.

We designed and built a system that allows us to

perform tests either manually or automatically, at a

constant rate of immersion or with an intermittent

motion. It is intended for detecting inhomogeneities by

immersing the thermocouple in different heat sourcesand takes advantage of water baths, oil baths, furnaces

and other equipment existing in our laboratory.

However it can, with some modifications, also be used

for detecting inhomogeneities by moving a concentratedheat source along the wire.

3 Apparatus

The apparatus is designed as shown in [5]. The central

part of the system is a custom-made thermocouple

support set-up with motorised movement control by

means of a stepper motor (Fig. 1). The frame is built by

using commercially available aluminium profile

elements. The main element is designed for a range of

movement of about 400 mm.


3/6

Measurement Device and Procedure for Thermocouple Inhomogeneity Detection 191

Figure 1: System for motorized linear movement

The carriage for thermocouple movement isdesigned to fit the grooves of the aluminium profile

element. The movement of the carriage is realized by a

spindle and stepper motor. The spindle is used to

transform the rotation of the motor to translation of the

carriage. The friction of the spindle also holds the

carriage in place when windings of the stepper motor

are not energized. The thermometer support is mounted

on the carriage. A number of different support elements

had to be developed since there are thermometers of

various shapes that need to be fitted to the carriage.

While the design of the support for vertical insertion is

fairly straightforward, a special support for horizontal

insertion of the thermometer had to be built. Sincethermometers are fragile instruments and need to be

handled with care, a special design of the support is

necessary in order to prevent the thermometer from

breaking while inserting it into the heat source. It should

compensate any deviation of the motion of thermometer

from the axis of insertion. The developed support

restricts the thermometer movement in one plane but

allows some rotation along x and z axes (Fig. 2). It also

allows some movement in directions of the two axes,

perpendicular to the axis of insertion, thus resulting in

the thermometer aligning itself with the opening at the

heat source.

A control circuit is built to control the stepper motor.The employed microcontroller is an Atmel

ATmega8535 8-bit microcontroller. Its main featuresare 8 Kbyte In-System Programmable Flash Memory,

512 Byte EEPROM, 512 bytes SRAM, 32 general-

purpose I/O lines, 8-chanel 10-bit ADC and up to 16

MIPS throughput at 16 MHz. It has a built-in serial

peripheral interface for communication with a PC

computer or other peripheral devices. To drive thestepper motor, an L297 stepper motor controller is used

in combination with an L298N bridge driver. This

combination allows a low-cost stepper motor control

system to be built with only few components and

simplifies software development. The L297 integratesall the control circuitry required to control bipolar or

unipolar stepper motors and provides the stepping

sequence for the L298N bridge driver. The L297

requires only a step clock and direction signals from the

microcontroller and generates control signals for the

power stage.

Figure 2: Thermocouple support

The software for the microcontroller is written in

such way that the load on the microcontroller is

minimized. The microcontroller is mainly used to

control the speed and direction of rotation of the stepper

motor and for the communication with the PC

computer. It receives the required instructions such as

action to be performed (e.g. movement, home

procedure, position request) and associated parameters(e.g. speed, direction and distance to be moved). It also

replies to requests about status information (e.g. current

position, speed). All instructions that are sent to the

microcontroller are processed in real-time. Besides

communication protocol and stepper motor control

functions, no movement algorithm is pre-programmed

on the microcontroller. The communication between

microcontroller and PC computer is via an RS-232

interface. The system therefore requires no dedicated

hardware and can be connected to any PC computer. On

the PC, a program in LabVIEW is written to serve as a

user interface and as communication interface. The

software on the PC controls the thermometer movementas well as records the measured data from the voltmeter.

Data acquisition by the measurement instruments and

positioning of the thermometer can thus be

synchronized. By using low-level instructions for the

control of the system, it is possible to reduce the burden

on the microcontroller and at the same time simplify the

software development. The programming of the

movement is done in LabVIEW thus enabling the user

to simply change the movement procedure, without the

need to reprogram the microcontroller each time

changes are made. By allowing the user to program

custom procedures in LabVIEW, even the most

complicated movement procedures can be realized byusing the simple pre-programmed instruction set of the

microcontroller.


