See discussions, stats, and author profiles for this publication at: [Link]

net/publication/265991600

Measuring and control the hysteresis loop by using analog and digital
integrators

Article · July 2008

CITATIONS READS

24 6,473

2 authors:

Zoltán Pólik Miklos Kuczmann

Balluff Elektronika Kft. Széchenyi István University
16 PUBLICATIONS 52 CITATIONS 156 PUBLICATIONS 1,092 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Zoltán Pólik on 01 April 2015.

The user has requested enhancement of the downloaded file.

JOURNAL OF OPTOELECTRONICS AND ADVANCED MATERIALS Vol. 10, No. 7, July 2008, p. 1861 - 1865

Measuring and control the hysteresis loop

by using analog and digital integrators
Z. PÓLIK*, M. KUCZMANN
Laboratory of Electromagnetic Fields, Department of Telecommunication, Széchenyi István University, Egyetem tér
1.H-9026 Győr, Hungary

We have managed to build up a computer controlled measurement system applying National Instruments Data Acquisiton
card and National Insturments LabVIEW software package. First, in a previous study, the specimen under test has been
simulated by the Jiles-Atherton model of hysteresis, but simulated measurements have been performed using LabVIEW
functions. Our main goal now is to control the magnetic flux density measured by using a toroidal shape core with a
feedback controller. In this paper, we show the developed procedure which is able to measure the hysteresis characteristics
applying analog and digital integrators, and to control the waveform of magnetic flux density determined from the induced
voltage of the secondary coil wound on the toroidal shape core. Our experience is that the developed controller is robust.
This work is an initial step of the vector hysteresis measurement system.

(Received April 1, 2008; accepted June 30, 2008)

Keywords: Scalar magnetic hysteresis, Computer controlled hysteresis measurement system

1. Introduction Two procedures have been developed to measure the

magnetic flux density based on the induced voltage of the
The exact simulation of the scalar hysteresis secondary coil of the toroid. The first is a digital procedure
characteristics of ferromagnetic materials, i.e. to know the which calculates the magnetic flux density with the left
magnetization M (t ) as a function of the magnetic field integration rule. The other one is an analog method with
intensity H (t ) is very important in the computer aided passive RL or RC integrator circuits.
design used in electrical engineering practice. Simulation
of this nonlinear system having memory is very difficult,
because it is hard to find a good structure for the
simulation. However, there are well known models to
simulate the scalar hysteresis characteristics, for example
the Jiles-Atherton model, the Stoner-Wohlfarth model, and
the Preisach model [1,2,3,4,5,6,7]. To compare the known
hysteresis models and the measured characteristics is very
important, because we want to see that the theoretical
models, simulations and procedures are working properly.
The implemented procedure developed in LabVIEW
is able to measure the scalar hysteresis characteristics. The
block diagram can be seen in Fig. 1. The control signal of
the amplifier can be defined by mathematical formula; and
Fig. 1. The block diagram of the measurement system
the amplifier can follow the control signal. The amplifier
2. Measurement of Field Quantities
is a voltage controlled current generator, which excites the
primary coil of the toroidal shape core, which current is
known. The magnetic field intensity generated by the It is well known that the measurement of magnetic
excitation current results in magnetic flux inside the core. field intensity H (t ) inside a toroidal shape coil is based
The magnetic flux density can be measured by the induced
on Ampere's law, because the current of the primary coil
voltage of the secondary coil and the magnetic field
intensity can be measured directly from the current of the i1 (t ) is in direct connection with this electromagnetic
amplifier. The amplifier has an output voltage which is quantity (the primary coil has N1 turns), i.e. [8,9,10]
proportional to the current. The measured results have
been postprocessed by the developed LabVIEW program, , now
∫ H(t ) ⋅ dl = ∫ J (t ) ⋅ dA ∫ J(t ) ⋅ dA = N i (t ) . (1)
and the hysteresis characteristics have been plotted by l A A
1 1

