GB2586435A

GB2586435A - Electrical grid monitoring system

Info

Publication number: GB2586435A
Application number: GB1904860.2A
Authority: GB
Inventors: Kulkarni Akshat; Kulkarni Yash
Original assignee: Orxa Grid Ltd
Current assignee: Orxa Grid Ltd
Priority date: 2019-04-05
Filing date: 2019-04-05
Publication date: 2021-02-24
Anticipated expiration: 2039-04-05
Also published as: GB2586435B; GB201904860D0

Abstract

System for monitoring an electrical grid, comprising: one or more data collection modules 100 connected to the grid and configured to collect a plurality of types of data, such as incoming or outgoing power, current or voltage, or phase angles, asset body temperature, On Load Tap Changer position or ambient temperature, from the gird; and a processing module (400 figure 5), which may be centralised, configured to process the data and predict the occurrence of an anomaly, wherein the processing module uses a machine learning algorithm trained on data of historical anomalies. The processing module may be configured to retrain the algorithm when an anomaly is detected but not predicted, or vice versa. The prediction may include probability of an anomaly occurring, predicted start- and end-times of an anomaly and predicted location. The anomaly may include one or more of power outages, voltage or current excursions outside a predetermined range, failure of grid equipment, or transformer losses above a predetermined threshold.

Description

ELECTRICAL GRID MONITORING SYSTEM

The disclosure relates to systems for monitoring electrical grids. In particular, the disclosure relates to the detection of anomalies in the electrical grid.

An electrical grid can be regarded as an interconnected network for delivering electricity from producers to consumers. Typically, an electrical grid comprises generating stations that produce electrical power, high voltage transmission lines that carry power from sources to remote distribution centres, and distribution lines that connect. individual consumers.

The electric power which is generated by a generating station is stepped up to a higher voltage at which it is carried by the high voltage transmission lines. Typically, the transmission lines will move the power long distances, sometimes across international boundaries, until it reaches a primary distribution centre, such as a substation. On arrival at a primary distribution centre, the power is stepped down from a transmission level voltage to a distribution level voltage. As it exits the primary distribution centre, it enters the distribution network. Finally, upon arrival at a secondary distribution centre, such as a distribution transformer, the power is stepped down again from the distribution voltage to a required service voltage.

Challenges facing the electricity sector include: ageing infrastructure and large capital expenditure required to maintain and/or over-haul the electrical grid network; inability to integrate distributed energy resources (DERs, such as solar generators, wind generators, and batteries); large losses in the network, such as theft losses. Monitoring the performance of the electrical grid network is important for addressing these challenges.

It is an aim of the present disclosure to at least partially address some of the problems discussed above.

SUMMARY OF INVENTION

According to a first aspect of the disclosure there is provided a system for monitoring an electrical grid, comprising: one or more data collection modules connected to the grid at predetermined locations and configured to collect a plurality of types of data from the grid; and a processing module configured to process the data and predict the occurrence of an anomaly, wherein the processing module uses a machine learning algorithm trained on historical data of historical anomalies.

Optionally. the processing module uses machine learning algorithms trained on historical data of historical anomalies.

Optionally, the historical data is collected by the data collection modules.

Optionally, the processing module is configured to retrain the machine learning algorithm when an anomaly is detected but not predicted, and/or when an anomaly is predicted but not detected. Optionally, the machine learning algorithm is retrained based on data obtained immediately prior to the occurrence of the detected anomaly or a predicted occurrence of the predicted anomaly.

Optionally, the machine learning algorithm is configured to determined one or more of: a probability of an anomaly occurring, a predicted start time of an anomaly, a predicted end-time of an anomaly, a predicted location of an anomaly According to a second aspect of the disclosure there is provided a system for monitoring an electrical grid. comprising: a plurality of data collection modules connected to the grid at predetermined locations and configured to collect a plurality of types of data from the grid; and a central processing module configured to process the data and predict the occurrence of an anomaly.

