Development of Fiber Optics
Sensor for Partial Discharge
Measurement
Commercially Available Instruments
For Measurement Of PD Using FOS.
MPD 600 (M/s Omicron make, CA)
FO System (Power Dignostix System, Germany )
MPD 600
The MPD 600 PD measurement
and analysis systems is already
in use worldwide in the
following areas:
For type testing and quality
control at well-known
manufacturers of cables,
transformers, generators and
motors, converters, switch
boards and electrical
components.
For the regular maintenance of
equipment in utilities and
industrial plants.
For research at laboratories.
Advantage of MPD-600
Modular Design of easy set-up.
Long battery life for uninterrupted testing.
Fiber optics for improved safety & sensitivity.
Adjustable frequencies for noise suppression.
Synchronous multi-channel measurement for accurate PD
source separation.
Measures up to the UHF range.
Disadvantage
MPD-600 do not use Optical Sensor; it’s use acoustic
sensor for measurement of PD.
FO System
The FO system allow easy installation of fiber optics
transmission to solve noise, safety & signal quality issue.
It’s consists of
1. Transmitter
2. Optical fiber
3. Receiver
Operation
The output signal of the signal source or sensor (for example, partial
discharge) is fed to the input of the Fiber Optic Transducer(FOT).
The transmitter then either modulates the infrared light emitted by
the transmitter diode according to the value of the input signal
(amplitude modulation, FOS1), or digitizes the signal and transmits
a digital telegram.
This optical signal is transmitted via the fiber to the receiver, which
converts it back to voltages for further processing.
Cont…
The transmitter is installed in a cast-aluminum enclosure. The FOT
is fitted with a NiMH-battery for a mini mum operating time of
twenty hours (FOT1 and 2), whereas the FOT3 offers up to two
years of operation.
Several options are available for the FOS e.g. pre-dividers, different
racks, and adaption of the FOT/FOR for special application.
Disadvantage
The FO system of (PDS, Germany) also don’t have
Optical sensor, Its also use Acoustic sensor.
The velocity of sound decrease as the oil temperature is
increase.
It’s also affect the sensitivity of the acoustics sensor.
Fiberoptic Sensors for High-Voltage
Applications
Transformers, cables, rotating generators, solid-state substations and
renewable energy production facilities are among the most
important power system components.
Unscheduled outages of these high voltage components due to
unexpected failures are catastrophic in many cases.
Advantage of FOS
FO Conveniently providing robust fiber optic transmission of
analog signals in high-voltage environments.
FO system solves signal integrity problems and safety issues found
with severe electromagnetic conditions and long distances.
Fiberoptic sensor is used because of its small size, no electrical
power is needed at remote location also its withstand high temp.
Acoustic monitoring of partial discharges
In Transformer the discharge will produce detectable acoustic
emission inside the transformer.
Several methods are presented where the ultrasonic pressure wave
is detected by running an optical fiber through the transformer.
To detect PD in cable, we run an optical fiber along an installed
high-voltage cable.
Partial discharge in oil filled high voltage
equipment
Partial discharge can cause creation of light and sound, increase in
dielectric losses, propagation of electromagnetic waves, chemical
reactions, creation of high gas pressure and generation of electrical
pulses caused by charge transfer.
The PD sensors detect ultrasonic pressure waves or similar, using
interferometric technique. Sensors seem to be feasible but complex,
and a major drawback is the sensitivity to vibrations and other
sound-like sources.
Partial Discharge Sensor
Diaphragm-based extrinsic Fabry-Perot (FP) interferometric
sensors.
The sensor can detect the acoustic waves generated by PD activities
inside a transformer.
Fig 1. Schematic of the Diaphragm-Based PD Sensor Head
PRINCIPLE OF OPERATION
comparing to 𝐸1 Let an electrical field of the incipient light is E, and
the electrical field of the reflections from the fiber end and inner
surface of the diaphragm are 𝐸1 and 𝐸2 respectively. Electrical field
𝐸1 can be expressed as:
Ƹ 1𝑒𝑗
𝐸1 =𝐸1 𝑧=𝐴 𝜔𝑡−𝛽𝑧+∅ 𝑧………….(1)
Ƹ
Where β is the propagation constant of the light in the fiber. 𝐴1 ,ω, &
𝜙 are the amplitude, angular frequency and initial phase of the light
respectively.
