GB2601468A - High-resolution condition monitoring of transformers using micro-synchrophasors - Google Patents

Abstract

An apparatus for monitoring a transformer by collecting high-resolution measurement data comprises: a grid data unit having a micro-synchrophasor measurement unit (micro phasor measurement unit or micro=PMU) to operate in the frequency domain and a power quality monitor (PQM) to operate in the time domain, wherein the grid data unit applies the same synchronised timestamp to the collected data points from both signals. The transformer may be an auxiliary transformer on an electricity distribution site e.g. a solar farm. A signal from a shared antenna e.g. a GPS receiver may be used for the time stamping. The transformer may supply single phase electricity to equipment at the distribution site.

Description

High-Resolution Condition Monitoring of Transformers Using Micro-Synchrophasors Six utility-scale solar farms in England are being equipped with a new type of high-resolution condition monitoring apparatus to enhance their performance, reduce operating costs, avoid outages, and extend service life. By collecting exceptionally accurate voltage and current phasor measurements using newly devised micro-synchrophasors (i./PMU or microPMU) plus a wide range of power quality, supraharmonic and environmental measurements, the condition and function of solar power generation transformers can be scrutinized in unprecedented ways and resolutions.

This condition monitoring data enables the effects from three specific phenomena on transformers to be studied in exceptional detail: a) harsh load factors caused by daily solar radiation profiles, b) significant harmonic currents injected by solar power inverters, and c) the diurnal power flow reversal and associated thermal loading cycle.

A new type of high-resolution condition monitoring apparatus is being installed on six utility-scale (>1 MW) solar farms in England. The goal is to enhance solar farm performance, reduce operating costs, avoid costly unplanned outages, and extend operating life. Power transformers are major component of any large solar farm and the focus of this study. C\I

C\I By using the above-described GDU collecting exceptionally accurate voltage and current phasor measurements using newly devised micro-synchrophasors (pPMU or microPMU) plus a wide range of power quality, supraharmonic and environmental sensor measurements, the condition and function of solar photovoltaic (PV) power generation equipment can be scrutinized in unprecedented ways.

The effects of three specific phenomena on transformers to be studied in exceptional detail; first the harsh load factors caused by daily solar radiation profiles (see Figure 1) [1], second the significant harmonic currents injected by close coupled solar power inverters, and third the diurnal cycle.

Encouraged by attractive British government financial subsidies, during 2011-2017 over 8GW of utility-scale (>1MW) solar capacity across approximately 1,200 sites was hastily constructed. Their average capacity is 5MW. This has resulted in an estimated UK fleet of 3,500-4,000 grid connected 11/33kV:0.4kV transformers in the 1.0 to 3.0 MVA size range. Solar farm development is cost sensitive and these transformers are significant capital cost item. To meet subsidy deadlines and minimize capital costs, it is understood that low-cost, standard distribution system transformers were commonly installed, with little consideration given to correct dimensioning for the expected operating conditions.

It is suspected these transformers are suffering statistically significant rates of unrevealed incipient faults. Although fleet failure rates are hard to come by, there are anecdotal reports of relatively young transformers failing prematurely. Sometimes damaging, even catastrophic fires have occurred with associated unplanned outages and replacement/restoration costs.

a) Increase our understanding of how solar power generation profiles, harsh operating environments, and poor equipment specification may contribute to premature solar-farm transformer failures, and how the early indicators of these failures may be identified, classified, prioritised and applied [2, 3].

b) Investigate the supraharmonics injected by large solar inverters [4], and how these might affect transformers undergoing large, frequent load swings.

c) Investigate the diurnal power flow reversal and associated thermal loading cycle effects on power transformer operation and failure modes.

The Grid Data Unit (GDU) measurement and monitoring apparatus consists of electrical signal acquisition, environmental sensing, signal analysis, data telemetry, and data storage.

High-resolution, medium-voltage data collection systems are being installed on six utility-scale, ground-mount, central-inverter solar farms across England (see Table 1 below). Each Grid Data Unit (GDU) consists of: a micro-synchrophasor (microPMU) operating in the frequency-domain; a certified IEC 61000-4-30 Ed 3 Class A power quality monitor (PQM) operating in the time-domain; an array of ambient environment sensors; and ultra-precise GPS antenna that enables sub-100 nanosecond timestamping; 256GB data-buffering solid-state memory; and secure 3/4G LTE cellular data telemetry system (plus provisions for 5G).

Underlying analogue to digital conversion occurs at a sampling frequency of 4MHz which is internally processed into 512 samples/cycle; this can be down-sampled into data-rate required.

In addition to micro-synchrophasor measurements, the system collects a large set of power quality measurements plus harmonics to the 50th order, inter-harmonics, and supraharmonics (2-150kHz). Transient events are captured at 1 or 4MHz by trigger settings C\I on one of 40 digital oscilloscope channels available. Ambient environmental data (i.e. C\I temperature, relative humidity (RH), barometric pressure, precipitation, wind speed, irradiance, etc.) plus other factors can be recorded. Ambient environmental sensors may be O attached to the GDU or one or more of the operative pair whichi s described below. Such sensors measuring interior to the enclosure and/or exterior (outdoor) temperature, relative humidity, barometric pressure, precipitation, vibration and or seismic activity. Additional o sensors may determine wind speed and air turbulence (e.g. rotating cup or ultrasonic type anemometer, LIDAR, etc.) near the ground or aloft. Further sensors may determine solar irradiance across various spectral bands (IS, visible, and or UV). Further connected sensors may determine various properties of air quality such as aerosol content, particulates, or gaseous species content. Additional detectors and sensors may be attached to monitor radiation (e.g. Geiger counter type device) or cosmic rays (e.g. a muon detector.) Further sensors may involve ultrasonic, laser-based or electromagnetic radar emitters, receivers or transceivers. In some examples, the environmental sensors are connected to the data recording with timestamp capability of the GDU.