4/6


The automated system consisting of the PC

computer, controller interface and thermometer support

is shown in Fig. 3. The system can also be used for

other applications where accurate linear movement is

required [6].

Figure 3: System for automated inhomogeneity detection

4 Results Of Thermocouple InhomogeneityMeasurements

The developed system was used for measurements of

thermocouple inhomogeneity on the existing laboratory

equipment: in a water bath at 50 C, in an oil bath at 100

C and in a heat-pipe at 600 C. Three different types ofthermocouples were used for the tests: K-type, T-type in

the baths and R-type in the furnace.

First the K-type and T-type thermocouples were

simultaneously immersed in the water bath at 50 C.The set-up of the measurement system on top of the

water bath is shown in Fig. 4. Both thermocouples were

fixed to the carriage. The test was performed by

withdrawing the thermocouples from the bath in 5 mm

steps with a rate of 2 mm/s. The time interval between

two steps was 10 minutes to allow temperature tostabilize. On each step, 60 measurements of the thermal

emf were made and the median value was calculated.

Thermal emf was measured with an HP 34420A

nanovoltmeter, cold junctions of the thermocoupleswere immersed in an ice bath and stability of the water

bath during the test was measured by an additionalSPRT kept at constant immersion. Stability of the water

bath was 10 mK.

The procedure was repeated with the same K-type

and T-type thermocouples in an oil bath at 100 C.Stability of the oil bath measured with SPRT was 15

mK. Results of inhomogeneity measurements for the T-

type and K-type thermocouples in the oil and waterbaths are shown in Fig. 5 and Fig. 6, respectively.

Figure 4: System set-up on water bath

Figure 5: Inhomogeneity of the T-type thermocouple measu-

red in the water bath at 50 C and in the oil bath at 100 C

It can be observed from the results that boththermocouples exhibit significant inhomogeneities, the

T-type thermocouple at immersion from 140 mm to 190

mm and the K-type for immersion more than 200 mm.

It can also be observed that the effect of

inhomogeneities is much smaller at lower temperatures.

This is a result of a small temperature difference

between the ambient temperature and temperature of the

medium inside the water bath which is about 25 K. In

the oil bath at 100 C, the temperature difference at theoil/air interface is about 75 K. It produces higher

temperature gradients, amplifies the effect of Seebeck

coefficient inhomogeneities and produces higher

thermal emf variation.


5/6

Measurement Device and Procedure for Thermocouple Inhomogeneity Detection 193

Figure 6: Inhomogeneity of the K-type thermocouple

measured in a water bath at 50 C and an oil bath at 100 C

Inhomogeneity of the R-type thermocouple was

tested in a furnace with a non-pressure controlled

Caesium heat-pipe at 600 C. The thermocouple was

inserted into the furnace horizontally. The time interval

between movements was at least one hour to allow

temperature to stabilize. Temperature equilibrium inside

the heat-pipe was disturbed during movement of the

thermocouple and required long settling times. The new,

stable temperature after each movement of the

thermocouple was measured with a reference

thermometer at a fixed location inside the heat-pipe and

was not necessarily identical to previous stable

temperature. To compensate for this temperature

change, a difference between the reference thermometer

value and the thermocouple under test value wascalculated. Due to the long time interval, the step of

withdrawal was increased to 10 mm. Results are shown

in Fig. 7. There are no significant variations, which

indicates a homogeneous thermocouple.

5 Conclusion

The presented apparatus was successfully tested by

measuring inhomogeneities of thermocouples indifferent heat sources. Results are given for tests

performed by immersing different types of

thermocouples vertically in baths as well as horizontallyin a tube furnace. By analysing measurement results

obtained with the developed system, identification of

any thermocouple inhomogeneity is simple.Further development will be necessary to reduce

vibrations caused by the stepper motor during

movement of the carriage. As such vibrations can cause

noise affecting measurement results, our measurements

were made only when the thermometer was not moving.