Electrical field 𝐸2 travel back and forth inside the FP cavity and has
phase shift of 4𝜋𝑑ൗ𝜆0.
Cont….
The electric field 𝐸2 can be expressed as:
4𝜋𝑑
𝑗 𝜔𝑡−𝛽𝑧+𝜙− +𝜋
𝐸2 =𝐸2 𝑧=𝐴
Ƹ 2𝑒 𝜆0 𝑧Ƹ ………..(2)
Where 𝜆0 is the wavelength of the light in the air, d is the cavity
length of FP.
𝐴2 is the amplitude of the light and extra phase shift 𝜋 is arise.
Cont…..
Since the reflections at both FP surfaces are small, the multi-
reflections inside the FP cavity can be ignored. The interference
signal is detected by a photodetector and the output of the detector
is then given by:
2
𝐼≈ 𝐸1 + 𝐸2 = (𝐸1 + 𝐸2 )(𝐸1 + 𝐸2 )⃰ …………..(3)
=𝐴1 2 +𝐴2 2 -2𝐴1 𝐴2 cos( 4𝜋𝑑
𝜆
)
0
After filtering out the direct current (DC) signal, the useful
alternative current (AC) is given by:
U= 𝛾 cos(4𝜋𝑑ൗ𝜆0)…………………..(4)
Where 𝛾 is the constant and from the equation.
Cont…..
The useful AC signal is function of both cavity length d,
wavelength of the light 𝜆0 .
The sensitivity of the sensor, which defined as the intensity change
caused by unit cavity length change, is given by:
𝜕𝑈
S (d)= =𝛾 ′ sin (4𝜋𝑑
𝜆0
)…………………(5)
𝜕𝑑
Where 𝛾 ′ =4𝜋𝑑𝛾 Τ𝜆0
Eq. (5) shows that the sensor sensitivity to the gap-length is
sinusoidal function of the gap length and reaches its maximum
when..
1 𝜆0
d = 𝑑𝑄𝑚 =( m+ ) m=0,1,2,3..... ….(6)
2 4
These 𝑑𝑄𝑚 are called the operating points or Q points of the sensor.
Due to the thermal expansion and the static background pressure of
the transformer oil, the original cavity length d or the working point
can slowly drift from the optimized point, causing degradation in
the sensor sensitivity.
The working point must be carefully controlled during the sensor
operation. Previously a tunable laser or filter is used to adjust the
wavelength. However, the required tunable lasers or tunable optical
filters are costly and the multiplexing of sensors is difficult.
Two-Wavelength Quadrature Detection is used to avoid this
problem.
Two-Wavelength Quadrature Demodulation
Two-wavelength quadrature demodulation uses two lasers with
different wavelengths as the light sources. Let an arbitrary
wavelength λ=𝜆0 +Δλ, where Δλ is the deviation from a reference
wavelength 𝜆0 and satisfies Δλ << 𝜆0 .
Then from Eq. (4), the normalized response of the sensor is
approximately a cosine function of Δλ which is given by:
1 Δ𝜆
U( 𝜆0 +Δ𝜆,d) = cos 4𝜋𝑑( −
𝜆0 𝜆0 2
= cos(𝜃 − 4𝜋𝑑. 𝜆0 2 Τ2𝑑)……………(7)
4𝜋𝑑
Where 𝜃 =
𝜆0
Cont……
The period of the sinusoidal wave is:
𝜆0 2
𝜆𝑑 = …………………(8)
2𝑑
Figure 1.(a) and (b) shows the normalized sensor response as a
function of the laser wavelength change Δλ at two different cavity
length d and d ′.
In two-wavelength quadrature demodulation, the wavelengths of the
lasers,𝜆1 and 𝜆2 are selected in such a way that:
𝜆0 2 𝜋 𝜆0 2
𝜆2 − 𝜆1 = x = ………………(9)
4𝜋𝑑 2 8𝑑
Fig-2.(Two-laser quadrature detection (d~d’) (a) the best condition;
(b) the worst condition)
Compared with the traditional Q-point control methods which
involve tunable lasers or tunable optical filters, two-wavelength
demodulation significantly reduced the total system cost.
