CN111596171A - Intelligent distribution network fault diagnosis and positioning integrated system under artificial intelligence deep learning - Google Patents

Abstract

An intelligent distribution network fault diagnosis and positioning integrated system under artificial intelligence deep learning belongs to the field of fault diagnosis. The system comprises a fault judging and line selecting device, a fault indicator and a wireless communication part; the fault indicator collects transient current on a line, when zero sequence voltage and zero sequence current are detected to exceed set values, the fault judging and line selecting device starts zero sequence fault diagnosis, identifies whether the single-phase earth fault is stable or not, then identifies a specific fault line and a fault phase, meanwhile recalls fault indicator current of the fault phase, positions specific fault points according to the principle that current magnitude changes suddenly before and after the fault points, and outputs signals to the monitoring equipment. The method comprises the steps of establishing a multi-dimensional urban distribution network multi-system coordination control and application mechanism by constructing an urban distribution network running state analysis and evaluation model based on artificial intelligence deep learning and high latitude space-time data driving; three-stage integrated hierarchical progression is realized, and the fault type and the fault point section position are accurately identified.

Description

Intelligent distribution network fault diagnosis and positioning integrated system under artificial intelligence deep learning

Technical Field

The invention belongs to the field of fault monitoring and diagnosis, and particularly relates to a fault diagnosis and positioning system for a power distribution network.

Background

With the rapid development of modern economy, the demand of society for electricity is increasing day by day, and the density of power distribution networks is also increasing day by day.

The realization of Distribution network Automation (Distribution Automation) is a key means for ensuring power supply reliability, realizing economic operation of a power grid and improving the efficiency of the power grid, and is also an important factor related to social production level and quality of life of people.

At present, the Neutral point of a 3-66 kV power distribution network in China is mostly in a low-current grounding mode, so that the power distribution network can be called a low-current grounding System or a Neutral point non-direct grounding System, which includes a Neutral un-grounded System (NUS for short), a Neutral resistor grounded System (NRS for short) and a Neutral over arc suppression coil grounded System (NES for short).

The distribution network in China mostly adopts a radial structure formed by overhead lines and cable lines (mostly used in an urban power supply system), the number of the feeder lines is directly related to the centralized level of power consumption of users, and the number of the feeder lines of the large urban distribution network can generally reach 20-30. With the further transformation and construction of urban and rural power distribution networks, the total length of the lines and the utilization rate of the cables are greatly increased, the capacitance current of the distribution network lines to the ground caused by line distribution capacitance is rapidly increased, and the capacitance current of some power distribution networks can reach 60-70A, even 100A. At the moment, if a single-phase grounding fault occurs on the line, the grounding point is easy to generate arcing, and the safety of the power system is seriously threatened. The neutral point of such a distribution network must be earthed via the arc suppression coil, according to regulations.

In the current development direction of the power system, the low-current grounding mode is applied to a neutral point grounding mode of a medium-low voltage distribution network, can avoid power supply interruption caused by single-phase grounding faults, has strong advantages of being suitable for the distribution network, and cannot be replaced by other grounding modes. Therefore, for a considerable period of time in the future, the low current grounding mode will still be important in power distribution network applications.

According to statistics, in the operation process of a power system, power failure accidents caused by power distribution network faults account for more than 95% of total power failure accidents, wherein 70% of accidents are caused by single-phase earth faults or bus faults. When a single-phase earth fault occurs in a power distribution network, as a zero sequence network has no direct earth point, fault current only circulates through a line-to-earth capacitance loop, fault characteristic quantity is weak, and a neutral point can compensate the fault current after being grounded through an arc suppression coil, so that the fault characteristic is further reduced, and a reliable fault line selection and positioning method is lacked because of difficult characteristic extraction. With the improvement of the requirement of people on the automation level of the distribution network, the problem of fault location of the distribution network is more urgently needed to be fundamentally solved.

In China, related research records are available for the problem of fault line selection of a distribution network as early as 1958, various fault line selection methods are proposed in sequence, and related line selection devices are developed at the same time. Since the 20 th century and the 80 th century, with the continuous maturity of microcomputer protection technology, various online automatic line selection devices were developed and put into use by different manufacturers, but from the opinion returned from the user aspect, the reliability of line selection is not high, the effect is not good, and the line selection technology is not fully mature.

However, the fault section positioning technology which is the current focus research direction also has many problems, such as unreliable partial positioning principle, inaccurate synchronization of signals of different monitoring points, difficulty in acquiring fault signals, and the like. In addition, most of the existing methods still stay in the theoretical research stage, and the technologies actually applied to the field are few, so that the small-current ground fault positioning technology has not made a substantial breakthrough.