LabVIEW functions as well. Results can be saved into

files for further use.
1862 Z. Pólik, M. Kuczmann

The magnetic field intensity vector and the magnetic the induced voltage. Exact measuring of this signal could
flux density vector are supposed to have component only be possible with very high sampling frequency, but it is
along the polar axis of the core. This is a consequence beyond possibility with the used DAQ measurement card.
Equation (1) can be simplified as If the frequency of the current signal is lower than 0.5 Hz,
the induced voltage is very small and noisy. Some BH-
H (t ) 2rπ = N 1 i1 (t ), (2) loops measured with digital integrator at 2 Hz can be seen
in Fig. 2.
where r is the mean value of radius of the core, and
l = 2rπ is denoted in Fig. 1. From expression (2), the
mean value of the magnetic field intensity H (t ) can be
determined as
N 1 i1 (t )
H (t ) = , (3)
2 rπ
and the waveform of current can be measured directly
from the amplifier, because it has an output which voltage
is proportional to the generated current.
The magnetic flux density B(t ) is depending on the
Fig. 2. BH-loops at 2 Hz with digital integrator
hysteresis characteristic of the specimen which can be
measured through the magnetic flux Φ (t ) inside the core.
The time variation of flux results in induced voltage in the
secondary coil having N 2 turns according to Faraday's
law, [8,9,10]

d d
u 2 (t ) = − Φ (t ) N 2 = − B(t ) S N 2 , (4)
dt dt
Fig. 3. RL passive integrator

since

Φ(t ) = ∫ B(t ) ⋅ dS = B(t ) S , (5)

S

where S is the cross section area of the core. The magnetic

flux density can be determined from equations (4) and (5),
Fig. 4. RC passive integrator
t
1
B(t ) = B0 + ∫ u ( τ ) dτ ,
2 (6) The analog integrators have many advantages. For
SN 2 0 example, the integration which is a time consuming
method with PC must not be done, they can work higher
where B0 is a constant. frequencies (here between 50 Hz and 1 kHz) and when
The magnetic flux density B has been determined by using RC or RL passive integrators, it is not necessary to
digital and analog methods. In the digital way the left use filters and other signal processing algorithms. The
integration rule has been used to calculate the magnetic schematic draw of the RL and RC passive integrators can
flux. In this case there is a problem on the input channel of be seen in Fig. 3 and Fig. 4. Some BH-loops measured
the used DAQ measurement card. Unfortunately, it has a with RC passive integrator at 200 Hz can be seen in Fig. 5.
noise with high frequency and some mV in amplitude. If The output of the RL integrator can be determined as [8],
we calculate B, this noise is added to the induced voltage
t t
and the outcome was “wrong”. That is why we used the 1 R
FFT (Fast Fourier Transform) to see how to create a filter v(t ) ≅ −
τ ∫ u(ξ )dξ = − L ∫0
u (ξ )dξ , (7)
to cut the added noise from the measured induced voltage. 0
After that we could construct the filter which is a
Butterworth one. Finally the program subtracts the offset the output of the RC integrator can be determined as
B0 from the signal. It is our experience, that we can use
this method between 0.5 Hz and 200 Hz. In higher
frequencies there appear impulses with wide spectrum in
 Measuring and Control the Hysteresis Loop by using analog and digital integrators 1863

t t 3. About the LabVIEW Program

1 1
v(t ) = −
τ ∫ u(ξ )dξ = −
0
RC ∫0
u (ξ )dξ .(8)
Applying the developed GUI, one can set the
parameters of the excitation signal (signal type, amplitude,
frequency, sample per period, number of periods), the
parameters of channels and the four parameters of the
specimen.
The measured data, i.e. the measured signals
according to the magnetic field intensity and to the
magnetic flux density (see equation (3) and (6)), the post
processed data as the magnetic field intensity, the
magnetic flux density and the resulting hysteresis
characteristic can be plotted as well. Measured results can
be saved to file for further use. This is very important,
because measured data can be applied to identify
Fig. 5. BH-loops at 200 Hz with RC passive integrator hysteresis models.
The signal generation rate is equal to the value of
At last, some BH-loops can be seen in Fig. 6 at 200 sample per period multiplied by the number of periods.
Hz, measured via RL passive analog, digital, and RC The sapling frequency of excitation and measurement have
passive analog integrators. It can be seen that the loops are been set to be equal, which is the frequency multiplied by
exactly the same in the stationary state. They are different the value of sample per period. From these data the signal
only in their transient. of exciting current can be set up. The realization of this
process in LabVIEW software environment can be seen in
Fig. 7. The induced voltage before and after the filtering
procedure can be seen in Fig. 8 and Fig 9.
Next the Data Acquisiton card have been reset, then
output and the input channels, the sample clock of signal
generation and measurement must be set, and finally the
signal generation and the measurement can be started.
These tasks happen simultaneously.

