Optical Instrument Transformers

Author: Farnoosh Rahmatian, NuGrid Power Corp., Canada

Measuring voltages and currents using light!  To some of us, with traditional electrical measurement technology background, it sounds like magic.  It definitely sounded intriguing to me as an undergraduate electrical engineering student about 30 years ago.  I had an implicit trust that there is science behind it, not just magic.  Once I read a bit more and learned about the Faraday Effect (magnetic field affecting the polarization of light) and the Pockels Effect (electric field affecting the speed of light in certain materials), it all seemed reasonable.  After all, these folks lived over a hundred years ago, (Michael Faraday, 1791 - 1867, and Friedrich C. A. Pockels, 1865 - 1913) and were well respected scientists, so it must be science.  It wasn’t until I worked on it (together with some very intelligent people) that I realized that making a good optical voltage or current sensor is indeed based on science, but it is also an art.

System Description
Optical instrument transformer systems typically consist of three major parts: the optical sensors or transformers (primary sensors), the sensor electronics (signal processing unit or the secondary converter), and the cabling system between the two (typically optical fibers).  Figure 1 shows a schematic of a typical optical sensor system. Figure 2 shows a picture of the optical transformers of an actual 550 kV class combined optical voltage and current sensor system.
The primary sensors, or optical transformers, as labeled by IEEE Std. 1601, are usually passive, consisting of glass and insulating material.  The electronics module, typically installed in the substation control house, or in a substation yard cabinet, contains the light source, photo-detectors, and digital processing unit.  Light is sent via the cabling system (usually optical fibers), from the electronic chassis to the sensor head (optical transformer), it is affected by current or voltage, and is returned via the cabling system back to the electronic chassis.  The return light is detected, analyzed, and deciphered to extract voltage or current present at the location of the sensor heads.

Optical Current Sensing through Faraday Effect
Most optical current transformers are based on the Faraday Effect, also known as the magneto-optic effect, discovered by British scientist Michael Faraday in 1845:
For a plane-polarized light wave propagating through the optical medium in a direction parallel to the applied magnetic field, the polarization plane of the light rotates.
The angle of that rotation is expressed by:  = VHdl where:
V = Verdet Constant (V ~5e-06 rad/A for fused silica)
H = magnetic field strength
l   = length (distance that light must pass through medium)

In summary, magneto-optic effect helps measure an integral of magnetic field over certain distance where the sensing fiber is aligned with the magnetic field.
Separately, Ampere’s law says integration of magnetic field around any closed path is equal to the current flowing through that path:  I = Hdl .

Combination of the two relationships here, using a magneto-optic medium (e.g., fused silica glass) to integrated magnetic field over a closed loop, will yield polarization rotation directly proportional to the current; i.e.,   = VNI
I    = current in the conductor
N = # turns (of sensing fiber coil or optical path in machined bulk glass)
Multiple turns of light around the conductor can help adjust sensitivity, easily achieved in fiber optic sensor heads.
Early optical current sensors were manufactured as machined glass blocks (bulk glass sensors) with the current carrying conductor passing through the middle.  Optical fibers are coupled to the glass block with a beam of light internally reflected to pass around the conductor exiting through a polarizer and returning to a photo detector.   The change in the amount of light received at the photodetector is proportional to the current in the conductor.   These earlier polarimetric sensors worked fine, but sensitivity was somewhat limited as only a turn or two of light could pass around the conductor.  Newer approaches offered more flexibility in use and better rejection of external effects such as vibration.

The new designs evolved with optical fibers replacing glass blocks.   Interferometric sensing, where two signals of opposite polarities are used, replaced polarimetric sensing.   The magnetic field influences the optical signals in opposite directions with the current being proportional to the difference between the signals.  Fiber optic sensing medium allowed multiple turns (sometimes hundreds) improving sensitivity and reducing noise while the interferometric technique allowed excellent vibration and temperature performance.  
High performance, linear interferometric fiber OCTs are very flexible and easily adapted to different applications with adjustments to performance and form factor as required.     The sensing head can be adjusted to any size with any number of turns.  The same design from a signal processing standpoint can be physically adapted to use for mounting on transformers, circuit breakers bushings, gas insulated apparatus or as a free-standing instrument transformer.  

