WO2016113447A1

WO2016113447A1 - Distributed wireless system and method for the classification and localisation of failures in an underground electrical distribution network

Info

Publication number: WO2016113447A1
Application number: PCT/ES2016/000005
Authority: WO
Inventors: Carlos LÉON DE MORA; Antonio GARCÍA DELGADO; Francisco Javier Molina Cantero; Joaquín LUQUE RODRIGUEZ; Julio BARBANCHO CONCEJERO; Enrique Personal Vazquez; Diego Francisco Larios Marin
Original assignee: Universidad De Sevilla
Priority date: 2015-01-16
Filing date: 2016-01-14
Publication date: 2016-07-21
Also published as: ES2577881A1; ES2577881B2

Abstract

The invention relates to a distributed wireless system and method for the classification and localisation of failures in an underground electrical distribution network, comprising sensor devices (3), forming a network of wireless sensors (2), distributed in the underground electrical distribution network (1) and coupled to conductors (7) of the network (1) such that all the sections of conductors between bifurcations have an associated sensor device (3). The sensor devices (3) comprise means for measuring the current (34) that circulates via the conductor (7), being synchronised with one another and configured to identify the type of failure occurring and the localisation thereof by means of the interchange of messages, between the different sensor devices (3), with information relating to the synchronised current measurements and by means of the analysis of the phasor information of said synchronised current measurements, taking into consideration the topology of the network.

Description

DESCRIPTION

Field of the Invention

The present invention falls within the field of methods and equipment for locating faults in electrical distribution lines, and more specifically in underground medium voltage electrical distribution lines.

Background of the Invention

It is a fundamental task of the electricity companies to guarantee the supply of energy to the users in conditions of continuity and quality. Thus, the location of faults in power lines becomes a priority task for these companies. Once a fault has been detected, automatically estimating its position from the control center provides great advantages from the point of view of the service quality, since it reduces both the restoration time and the number of maneuvers needed. Therefore, it is very attractive, in terms of of quality and economy, to the implementation of location systems of this nature.

The development and application of microprocessor-based systems in electrical engineering (IEDs) marks the appearance of the first automatic fault location techniques [1, 2j. Since then numerous methods of automatic fault location have been developed that fall into four categories:

■■ Methods based on traveling waves ("traveiiing waves"). These types of methods analyze the high frequency signals (> 500kHz) ₍ in currents or voltages, produced by the impulses generated if the fault is triggered [3] These impulses travel at high speeds through the electric lines. The detection of the arrival of the Impulse to two distant points allows to estimate the position of the fault.These techniques do not require a complete knowledge of the characteristics of the lines, however, high wave speeds cause great errors in the location of the fault.

- Methods that use relatively high frequency components of voltages or currents. To estimate the position of the fault, these methods use the measurements obtained with relatively high frequencies (10KHz), which are analyzed using techniques based on frequency domain analysis.

- Impedance methods. These methods use the fundamental components (rates) of tensions and intensities at terminal points of the line. They are the most used in the location of faifas in distribution networks and consist in calculating the impedances of the lines, as seen from the terminals of the same, before and during the lack. A good description of the different methods is found in | 5] and its extensions to different cases of failure in [6],

- Methods based on the timely detection of an overpowering. These methods are based on the use of Missing Circuit Indicators (FCi), as explained for example in US7G23891 -B 1. These elements are able to detect, locally, overheating events, roemorizinoolos for a while or until the device is checked. Based on this information, in systems with radio distribution !, it is possible to determine the section of line in which the fault originated. However, in systems with distributed generation where the power is not unidirectional, these systems are not able to perform the location. In these cases it is necessary to use directional FCis {as indicated for example in US79691 55-B2) which, using a measure of the voltage phase, are able to determine the direction of the power flow caused by the taita . Following these directions it is possible to determine the section under fault, even in these networks with distributed generation. The information generated by this type of devices is usually visual (light pilot or mobile element) that allow the operator to know, by means of a Direct Inspection, the status of the device. There is a system in the market with wireless communication that allows access to information remotely. These systems are based on short distance radio links, or GSM or GPRS communications.

To date there are numerous patent documents related to fault location systems in power lines:

- ■ VVO2007032697-A1 presents a method to locate faults by dividing the lines of a transmission or distribution system into sections and assuming a hypothetical location in at least one of these sections, based on the measurements of the currents, in the fault conditions and also in pre-failo, in all seasons terminals of the system, and also, of the measurement of the phase of the line voltage, in the conditions of failure and pre-failure, in one of the terminal stations of the system.

- AU2008200131 -A1 describes a system for locating the point of failure, using several slave stations that collect data on the pre-failure situation of the distribution line and a master station, where all this information is received and processed, deducting a single point of failure, from a greater set of possibilities compatible with the measurements.

- WO2013G9 G28-A1 proposes a deployment of a sensor system for recording electrical parameters, in low voltage networks These measuring elements use a GPS system for synchronization and will be connected at the ends of the distribution lines (one per line) and will communicate with the base station through a broadband connection.