The GDU itself is condition monitored with internal temperature, humidity, and vibration sensors plus condition status relays on the key DIN rail mounted components in the panel box assembly.

GDU sensor boxes from Neuville Grid DataTM uniquely combines newly devised microsynchrophasors (pPMU) with power quality monitors (PQM) wherein the two devices share a common timestamping method to facilitate effective combination and comparison of collected data points. In some example, the GDU further includes a metal enclosure; ambient environmental condition sensors; sub-100 nanosecond timing; and next-generation secure telemetry via redundant/alternate communications propagation paths. The collected, transmitted, stored and delivered electrical measurement data features 100-100,000x improvement on the state of the art. In the frequency domain, the micro-synchrophasor function provides 0.001 degree phasor angle accuracy giving voltage and current measurements accurate to 2 parts per million (PPM). In the time domain, the power quality function provides an array of high-accuracy measurements according to the IEC 61000-4-30 Ed 3 Class A standard plus supra-harmonics in the 2-150 kHz range. Data-points are time stamped with global navigation satellite system (GNSS), such as GPS, derived fimestamp with an accuracy of 50-80 nanoseconds depending on geo-atmospheric conditions relative to the observed GNSS satellite constellation. A more effective and efficient arrangement of components within the apparatus in relation to a shared antenna array and environment sensors is provided. The GDU also incorporates a multi-core, multi-thread capable, coprocessor for local-node, comparative paired-node, mesh-network, fog-computing, or edge-processing data analysis processing. Neuville's GDU hardware design specifically but not exclusively improves upon the GridAnalyzer TM device from Power Standards Lab of Alameda California.

The GDU may comprise one or more of the same type of physical devices operating in different firmware modes (e.g. the PQube3 (Registered Trademark) device from Power Standards Lab of Alameda California). In some examples, the devices may be different types of signal analysers that may or may not be from the same manufacturer. In some examples, fundamentally different methods of electrical signal processing can be implemented in a single physical device featuring capable sub-functions within its firmware or via further software processing on a separate computation device. Any number of physical devices with C\I the same or different functionality may be provided. C\I

Newly devised pPMUs from PSL such as the PQube3 operating in a pPMU mode, for example, are able to provide high accuracy (0.003°) phasor angle (5) measurements of O current and voltage every half-cycle. Small, compact and robust, PQube3 pPMUs offer 333x accuracy improvement at roughly 1/10th the cost of a standard PMU. The 333x typical accuracy improvement make pPMUs particularly useful to the much larger market of MV power distribution grids. The PQube3 device, for example, can switch firmware modes and collect power quality measurements (PQM) at 512 samples per cycle (26/31kHz) including waveforms and 50th order harmonics. This is a 50,000x improvement on existing utility SCADA systems. Under PQM mode, it also supports a 40-channel digital oscilloscope capturing disturbance events at up to 4 MHz; 2,000 parameter triggered meters sampling at 2Hz; collection of supra-harmonics; and a revenue grade energy meter. All data is GPS time-stamped and the measurements thus synchronized across distant points to sub-100 nanosecond accuracy -a 100,000x refinement of the temporal dimension over existing SNTP based systems. This permits an enlarged and enhanced range of powerful network analysis techniques and cost-saving end-use applications. On-board analysis and reporting tools make even a single pair of pPMU potent electrical instruments. When deployed across a power grid, a networking effect comes into effect. The economic value and technical impact of the pPMU network will grow in a nonlinear exponential fashion according to a lessor variant of Metcalf's law.

The GDU may comprise an operative-pair of PQube3 devices with one PQube3 device operating in a micro-synchrophasor mode and the other PQube3 device operating in the PQM mode. In some examples, the operative pair may further comprise additional signal analyzing devices. In an example, the operative-pair processes an electrical source signal simultaneously in the frequency-domain via the micro-synchrophasor method and in the time-domain using power-quality measurement techniques to collect time-domain data points and frequency domain data points. A congruent timestamp is applied to the collected time-domain data points and frequency-domain data points. In some examples, the timestamp applied to the collected time-domain data points and frequency-domain data points is derived by the same method. In an example, one or more operative-pairs within an apparatus and/or further adjoining apparatus may process different source signals drawn from instrument transformers' secondary outputs or directly sensed.

Incorporating additional collocated operative-pairs of signal analysers into the apparatus facilitates cost-efficient sharing of a common power supply arrangement, control signals, Global Navigation Satellite System (GNSS) antenna/signal, anti-tamper/anti-theft protections, and/or telemetry arrangements. Sharing the same GNSS (GPS) antenna for example saves: procurement, installation, and maintenance costs; wind-load, weight & space on the antenna array mast; etc. It also provides exactly the same timing signal which ensures chronological unity of measurement (i.e. perfect time alignment). Prior solutions using two adjoining GNSS (GPS) antennas will normally provide slightly different results caused by instrument drift and other subtle differences leading to timestamp divergence.

