Distributed Electrical Sensing for Digital Substations

Authors: Philip Orr, Neil Gordon, Synaptec, UK, and Campbell Booth, University of Strathclyde, UK

Modularization, reduced outage times, standardization and interoperability of equipment, and reduction in copper and CO2 are just a few of the benefits that the roll-out of digitized substations will deliver using technologies developed in neighboring fields.
At the physical layer, the key enabling technology is of course fiber-optic communications, which forms the backbone of a digital substation and is the primary factor in delivering improved safety and reduced environmental impact.
While it is generally well-understood by our industry that single-mode fiber is the “gold standard” medium for long-distance low-loss communications, less well understood is the direct application of such fiber to instrumentation providing accurate measurement of primary electrical system parameters. However, since the early days of optical fiber in the 1970s, development of instrumentation techniques based on the medium has been pursued in parallel with the telecommunication applications. Notably, this has resulted in the techniques behind Non-Conventional Instrument Transformers (NCITs) which are now produced by a range of vendors and provide measurement accuracy and bandwidth far beyond those achievable by conventional measurement transformers, and which are now recognized as a core part of digital substation technologies. More recently, the power industry has been introduced to Distributed Acoustic Sensing (DAS) and Distributed Temperature Sensing (DTS) which, while not having direct application to digital substations, provide means for real-time temperature and vibration mapping of lines and cables using the embedded fiber and thereby support fault location and dynamic rating of conductors.

In this article, we will discuss certain remaining challenges in the fields of substation automation and digitization, and intra/inter-substation instrumentation, which we believe can be addressed by further leveraging the optical fiber which has become critical to the power sector. We will introduce the technique of passive multiplexing of sensors using standard optical fiber and show how this can be used to support modern protection, monitoring, and control systems by addressing common features of certain contemporary power system challenges. We will discuss two key applications that serve to illustrate practically how these new techniques may be applied - in multi-zone or multi-terminal protection, and in acquiring wide-area synchrophasors for DER monitoring and fast response - and round off with information on how the technology is being trialed today in the UK.
The purpose of this article is transparently to convince the reader that while fiber optics may already be the backbone of modern substation communications, and the basis of new measurement techniques, the potential of this medium to advance our field - and the power of IEC 61850 - is yet far from exhausted.

Instrumentation Challenges
As the reader is well aware, IEC 61850 is a powerful and flexible industrial communication standard that enables interoperability and the sharing of data and signaling. It remains a hot topic in this magazine, in vendor brochures, and in the academic literature, and is constantly being refined to suit an evolving industry. As touched upon earlier, it brings many benefits to substation architecture – primarily in the areas of cost, environment, and flexibility - but it does not alone solve all immediate challenges. For example, to some extent it reduces the wiring required to instrument substations, and to some extent it permits the sharing of time-synchronized samples over distances, but certain limitations of the platform are apparent. The following are examples of challenges in the realm of protection, automation and control that digital substation technologies do not yet readily address to the satisfaction of network operators:

  • Cost barriers to implementing wide-area synchrophasor systems, which are so significant that despite the well-articulated benefits - roll-out is being restricted
  • Cost and inefficiency of multi-terminal and multi-zonal protection schemes and wide-area backup protection
  • Inaccuracy and cost of line parameter calculation for real-time dynamic ratings and fault location
  • Cost barriers to fast, granular fault detection on mixed overhead/underground circuits
  • Cost barriers to discrete faulted section identification on linear multi-section circuits such as turbine arrays

A common theme amongst these challenges is the cost and complexity of instrumentation and control networks, particularly over extended distances or when the number or topology of instruments (i.e. sensors and actuators) is nontrivial. The key point is that while sampled values are a convenient way to structure and share data on a digital communication network, they require conventional telecommunication and LV infrastructure to be in place to support local active devices at each instrument.
This architecture does not lend itself to certain applications which are coming to the fore, not least of which is the integration, monitoring and control of distributed energy resources.
Therefore, building upon the 61850 bedrock we are now laying down, new technologies must be developed and explored that extend the reach and scope of this platform.