- US201302Q59GG-A1 proposes an electrical distribution network management system, based on a deployment of sensors in aerial electrical distribution networks, recording information of an electrical and mechanical nature Each of these sensors communicates directly with a receiving station, which registers and processes the information of the nodes, trying to determine the characteristics of the installation, facilitating maintenance oppressions.

In the present invention, the use of a non-intrusive distributed system based on wireless sensor networks (WSN) is proposed as an alternative to locate faults in underground distribution electrical networks. A network of wireless sensors is made up of a series of small, low-power devices (nodes), capable of wireless communication between them and collaborating to analyze a common phenomenon.

Sensor networks have many applications [7], one of the main ones being ubiquitous and pervasive monitoring of an environment. The application of wireless sensor network technology in the field of electrical distribution networks is an emerging alternative, with numerous contributions in recent years, both of a general nature [8, 9], as applicable in specific aspects, such as : surveillance of overhead lines (10), monitoring of distributed generation systems (11, or management of charging systems of hybrid electric vehicles, etc.) However, in no case has the networks been applied from wireless sensors to the iocaiization of faults in underground power lines, which constitutes one of the novelties of the present invention.

Precisely, this is one of the points where there is an important innovation with respect to the current faita location devices. So far, there are only two variants for these location systems: those that use local and isolated processing (based exclusively on local information) and those that use centralized processing (in which all information is concentrated and processed in a single node ). The present invention, on the other hand, is based on collaborative processing between nodes, which gives the system the ability to detect, locate and classify taitas in systems with distributed generation, based solely on current information, not requiring a measure of the voltage phase (unlike the directional FCIs that do need it), as it is inaccessible in shielded cables (common in underground distribution lines). This fact shows the capacity of the collaborative system of the present invention, thanks to the qua! a fault due to a stem can be found in the internal insulation (in which the return of the taita current is carried out through the mesh), and which solve the current FCI type systems.

Bibliographic references

[11 A. Girgis, C. Fallón, and D Luhkeman, "A fauit iocation technique for rural distribution feeders." industry Applications, IEEE Transactlons on, voi. 29, 1993, pp. 1 1 70-1 175,

Í2 i T. Takagi, Y Yamakcshi, M. Yamaura, R. Kondow, and T. atsushima, "Development of a New Type Fauit Locator Using the One-Terminal Voitage and Current Data," I EEE Transactions on Power Apparatus and Systems, vol. PAS-101, 1982, p. 2892-2898,

Í3] Wenjin Dai, Mm ^' Fang, and Lizhen Cui, Traveling Wave Fauit Location System ", Inteliigent Control and Automation, 2006. CíCA 2006, pp. 7449-7452

[4] E. Rosoxliovvskl, J. izyko ski B. Kasztenny, and M M. Saha, "A new distance reiaying a! gorithm based on complex differentia! equafion for symmetricai components," Electric Power Systems Research, voi. 40, Mar. 199?, P. 1 75-180

[5] J. Mora-Florez. J. Melendez, and G Carriüo-Caicedo, "Comparison heard impedance based fauit location ethods for povver distribsjtíon systems," Electric Power Systems Research, vol. 78, Apr 2008. pp. 657-666. [8] Salim,. Resener, A Fiiomena, K. Rezende Caino de Oíiveira, and A. Bretas, "Extended Faiilt-Location Formation for Power Distribution Systems," Power Deüvery, iEEE

Transactions on _: voL 24, 2009, p. 508-516.

[7] Akyildiz, i.F, Weilian Su; Sankarasubramaniam. Y.; Cayirci, E. in "A survey on sensor networks," Communications Magazlne, IEEE, vol 40, no.8, pp.102-1 14, 2002

[8] V.C. Gungor B.Lu and G.P. Hancke, Opportunifies and C allenges of Wireiess Sensor Networks in Smart Grid, "IEEE Trans. Ind. Electron., Voi. 57, n.10, 2010.

[9] M.Eroí-Kantarci and H.T. Mouftah, "Wireless multimedia sensor and actor networks for the next generation power grid." Ad Hoc Networks, vol.9, n.4, 201 1.

[0] Y. Yang, D. Diván, R.G. Hariey and T G. Habbeter, "Design and implementation of power ine sensornet for cverhead transmíssion iines," IEEE PES Genera! Meeting, 2009.

[1 1] i.S.AI-Anbagi, H.T. ouftah and .Eroi-Kantarci, "Design of a deiay-sensitive WSN for wind generation moniíoring in the smart grid," 24th Canadian Conference on Electric! and Computer Engineeríng (CCECE), 201 1.

[12] Y. Hong and A.Scaglione, ^* A scalabie synchronization protocol for iarge scale sensor networks and its applications, "! EEE J. Sel. Areas Commun ,, vol.23, n.5, pp. 1085-1099 , 2005.