In some examples, the operative-pair may be implemented with appropriate compensation, as logical or functional pairs over a communications network but physically separated by a noticeable distance between the emplacements of each device forming an operative-pair.

C\I Such a capability allows the dynamic compensation for a failed device or adaptive C\I configuration of widely emplaced devices to suit functional objectives or to flexibly meet changing conditions.

O In one example illustrated in Figure 3, a method 200 is provided for monitoring a power grid and collecting high-resolution electrical measurement data comprising: collecting 101 a first o set of data points using a micro-synchrophasor operating in the frequency-domain; collecting 102 a second set of data points using a power quality monitor operating in the time-domain; and applying 103 a synchronised timestamp to the collected first and second sets of data points. The same synchronised timestamp will be applied to both sets of data to form a time synced array of multiple measurements derived from two fundamentally different approaches.

The examination and processing of the same electrical signal by two difference means (frequency domain synchrophasor and time domain power quality) facilitates the use of two different classes of analytical techniques. The combination of the two processed signals facilitates the use of a broader range of established methods, new or novel techniques, and as yet un-devised analytical methods.

The provision of a time synched array of multiple measurements derived from two fundamentally different approaches creates analytical opportunities for further development and exploitation. An analytical example being to derive the presumably same or similar frequency, voltage, and measurement values via both methods and check for reasonable agreement or deviation of the results within a statistical tolerance. A simple lack of agreement or divergence/cycling over time in agreement being indicative of an anomaly worthy of further examination or correction.

In an example, the operative-pairs of devices are functionally combined into a single unitary device. This placement and integration may use integrated circuits and other components onto a single circuit board or reduced number of circuit boards featuring shared, but not limited to: power supply, memory storage, telemetry modem, input/output connectors, protection features (e.g. fuses) plus environmental, anti-tamper, or self-condition monitoring sensors (e.g. relative humidity and temperature to detect condensing moisture within the enclosure and over-temperature operating conditions). Such a unitary design could be configured with an appropriate connector(s) for slotted rack configuration and/or easy replacement. In some examples, the apparatus may include additional signal processing devices such as an oscilloscope, and these devices may be directly connected or remotely operated to facilitate the wired or wireless monitoring of measured signals. Such additional devices may be attached via a connector or coupler that passes through the exterior of the enclosure to integral electrical leads and probes pre-installed in the enclosure. Such device can be provided with electrical power from the GDU's power supply via a socket or plug that also passes through the enclosure wall. Such connectors and plugs may be protected with caps and or a hinged cover when not in use. Implementing the probes and connection leads within the apparatus removes the need to compromise the integrity of the device when additional signal processing functionality (e.g. an oscilloscope) is required, as there is no need to make new connections through the enclosure.

In an example, the operative-pair may perform the functions of an energy meter based upon C\I calculations via either its power quality or micro-synchrophasor measurement capabilities. In C\I some examples, both the power quality monitor and the micro-synchrophasor measurements are compared to check for reasonable agreement to give further assurance the reported amount of energy generated, transmitted, delivered or consumed is correct and accurate to within a given statistical tolerance of accuracy. It also provides redundancy and diversity in the event one device or method fails or becomes unavailable.

In some examples, the apparatus may also contain a shared power supply, data storage, and telemetry (modem) equipment. In an example, each operative pair has its own separate and not shared power supply, data storage and telemetry equipment. An enclosure box comprising the apparatus may be arranged with single power supply arrangement that supports multiple operative-pairs. An enclosure comprising the apparatus may be arranged with single telemetry and antennas array arrangement that supports multiple operative-pairs. Where equipment is shared, cost savings and simplicity of design results. In some examples, redundancy and diversity of systems will be implemented for avoidance of single points of failure in the system architecture.

In some examples, acoustic sensors attached to or otherwise positioned to listen to monitored equipment (e.g. transformers, inverters, motors, rotating machinery, etc.). In an example, the acoustic sensors are connected to the data recording with timestamp capability of the GDU.

In an example, the apparatus components are connected to one or more antenna and may include signal processing/converters for more than one global navigation satellite system (GNSS), e.g. the American Global Positioning System known as GPS. The GNSS antenna receiving high-precision locafional and timing information. Available GNSS include but are not limited to: American Global Positioning System known as GPS; Russian Global Navigation Satellite System called GLONASS; China's BeiDou Navigation Satellite System; the European Union's Galileo system; India's NAVIC; and Japan's Quasi-Zenith Satellite System. Alternate timing and locafional information could come from other radio-navigation systems or a clock internal to the enclosure of sufficient stability, precision and synchronization. In some examples, multiple GNSS inputs are used to provide redundancy and diversity of systems. Continuous monitoring of mutual agreement between the multiple GNSS inputs supports detection of divergence or other discrepancy indicative of an anomaly, problem, failure, malicious tampering, or cyber-attack on one or more GNSS either globally or with local/targeted effects.