Further work could be towards different inhomogeneitydetection techniques by using alternative movement

procedures and modifying the thermocouple support to

serve as a platform for a moving heat source.

To end with, we recommend the developed systemfor analysis of reliability of thermocouples where

knowledge of uncertainty and homogeneity is required.

This will contribute to an improved calibration integrity.

Figure 7: Inhomogeneity of the R-type thermocouple

measured in a heat-pipe at 600 C

6 References

[1] R. P. Reed, The effect of interrogating temperatureprofile in the Seebeck inhomogeneity method of

test (SIMOT), D. C. Ripple, Proceedings of

International Symposium on TemperatureMeasurement and Control in Science and Industry,

Volume 7, American Institute of Physics, New

York, 2003, pp. 491-496

[2] D. Zvizdic, D. Serfezi, L. Grgec Bermanec, G.

Bonnier, E. Renaot, Estimation of uncertainties incomparison calibration of thermocouples, D. C.

Ripple, Proceedings of International Symposium

on Temperature Measurement and Control in

Science and Industry, Volume 7, American

Institute of Physics, New York, 2003, pp. 529-534

[3] R. P. Reed, Thermoelectric inhomogeneity testing:Part I-Principles, J. F. Schooley, Proceedings of

International Symposium on Temperature

Measurement and Control in Science and Industry,

Volume 6, American Institute of Physics, New

York, 1992, pp. 519-524

[4] R. P. Reed, Thermoelectric inhomogeneity testing:Part II-advanced methods, J. F. Schooley,

Proceedings of International Symposium on

Temperature Measurement and Control in Science

and Industry, Volume 6, American Institute of

Physics, New York, 1992, pp. 525-530

[5] J. V. Nicholas, D. R. White, Traceable

Temperature, Chichester, John Wiley&Sons Ltd.,

1994, pp. 272-273.

[6] A. Miklavec, M. Hiti, J. Bojkovski, V. Batagelj,

Merilni sistem za merjenje nehomogenosti

termo*lenov in temperaturnega profila v fiksnih

to*kah, B. Zajc Zbornik trinajste mednarodne

Elektrotehnike in ra?unalnike konference ERK2004, Ljubljana, 2004, str. 423-426


6/6


Miha Hiti received his B.Sc. degree in Electrical

Engineering in 2002 from the University of Ljubljana,

Slovenia. As a member of the National Young

Researcher Scheme he works in the Laboratory of

Metrology and Quality at Faculty of Electrical

Engineering.

Jovan Bojkovski received his B.Sc. and Ph.D. degrees

in Electrical Engineering from the University of

Ljubljana, Slovenia in 1994 and 2002, respectively.

Since 1992 he has been a Research Staff Member at the

Laboratory of Metrology and Quality at the Faculty of

Electrical Engineering. His current research interests are

temperature calibration, International Temperature

Scale 1990, thermometers, relative humidity and on-site

calibrations. He is chairman of TC-Temperature at

EUROMET.

Valentin Batagelj received his B.Sc., M.Sc. and Ph.D.

degrees in Electrical Engineering from the University of

Ljubljana in 1998, 2001 and 2004, respectively. Since

1998 he has been a Research Staff Member at the

Laboratory of Metrology and Quality at the Faculty of

Electrical engineering. His current research interests are

automation of measuring systems in calibration

laboratories and temperature calibrations on the primary

and secondary level.

Janko Drnovek received his M.Sc. degree in 1979

from the Imperial College of Science and Technology of

London and his Ph.D. degree in Electrical Engineering

in 1986 from the University of Ljubljana. Currently he

is Head of the Laboratory of Metrology and Quality and

Professor at the Faculty of Electrical Engineering which

he joined in 1980. Between 1982 and 1986 he was

project leader for laboratory measuring systems at Iskra

Kibernetika, Kranj. Currently he is President of the

Metrology Council at the Metrology Institute of the

Republic of Slovenia and the Slovenian delegate to

EUROMET.

Documents

Homogeneidade Sirilanca.pdf