[13] A.Marco, R. Casas, J.L. Sevillano, V. Caoarasa, J.LFalco and .S.Obaidat, "Muiti-Hop Synchronszation at the Application Layer of Wireiess and Sateliite Networks," IEEE Global Telecommunications Conference, 2008.

[14] D F. Larios, J.M. ora-Merchan, E. Personal, J.Barbancho and C. Leon, "Implementing a Distributed WSN Based on IPv6 for Ambient onitoring," Int. J. Distrib. Sens. Networks,

2013

The present invention relates to a system and method for locating faults in medium voltage electrical distribution underground lines. Specifically, a system designed as a network of wireless sensors with multiple non-intrusive current sensors, which allows the detection, classification and location of the taitas.

The distributed wireless system for the classification and location of faults in an underground electrical distribution network comprises a network of wireless sensors distributed in an underground electrical distribution network The wireless sensor network is formed by a plurality of sensor devices that constitute the sensor nodes of the network. Each sensor device is magnetically collected to a conductor of the underground electrical distribution network to be monitored, the sensor devices of the network being distributed so that all sections of conductors between branches have collected at least one sensor device

Each sensor device has a wireless communication module for communication with other network sensing devices located within its reach, data processing means and current measurement means that circulates for the conductor to which it is associated.

The sensing devices are synchronized ^'and configured to, once detected fault event, identify the type of originated Falia and ei point of the network of underground power distribution where the fault has occurred through the exchange of messages between different sensor devices, with information on the synchronized current measurements and through the analysis of the phasor information of said synchronized current measurements, taking into account the network topology.

To perform the location of the fault, the sensor devices are preferably configured for;

- determine the line segment in which the fault is found by comparing the input and output currents of the different segments that make up the underground electricity distribution network,

· ■ classify the type of fault in the affected segment, determining the line or lines of the segment subject to the fault and the type of insulation fault produced, and in particular if the fault is caused by a grounding or a loss of insulation inside the cable,

- estimate the location of the fault within the segment in which the fault occurred taking into account its classification. The sensor devices are preferably configured to perform the joint synchronization using the zero-pass detection of the current measurements of the sensor nodes, and taking into account the situation of the sensor nodes within the underground electrical distribution network to be monitored and that the fasonai sum of currents in all nodes is equal to zero.

In a preferred embodiment the sensor devices have a power coil that surrounds the conductor and through which it receives the power supply.

The current measurement means of the sensor devices preferably comprise at least one Rogowski coil.

The sensing devices can be located in the underground air distribution network, so that they can monitor the entire power flow of the underground electrical distribution network

The system may additionally comprise a control unit and a gateway with wireless communication capability configured to collect information from the wireless sensor network and transmit it to the control unit

Another aspect of the present invention relates to a method of classifying and locating faults in an underground air distribution network. The procedure includes:

- establishing a wireless sensor network in an underground electrical distribution network, said wireless sensor network being formed by a plurality of sensor devices synchronized with each other and constituting the sensor nodes of the network; where each sensor device is associated with a conductor of the underground electrical distribution network to be monitored, the sensor devices of the network being distributed so that all the sections of conductors between branches have at least one sensor device associated;

■ · measure, by each sensor device, the current flowing through the conductor to which it is associated;

- before the detection of a fault event, Identify the type of fault originated and the location of the fault by means of the exchange of messages, between the different sensor devices, with information of the synchronized current measurements and by the analysis of the fasonai information of these synchronized current measurements, taking into account network topology

To achieve the detection and location of faults, in addition to the measurement itself of the current flowing through the lines, it is necessary that the sensor network provide two basic services: a method of formation and routing of messages, and a mechanism of time synchronization that allows to share a global clock between all the devices. Both are classic problems of sensor networks, for which multiple solutions have been proposed [12, 13, 14], but even today it is not a closed line of research, since there are no optimal algorithms for all applications.

In order to synchronize the nodes, the present invention includes a method of global time synchronization using the distributed system of current measurements of the sensor network. Thanks to these distributed measures and through a collaborative method, a global clock can be estimated and calibrated for the sensor network that supports, among other things, the implementation of information routing methods that minimize energy consumption electric

The new system of location of faults in underground electrical distribution lines is therefore based on the collaborative analysis of the net flows of currents acquired through the use of a network of non-intrusive sensors that communicate wirelessly, forming a network of sensors Wireless (WS) The advantage of this approach with a sensor network is that it does not require any type of module information and / or voltage phase, parameters not accessible without perforation in the shielded cables presently present in the underground power distribution lines . The system only requires the current measurements acquired directly by the nodes, which ai Exchange information, are able to detect, classify and locate the faults, both in traditional distribution systems (with unidirectional current flows), and in systems with generation distributed, where the power flow can change direction. This is another advantage and innovation of the present method and system with respect to its predecessors

As an additional advantage, no other system is required to extract the information from the network, that is, the system takes advantage of its own wireless communication structure, to inform the control center (on demand) of the different parameters of the network. However, the use of this technology is bound to adapt different aspects of the routing and synchronization of communications, to the characteristics of the problem to solve. In the present invention, techniques based on the use of implicit information to the own electricity network to be monitored for obtaining route tables and time synchronization are defined, which results in a novel approach compared to existing systems in ia Literature, which primarily employs only information obtained from the exchange of messages.