The low-current grounding mode increases the structural complexity of the power distribution network, and makes fault line selection and positioning become a recognized problem. With the transformation of the power system from the traditional power grid to the smart power grid, it is very important to realize the self-healing function of the smart power grid, and the problems of rapid fault detection and reliable fault positioning need to be urgently solved.

Disclosure of Invention

The invention aims to solve the technical problem of providing an intelligent distribution network fault diagnosis and positioning integrated system under artificial intelligence deep learning. The fault detection method is based on a fault indicator, a wireless communication technology and a ground fault line selection technology, and can automatically and efficiently detect the section where the fault (line disconnection, interphase short circuit and single-phase grounding) is located. A fault positioning software system in the fault line selection upper computer is matched with a fault indicator with a communication function on a line, and indication information of a fault position and fault time can be given on a geographic information system diagram of a monitoring center within a few minutes after a fault occurs, so that accurate fault positioning service is provided when a 10kV distribution network line has ground fault, short circuit and disconnection fault, meanwhile, blind area-free monitoring of the distribution network is realized, maintenance personnel can be helped to rapidly arrive at the site, the fault is eliminated, normal power supply is recovered, and power supply reliability is greatly improved.

The technical scheme of the invention is as follows: the utility model provides an intelligent distribution network fault diagnosis location integrated system under artificial intelligence degree of depth study, characterized by:

the intelligent distribution network fault diagnosis and positioning integrated system comprises an in-station fault judgment and line selection device, a fault indicator on an out-station line and a wireless communication part;

the in-station fault judging and line selecting device is arranged in a transformer substation and used for monitoring zero sequence voltage and zero sequence current of a neutral point of a transformer in real time, when the zero sequence voltage and the zero sequence current are detected to exceed a set value, zero sequence fault diagnosis is started, whether the single-phase earth fault is stable or not is identified, then a specific fault line and a fault phase are identified, meanwhile, fault indicator current of the fault phase is recalled through a wireless communication front-end processor, the position of the specific fault point is positioned according to the principle that the current magnitude of the fault indicator before and after the fault point changes suddenly, and a signal is output to monitoring equipment;

the fault indicator on the outside line is a main body for detecting load current and fault current, is directly hung on the line, obtains working current through line induction power taking and a backup battery, and collects transient current on the line through a current transformer coil. The fault indicator is provided with a short-distance wireless communication module and can upload real-time data, historical data and fault data;

the wireless communication part comprises short-distance wireless equipment and VPN 4G equipment, short-distance wireless modules are arranged on a communication front-end processor in the station and a line fault indicator outside the station, and the function of the communication front-end processor in the station establishes a wireless link with the line fault indicator through the wireless modules to concentrate data; the wireless module on the out-of-station fault indicator functions as a wireless repeater and transmits local data; meanwhile, data interaction is carried out on the on-site data through the VPN 4G module and a remote master station;

a short-distance wireless transmission module and a VPN 4G wireless module are used for organically combining a fault indicator on a line with ground fault line selection equipment, relay protection equipment, a background main station and the like at a transformer substation end, and comprehensively judging short-circuit faults and ground faults of the line.

Specifically, the in-station fault judging and line selecting device comprises a positioning decision device, a data concentrator, a grounding detection comprehensive diagnosis device and a background/GIS;

the positioning decision device is used for positioning the whole fault;

the data concentrator is responsible for communication between each wireless communication module and the positioning decision device;

the grounding detection comprehensive diagnosis device is responsible for judging a single-phase grounding branch of the system;

and the background/GIS is responsible for displaying and storing fault positioning results.

Specifically, the fault indicator on the outside-station line is used for fault judgment and positioning, and provides real-time running conditions of line current for all background services.

Specifically, each group of the fault indicators is 3, and the ABC three phases of the line are respectively monitored; each fault indicator comprises a short-distance wireless module, the short-distance wireless modules can be mutually cascaded to form a network, and when a fault line selection call is received, data can be relayed and transmitted back to a fault line selection device;

the fault indicator is an external hanging fault indicator and is suspended in a grading or rod-dividing mode, each grade or rod is divided into A, B, C three-phase lines, and each phase line is correspondingly suspended with one fault indicator; each fault indicator adopts an independent identification code, the current number and harmonic data on the line are transmitted to the main station in a wireless cascade networking mode, and the main station analyzes the data and compares the codes to obtain the type and the position of the fault.