In either of the above aspects, optionally, the anomaly includes one or more of: a power outage, a voltage or current moving outside a predetermined range, the failure of grid equipment, and transformer losses above a predetermined threshold.

In either of the above aspects, optionally, the predetermined locations include: a primary distribution centre, a secondary distribution centre and distribution power lines.

In either of the above aspects, optionally, the types of data include: incoming power at the location of the data collection modules, outgoing power at the location of the data collection modules, the current at the location of the data collection modules, the voltage at the location of the data collection modules, the phase angles at the location of the data collection modules, the power factors at the location of the data collection modules, the ambient temperature, asset body temperature, oil temperature, internal winding temperature, On Load Tap Changer Position, at the location of the data collection modifies, a circuit breaker operation status, protection relay status), at the location of the data collection modules.

In either of the above aspects, optionally, the processing module is configured to predict the occurrence of an anomaly further based on data including one or more of: location data, public holiday data, weather data.

In either of the above aspects, optionally, the processing module is configurable by a user. Optionally, the configurable parameters include thresholds determining what In either of the above aspects, optionally, comprising a user interface module through which a user is notified of a predicted anomaly According to a third aspect. of the disclosure there is provided a method of monitoring an electrical grid, comprising: connecting one or more data collection modules to the grid at predetermined locations, collecting a plurality of types of data from the grid using the one or more data collection modules; and processing the data and predicting the occurrence of an anomaly, wherein the processing uses a machine learning algorithm trained on historical data of historical anomalies.

According to a fourth aspect of the disclosure, there is provided a method of monitoring an electrical grid, comprising: connecting a plurality of data collection modules to the grid at predetermined locations; collecting a plurality of types of data from the grid using the plurality of data collection modules; and processing the data and predicting the occurrence of an anomaly using a central processing module.

BRIEF DESCRIPTION OF THE DRAWINGS

The disclosure will be described below by way of non-limiting examples, with reference to the accompanying drawings, in which: Figure 1 schematically shows typical electrical grid network; Figure 2 schematically shows a first data collection module for monitoring the electrical grid at a primary distribution centre; Figure 3 schematically shows a second data collection module for monitoring the electrical grid at a secondary distribution centre; Figure 4 schematically shows a third data collection module for monitoring a location on the distribution line; Figure 5 schematically shows a system comprising the data collection modules and a processing module; Figure 6 is a flow diagram showing an example process performed by the processing module.

DETAILED DESCRIPTION

Figure 1 shows an electrical grid network comprising lines and nodes (boxes in Figure 1). The lines transmit electricity between the nodes. The nodes generally comprise electrical generators 1,7, distribution centres 2.3. and end consumers 4. The lines generally comprise transmission lines 5 and distribution lines 6.

The electrical grid network can be divided into sub-networks, namely a transmission network TN (or transmission grid), including high power generators 1 and high voltage transmission lines 5, and a distribution network DN, including low power generators 7, low voltage distribution lines 6, secondary distribution centres 3, and end consumers 4. The transmission network TN and distribution network DN are bridged by a primary distribution centre 2.

High power generators 1 may include relatively large power plants such as coal power plant, nuclear power plants or hydroelectric power plant. These may generate around 150 MW or more of electrical power. Transmission lines 5 may carry voltages of 50kV and higher, e.g. 400kV or 275kV.

Primary distribution centres 2 may step down the transmission voltage to a lower distibution voltage of less than 50kV, e.g. 33kV, 11kV. Low power generators 7 may include relatively small power plants, such as solar generator farms or wind generator farms. These may generate around 150MW or less of electrical power. Secondary distribution centres 3 may step down the distribution voltage to yet a lower voltage e.g. 415V or 230V, suitable for end consumers.