The method of the present invention therefore allows the monitoring of a medium voltage underground electrical network by means of e! Use of a network of wireless sensors that with a non-intrusive procedure allow the measurement of the currents that circulate through the network to be monitored. Using these measures of current and the exchange of messages, the network is able to monitor the state of the lines, as well as to classify and locate the possible failures that may occur in them.

The system consists of a deployment of sensor nodes, distributed from ta! In this way, it is necessary to ensure that all cable frames between branches have at least one current meter collected. The communications structure is automatically formed by exchanging messages, based on prior knowledge of the structure of the electricity grid.

The method of synchronization of the sensor nodes takes advantage of the previous knowledge of the topology of the electrical network and the situation of the sensor nodes to synchronize, with high precision, a common globaí clock among all the nodes of the monitoring network This synchronization of time is necessary to be able to detect and locate possible faults or defects that may occur on the monitored underground power grid. To perform the synchronization, the system uses the knowledge that the sum of fasoriai currents in all nodes is equal to zero. From this and assuming, in a first approximation, that the time lag in the reception of the messages is negligible among all the nodes of the fork, a mathematical model is established that uses the zero steps of the current signal to calibrate, with high precision, the clocks of each of the nodes. The complete synchronization method is subsequently uncovered in detail.

The method for the classification and location of faults in underground electrical distribution lines allows the fault to be located once it has been detected (by means of the protection relay, located at the head of the line), based on the analysis of the fasoriai current information recorded by the different sensor nodes and known topology of the distribution network, including the location of each node within it. The location method based on this information is structured in three phases: determination of the line section between two transformation centers where the fault is found, classification of the type of fault and, finally, the estimation of the specific point of that section, in which the fault originated. The knowledge of this position allows the isolation of the fault and the restoration of the electricity supply to be carried out more quickly and accurately, considerably improving the continuity of the supply.

^ íg g _. dgsc ipgion of the di ^; spark plugs

Next, a series of drawings that help to better understand the invention and that expressly relate to an embodiment of said invention which is presented as a non-limiting example thereof is described very briefly.

Figure 1 represents the general structure of a wireless sensor network.

Figure 2 shows the connection of a sensor node to the conductor to be monitored.

Figure 3 shows the internal structure of a sensor node.

Figure 4 shows the structure of a sensor network for monitoring the underground electrical network, where each box can have one or more sensor nodes.

Figures 5A, 5B and 5C represent a model of the simple underground distribution network on which a deployment of nodes in the different boxes has been made. Figure 5C is a continuation of Figure 58 and this in turn is a continuation of Figure 5A.

Blind description of the invention

The method and system of the present invention are based on the deployment of a wireless sensor network that allows the detection, classification and location of anomalies in underground electrical distribution lines.

Figure 1 shows a network of wireless sensors 2 distributed in an underground electrical distribution network 1 and formed by a plurality of low consumption sensor devices 3 (or sensor nodes), with wireless communication capability between them and collaborating to analyze a common phenomenon. The wireless sensor network 2 communicates with a control center 4, formed by a gateway 5 with wireless capability (in communication with sensor devices 3) and connected to a counter unit! 6 (eg a computer).

Each sensor device 3 of the network, also called a sensor node, is designed to operate autonomously by acquiring its energy directly from the mea, from a non-intrusive connection with it, since the device only has to be collected around the distribution network conductor, without the need for drilling or sectioning itself, as shown in Figure 2, which represents the way of connecting a sensor device 3 to the conductor 7 moniton ar.

Figure 3 shows the internal structure of the sensor devices 3. Each of these nodes is formed by a wireless communication module (Le. A radio transceiver 30} that allows communication within the underground pipeline itself with other sensor devices 3 , data processing means (for example, a microcontroller 31) that processes the information, a power supply 32 and an input-output subsystem 33, responsible for the monltorlzación of the environment through the different measurement systems that compose it , among them a current measurement module 34 that circulates through the conductor 7 of the distribution network, and an I / O interface for other sensors 35 useful for the distribution company (such as temperature sensor, humidity sensor , light intensity sensor, vibration sensor, sound sensor, etc.), which would provide more information about the environment or working conditions of the in Thus, the sensor device 3 of the invention has the ability to integrate additional sensors for the monitoring of the installation, which allow the use of the communications infrastructure to provide additional information to the company that owns the installation. The sensor devices 3 are designed to, apart from being economical, allow rapid network deployment. Fundamentally the device network executes the following methods:

- Method for the calibration of the global system clock by means of a hybrid algorithm based on wireless communications and fasonal estimates of the currents that circulate through the conductors of the underground electrical network to be monitored.