Furthermore, the intelligent distribution network fault diagnosis and positioning integrated system completes detection on the interphase short-circuit fault through the fault indicators hung on the line, can judge that the line is broken or the interphase short-circuit fault occurs when the fault indicators detect that the line current has sudden change and meet corresponding criteria, and then collects data of each fault indicator through the upper computer of the fault diagnosis system and completes fault positioning; for the earth fault, differential changes of current signals before and after a fault point in a line are monitored on line through a fault indicator, and positioning is completed by matching an upper computer of a comprehensive fault diagnosis system installed in a transformer substation with the fault indicator on the line.

Further, the corresponding criteria include an interphase short circuit criterion and a grounding criterion;

the interphase short circuit criterion comprises the following steps:

1) the positive sudden change current in the line is larger than a set value Iset and lasts for a period of time, namely It is larger than or equal to Iset;

2) 2S after the current is suddenly changed, the circuit breaker is detected to be tripped, the line current is reduced to 0, and in other words, Ins is 0;

wherein It is an abrupt change current starting value, Ins is a line current measured n seconds after current abrupt change, and at the moment, the circuit breaker has tripped and the line current is reduced to 0;

the grounding criterion comprises the following steps:

the method comprises the steps that a line selection diagnosis device is judged by utilizing ground faults in a transformer substation, zero sequence voltage of a neutral point of a transformer is monitored in real time, when the fact that the variation value of the zero sequence voltage exceeds 15% -30% is detected, zero sequence fault diagnosis is started, and whether the zero sequence voltage is a single-phase ground fault or not is identified; if the single-phase grounding is achieved, grounding and line selection are carried out through a fault diagnosis device in the station, and the grounding fault is positioned on a certain phase of a certain line; and when the single-phase earth fault is judged to be a stable fault, the fault earth line selection device recalls real-time current data of the fault indicator on the fault phase line through the communication front-end processor, compares data difference before and after a fault point through comparing the acquired signals, and judges that two fault indicators with sudden change can judge the specific section with the fault.

Specifically, the default value of the duration is 2 power frequency cycles, i.e., 40 ms.

Furthermore, the intelligent distribution network fault diagnosis and positioning integrated system calls the recording state information of the fault branch fault indicator when the system has a stable earth fault, and determines the fault point through fault integrated positioning judgment.

Furthermore, the intelligent distribution network fault diagnosis and positioning integrated system establishes a multi-dimensional urban distribution network multi-system coordination control and application mechanism by constructing an urban distribution network operation state analysis and evaluation model based on artificial intelligence deep learning and high latitude space-time data driving;

the control and application mechanism of multi-dimensional urban distribution network multi-system coordination at least comprises:

and (3) time dimension: acquiring transverse data, calculating edges and controlling a sequence in a longitudinal hierarchical manner; short-period fault judgment and positioning, long-period state evaluation, trend analysis and auxiliary decision making; the assistant decision includes the measures and schemes for managing the problems of earth fault capacitance current compensation, overvoltage, system flashover and the like;

and (3) space dimension: judging faults in the station, confirming fault branches and accurately positioning the line sections outside the station; data acquisition and intelligent study and judgment in the station, and data support and disposal linkage outside the station;

target dimension: fault discrimination, branch confirmation, section positioning, auxiliary decision and optimization promotion;

the intelligent judging and positioning service framework comprises: from 'fault discrimination' to 'fault branch confirmation' and then to 'section positioning', three-stage integrated hierarchical and hierarchical progressive reduction of fault points is realized, and fault types and fault point section positions are accurately identified.

Furthermore, the intelligent distribution network fault diagnosis and positioning integrated system realizes intelligent study and judgment and fault positioning through the following modes:

1) recording all zero sequence loop signals in the whole process by using a real-time wave recording technology of full time domain, full frequency band and full type monitoring, and carrying out comprehensive fault on-line diagnosis;

2) and a unique fault waveform feature library is established, and the accurate identification of the fault is realized by comparing the field fault recording and broadcasting data with the fault feature library. Meanwhile, the waveform with specificity can be added into the original feature library so as to enrich the feature library;