Distributed Electrical Sensing

Distributed electrical sensing is a fiber-optic instrumentation technology that originated primarily at the University of Strathclyde (Glasgow, UK) during the 2000s, and was further developed for use in power systems by Synaptec in the UK with the support of transmission system operators SSE and ScottishPower. Originally developed for measurement of voltage and current on remote subsea plant such as pumps, the technology now underpins a range of new PAC products in transmission, distribution, and offshore renewables.
The key novelty of this technology is the decoupling of instrumentation (including both sensors and actuators) from telecommunications. This is achieved by enabling passive - i.e. power supply free – instrumentation networks which can be spread over very wide areas. Once this proposition is clearly understood, a reader versed in PAC will immediately start to consider exciting new PAC topologies and applications.
The technology utilizes the ubiquitous optical fiber sensor, the fiber Bragg grating (FBG). FBGs are periodic perturbations of the refractive index along ~10 mm of fiber core. This small, permanent perturbation acts to produce a strong optical reflection at a specific wavelength, known as the Bragg wavelength (denoted λB).

This wavelength is proportional to only two parameters: the average refractive index of the grating, n, and the period of the grating, d; a relation that is usually expressed as λB = 2nd. In other words, a wavelength twice the optical period of the grating will be strongly reflected: an FBG is simply a mirror that reflects only a single color of light. (Figure 1)
FBGs are commonly used as sensors for temperature or strain, since n is strongly proportional to temperature and d is, intuitively, proportional to longitudinal stretching or compression of the fiber. Thus, we arrive at an optical strain gauge that can be interrogated passively over the very low loss medium of optical fiber.
Finally, by fabricating multiple FBGs that reflect at different nominal wavelengths, we arrive at a passive, long-distance sensor array; and by analyzing the reflected optical spectrum from this array, each sensor can be interrogated discretely and simultaneously.

Using the same high-performance piezoelectric actuators as NASA’s Mars rovers, Synaptec has developed techniques to sensitize FBGs to both voltage and current in a way that retains the passivity, networking, and reach inherent to FBG-based sensing.
By converting the electrical quantity firstly to microscopic strain in the piezoelectric material and then to a wavelength shift, Synaptec’s core platform is able to measure and control at 50 points over distances of up to 100 km.
This PAC platform is compatible with a range of electrical and mechanical sensors and control modules which can be ‘mixed and matched’ to provide bespoke measurement or control functions for electrical networks of various topologies. Sensor types include AC and DC voltage (LV to HV), AC current, temperature, vibration, and static strain or tension – each measured simultaneously on a unified, time-correlated platform. Sensors are IEC 60044/61869 compliant and, of course, the merger or controller interfaces with other PAC equipment via IEC 61850. (Figures 2,3 and 4)
Importantly, since this instrumentation network is passive and analogue, much of the complexity of time synchronization between measurements is eliminated. In a passive, photonic sensor network, the time of flight of signals from the sensors to the merger is defined by the speed of light in fiber - 68% of c - which interestingly is about the same as the speed of a travelling wave in a conductor.
This speed varies slightly across the year as the ambient temperature changes, which can be calibrated by a simple photonic ‘ping’ of each sensor to well within an accuracy of one microsecond, thus eliminating the need for GPS or equivalent reference clock access at measurement points.

Let’s recap. What we have just described is a new instrumentation platform with the following unique feature set:

  • Uses standard single-mode telecoms fiber - often pre-installed
  • Has a measurement range of 100 km from the merging unit
  • Accommodates up to 50 sensors or controllers per fiber
  • Single-ended scheme - or double-ended for redundancy
  • All measurements time-synchronized
  • Measures voltage, current, temperature, strain, tension
  • IEC 61850 compliant

Given this, the question is, what can we do with this platform?

Key applications

Clearly, many potential applications exist for distributed passive sensor networks in the area of protection, automation and control, including:

  • Protection (main and backup) using distributed measurements - for lines, plant, and busbars
  • Fault location, identification and characterization through time-correlated electrical and mechanical measurements
  • Plant protection and condition monitoring using electrical and mechanical measurements
  • Relatively time-synchronized samples or phasor measurements without the need for satellite/GPS signals at each measurement location
  • Measurement-based calculation of line parameters for fault level and dynamic ratings

Here we limit the discussion to two key applications which we believe can address the challenges listed earlier.