Figure 2 schematically shows an example of a first data collection module 100 for collecting data at a first location on the grid, e.g. a primary distribution centre 2. The first data collection module 100 may comprise a voltage detection module 101 to detect voltage data and/or a current detection module 102 to detect current data from a primary coil and/or a secondary coil of an electricity feeder 21 at the primary distribution centre 2. The voltage and/or current data may be in the form of time-series data. The data may be sampled at a micro-second rate.

The voltage detection module 101 may be configured to measure voltage data by using the principle of ohms law and a voltage divider circuit of the voltage detection module 101. Primary distribution centre 2 secondary voltage (which is to be measured) is applied to a voltage divider circuit with pre-set resistors. The drop of voltage across the resistor in the voltage divider circuit is supplied to an analogue to digital converter of the first data collection module 101 to get a count value reflecting the corresponding voltage drop value across the resistor in the voltage divider circuit. This count is used in Ohm's law formula to calculate the primary distribution centre secondary voltage count. The primary distribution centre secondary voltage count requires to be further multiplied by a predefined location specific multiplication factor to determine the actual high voltage of the electricity feeder 21.

The current detection module 102 may be configured to measure current data by using the principle of magnetic induction and Rogowski coil of the current detection module 102. A Rogowski coil is a helical coil having a toroidal shape with a current carrying cable encircled by the toroid. The outputs of the helical coil are provided to an integrator circuit of the current detection module 102, which provides a (e.g. digital) count proportional to the current carried by the cable through the toroid. The output count requires to be further multiplied by a predetermined location specific multiplication factor to determine the actual current of the electricity feeder 21.

The first data collection module 100 includes a processor 103 for calculating power (e.g. active, reactive and/or total power), power factor and/or phase angle from the voltage and current data. The processor may be implemented using a microcontroller or a microprocessor, for example.

The first data collection module 100 may also collect further data from the primary distribution centre 2 including: the status of circuit breakers, master trip relay, earth fault relay, overcurrent relay, air break switches and isolators will be monitored. This data may be collected by a sub-module forming part of the first data collection module 100.

The first data collection module 100 may further comprise a communication module 104 for transmitting and/or receiving data. The communication module 104 may transmit and/or receive data wirelessly, e.g. via WiFi and/or RF and/or a cellular network (e.g. a 30, 40 and/or 50 network). Data transmitted may include the voltage data, current data, digital inputs data and parameters calculated from these by the processor 103, and/or any other type of data described above. The communication module 104 may transmit the data to a remote processor (e.g. a cloud based processor). The first data collection module 100 may also store all collected and calculated data as backup on Secure Digital (SD) card or flash in event of loss in communication. The stored data along with error log files may be retrieved once connection with remote processor is restored. After successful retrieval, backup data may be deleted on collection module 100. The collection module 100 firmware maybe be upgraded through Over The Air upgrades from the remote processor.

Once a new firmware is received, the module reboots and updates its firmware to new version if firmware works fine else, rolls back to previous running version.

Figure 3 schematically shows an example of a second data collection module 200 for collecting data at a second location on the grid, e.g. a secondary distribution centre 3.

The second data collection module 200 may comprise a voltage detection module 201 to detect voltage data and/or a current detection module 202 to detect current data from a primary coil and/or a secondary coil of a transformer 31 of the secondary distribution centre 3. The voltage and/or current data may be in the form of time-series data The data may be sampled at a micro-second rate.

The voltage detection module 201 may be configured to measure voltage data by using the principle of ohms law and a voltage divider circuit of the voltage detection module 201. Primary distribution centre 2 secondary voltage (which is to be measured) is applied to a voltage divider circuit with pre-set resistors. The drop of voltage across the resistor in the voltage divider circuit is supplied to an analogue to digital converter of the first data collection module 201 to get a count value reflecting the corresponding voltage drop value across the resistor in the voltage divider circuit. This count is used in Ohm's law formula to calculate the primary distribution centre secondary voltage count. The primary distribution centre secondary voltage count requires to be further multiplied by a predefined location specific multiplication factor to determine the actual high voltage of the transformer 31.