In some examples, mounted on one or more elevated masts is an array of antennas associated with GDU operative-pair functions such as telemetry communications and the obtaining of GNSS timing and locational data. In some examples, this includes one or more of each of the following types: GNSS (e.g. a modified GPS antenna able to obtain high-precision timing signals); omni-gain and directional cellular (e.g. 2/3/4/5G), satellite (e.g. lmarsat or low-earth orbit (LEO) constellation); directional microwave communications link; omni and or directional (e.g. yagi type) radio-frequency (RE); directional light-wave carrier; etc. In an example the antenna array is provided with protective lighting arrestors and an enclosing protective shroud that is transparent to the relevant spectra (e.g. shroud made of a plastic (e.g. polypropylene, ABS or PVC) or fiber reinforced plastic (GRP fiberglas) composite). The protective shroud protects the antenna array components from C\I environmental degradation and deters theft. Any powered masthead devices (e.g. (\I transceivers, emitters and receivers) can be provided electrical power via either mains-supply or by one or more power-over-ethernet (PoE power supply) electrical supply O connections alongside network cabling connections. Environmental sensor mountings for irradiance, wind, temperature, humidity, precipitation, barometric pressure may be provided. The environmental sensors in some examples form a unitized sub-assembly. The antenna o mast and array can also provide a vantage point for mounting one or more security cameras or other security surveillance sensors. In some examples, the mast accommodates a picocellular communications transceiver, other telemetry-capable antennas, and/or Wi-Fi, VVIMAX, LoRa and other LAN antennas. In some examples, mast-head omni-directional antennas and one or more directional antenna(s) are geographically oriented toward nearby communications relay/downlink tower(s) or other correspondent transceiver antennas with the strongest signal or preferred telecommunications provider.

In an example, the enclosure of the apparatus is made of metal or having a conductive coating or similar material property create a Faraday cage around the apparatus to exclude radio-frequency interference (RFI), electro-magnetic interference (EMI) or similar radiated or conducted emissions (particularly those associated with alternating power circuits with a grid frequency of 50/60/400 Hz) from entering the enclosure and hampering or otherwise interfering with the proper operation of the equipment contained within. Wires or other conductive pathways into the enclosure may be protected by appropriately conductive glands/grommets and encircling ferrite beads or other effective means. By making the enclosure out of metal or other material with high thermal conductivity, internally generated heat can be more readily dissipated via radiance and convection. Retained heat build-up and resulting excessive interior temperatures within the enclosure can hamper proper GDU operation and degrade components such as capacitors leading to faulty operation, stoppage or shortened service life. In some examples, the enclosure is made of steel or aluminum. 6.

Metal enclosures are durable; generally having longer service lives from being more resistant to degradation from ultra-violet light and other ill-effects. Metal enclosures also feature a much higher degree of fire protection than plastic ones. The metal enclosure also has a high degree of physical intrusion protection against dust, water and probing tools or fingers. Existing methods of enclosing micro-synchrophasors make use of polycarbonate plastic boxes which can be less expensive and easier to modify but are susceptible to EMI/RFI problems and heat build-up.

In an example, the apparatus has an enclosure having a removable lid or cover that may be hinged. In some examples, the enclosure has an integral locking device or hasp lock. In some examples the enclosures have mechanisms that detect, record and alert by data transmission a central operations centre when the enclosure is opened, moved or tampered with. For cabinets, enclosures or other structures containing electrical devices like microsynchrophasors and power quality monitors or other equipment -anti-tamper security features like photo-cell and environmental sensor (temp, RH, barometric pressure, tilt, vibration) that detect the opening of the enclosure, its molestation, and or movement. Triggers may be internally mounted to the enclosure or integrated onto a circuit board in such a position that a camera that images a scene out of the opening -presumably taking an image of the person or persons who opened the cabinet. Learning software able to detected departure from normal patterns and send alert by one or more means of data communication plus making of an annotation in recorded data files is implemented in some examples. In an C\I example, this system includes the performing of physical intrusion detection of the enclosure: C\I opening, GPS location movement, abrupt change of enclosure's internal temperature or humidity conditions, photocell detection of enclosure opening, tilt/jarring, etc. CD In some examples, the apparatus utilizes distributed mesh fog hive data processing, where each node of operative-pairs is able to independently process data locally and/or o cooperatively in conjunction with other GDU nodes and optionally other data stores. Such a system facilitates the sharing between topologically neighboring nodes of analytical processing tasks or results, such as local state determination. Co-processor with a single or multiple processors featuring single or multiple: computational cores, graphics processing units (GPU), float-point processing units (FPU) can be used. Such co-processor(s) may or may not be capable of multi-thread processing. Such a meshed, edge-processing or fog computing capability can perform a variety of processing and analytical functions related to GDU gathered grid data, equipment condition monitoring inputs, or on other information to serve non-electrical purposes needing computational capacity.