Optical Voltage Sensing through Pockels Effect

Optical VTs are generally based on the Pockels Effect.    Light travels through electro-optic materials where the speed of light is changed in the presence of the electric field.   The birefringence, a change in refractive index in opposite directions for light polarized in different (usually orthogonal) directions, is linearly proportional to the magnitude of the electric field. 

Figure 3 shows the Pockels Effect where linearly polarized light is converted to a circular polarization through a waveplate.   The circularly polarized light becomes elliptically polarized as a result of the electric field present, and its ellipticity alternates with the AC voltage.  Ultimately, the amount of light passing through the end polarizer depends on the electric field present at the location of the electro-optic crystal.  

The assemblies of electro-optic crystals, lenses, filters and polarizers are called Pockels cells.  Technically a Pockels cell is a voltage sensor, integrating electric field from one end of the crystal to the other.  Small Pockels Cells positioned between two distant points at different voltages are effectively point electric field sensors. 

The challenge with electric field measurements is that they are not always proportional to voltage.   External influences, insulator contamination, ice, snow, conductor arrangements etc. can all change electric fields at localized points even as the voltage remains the same (i.e., electric field distribution can change between the two points across which voltage is to be measured).  

Some implementations of optical VTs apply the entire voltage of interest across the Pockels cell yielding a very robust voltage measurement (integrating electric field continuously along the entire voltage drop path).   From a practical standpoint, this requires large expensive crystals and/or an SF6 gas insulation system.  It works well in the well-controlled field environment of a gas insulated switchgear and when the voltages of interest are smaller than the linear range of the Pockels Cell’s transfer function (usually less than 10 to 20 kV for typical Pockels materials).  For larger voltages, applying the entire voltage across a Pockels cell brings insulation and signal processing challenges, as well as significant cost and design complications.     

One implementation that has been used with success takes an array of Pockels Cells distributed through an insulating column and combining the measurements together to yield voltage (effectively using sample electric field measurements along the path to perform numerical electric field integration {summation} between two ends of the insulating column across which voltage is to be measured).    A perturbation in the electric field distribution can affect the measurement of an individual electric field sensor, but the combined result remains a robust voltage measurement.  In other words, electric field distribution profile can change, but increase at one location is effectively compensated by decrease of electric field at another location.   The distributed array sensor trades off a little signal processing complexity for a much simpler insulation system using only dry nitrogen (avoiding oil and SF6).  Figure 4 shows an example of an optical VT using distributed electric field sensors.

One of the key advantages of the distributed OVT design is the safety and elimination of SF6.  Figure 2 shows the very first 550 kV class optical VT (combined with OCT) in the world, installed in a 550 kV substation for monitoring a major utility to utility intertie. 

Deployment - Optical instrument transformers offer a number of attractive features:

  • Small size and weight -  allowing zero-foot print installations, or easy integration into other equipment and switchgear
  • Linearity - providing same accuracy over a very wide dynamic range (wide range of current or wide range of voltages). 

Aside from enabling support of multiple applications concurrently (e.g., metering and protection), linear sensors can simplify substation design by allowing a simple template design (scalable architecture) for multiple applications at various current/voltage levels

  • Wide bandwidth - useable for measuring AC, harmonics, transients, and possibly DC signals
  • Seismic performance due to being light-weight and compact
  • User-adjustable sensitivity (turns-ratio is a setting), helping with both efficient spares-keeping and streamlined integrated switchgear manufacturing processes
  • No iron-core saturation
  • No open secondary circuits
  • Galvanic isolation from HV line with no oil or SF6 (depending on design)
  • Integrated digital output, e.g., in 61850-9-2 format.
  • Multifunction - e.g., metering & protective relaying capability in one device
  • Self-monitoring and communicating

The main point of the above list is to communicate some of the reasons why optical instrument transformers have been used.  It is important to realize that these features are “possible” in optical current and voltage sensors; that doesn’t mean that every optical current or voltage sensor product offers all these features; similarly, it doesn’t mean that a specific optical sensor product implements these features perfectly.  Like any other technology, there are different products available on the market with varying levels of performance and quality.  There is no replacement for good judgement and engineering in choosing the right solution for the right application.  