The optical power of the lasers is more efficiently used than tunable
filter-based Q point control in which only a small part of the light is
used as the probe light.
When the environment induced sensor cavity length change is too
large, In such case the worst sensitivity of the whole system may be
reduced.
Cont…
Fig 3.(Tow-laser quadrature detection (d>d’) (a) the worst condition
without background pressure; (b) the worst condition with
background pressure.)
Cont….
From Fig-2:
𝜆𝑑 𝜆2
Δ𝜆 = 𝜆2 − 𝜆1 = = ……………(10)
4 8𝑑
& from Fig-3:
𝜆′ 𝑑 𝜆2
Δ𝜆 = 𝜆2 − 𝜆1 = = ………….(11)
6 12𝑑′
where 𝜆′ 𝑑 and d′ are the changed fringe period and sensor cavity
length respectively.
From Eq.(10) & (11) we can write that:
𝑑′ 𝜆𝑑 2
= = ……………………..(12)
𝑑 𝜆′ 𝑑 3
Cont….
The static-pressure-induced cavity length change must be smaller
than 1/3 of the original cavity-length to guarantee the sensitivity is
better than 0.5.
Therefore, from Eq. (10) the maximum cavity length due to this
requirement is:
𝜆2 15502
𝑑𝑚𝑎𝑥 = = =43μm. ………….(13)
8.Δ𝜆 8𝑋7
Typically the CWDM add/drop multiplexer has 20dB attenuation at
7nm away from the center wavelength (the central wavelength is
1550nm.).
From Eq. (8) the minimum cavity length is:
𝜆0 2 15452
𝑑𝑚𝑖𝑛 = = = 24 μm………………(14)
2𝜆𝑑 2𝑋50
Where 𝜆0 = 1545 nm & 𝜆𝑑 = 50nm
Cont….
The static-pressure-induced cavity length change must be smaller than 1/3
of the original cavity-length to guarantee the sensitivity is better than 0.5.
Therefore, from Eq. (10) the maximum cavity length due to this
requirement is:
𝜆2 15502
𝑑𝑚𝑎𝑥 = = =43μm. ………….(13)
8.Δ𝜆 8𝑋7
Typically the CWDM add/drop multiplexer has 20dB attenuation at 7nm
away from the center wavelength (the central wavelength is 1550nm.).
From Eq. (8) the minimum cavity length is:
𝜆0 2 15452
𝑑𝑚𝑖𝑛 = = = 24 μm………………(14)
2𝜆𝑑 2𝑋50
Where 𝜆0 = 1545 nm & 𝜆𝑑 = 50nm
Sensor Structure and Fabrication
Dimensions of Key Components of The Sensor Head:
I.D (μm) O.D (mm) Length (mm)
Ferrule 127 1.8 6
Sleeve 1815 2.8 3
Figure-4 (Sensor Head Structure and Bonding (a) Schematic of Sensor Head; (b)
Cross-Section View of the Sensor Head After Laser And Sol-Gel Bonging)
Cont…..
Based on the diaphragm theory the sensitivity of sensors depicted in
Figure 4 is given by:
𝑎 4
𝛿= 2.524𝑋10−6 3 (nm/kpa)…………..(15)
ℎ
Where a and h are the radius and thickness of the diaphragm in μm
respectively.
The resonant frequency of the diaphragm is given by:
ℎ
𝑓0 =2.72x106 (kHz) …………………..(16)
𝑎2
Optical Receiver Circuits
Single Channel PD Detection
To check the efficiency of the quadrature detection, A PD detection
module with single channel (single PD sensor) was built.
Single Channel PD Sensor Test System
Spectra and Output Signal from a PD
Sensor in Three Conditions
Conclusion
Fiber optic sensors are inherently advantageous for application
within transformers because of their immunity to the strong
electromagnetic interferences present inside the transformer.
Due to the fact that they are an excellent dielectric – and hence can
be safely inserted into the high electric fields in a transformer
without compromising the insulation.
The internal fiber optic sensor is not subjected to the attenuation
through the tank walls