Existing relational databases cannot efficiently handle the quadrillions of data-points amounting to petabytes streaming from a network of Neuville GDUs. A time-series data-base (TSDB) such as that referred to as the Berkeley Tree Database is therefore incorporated and modified into a novel time-series matrix data-base (TSMDB) to provide a low-cost, highly-scalable, Cloud-based or private server-based solution. The base Berkeley Tree Database offers a 1,400x improvement on existing commercial methods In known time-series matrix databases (TSBD), each recorded value (scientific measurement, financial trading quantity, etc.) is paired with timestamp or timing mark from an epoch starting reference into "tuple" consisting of the measured value or assigned value paired with a temporal or chronological value. Neuville improves upon this data-array architecture by incorporating multiple measured values matched to a single timestamp in single row of entries; not a tuple but a multi-tuple. As part of the larger Neuville apparatus, this ensures the aligned recording (qualitative assurance) of multiple simultaneous measurements (e.g. voltage, amperage, and frequency) from a single or multiple electrical source signal(s) with a unifying and harmonized timestamp. Thus, forming a time-series matrix database (TSMDB) data structure that is more storage space efficient, faster, and more suitable to handling and analytical processing of time aligned measurements or quantified values or condition states (e.g. switch position).

In known time-series matrix databases (TSBD), each recorded value (scientific measurement, financial trading quantity, etc.) is paired with timestamp or timing mark from an epoch starting reference into "tuple" consisting of the measured value or assigned value paired with a temporal or chronological value. Neuville improves upon this data-array architecture by incorporating multiple measured values matched to a single timestamp in single row of entries; not a tuple but a multi-tuple. As part of the larger Neuville apparatus, this ensures the aligned recording (qualitative assurance) of multiple simultaneous measurements (e.g. voltage, amperage, and frequency) from a single or multiple electrical source signal(s) with a unifying and harmonized timestamp. Thus, forming a time-series matrix database (TSMDB) data structure that is more storage space efficient, faster, and more suitable to handling and analytical processing of time aligned measurements or quantified values or condition states (e.g. switch position). C\I

C\I As shown in figure 4, the TSDB system may be built upon the Berkeley Tree Database BTrDB 200 to provide the following features: - Uniquely it runs 1,400x faster than the best commercially available solution for handling time-series data; - Can collect and store multiple concurrent high-bandwidth, unordered data streams - Achieves 2.9 compression ratio with a demonstrated throughput of 53 million inserts and 119 million queries per second.

- Can handle 1,000 pPMU sensor nodes with a single server; - Can locate a handful of voltage sags among 3.4 billion data points in under 200ms; - Easily implemented on easily scaled, standard Amazon web servers - Plus adaptable to deftly handling electrical, financial, loT, process, environmental, and many other types of time-marked data.

For electric grid data, this novel TSDB structure permits mulfiscale data analysis, enabling prediction and anomaly detection across the range of voltage, current, and time scales that affect transmission & distribution grid performance.

Hyper-efficient data-engine requires less than 5,000 lines of code. Neuville is commercializing it with an Application Programming Interface (API), secure telemetry linkages, revenue mechanisms, access controls, and other proprietary enhancements.

Most TSDBs are limited to millisecond precision and are therefore unsuited to synchrophasor data. Most are also not well suited to the enormous amounts (petabytes holding quadrillions of datapoints) of electrical grid data Neuville intends to handle. BTrDB provides a solution to both of these significant challenges.

In some examples, the apparatus comprises an integrated system consisting of electrical sensors, operative-pairs of signal analyzers in enclosures, APN plus VPN secured telemetry, a time series database that ingests, stores and retrieves collected time-series data.

The time-series data-base implemented in the system may feature an application programming interface (API) that facilitates interaction with third-party software, systems and hardware devices; revenue gathering and recording software and database mechanisms; access security; and generally supporting Data-as-a-Service method of disseminating grid data collected and stored. The telemetry package provides secure, confident transmission of power system measurements.

Software and related communications plus data-structures and security features permits the remote or local configuration and management of the GDU signal processing functions and device operative-pairs, plus data-storage handling, security functions, self-conditional monitoring, and telemetry. In some examples, the software and related features are able to monitor and manage a fleet of equipment across a wide geography, and in an example, watches for and provides alerts to any tampering detected with operative-pair configuration, firmware, stored data, data-storage mechanisms, analytical results, control functions, C\I messaging, security functions, enclosure, antenna array, and/or telemetry settings. C\I

The operative-pairs of micro-synchrophasors and power quality monitors may adhere to and O perform firmware functions/data formatting in conformance with IEC or other technical standards. This makes possible technical interoperability among other open-standard adhering software, firmware, systems and hardware.

CD Such an apparatus should allow condition monitoring and life remaining estimation methods normally used only on high-value high-voltage transmission grid transformers to be brought down and cost-effectively applied on relatively low-value, medium-voltage distribution grid tied transformers.

The PQM is able to collect an array of high-accuracy measurements which allows investigating as to whether, and if so how, supraharmonics interact with and affect close coupled solar power inverter-transformer behaviour [5]. It will also collect supraharmonic data indicative of conducted emissions transiting the distribution grid.

Defined as any type of waveform distortion of voltage and current between 2 and 150 kHz [6], supraharmonics are a relatively new realm of electric power system analysis. Power electrics like solar PV inverters and even common household appliances [7] commonly give off conducted emissions in the supraharmonic range. Only three original equipment manufacturers (OEM) are known to the authors as having supraharmonic specific detection equipment on the market.

The micro-synchrophasor (microPMU or pPMU) device features a better than two order of magnitude improvement on state-of-the-art conventional Phasor Measurement Units (PMU) synchrophasor measurements.

Frequency domain PMUs have been collecting synchrophasor measurements for about 30-years. They are used by high-voltage transmission systems to monitor bulk power flows with an angular accuracy of ±1.0° under the IEEE/IEC 60255-118-1-2018 (previously 037.118.1) standard.