- Collaborative method for the classification of faults in power lines based on the exchange of messages with the information of the net charge flows between the different nodes that form the electricity distribution network - Co-operative method for the location of the fault in power lines based on the exchange of messages with the information with the synchronized current measurements of the nodes, and in the previous classification of the fault. This location allows to drastically reduce e! number of maneuvers and the interruption time of the network service, facilitating maintenance tasks (operators know exactly where the repairs should take place).

The different devices that are part of the invention are described in detail below, and the functionality of said devices that allows the analysis of the faifaes in the underground distribution lines, as well as the proper functioning of the communications between the devices.

The present invention raises a system whose development is deployed over a network of wireless sensors. The network is formed by a series of devices, called nodes, which are responsible for interacting with each other, so that through their interaction they allow the monitoring of their environment. In the present invention the following components are distinguished:

- Control unit 8: It is an element, generally located in the control center 4, with which the user interacts and allows him to retrieve the information of the sensor nodes regarding the state of the underground electrical distribution network under analysis, indicating, in the event of a breakdown, the section in which it occurred and its location within! same, making it easier for them to operate their tasks of isolation and repair of the fault. In addition, it is responsible for storing the corresponding historical records of detected faifas.

- Gateway S: This element is a node that on the one hand allows communication with the control unit 6 through a standard port and on the other it integrates a radio transceiver compatible with that used in the other devices that make up the network. This device is responsible for redirecting the inquiries of the operators through the wireless sensor network 2. Thanks to this device, the control unit 6 can collect the information on the operating status of the underground electrical distribution network 1, as well as the internal status (remaining battery, error rate, etc.) of each of the sensor devices 3. In addition, the gateway 5 is also responsible for making the necessary steps to ensure that the route tables of all the routes are kept updated network nodes, so that access is guaranteed through adequate retransmission using intermediate nodes to any element of the network.

~ Sensor devices 3 or nodes: It is understood by sensor nodes, or simply nodes, a series of low-cost and low-cost sensor devices 3 that allow interaction between wires through the use of wireless communications and that allow a localized part of the problem to solve, so that through its interaction you can find the global solution of the problem. The nodes proposed for the present invention consist of the following elements or subsystems (see Figure 3): o Wireless communications subsystem: It will consist of a low consumption radio transceiver 30. The frequency of the carrier signal will be chosen based on the network topology, seeking a minimization of transmission losses. Since the pipes that interconnect the boxes are surrounded by a plane of mass, it can be assumed that for certain frequencies this channel will act as a waveguide, allowing a large range with low losses. These frequencies, together with an appropriate modulation, will be the one chosen for the communications between all the elements that make up the network or Power Subsystem: The sensor devices 3 receive the energy for their correct operation through a supply coil 38 (Figure 2.) surrounding the conductor 7, by means of which all the electronic circuits of the sensor can be fed together with suitable electronics _. Including the radio transceiver, without the need to run or drill the insulation of the cable to which it is associated, Thanks to this, a very fast and non-intrusive deployment can be achieved, which does not damage the characteristics of the distribution network to be monitored.

In addition to the energy collection system, if the power system consists of a backup energy storage system, such as a battery or a high capacity capacitor that will allow these devices to continue operating after the appearance of a fault in the distribution network, situation in which there is no longer a per-conductor energy flow and, therefore, the node could not be fed through the supply coil 36. o I / O Subsystem 33: Allows the monitoring of the environment through the different measurement systems that compose it. In this case the most important magnitude is The current through conductor 7 of the distribution network, measured through a current measurement sensor, preferably one or several Rogowski coils 37 (Figure 2), whose behavior is defined by:

Where n is the number of turns, A the area of the section of the torch formed by the coil and ο, vacuum permeability (4 ·· π · 10 ^"? (Vs) / (A m))

Thus, each coil allows the estimation of the current flowing through a conductor. Depending on the characteristics of the installation, nodes are provided with up to a maximum of one coil per conductor, typically three in conventional three-phase installations. However, for those installations whose onology requires it, the present invention can also be developed considering nodes with a single coil, so that several are installed per cassette, thereby achieving greater redundancy in communications and, therefore, a greater reliability of them. o Processing subsystem: Pre-assigned preferably by means of a microcontroller 31, it is responsible for the management of communications and the processing of the information necessary for the realization of the different methods or procedures described below. In addition, it is responsible for the aggregations of data necessary to minimize the use of the wireless communications channel, thereby increasing the reliability of! system.

In order to obtain a correct functioning of the system, the deployment of the sensor network is carried out within the boxes of the underground distribution lines to be monitored, preferably complying with the following premises:

- Both the power elements and the sensors are wrapped around the conductors 7 to be monitored (Figure 2), with no action required on it (eg, perforation or cutting of its insulator).

- A current measurement sensor is placed for each of the lines that cross the box 8. Thus, for example, in a three-phase system it is necessary to place of three current sensors Depending on the type of nodes used, this may require the installation of one or more nodes per cabinet.