Key application 1: Multi-zone differential protection for hybrid lines and offshore arrays: Hybrid or mixed circuits contain both underground and overhead sections. These circuits are often protected as a single zone, due to the cost and complexity of remote instrumentation at the overhead/underground interface locations, which generally leads to a restriction on the use of auto-reclose procedures to guard against reclosing on to an underground cable fault. This means that retrospective fault location methods (e.g. by time-domain reflectometry, analysis of fault records, or manual line inspection) may be employed to establish the fault location, with a risk that the fault was non-permanent on an overhead section and location and repair is not necessary. One solution to this problem may be found in distributed electrical sensing, since deployment of passive long-distance sensors avoids the need for dedicated cable fault detection schemes and eliminates the risk of reclosing on to a cable section. With the availability of discrete current measurements at any transition point up to 100 km from a substation, well-established differential current algorithms can be employed centrally to identify the faulted section and take appropriate action.
This has the immediate operational benefit of enabling auto-reclose functionality to deal with transient faults in overhead sections while intelligently blocking auto-reclose for underground faults. This unprecedented combination of discrete sensing, long-range and faulted section identification means outage time, asset damage, repair time and even reputational costs can be greatly reduced. (Figure 5)

The very same approach can also be taken in the case of distributed generation. Consider for instance an offshore wind turbine array: while protection and circuit breakers or other methods of fault isolation (e.g. using converters with blocking capability) may be employed in each turbine (generally at high cost to the developer due to requirements for power supplies and telecommunications), the intra-array cable linking a string of turbines is traditionally treated as a single protection zone, because the cost of discretely protecting each cable section is prohibitive. When faults occur along an intra-array cable, the entire string is isolated and must remain offline while time and revenue is lost through manual location of the faulted section. Deployment of distributed electrical sensing in each transition piece can provide an affordable means of monitoring each intra-array cable individually, quickly identifying a faulted section and allowing turbines to continue operation up to this point. This has the potential to significantly reduce loss of both generation and revenue, yielding significant financial and operational benefits with respect to outage time, asset damage, and resulting costs. (Figure 6)

One further topic to which multi-zone distributed sensing has clear relevance is in the protection of multi-ended circuits. Multi-ended circuits traditionally require multiple relays and dedicated communications schemes, typically introducing non-negligible latency and restricted bandwidth due to the use of digital communications between each circuit terminal.
Distributed electrical sensing has the potential to centralize and improve the robustness of multi-ended protection and monitoring schemes by introducing a single-location differential current instrumentation scheme, with one relay/IED gathering measurements from several locations via the distributed sensing scheme. Beyond providing near-instantaneous identification of faulted feeders or circuit sections, this scheme potentially enables the discrete protection and isolation of each line in a multi-terminal system (if suitable breakers/disconnectors are available), resulting in reduced outage area and loss of revenue in the event of a fault. (Figure 7)

Key application 2: Satellite-free wide-area synchrophasor aggregation for fast response: Enabling a fast-acting response to major power system events (e.g. events that have the potential to cause major frequency disturbances, instability and ultimately blackout) is becoming critical to stable grid operation, particularly in the future, where low-carbon, low-inertia, “weak” sources of generation are expected to become prevalent. Large-scale Phasor Measurement Unit (PMU) monitoring schemes are increasingly being utilized to enable new system functions such as fast-acting frequency control in low- and variable-inertia systems and, in general, highly distributed control schemes. The use of data from PMUs will therefore underpin the real-time operation of future power systems, and this is the subject of recent investigations by CIGRE and the IEEE Power System Relaying and Control (PSRC) committee.
Large-scale PMU monitoring requires time-synchronization to enable comparison of phase and frequency. Generally, this depends on use of the GPS satellite network for time-stamping by each PMU, which has inherent vulnerabilities and a need for supporting auxiliary power and telecommunications with each PMU, requiring significant investment. To mitigate these costs and risks, satellite-independent means of reference timing distribution to substations are being explored, for example the investigation of NPL Time in the UK by the National Physical Laboratory.

An alternative solution may be found in distributed electrical sensing, which allows a satellite-independent PMU merger to achieve wide-area or remote line end acquisition of PMUs. This approach makes use of the passive time-synchronisation capability of a distributed photonic sensing platform, enabling acquisition of PMUs in previously cost-prohibitive situations where PMUs and associated power supplies and communications are not readily available, e.g. on distribution systems or at certain DER locations. (Figure 8)
Leaders in this field are demonstrating regional aggregation of phasors to support fast-response systems in collaboration with National Grid in the UK.
Groups delivering fast response analytics and control systems are aware of the challenges in acquiring the underlying PMU data, and therefore are generally supporters of emerging instrumentation technologies that can assist.