Micro-synchrophasors bring PMU capabilities and more to noisy distribution level systems [8]. They feature a total vector error (TVE) of ±0.01% and typical operating angular accuracy of ±0.003° in the measurement of voltage and current phasor angles; enabling voltage and current measurements with an amplitude resolution of 0.0002%FS or two parts per million (2 PPM).

The result is half-cycle current and voltage phasor angle RMS values for all three phases timestamped to sub-100 nanosecond accuracy.

It has been field demonstrated in California, that micro-synchrophasors can enable the detection of unsuspected distribution utility transformer oil leaks at a distance of 20km by tap-changer electrical signature analysis. Various methods for detecting similar events using pPMU data are described in the literature [9, 10].

Utility-scale English solar PV facilities are rarely of the same design, but Figure 3 depicts a representative 5MW installation that includes three 1.5MVA step-up transformers plus three 10kVA auxiliary transformers on an 11kV grid connection.

Table 1: Characteristics of first six solar farms being studied The twenty 11/33kV power transformers across the farms include both cast-resin (4 of 20) and oil-filled (16 of 20). As typical of British solar farms, all of the transformers in this study are aluminium wound, Delta/Star configured, Dyn11 type units but these may not be limited to this. None are British made. All are nominally Oil Natural Air Natural (ONAN) but most units have been placed inside small enclosed spaces close by their paired inverter with fans drawing in outside cooling air through simple dust filters.

Except for the Taunton site which uses 15kW string inverters and incidentally has the largest transformer (2200kVA) being studied, the associated grid-tied inverters are all of the large (>500kVV) central type.

Peculiar to solar farms, the transformers undergo a diurnal flow reversal twice a day, as the solar farm transitions from generating power dawn-to-dusk to over-night consumption of grid supplied power. The instant and long-term effects of this diurnal flow reversal and Site Name County Connection inverter Commissioned DNO kV Type Capacity kWpDC MV Power Transfomwm Fen Road Lincolnshire Jul-14 WPD 11.0 Central 1,441 2x 340V:11kV 1250kVA oil cooled Kenninghall Norfolk Mar-15 UKPN 33.0 Central 8,000 5 x 400V:33kV 1400kVA oil cooled Langford Pond Farm Taunton Wroxton Bedfordshire Norfolk Somerset Oxfordshire Mar-15 UKPN 33.0 Central Jul-14 UKPN 11.0 Central Mar-15 WPD 11.0 String Mar-14 WPD 11.0 Central 4 x 400V:33kV 1800k VA oil cooled 2 x 400V:33kV 1400k VA oil cooled 2 x 400V:33kV 1250k VA oil cooled 2 x 360V:11kV 1650k VA cast resin 1 x 400V:11KV 2200k VA oil cooled 2 x 360V:11kV 1650k VA cast resin total 29,750 20 MV transformers ID, associated thermal loading cycle on the power transformers deserves is part of the study. The power flow reversal data will also serve smart grid analysis and operation.

Power is required 24/7 to maintain inverter readiness, keep transformers energized, support SCADA monitoring equipment, power security surveillance cameras, etc. To allow for daytime export of power and overnight import, a site's "connection agreement" with the local distribution network operator (DNO) has vastly different limits on export vs import. The Langford solar farm for example has a 10.0 MVA export connection and just a 0.060MVA import connection (a ratio of 167:1). The overnight grid supplied energy imported costs 2-3x the value of the exported solar energy. This brings attention to minimizing transformer low load losses.

The six sites have points of connection to the 50Hz distribution grid at 11kV (4 of 6) and 33kV (2 of 6). Smaller sites are usually on 11kV; larger capacity sites being on 33kV. High-voltage transmission grid connections for English solar farms remain exceedingly rare. Of the six sites selected, half are in the Western Power Distribution (WPD) and half in the UK Power Networks (UKPN) DNO licence areas. More sites may be added to the study.

Each farm also has one or more, small typically 10-50kVA, step-down 11/33kV: 0.240kV auxiliary transformers supplying single phase mains power onsite. These auxiliary transformers can act as a "house mains" to a solar farm, for example, and provide onsite power to the GDU, and other onsite equipment such as supervisory control and data acquisition (SCADA) systems, etc. Particularly when supplying the customer substation with C\I power, these often-overlooked auxiliary transformers can be the site's Achilles' heel if they C\I fail. Auxiliary transforms might fail because they are provided with power that is tapped from the three phase power. The power provided to the auxiliary transformers is typically of a low quality and susceptible to load swings which may cause the transformer to burn out and fail.

o Once any backup battery power is exhausted, the SCADA system and other onsite equipment go offline and the site has an outage. Some sites do have pre-installed 240V receptacles for temporary use of portable generators. A novel feature is monitoring the 240V o mains power supply by connecting the GDU to the 240V site mains power supply (see figure 2). This allows specific monitoring of the auxiliary transformer output characteristics which helps for ongoing analysis and monitoring of the auxiliary transformers. This may help to predict and/or prevent failure of these transformers which are typically low cost due to a rush to meet subsidy deadlines and minimize capital costs. Further, little consideration was given to correct dimensioning for the expected operating conditions so the auxiliary transformers are not always suitable for the site and operating conditions of the site. Failure of these transformers may therefore be high which can result in power failures for onsite equipment such as SCADA as well as result in onsite fires. Operations of the farm may therefore be reduced or stopped by transform failure.