- Sensor devices 3 are placed in all those boxes 8 (see Figure 4, which represents a structure of a sensor network) of the underground electrical distribution network 1 to be monitored that meet the following premises:

^" When the separation between the box 8 with sensors before and after the current box 8 exceeds the maximum coverage distance of the Wireless communication system,

^• When a fork of drivers occurs. In this case, sensor devices 3 must be placed at the exit of each fork although the coverage conditions do not require it. For example, in the case that a three-phase line separates into two branches, six sensor elements are necessary, three for each branch, placing one in each of the phases of the three-phase line.

In order to monitor a medium voltage underground electrical distribution network, it is necessary to know in advance the situation of the nodes and the topology of the electrical network to be monitored. Based on this information, the devices that make up the invention impose the following methods to monitor the power grid:

- Method for the formation of the network and its mine. The algorithm for network formation is based on a flood algorithm, e: what will start on the bridge device of the gateway 5, generating a series of tricks that will be relayed to all the nodes in order to obtain the number of minimum jumps for communication between all nodes. This process is called wireless network discovery, executing the following procedure:

• A given node, either at the request of the Information system or because it has been a long time without exchanging information with other nodes, sends a message in redcast that will be received by a series of nodes, informing of its battery level.

• A series of nodes receive this message. These nodes are called neighboring nodes of the issuance. With this information, the nodes update their table of local routes, with which they determine the number of jumps between the node that originated the message and the operative bridge or gateway 5.

• Based on this information and assuming the physical topology of the electrical network to be monitored, the giobal map of routes to be used over the network is calculated by using a minimum weight overlay tree algorithm designed for wireless sensor networks. The weight of the covering tree takes into account various parameters, such as the topology of the electrical network to be monitored, the battery level of the nodes or the physical distance between sensor nodes to determine the optimal communication paths.

Cyclically, depending on the internal changes that occur in the nodes, the information system is responsible for updating these routes in case there are not enough communications in the network to keep the route tables updated, maximizing the functionality of the network and the giobal battery life

- Method for time synchronization; To correlate the information associated with different measurement points collected by different nodes, a time synchronization between them is necessary. For this purpose, a method is used that combines the detection of zero steps of the current measurement, together with the sending of beacons and the prior knowledge of the situation of the sensor nodes within the electrical distribution network to be monitored. The method complemented is as follows:

^« The nodes are continuously measuring the current of the cable or cables to which they are associated. From the shape of the measured wave, each node calculates from a global internal clock a time stamp (timestamp) in which the zero crossing of the positive slope of the wave occurs. That is, the nodes are continuously measuring the current and processing it so that they always have in memory the magnitude and phase of the current, referred to their internal clock.

• If a request for synchronization of the information system occurs, or if more time has elapsed than the time allocated for synchronization, a node sends a synchronization message to the child nodes {those hierarchically inferior in the distribution network architecture electrical) and considers as last synchronization time its last crossing through the zero of positive slope detected. • Due to the lags and the operation of the radio transceivers, the broadcast time is unknown, but it can be assumed, due to the speed of transmission of the communications via radio, that all the child nodes will receive the message In the same instant of time.

^• Each of the child nodes considers at this point the reception time as the origin of synchronization time, and they send to the parent node (the hierarchically superior one in the architecture of the electricity distribution network, and which has originated the synchronization process ) a message with the current module and the time lag between its last zero crossing of positive slope and the reception time of the synchronization message.

^{4 The} parent node, after receiving the information from all of its child nodes, adds the current magnitudes of all these fascinatingly and determines the phase resulting from the aggregation. With this Information, the error introduced due to the unknown time of receipt of the message is estimated from the time lag between the aggregate phase and the phase considered as the origin of time.

• Once that phase is measured, the parent node sends a message in redcast to the nodes informing of the overall time in which the reception of the message originated.

• Each of the child nodes uses that information to calibrate their internal global clock.

With this algorithm, high-precision time synchronization can be ensured, taking advantage of the fact that the architecture of the electricity distribution network is known and that the time difference in the reception by the child nodes of a message spread by the parent node is negligible.

- Method for the classification and location of taitas in underground electrical distribution lines: Once a faifa event has been originated, it is analyzed through the analysis of the fasorlal information recorded by the different nodes and the network topology known, which type is missing. it has originated and it is estimated at what point of the network this has been produced. As an example, Figure 5 (which due to its extension has been subdivided into Figure 5A - which includes the substation and the first and second segments-, Fsgura SO - which includes an intermediate segment f - and Figure SC - which includes the last segment z-) represents a simple underground electrical distribution network 1, without bifurcations, in which the deployment of the sensor network has been carried out in the different boxes 8. The underground electrical distribution network 1 shown in Figures 5A, 58 and 5C comprises a substation 9, a piuraiity of segments 1 G (a segment is the section of the power line that separates two transformation centers 1 1) with their respective transformation centers 1 1. In the segment f a faita is located. The location is carried out collaboratively by all the nodes of the ed, while the classification is carried out by the nodes that are right in front of the transformation centers. The unimmented procedure is divided into the following phases:

• Determination of the segment in which the faita originated, segment f. This first stage consists in comparing the input currents of segment f with the output currents of segment f. The input current of segment f is calculated from the current data measured by the sensor devices 3 of the segment immediately before the transformation center f-1, segment f-1, and the own consumption of! Transformation center f-1 The output currents of segment f are calculated from the current input values of each of the transformation centers that hang from that line section, including the own transformation center f, Additionally Another way to calculate the output currents of segment f would be through the consumption recorded by the nodes of the next segment f + 1, adding the current that is derived by the transformation center f. Thus, if the difference between this input and output current to the segment f exceeds a threshold, the detection of the faita in the

Where i E is the current entering segment f,

(¾'.V i {f +! I ⁺ are two different ways of calculating the current

leaving the segment f; where x is the concrete line being analyzed, I _S EN Χ <Τ · 15 is the current measured by any of the sensor nodes of the segment (f-1) of the line x, i SE is the current measured by any of sensor nodes segment (f + 1) for the line x, i _C r 'MJ is current IA measured in ei transformer (f-1) for the line x, and finally take that ail values from f to z , to add in the summation all the currents derived by the trasiormation centers that hang from segment f. This analysis is performed for each of the phases (A, B and C) independently, to know the existence of a fault. s Once the segment 10 under fault is determined (segment f in Figure 58), it is possible to classify the type of fault based on the lines (phases A, B, C) from the input and output data of the segment. affected and the type of fault that has occurred in the isolation. This analysis is divided into two parts -

* Determination of missing lines: this overcurrent analysis allows to determine that line x is subject to missing. From the evaluation of (3) on the different lines, we can determine simple typologies (a single affected line), double (two affected lines) or triple typologies (the three affected lines) of the fault

[4 ,,,,,, ,,, -i _ÍW J> threshold

Missing on

or (3)

line x

_? , - 1 _{{Ί Λ!} ^> threshold

<· Determination of the type of insulation fault: In this phase it is distinguished whether the fault is caused by a grounding, originating at the same point of the fault, or on the contrary due to a loss of insulation inside the cable and returning the current through the mesh. Based on Figure 5B, it represents the current that would circulate through the conductor of the line x cable, in the segments of segment f before the missing point. _> -? represents the current that would circulate through the wire mesh of the line x, in the segments of the segment f before the point of failure, _v represents the fault current in the line x (null if the fault is internally insulated), e! _¾. : ¾ represents the current that is derived from the conductor of the line x to its mesh. Analogously. _a and í ' _Sxf represent the currents that would circulate through the conductor and through the cable mafia of line x, in the segments of segment f after the point of failure. This analysis can be done through the cclaborative analysis of the sensors of the affected segment, IF the return of the current is through the mesh _; All these sensors must measure approximately the same (equation 4).

In ia equation (4) I _{sm f ≠} represents any of the currents measured by the sensor devices 3 located in the segments of the segment prior to the point of lack of the line x, e Ι ^, _(ΐ ^ represents any of the measured currents by ios sensor devices 3 located in manholes segment _f;. subsequent ai fault point of the line x tant, if this condition is met n defined in (4) for all sensors í _s segment is determined that lack it is for a fault in the internal insulation, otherwise, it is determined that the fault is due to a direct grounding at the point of failure.

»Once the segment under faifa has been determined and its type determined, the next step is to estimate the point at which the fault is found within the segment. In this respect, two important variants are distinguished:

«Insulation fault with grounding at the point of failure (negative case of the condition defined in equation 4): In this circumstance, the determination of the section under fault is reduced to Go evaluating the current measurement of each node, in each arc of segment f, with the measurement recorded by the next node. For an x line, when there is a difference between these two currents greater than a threshold (equation 5), it can clearly determine that the fault is between the defined sub-division between those nodes je i + 1

[Missing detected in subsegment

.S¾V ÍÍ.¡; -> threshold ^™ í (5)

i defined between the boxes i and i-H

^« Insulation fault without grounding at the taita point (current return through the mesh, affirmative case of the condition defined in equations 4): In this circumstance the segment sensors do not detect the current ius * (current that is derived from the conductor of the line x to its mesh) caused by the fault (equation 6), because the net current flow is zero. For this alternative, the use of a basal method is proposed in the calculation of the resistance of the anterior (pR _s «) and posterior ((1 -p} 'Rsh) mesh to the point of failure (equation 7), where R _S represents the total resistance of the mesh in the segment of segment í, and Γ & «· represent the comments that would circulate through the mesh of the cable of line x in the segments of segment f anterior (is _x ») and posterior (i ¾ < ) to the point of failure, respectively, 'f _t represents the voltage (referred to ground) on ia maiia in the fault point in the Iine x. Finally, to determine ia normalized position "p" of the fault within the segment / (equation 8).