The current detection module 202 may be configured to measure current data by using the principle of magnetic induction and Rogow ski coil of the current detection module 202. A Rogowski coil is a helical coil having a toroidal shape with a current carrying cable encircled by the toroid. The outputs of the helical coil are provided to an integrator circuit of the current detection module 202. which provides a (e.g. digital) count proportional to the current carried by the cable through the toroid. The output count requires to be further multiplied by a predetermined location specific multiplication factor to determine the actual current of the transformer 31.

The second data collection module 200 includes a processor 203 for calculating power (e.g. active, reactive and/or total power), power factor and/or phase angle from the voltage and current data.

The second data collection module 200 may be configured to further measure operational parameter values of the transformer including one or more of: ambient temperature data, equipment temperature data, oil temperature, internal winding temperature, and transformer tap positions. This data may be collected by a sub-unit forming part of the second data collection module 200. Accordingly, the second data collection module 200 may further comprise a temperature sensor that provides a resistance value corresponding to the measured temperature. The resistance value may be based on the sensor manufacturers datasheet and may be detected by measuring the voltage drop across the contacts and using of Ohms law to calculate the resistance value. The second data collection module 200 may also include a universal connection point to measure 4-20mA readings corresponding to digital temperature sensor information.

Transformer tap position may be also detected using the same principle of Ohm's law to measure the resistance across the transformer tap changer module. The manufacturers datasheet provides a correlation of the resistance value with the position of the transformer tap.

The second data collection module 200 may also collect digital inputs data from a secondary distribution centre where the status of transformer fan, bucholz relay, pressure release value and magnetic oil level may be monitored.

The second data collection module 200 may further comprise a communication module 204 for transmitting and/or receiving data. The communication module 204 may transmit and/or receive data wirelessly, e.g. via WiFi and/or RE and/or a cellular network (e.g. a 3G, 4G and/or 5G network). Data transmitted may include the voltage data, current data, parameters calculated from these by the processor 203, and/or any other type of data described above. The communication module 204 may transmit the data to a remote processing module (e.g. a cloud based processing module).

The second data collection module 200 may also store all collected and calculated data as backup on Secure Digital (SD) card or flash in event of loss in communication. The stored data along with error log files may be retrieved once connection with remote processor is restored. After successful retrieval, backup data may be deleted on collection module 200. The collection module 200 firmware maybe be upgraded through Over The Air upgrades from the remote processor. Once a new firmware is received, the module reboots and updates its firmware to new version if firmware works fine else, rolls back to previous running version.

Figure 4 schematically shows an example of a third data collection module 300 for collecting data at a third location on the grid, e.g. a location on a distribution line 6. The third data collection module 300 may comprise a current detection module 301 to detect current on one phase of electricity feeder 61 and/or a time module 302 consisting of real time clock to accurate timing fault occurrence and watchdog timer to reset data collection module after fault occurrence. The third data collection module 300 may be installed on each of the three phases of distribution line 6. The third data collection module 300 includes a processor 303 that may measure short circuit permanent fault, transient fault and fault current, ground fault and fault current, load cut-rent, battery voltage, temperature and presence of electricity in feeder. The third data collection module 300 may also include indicator lights which change colour when there is a fault occurrence.

The third data collection module 300 may further comprise a communication module 304 for transmitting and/or receiving data. The communication module 304 may transmit and/or receive data wirelessly, e.g. via WiFi and/or RF and/or a cellular network (e.g. a 30, 40 and/or 50 network). Data transmitted may include any type of data described above. The communication module 304 may transmit the data to a remote processor (e.g. a cloud based processor).