Of particular relevance is distributed electrical sensing’s reduced and stable latency, lack of vulnerability to GPS synchronization issues, and reduced capital investment compared to conventional PMU deployments which may assist in overcoming cost barriers to the wider deployment of synchrophasor-based WAMPAC systems for wide area monitoring and provision of fast response when required.
Harnessing these benefits, the Synchromerger (i.e. synchrophasor merger) was recently developed - an instrumentation system that acquires time-stamped sampled values from multiple remote points (with no need for power supplies or dedicated communications systems at the remote measurement points) and integrates these into a single PMU stream at a central merging location.
This technology can also be used within the substation to facilitate dynamic real-time state estimation to detect any sort of abnormality within the substation, providing protection for multiple zones, and verifying the accuracy of system models and operating conditions.

Trial on GB Transmission Network
A major digital substation project was launched by ScottishPower in 2016 and will run until 2020. The Future Intelligent Transmission Network Substation (FITNESS) project will deliver the pilot GB live multi-vendor digital substation instrumentation system to protect, monitor, and control the transmission network using optical fiber to replace copper hardwiring. The objective of the project is to provide a live demonstration that the roll-out of digital substations will reduce the cost, risk, and environmental impact of transmission substations and increase flexibility, controllability and availability of the network.
As part of the FITNESS project, Synaptec, GE, and ABB are collaborating to test new instrumentation equipment, demonstrate interoperability between various 61850 devices, and prove the case for business-as-usual deployment of digital substation technologies and techniques. The role of distributed electrical sensing in this project is to provide instrumentation that extends the observability of the network and demonstrates passive and efficient sampled value provision inside and outside substation boundaries.
By demonstrating passive, aggregated measurements at local and remote line ends, one aspect of the project will seek to enable more robust differential current protection of hybrid overhead-underground circuits and the satellite-free acquisition of phase and frequency measurements from over a wide area, which when coupled with wide area monitoring (WAM) technology will potentially allow the transmission system to be operated closer to its operational limits.

Conclusion: As the power industry adopts the IEC 61850 design standard for substations, the integration of distributed sensors becomes a more natural fit for general monitoring solutions. Fiber is both a key component in digital substations, reducing copper wiring and improving safety, and also the primary enabling technology for secure, high-bandwidth, long-distance communications. Combining these two aspects enables an instrumentation platform that does not require power or digital comms at measurement locations, thus reducing the cost and complexity in delivering robust power system protection, automation and control.
The ability to protect complex circuits (multi-terminal, hybrid overhead/underground) using distributed measurements will be of great benefit in the future.
Furthermore, the distributed sensing technology permits data to be gathered from large areas to a single source with very low delay and without the need for dedicated communications, distributed time synchronization or PMUs (requiring local power supplies, and GPS signals), which in turn presents significant opportunities to implement wide-area, fast-acting monitoring, control and protection solutions. This will assist in allowing power systems to evolve to a low-carbon future, with massively distributed sources of power, while continuing to be operated in a safe, reliable and secure fashion.

Dr. Philip Orr is a founder and the managing director of Synaptec. He is an expert in photonic instrumentation for power systems and a member of the IEC SC86C working group on fiber optic sensors. Prior to founding Synaptec, Philip held Scotland’s first Royal Academy of Engineering Enterprise Fellowship. He remains an honorary staff member at the University of Strathclyde where he conducted his doctoral research on photonic instrumentation for nuclear fusion reactors. He is the recipient of a number of engineering prizes, including the Sir William Siemens Medal and the EPSRC Doctoral Prize, and the author of over 40 articles and 3 patents.

Dr. Neil Gordon is Lead Engineer at Synaptec, responsible for product development and engineering research. He has a strong background in photonics, having achieved his doctorate at Glasgow University in 2015 on advanced photonic engineering as a member of the international LIGO Scientific Collaboration that successfully detected gravitational waves for the first time. During his doctoral work, he was the recipient of the Thompson Experimental Prize for physics research and has co-authored over 70 papers in the diverse fields of gravitational wave engineering and astronomy and optical measurement technologies for power systems.

Prof. Campbell Booth is currently a Professor and Head of the Department for Electronic and Electrical Engineering at the University of Strathclyde and holds the position of Applications Director at Synaptec. He leads the technical team of the Power Network Demonstration Center (PNDC) in the UK and is on the organizing committees of the IET DPSP and PAC World conferences. He has published over 150 articles, engineering recommendations, white papers and several book chapters.

Let?s start with organization in protection testing