^■ * '"p - _S!' - 0 P) R, Í ^{^'...}

This planning is not only valid for simple topologies (for example, that of Figure 5), it is also extended to distribution lines that have bifurcations. In these more complex topologies, it will only be necessary add to the described approach the equations derived from these electrical knots

Thanks to the start of all these methods or functionalities, the monitoring network can easily determine the operating status of the underground power grid.

Claims

1 Distributed wireless system for the classification and location of faults in an underground electrical distribution network. characterized in that it comprises a wireless sensor network (2) distributed in an underground electrical distribution network (1), said wireless sensor network (2) being formed by a plurality of sensor devices (3) that constitute the sensor nodes of ia net; where each sensor device (3) is coupled to a conductor (7) of the underground electrical distribution network (1) to be monitored, the sensor devices (3) of the network being distributed so that all sections of conductors between branches they have connected to at least one sensor device (3), each sensor device (3) having:

^• a Wireless communication module {30} for communication with other sensor devices (3) of the network located within its scope;

- data processing means (31);

^« Means for measuring the current (34) circulating through the conductor (?) To which it is coupled;

the sensor devices (3) being synchronized with each other and configured to, once a fault event is detected, identify the type of fault originated and the point of the underground electrical distribution network (1) where the fault occurred through the exchange of messages, between the different sensor devices (3), with information on the synchronized current measurements and through the analysis of the fasoriaf information of said synchronized current measurements, taking into account the network topology.

2. System according to claim 1, characterized in that the sensor devices (3) are configured to:

- determine the line segment (1 0) in which the fault is found by comparing the input and output currents of the different segments that make up the underground electrical distribution network (1),

- classify the type of fault in the affected segment, determining the line or lines of the segment (10) subject to the fault and the type of insulation failure produced, and in particular if the fault is caused by a grounding or a loss insulation inside the cable;

- estimate the location of the fault within the segment (1 0) in which the fault has occurred taking into account its classification.

3. System according to any of the preceding claims, characterized in that the sensor devices (3) are configured to perform the joint synchronization using the zero-pass detection of the current measurements of the sensor nodes, and taking into account the situation of ios sensor nodes within the underground electricity distribution network {1} to moniiorízar and that the surnatono fasoriai of the currents in all nodes is equal to zero,

System according to any of the preceding claims, characterized in that the sensor devices (3) have a supply coil (38) that surrounds the conductor (7) and through which it receives the power supply,

System according to any of the preceding claims _: characterized in that the current measurement means (34) of the sensor devices (3) comprise at least one Rogowski coil (37).

6 System according to any of the preceding claims, characterized in that the sensor devices {3} are located in boxes (8) of the underground electrical distribution network (1), in such a way that they allow monitoring the entire power flow of the network underground electrical distribution (1),

System according to any of the preceding claims, characterized in that it additionally comprises a control unit (6) and a gateway (5) with wireless communication capability configured to collect information from the wireless sensor network (2) and transmit it to the control unit (6).

S. Procedure for classifying and locating faults in an underground electrical distribution network, characterized in that it comprises:

■ establishing a wireless sensor network (2) in an underground electrical distribution network (1), said wireless sensor network {2) being formed by a plurality of sensor devices (3) synchronized with each other and constituting the sensor nodes of the net; where each sensor device (3) is coupled to a conductor (7) of the underground electrical distribution network (1) to be non-linearized, the sensor devices (3) of the network being distributed so that all sections of conductors between branches have associated at least one sensor device (3);

- measure, by each sensor device (3), the current flowing through the conductor (7) to which it is associated; - before the detection of a fault event, identify the type of fault originated and the location of the faifa by means of the Exchange of messages, between the different sensor devices (3), with information of the synchronized current measurements and through the analysis of the fasoriai information of said synchronized current measurements, taking into account the network topology.

9. Method according to claim 8, characterized in that the location of the fault comprises:

- determine the segment (10) of the line in which the fafa is located by comparing the input and output currents of the different segments that make up the underground electrical distribution network (1),

- classify the type of fault in the affected segment, determining the line or lines of the segment (Q) submitted to taita and the type of insulation failo produced, and in particular if the failure is caused by a branch derivation or by a loss insulation inside the cable,

- estimate the location of the fault within the segment (10) in which the taita occurred taking into account the classification of the same

Method according to any one of claims 8 to 9, characterized in that the sensor devices perform the joint synchronization using the zero-pass detection of the current measurements of the sensor nodes, taking into account the situation of the sensor nodes within the underground eelectric distribution network (1) to be monitored and that the fasoria sum of the currents in iodes and nodes is equal to zero.

eleven . Method according to any one of claims 8 to 10, characterized in that the sensor devices (3) are arranged in boxes (8) of the underground electrical distribution network (1), in such a way that they allow to monitor all the power flow of the I retired underground electrical distribution (1).