The third data collection module 300 may also store all collected and calculated data as backup on Secure Digital (SD) card or flash in event of loss in communication. The stored data along with error log files may be retrieved once connection with remote processor is restored. After successful retrieval, backup data may be deleted on collection module 300. The collection module 300 unaware maybe be upgraded through Over The Air upgrades from the remote processor. Once a new firmware is received, the module reboots and updates its firmware to new version if firmware works fine else, rolls back to previous running version.

Figure 5 schematically shows an electrical grid monitoring system according to the disclosure. The monitoring system comprises one or more data collection modules, such as the first to third data collection modules 100, 200, 300 described above, connected to the grid at predetermined locations and configured to collect a plurality of types of data from the grid. The monitoring system also comprises a processing module 400 configured to process the data and predict the occurrence of an anomaly.

As shown in Figure 5, the processing module 400 may be a remote processing module, configured to wirelessly receive data from the data collection modules 100, 200, 300. For example, the processing module 400 may be a cloud based processing module hosted by a remote server.

In addition to receiving data from the data collection modules 100, 200, 3()), the processing module 400 may store additional data or receive data from other sources. For example, the processing module 400 may store meta-data associated with the data collection modules 100, 200, 300, such as the location of the data collection modules 100, 200, 300, the age of the equipment at the location of the data collection modules 100, 200, 300. Other stored parameters may include additional metadata comprising one or more of: information input by users of the system such as geographic position, social data and specific calculated factors, to ensure detection modules estimate the correct primary voltage and currents. Information such as email addresses and mobile numbers of operators for alerts and notifications, user access information, hardware device IDs and firmware versions, etc. may also be stored. Real-time data received from other sources to be considered may include public holiday data, weather data or data collected by any other sensor such as vibration, proximity, pressure, humidity, smoke etc. customer maintenance plans, load curtailment plans or electricity service interruption plan to differentiate planned shutdowns from network breakdown conditions. Other sources may be existing databases such as Meter management systems and Supervisory control and data acquisition systems.

Based on data received by the processing module 400, and/or stored data the processing module 400 uses one or more machine learning algorithms to predict the occurrence of an anomaly within the network. The anomaly may include one or more of: a power outage due to reasons such as electrical faults (eg. short circuit faults or earth faults) or conductor breaker, a voltage or current moving outside a predetermined range (a.k.a. a limit violation) leading to unstable or poor-quality electricity supply, the failure of grid equipment, and transformer losses above a predetermined threshold.

The processing module 400 may predict the occurrence of a power outage based on power data obtained at a node on the network, such as a primary distribution centre 2 and/or a secondary distribution centre 3. To predict a power outage at given node on the network, data is obtained by a data collection module at that node and optionally data obtained by data collection modules located at nodes that are downstream and/or upstream of the node of interest. In additional to power data, the processing module may predict the occurrence of a power outage additionally based on status of digital input monitoring relays (circuit breaker, earth fault, over current), sudden severe voltage fluctuations, sudden increase in Peak Demand leading to power shortage and environmental factors like storm and precipitation.

A power outage is determined to have occurred when the real power detected falls to, or below, a predetermined threshold (e.g. 0) for a set period of time or when the digital input monitoring relays (circuit breaker, earth fault, over current) changes status. A power outage is determined to have ended when the real power exceeds the predetermined threshold or when the relay status resets. When a power outage occurs, the processing module 400 stores information about the power outage, e.g. in a table. This information may include the start time of the power outage, the duration of the power outage (and/or end time), the reason for a power outage, the location of the power outage (e.g. the identity of the relevant data collection module and/or node), for example.

Based on data obtained prior to the occurrence of one or more faults, a machine learning algorithm is trained to determine the probability of the occurrence of a future power outage based on real-time data. A power outage may be considered to be predicted if the probability exceeds a predetermined threshold.

Figure 6 is a flow chart showing steps of a process performed by the processing module 400.