GDU installation is somewhat similar to a medium-voltage revenue meter. The equipment is installed in the customer substation or inverter station or even directly on a transformer. Ultra-precise, split-core current transformers sense the 120V/5A secondary outputs of revenue-grade 11/33kV current transformers (CT), while voltage transformers (VT) on all phases of the export cables provide proportional signals to the operatively-paired pPMU and PQM. Where possible protection CTs are also monitored for a wider dynamic range. The CTs and VTs being of the classic induction type; Rogowski coils will not work with microsynchrophasors. Supplemental High-Frequency Current Transformers (HFCT) with a dynamic range reaching 10MHz are planned. Hall effect DC sensors on the combiner box busbar are also employed. Subsequent installations may adapt bushing mounted voltage and current sensors.

In the example above, the GDU is attached to a single phase power signal for monitoring purposes but it can also work with dual-phase, three phase or just two phases of a three phase system.

The established means of attachment may be the emplacement of current and voltage instrument transformers such that they sense the target current and provide a proportionate signal via their secondary outputs. Revenue grade induction type instrument transformers are best but Rogowski coil and other types can be used in certain instances. Protection grade instrument transformers can also be utilized with somewhat reduced performance. For low voltage situations direct sensing of voltage can be applied. For medium and high-voltage applications, an intermediate instrument transformer is emplaced or in the case of existing installation adapted such that instrument transformers' secondary output leads carrying the proportionate signal (typically 120v and 5A) are sensed by GDU connected high-precision current transformers and voltage inputs to the operative-pairs. Such high-precision current transformers being of either aperture or split-core type. The GDU includes a suitable supply of electric power and internet broadband access.

Unlike known systems containing micro-synchrophasors, to make the apparatus suitable and safe for installation and operation in a utility substation, electrical generation plant, or other electrical facility the device incorporates one or more of the following with or without redundancy: C\I C\I -Isolation switch or switches that de-energize the system and disconnect it from external signal, data, and telemetry connections to include a switching arrangement that isolates both AC and DC power supplies; - Type-1 surge protection device plus a type-surge protection device (SPD) type or a combined a type-1+2 surge protection device type; - Residual current device (RCD) type electrical protection; - Singular or redundant switching power supply(s) of sufficient capacity and durability that convert incoming AC mains or parasitically obtained power into suitable voltage DC power; - Buffering short-duration capacitor type uninterruptable power supply; - Longer-duration battery type uninterruptible power supply (UPS) with battery management module; and/or - An arrangement whereby the incoming mains power and the resulting DC power are both conditioned monitored by an operative-pair.

According to the above, an apparatus and method of monitoring a transformer by collecting high-resolution measurement is described. The transformer may be an auxiliary transformer which may be provided on a solar or wind farm. The auxiliary transformer may provide single phase power to onsite equipment such as SCADA. The auxiliary transformer may tap into the power signal from an energy source such as a solar panel. The auxiliary transformer may tap into power signal after it has been inverted. The apparatus may comprise an operative pair of signal analysers, and each operative pair may comprise: a micro-synchrophasor measurement unit configured to operate in the frequency domain to process an electrical signal and collect a first set of data points; and a power quality monitor configured to operate in the time-domain to process the electrical signal and collect a second set of data points. The apparatus may be configured to apply the same synchronised timestamp to the collected first and second set of data points. Timestamping of the first and second set of data points may be synchronised to provide a time synched array of data points. The apparatus may comprise multiple operative pairs of signal analysers. The apparatus may be configured to measure a divergence between the first and second timestamped data points.

This unique, pioneering effort is employing an innovative data gathering apparatus never before applied to solar farm transformer condition monitoring. As there may be seasonal effects, it will be used to collect at least one-year, a full solar cycle, of extensive data. The desired capability is improved online condition monitoring and proactive intervention.

a) Up to 2-3 GB/day per site of highly structured, IEC standard format data delivered via secure telemetry to a time-series database is anticipated. This will amount to roughly two billion individual measurements per site per month depending on triggering electrical event activity. A sample dataset will be made available to industry and academic researchers upon request.

b) The wide range of transformer condition monitoring and diagnostic tests available today [11] include a number of electrical measurement based methods (e.g. transfer function and time-frequency domain dielectric response). Some of these methods may prove adaptable to the online evaluation of the data collected.

c) Improved dynamic methods of collecting and applying frequency and time domain information to transformer analysis that is cost-effective for relatively small transformers.

d) Detection of supraharmonic interaction (e.g. possibly ringing around 90kHz) between inverters onsite or between inverters on neighbouring solar farms over distribution grid circuits is expected. Such interaction potentially having adverse effects on the C\I inverters, transformers and other equipment.