Fig. 7. Generating the excitation signal

Fig. 8. The induced voltage before the filtering procedure

Fig. 6. BH-loops at 200 Hz with RL, digital, and RC integrator

Fig. 9. The filtered induced voltage

1864 Z. Pólik, M. Kuczmann

4. Controlling the Flux

It must be noted that the hysteresis characteristic can

be very steep, especially close to the coercive field, when
the rate of change of magnetic flux is very high. If the
variety of magnetic field intensity is controlled, then very
few H – B data can be measured from the steep part of the
curve. The aim is to get enough information about the
variation of magnetic flux from every part of the
characteristic. That is why we have started to develop a
procedure which aims to control the waveform of
magnetic flux.
The aim is to reach sinusoidal variation of the magnetic flux density, and the waveform
Fig. of current
11. The realized must be modified iteratively by an
algorithm
The reference signal is a given sinusoidal function,

⎛ B ( 0) ⎞ In Fig. 12 and Fig. 13 concentric loops can be seen by

B ref (t ) = B max sin ⎜⎜ ωt − arcsin ⎟⎟ ,(9) using the proposed simple but robust feedback,
⎝ B max ⎠ proportional controlling algorithm. In the first case the
amplitude of the magnetic flux density is 1 T, then it is 1.8
T.
where Bmax is the amplitude, ω is the angular frequency
of the magnetic flux density, and B(0) is the value of
magnetic flux density when t = 0 .
The difference between the reference signal (9) and
the measured magnetic flux density is the error which is
the input of the controller,

e(t ) = B ref (t ) − B(t ) .(10)

First, the controller is a simple proportional one, i.e. the

exciting current has been modified by α e(t ) , where α is
a small positive number. Here is α =0.05 has been used
after some trials. Fig. 12. Concentric loops by using the controlling procedure,
The control algorithm has been applied until f=5, 10, 20, 50, 100, 200 Hz

N −1
1
MSE =
N
∑e
k=0
2
k ,(11)

i.e. the mean square error is larger then a given limit (here
ε = 0,002).

Fig. 13. Concentric loops by using the controlling procedure,

f=5, 10, 20, 50, 100, 200 Hz
Fig. 10. The block diagram of controlling algorithm.

5. Conclusions

In our Electromagnetic Fields Laboratory we have set

up a measurement system to record the scalar hysteresis
charateristics which is based on National Instruments
 Measuring and Control the Hysteresis Loop by using analog and digital integrators 1865

LabVIEW software and National Instruments Data [2] P. Kis, A. Iványi, Parameter identification of Jiles-
Acquisition card. Atherton model with non-linear least-square
We have studied a control algorithm which is able to method, Physica B 343, 59 (2004).
generate the given signal of magnetic flux density. The [3] D. C. Jiles, J. B. Thoelke, Theory of Ferromagnetic
aim of further research is to apply the developed Hysteresis: Determination of Model Parameters
algorithms in measuring the hysteresis of the simulated from Experimantal Hysteresis Loops, IEEE Trans.
core. The controller is the simplest one, but we belive that on Magn. 25 (1989) .
this is a roboust method which can be used with minor [4] E. Sárospataki, M. Kuczmann, Realization of the
modifications. Jiles-Atherton Hysteresis Model Applying the
Our future plans are to identify and to compare known LabVIEW Software Package, Journal of Electrical
hysteresis models which can help in computer aided Engineering 57, (8/S), 40 (2006).
design with numerical simulations of electromagnetic [5] A. Iványi, Preisach Memorial Book, Akadémiai
fields. Kiadó, Budapest (2005).
[6] I. D. Mayergoyz, Mathematical Models of Hysteresis,
Acknowledgements Springer (1991).
[7] E. Della Torre, Magnetic hysteresis, IEEE Press, New
The research work has been sponsored by the York (1999).
Department of Telecommunications, Széchenyi István [8] K. Simonyi, Theoretical Electromagnetics,
University, Győr, Hungary. This paper was supported by Tankönyvkiadó, Budapest, (1999).
the János Bolyai Research Scholarship of the Hungarian [9] Z. Pólik, M. Kuczmann, Measuring and Control of the
Academy of Sciences (BO/00064/06). Scalar Hysteresis Characteristic with Analog And
Digital Integrators, Journal of Electrical Engineering,
58(4), 2007, to be printed P. Kis, M. Kuczmann, J.
References Füzi, A. Iványi, Hysteresis Measurements in
LabVIEW, Physica B 343, 357 (2004).
[1] A. Iványi, Hysteresis Models in Electromagnetic
Computation, Akadémiai Kiadó, Budapest (1997).
________________________
*
Corresponding author: polikzoltan@[Link]

View publication stats