In step S I, the machine learning algorithm may be trained based on historical fault data to output a model that captures relationship between the fault and collected data. In step S2, the processing module 400 obtains real-lime data from one or more data collection modules and other data sources mentioned above. hi step S3, the processing module 400 determines whether a power outage has occurred or not, and if a power outage has occurred, records the start time along with fault location in step 54. Other information about the power outage may also be recorded in step 54, such as information about the location of the power outage.

If in step 53 it is determined that no power outage has occurred, the process proceeds to step 58. In step 58, the machine learning algorithm is run based on the real-time data obtained in step S2, in order to predict a power outage. Step S9 determined whether or not a power outage is predicted, and if a power outage is predicted, information about the predicted power outage is recorded in step S 10. This information may include the predicted start time, location and/or probability. After this, or if no fault is predicted in step S9, the process returns to step S2.

After step S4, the process proceeds to step S5, at which it is determined whether the power outage was correctly predicted, if not the process returns to step S I and the algorithm is retrained based on the newly obtained data. If the power outage was correctly predicted, the process proceeds to step S6, where it is determined whether the power outage has ended, and when it has, the end time and/or duration is recorded in step S7. After the end time is recorded, the process returns to step S I and the algorithm is retrained based on the newly obtained data.

When a power outage is predicted in step S9, a user of the system may be alerted and information about the predicted power outage presented to the user via a user interface.

The user interface, may be provided by a PC, tablet computer or smart phone for example.

The processing module 400 may predict the occurrence of a parameter limit violation based on data about the relevant parameter obtained at a node on the network, such as a primary distribution centre 2 and/or a secondary distribution centre 3. To predict a limit violation at given node on the network, data is obtained by a data collection module at that node and optionally data obtained by data collection modules located at nodes that are downstream and/or upstream of the node of interest and/or data from other sources mentioned above.

Typical power quality standard limits to ensure optimal power flow set by regulators (e.g. Ofgem, in the UK) and enforced through penalties and fines include voltage limits, power factor limits and frequency limits. Accordingly, the relevant data is voltage data, power data and frequency data, each having a prescribed range with desired levels and upper and lower limits. In additional to this data, the processing module may predict the occurrence of a limit violation based on weather data (temperature, precipitation) and downstream renewables generation.

A limit violation is determined to have occurred when the relevant parameter (e.g. voltage, power factor, or frequency) detected moves outside a predetermined allowable range. This range is user defined. For instance, for voltage the declared tolerance for an electricity supply may be 230 volts -6%, +10%, allowing a voltage range of 216.2 volts to ID 253.0 volts; for power factor, ranges from 0.8 (bad power factor) to 1 (good power factor); and for frequency power networks may operate on a standard of 50 Hz +0.5. A limit violation is determined to have ended when the relevant parameter return to the predetermined allowable range.

When a limit violation occurs, the processing module 400 stores a range of information about the limit violation, e.g. in a table. This information may include the start time (e.g. day of week and month, time of day) of the limit violation, the duration of the limit violation (and/or end time), number and nature of customers impacted (e.g. urban, industry, hospitals or commercial etc.) and the location of the limit violation (e.g. the identity of the relevant data collection module and/or node), for example.

Based on data obtained prior to the occurrence of one or more limit violations, a machine learning algorithm is trained to determine the probability of the occurrence of a future limit violation based on real-time data A limit violation may be considered to be predicted if the probably exceeds a predetermined threshold.

The processing module 400 may perform a process essentially the same as that shown in Figure 6. However, steps S3 and S6, would, of course determine whether a limit violation had started or ended.

The processing module 400 may predict the failure of grid equipment (e.g. a transformer) based on data obtained at a node on the network, such as a primary distribution centre 2 and/or a secondary distribution centre 3. To predict the failure of grid equipment at given node on the network, data is obtained by a data collection module at that node. Such data may include voltage data, power data, ambient temperature data, transformer winding temperature and/or transformer oil temperature. Other information may also be used to predict the failure of grid equipment, such as the number of years of operation of the equipment, corrosion factor (i.e. likely rate of corrosion of an asset), dissolved gas analysis data (analysis of the gases being generated by an asset to diagnose issues), transformer winding resistance data (e.g. milliohms, tap position, winding temperature).