Also, while all these attractive features are possible, actual optical sensors come with deployment challenges.  One of the most significant challenges for using optical voltage and current sensors in the main stream protection applications in the electric power industry is the issue of the interface, particularly for current sensors.  Relays, including most electronic or digital relays installed in substations, use 1A/5A interface for current and 69V to 120V interface for voltage inputs.  While it is possible to build current amplifiers that produce 5A at the rated current and 100 A during faults (20 times rated current) into rated burdens of 0.25 ohm or even larger, large scale use of such amplifiers is impractical (i.e., inefficient, unreliable, and prohibitively expensive).  

The technical solution is a low-energy interface, either analog (voltages <10V) or digital, useable with electronic and digital relays, as long as the right interface is standardized for this purpose.   A number of standards are now available for low-energy analog interface between electronic instrument transformers and electronic relays including IEEE C 37.92, IEEE 1601, IEC 60044–7/8, and IEC 61869-6 through -11. The practical challenge is, however, that there are not many relays available on the market with low energy analog inputs.  There are a select number of relays available with these low-voltage inputs (<10 V), but the selection is not large enough to impact wide scale use of optical or electronic instrument transformers.

In fact, from a technical point of view, a more practical approach is to use a digital interface between the instrument transformers and intelligent electronic devices (IEDs) such as relays.  There are a number of standards available for this digital interface: IEC 60044-8 was one of the earlier standards partially used in the market; IEC 61850-9-2 (together with a UCA implementation guide, known as “9-2LE”) has been the most common digital interface used to communicate the output of digital instrument transformers to digital relays.  9-2LE is, nevertheless, a guide and not a standard.  The effort to transition 9-2LE to an implementation standard, while improving a few short-comings of 9-2LE, took about 13 years: IEC 61869-9, which is standard for implantation of 61850-9-2 interface between digital instrument transformers and digital IEDs, was published in 2016.  The good news, from a more practical perspective, is that there seems to be plans from most relay manufactures for providing their relay platforms with IEC 61869-9 input (in addition to the traditional 1A and 5A analog inputs).  Thanks to these recent developments, the next few years will be very interesting with respect to the adoption of digital instrument transformers and digital substations (not necessarily the same rate of adoption for both).  The use of standalone merging units and availability of relays with digital input are major contributors to the adoption rate.

Another natural challenge of using optical instrument transformers is the novelty, or lack of familiarity for most users.   Most utilities have used optical sensors in a very limited way, either in pilot trial installations, or for very specific applications such as ultra-high voltage (UHV) or HVDC applications. Real familiarity and efficiency for the technicians and engineers come with the everyday use of the technology or the products.  It is not realistic to expect a technician to remember how to deal with a device if he/she interacts with that device for only a week in two years!  It may help if a utility looks at the optical technology seriously and considers the business case for using it systematically through its infrastructure. The fiber-optic current sensor, in particular, is a great technology to consider for widespread adoption by the utilities of the future:

1.  It enhances the safety around the substation

2.  It can serve both protection and metering applications (as well as others)

3. It can be used for both AC and DC measurement (including high-frequency and low-frequency applications, such as measuring geomagnetically induced currents, harmonics, and transients);

4.  It can be used in both air-insulated and gas-insulated substations

5. It can be used for distribution, transmission, and generation applications

6. Because of the widespread applicability across the utility, the overall cost of training and familiarization for the employees can be very little (repeat use is the best natural training)

7. It is available from more than one established manufacturer

8.  It is a natural fit for the modern grid

Again, regardless of these potential advantages and benefits, new technology adoption in any high-reliability industry requires adherence to a thorough technology validation and adoption process.  This process usually starts with laboratory tests and validations, and follows through the stages of pilot installations, performance validation, debugging, standardization, training, limited-use deployment, and finally mass deployment.  

Collectively, various users have travelled down this path for the past 30 years or so, and we can see maturity and full acceptance of the technology in certain specific market segments such as HVDC, UHV (above 800 kV AC), and industrial high-current DC applications (>40 kA) such as aluminum smelters.  The adoption process continues for other applications, such as mainstream protection, automation and control applications in transmission and distribution systems, and with the availability of digital interface standards (and products adhering to these standards), the adoption rate can accelerate significantly in the coming years.