BIBLIOGRAPHY

[1] Piotrowicz, M. and Maratida, W 'Report on efficiency of field-installed PV-inverter with O focus on radiation variability" (IEEE Proceedings of the 20th International Conference Mixed Design of Integrated Circuits and Systems -MDCDES, June 2013, pages 440-443) [2] Christina, A.J., Salam, M.A., Rahman, Q.M., Wen, F., Ang, S.P. and Voon, W., "Causes of transformer failures and diagnostic methods -a review" (Renewable and Sustainable Energy Reviews, 82, 2018, pages 1442-1456) [3] NIurugan, R. and Ramasamy, R. "Failure analysis of power transformer for effective maintenance planning in electric utilities" (Engineering Failure Analysis, 55, 2015 pages 182-192) [4] Moreno-Munoz, A., Gil-de-Castro, A., Romero-Cavadal, E., Rennberg, S. and Bollen, M., "Supraharmonics (2 to 150 kHz) and multi-level converters" (IEEE 5th International Conference on Power Engineering, Energy and Electrical Drives -POWERENG, May 2015, pages 37-41) [5] Darmawardana, D., Perera, S., Robinson, D., Ciufo, P., Meyer, J., Klatt, M., and Jayatunga, U. "Investigation of High Frequency Emissions (Supraharmonics) from Small, Grid-tied, Photovoltaic Inverters of Different Topologies" (18th International Conference on Harmonics and Quality of Power (ICHQP), May 2018, pages 1-6) [6] Smith, J., Ronnberg, S., Bollen, M., Meyer, J., Blanco, A., Koo, K. and Mushamalirwa, D. "Power quality aspects of solar power -results from CIGRE JWG C4/C6.29" (24th International Conference & Exhibition on Electricity Distribution (C1RED), June 2017, pages 809-813) [7] Grevener, A., Meyer, J., Ronnberg S., Bollen, M. and Myrzik, J. "Survey of supraharmonic emission of household appliances" (24th International Conference & Exhibition on Electricity Distribution (CIRED), June 2017, pages 870-874) [8] von Meier, A., Stewart, E., McEachen, A., Andersen, M., and Mehranesh, L. "Precision Micro-Synchrophasors for Distribution Systems: A Summary of Applications" (IEEE Transactions on Smart Grid, June 2017, pages 1-11) [9] Konakalla, S.A.R. and de Callafon, R. "Optimal filtering for grid event detection from real-time synchrophasor data" (Procedia Computer Science, 80, 2016, pages 931-940) [10]Ardakanian, 0., Yuan, Y., Dobbe, R., von Meier, A., Low, S. and Tomlin, C. "Event detection and localization in distribution grids with phasor measurement units" (IEEE Power & Energy Society General Meeting, July 2017, pages 1-5) [11] Islam, M., Lee, G., and Hettiwatte, S. "A review of condition monitoring techniques and diagnostic tests for lifetime estimation of power transformers" (Electrical Engineering, April 2017, pages 1-25) "1-14.

Claims (14)

1. Claims 1. An apparatus for monitoring a transformer in an electricity distribution site by collecting high-resolution measurement data, comprising: a grid data unit comprising an operative pair of signal analysers located at the electricity distribution site, the operative pair comprising: a micro-synchrophasor measurement unit configured to operate in the frequency-domain to process an electrical signal and collect a first set of data points; and a power quality monitor configured to operate in the time-domain to process the electrical signal and collect a second set of data points, wherein the grid data unit is configured to apply the same synchronised timestamp to the collected first and second sets of data points, the apparatus further comprising: a transformer for supplying mains power to the grid data unit, wherein the grid data unit is adapted to monitor output characteristics of the transformer.
2. 2 The apparatus of claim 1, wherein the apparatus uses a signal from a shared antenna to obtain the synchronised timestamp to apply to the collected first and C\I second sets of data points.
3. 3. The apparatus of any preceding claim, wherein the transformer is an auxiliary transformer. cr)c\I
4. 4. The apparatus of claim 3, wherein the auxiliary transformer is to tap into the power signal from an energy source.
5. 5. The apparatus of claim 4, wherein the auxiliary transformer is to tap into the power signal after it has been inverted.
6. 6. The apparatus of any one of claims 3, 4, or 5, auxiliary transformer is to supply single-phase mains power to the electricity distribution site.
7. 7. The apparatus of any one of claims 3 to 6, wherein the auxiliary transformer is to provide single-phase power to other onsite equipment at the electricity distribution site.
8. 8 A method for monitoring a transformer in an electricity distribution site by collecting high-resolution measurement data, comprising: collecting a first set of data points using a micro-synchrophasor measurement unit of a grid data unit operating in the frequency-domain; collecting a second set of data points using a power quality monitor of the grid data unit operating in the time-domain: applying a synchronised timestamp to the collected first and second sets of data points; and supplying mains power to the grid data unit, wherein the grid data unit is adapted to monitor output characteristics of the transformer.
9. 9 The method of claim 8, further comprising using a signal from a shared antenna to obtain the synchronised timestamp to apply to the collected first and second sets of data points.
10. 10. The method of claim 8 or 9, wherein the transformer is an auxiliary transformer.
11. 11. The method of claim 10, wherein the auxiliary transformer is taps into the power signal from an energy source.
12. 12. The method of claim 11, wherein the auxiliary transformer is to tap into the power signal after it has been inverted.
13. 13. The method of any one of claims 10, 11, or 12, wherein the auxiliary transformer supplies single-phase mains power to the electricity distribution site.C\I
14. 14. The method of any one of claims 10 to 13, wherein the auxiliary transformer provides single-phase power to other onsite equipment at the electricity distribution site.