The processing module 400 may predict transformer losses above a predetermined threshold based on data obtained at a node on the network, such as a primary distribution centre 2 and/or a secondary distribution centre 3. To predict transformer losses above a predetermined threshold at given node on the network, data is obtained by a data collection module at that node. Such data may include ambient temperature, transformer body temperature, winding resistance data, power data, voltage data, current data, accumulated energy voltage sag (voltage variations between 10% and 90% of nominal voltage for less than one minute), current or voltage harmonic distortion (the measurement of the level of harmonic distortion present in power systems).

Claims

CLAIMS1. A system for monitoring an electrical grid, comprising: one or more data collection modules connected to the grid at predetermined locations and configured to collect a plurality of types of data from the grid; and a processing module configured to process the data and predict the occurrence of an anomaly, wherein the processing module uses a machine learning algorithm trained on historical data of historical anomalies.
2. The system of claim 1, wherein the processing module uses machine learning algorithms trained on historical data of historical anomalies.
3. The system of claim 1 or 2, wherein the historical data is collected by the data collection modules.
4. The system of any one of claims 1 to 3, wherein the processing module is configured to retrain the machine learning algorithm when an anomaly is detected but not predicted, and/or when an anomaly is predicted but not detected.
5. The system of claim 4, wherein the machine learning algorithm is retrained based on data obtained immediately prior to the occurrence of the detected anomaly or a predicted occurrence of the predicted anomaly.
6. The system of any preceding claim, wherein the machine learning algorithm is configured to determined one or more of: a probability of an anomaly occurring, a predicted start time of an anomaly, a predicted end-time of an anomaly, a predicted location of an anomaly
7. A system for monitoring an electrical grid, comprising a plurality of data collection modules connected to the grid at predetermined locations and configured to collect a plurality of types of data from the grid; and a central processing module configured to process the data and predict the occurrence of an anomaly.
8. The system of any preceding claim, where in the anomaly includes one or more of: a power outage, a voltage or current moving outside a predetermined range, the failure of grid equipment, and transformer losses above a predetermined threshold.
9. The system of any preceding claim, wherein the predeteiinined locations include: a primary distribution centre, a secondary distribution centre and distribution power lines. 10
10. The system of any preceding claim, the types of data include: incoming power at the location of the data collection modules, outgoing power at the location of the data collection modules, the current at the location of the data collection modules, the voltage at the location of the data collection modules, the phase angles at the location oldie data collection modules, the power factors are the location of the data collection modules, the ambient temperature, asset body temperature, oil temperature, internal winding temperature, On Load Tap Changer Position, at the location of the data collection modules, a circuit breaker operation status, protection relay status), at the location of the data collection modules.
11. The system of any preceding claim, wherein the processing module is configured to predict the occurrence of an anomaly further based on data including one or more of: location data, public holiday data, weather data.
12. The system of any preceding claim, wherein the processing module is configurable by a user.
13. The system of claim 12, wherein the configurable parameters include thresholds determining what constitutes an anomaly.
14. The system of any preceding claim, further comprising a user interface module through which a user is notified of a predicted anomaly
15. A method of monitoring an electrical grid, comprising: connecting one or more data collection modules to the grid at predetermined locations, collecting a plurality of types of data from the grid using the one or more data collection modules; and processing the data and predicting the occurrence of an anomaly, wherein the processing uses a machine learning algorithm trained on historical data of historical anomalies
16. A method of monitoring an electrical grid, comprising connecting a plurality of data collection modules to the grid at predetermined locations; collecting a plurality of types of data from the grid using the plurality of data collection modules; and processing the data and predicting the occurrence of an anomaly using a central processing module.