Use of Optical Sensors for Field Testing

Optical sensors can be very attractive tools for measurement and data acquisition during field (and laboratory) testing.  They can be linear (providing accuracy over a wide range of currents, voltage, and frequency), provide excellent safety through optical isolation from high voltage, be compact and light, and measure voltage and current non-intrusively (they don’t affect or load the system being measured); all very attractive features for a field testing sensor.

Figures 5 and 6 show use of optical voltage and current sensors for staged fault testing of a 500-kV series capacitor bank system.  To verify the design and ensure the protection operates effectively within the required timing, several actual faults were staged.  The faults were initiated by shooting a thin metal wire through two metallic hoops connected to the ground and high-voltage sides of a post insulator at the substation (see Figure 5).  The fault current is measured by running the ground wire connected to the lower hoop through a window type optical CT.  On the series capacitor bank platform, Figure 6, metal oxide visitors (MOVs) with cut-off voltage of about 200 kV protect the capacitors during the few cycles of the fault before the bypass breaker operates.  A lightweight 138kV class optical VT (lightning impulse withstand level of  650 kV, one-minute AC withstand of 275kV rms) with a ratio of roughly 200 kV to 10 V is used to measure the waveform seen by the MOVs.  A bucket truck was used to lift the Optical VT onto the platform.  MOV current is also measured using optical CTs. 

A very critical feature of the optical sensors was that the sensing heads were passive (no electronics) optical elements that could be placed on top of the 500-kV series capacitor platform, and they were connected to the electronics and data acquisition in a remote-control house at ground potential using optical fibers.  This was a major safety benefit as the substation ground grid itself would have significant voltage rise during the staged fault in the substation. 

Figure 7 shows waveforms measured during one of the stage faults, showing MOV voltage, MOV current, MOV energy (calculated from instantaneous voltage and current waveforms), as well as fault current.  Wide frequency band and dynamic range of the optical sensors are apparent from the waveforms.

Figure 8 shows an example of a mobile 500 kV class optical VT used as a reference for calibrating 500 kV CVTs (Capacitive Voltage Transformers) and to measure harmonics on HV lines without requiring an outage.  A typical VT or CVT pulls about 0.5 A to 1 A current (inductive or capacitive current) from the high voltage line.  The optical VT used here takes less than 1 mA from the HV line, making it possible to be used for live connection to HV line (or bus) without melting (or welding to) the connection point.  It is also rated for permanent operation at 500 kV class (e.g., lightning impulse withstand level of 1800 kV, one-minute power frequency withstand of 800 kV, etc.), so it can be left connected for as long necessary in all weather conditions.  The optical VT is built into a trailer for horizontal transportation, with remote controlled hydraulic jacks to raised it vertically to connect to live 500 kV bus work in the substation.

Summary and Conclusion

Optical instrument transformers have been evolving for the past few decades.  They offer a number of attractive features and have been used mostly in niche applications until a few years ago.  With the growth of digital substations and introduction of digital interface standards such as IEC 61850-9-2, IEC 61850-9-3, and IEC 61869-9, we expect significant acceleration in the adoption of optical sensors for main stream power system applications.  

Various informative publications such as IEEE Std. PC37.241, IEEE Guide for Application of Optical Instrument Transformers for Protective Relaying, expected to be available in early 2018, will be helpful for the industry practitioners in learning and using optical sensors.


Farnoosh Rahmatian (Ph.D., P.Eng.) is a co-founder and the president of NuGrid Power Corp. He was previously a Director of Measurement Systems at Quanta Technology and a co-founder and Vice President of Engineering at NxtPhase Corporation.  His present focus is on optical sensors, synchrophasor systems, digital substations, and on-site live calibration of high voltage instrument transformers.  He is a Professional Engineer and a Fellow of the IEEE "for contributions to optical voltage and current sensors ..."  He is the VP of Technical Activities for IEEE Power & Energy Society (PES), active at PES PSRC and PSIM committees (with focus on optical sensors and synchrophasors). He is also active at CIGRE, IEC, CSA, and NASPI.  Farnoosh has over 80 technical papers and 11 patents with focus on optical measurement technology.

Power. Flexible. Easergy.
Let?s start with organization in protection testing