3) based on high sampling rate and full-synchronous real-time wave recording, the system calculates whether the zero-sequence voltage of each bus of the system is out of limit in real time, and if the zero-sequence voltage is out of limit, whether the voltage is abnormal caused by faults such as PT disconnection and the like is judged firstly to prevent misjudgment of line selection. After confirming that the PT input signal is normal, judging that the power grid is grounded in a single phase, and immediately starting to judge a grounding branch;

for any single-phase earth fault, the system selects an earth branch by adopting a proper earth line selection algorithm aiming at the signal according to the characteristics of the earth signal; the grounding line selection algorithm of the system comprises the following steps: transient current method, transient energy method, steady state current method, steady state energy method and artificial intelligence deep learning method: the judgment on special and high-resistance grounding faults is effectively increased through deep learning of wave recording data;

4) establish corresponding mathematical model to arc suppression coil, establish the relation of record ripples data and arc suppression coil dynamic characteristic index through mathematical model to calculate arc suppression coil dynamic characteristic, include: the real-time online analysis of action time, dynamic response speed and tracking sensitivity, the dynamic monitoring of capacitance current detection precision and arc suppression coil adjusting range, the identification capability of grounding state and series resonance state and the like, and the calculation and monitoring of the dynamic characteristics of the arc suppression coil also ensure the accuracy of line selection in an arc suppression grounding system;

5) monitoring the overvoltage phenomenon of the power grid, recording overvoltage wave recording data, diagnosing the type and the area of the fault of the overvoltage phenomenon exceeding 10 seconds on line, and recording the overvoltage phenomenon less than 10 seconds. Analyzing the insulation coordination of each device of the transformer substation according to the actual overvoltage condition of the transformer substation;

6) the tripping of a branch outlet can be realized, and the tripping and selective tripping functions of a branch wheel can be set by matching with the operation requirement; the system can realize the acceleration function after the fault by matching with the line protection and the reclosing, thereby ensuring the safe and stable operation of the system;

7) the wave recording data are analyzed through background wave recording data off-line analysis software, the conditions of wrong connection and reverse connection of signals can be found through the analysis of the wave recording data, and the adjustment can be carried out through the software;

8) the system adopts a centralized and distributed optical fiber communication mode, the hardware is modularized, and the system data acquisition expansion unit and the core controller can be connected through optical fibers, so that reliable electrical isolation is realized, and the safety of system operation is improved;

9) accessing a background analysis workstation through an IP network, uniformly managing and analyzing wave recording data of each device by the background analysis workstation, and monitoring the operation of each device;

10) the system adopts a centralized and distributed optical fiber communication mode, the hardware is modularized, and the system data acquisition extension unit and the core controller can be connected through optical fibers, so that reliable electrical isolation is realized, and the safety of system operation is improved.

Compared with the prior art, the invention has the advantages that:

1. by adopting the technical scheme, the fault section can be automatically positioned without manual intervention, and the work load of line patrol is reduced;

2. the detection precision of the current transformer coil is better than 5% through redesigning the current transformer coil;

3. all fault indicators on a line are combined into a network to run by utilizing a 433MHz wireless communication module, the accuracy of fault discrimination and fault status can be greatly improved from the global level through background software, and the phenomena of false alarm and missing alarm are avoided;

4. the comprehensive fault diagnosis system is adopted as the upper control part of the fault positioning system, so that the management of the electronic parts of all fault indicators is realized, the control of an injection signal source is realized, and the final positioning algorithm of fault positioning is realized

5. The system device in the technical scheme has the fault diagnosis functions of single-phase grounding, resonance and the like;

6. on the basis of ground fault discrimination, a mainstream artificial intelligence algorithm library is innovatively provided, high latitude space-time data is established, different types of fault simulation are carried out, and deep learning-based feature self-learning is realized, so that the accuracy of ground fault discrimination is greatly improved.

Drawings

FIG. 1 is a schematic diagram of the hardware configuration of the system of the present invention;

FIG. 2 is a system hardware layout of the present invention;

FIG. 3 is a schematic diagram of the networking mode of the system of the present invention;

FIG. 4 is a block diagram of the workflow of the fault routing apparatus of the present invention;

FIGS. 5 a-5 c are schematic views illustrating the installation process of the fault indicator with a semi-self-locking clamping mechanism according to the present invention;

FIGS. 6 a-6 b are schematic views illustrating the disassembly process of the fault indicator with a semi-self-locking clamping mechanism according to the present invention;

fig. 7a to 7b are schematic views illustrating the assembly and disassembly of the detachable battery compartment of the fault indicator according to the present invention.

Detailed Description

The invention is further illustrated with reference to the following figures and examples.

In fig. 1 and fig. 2, the technical solutions of the present invention provide an intelligent distribution network fault diagnosis and location integrated system under artificial intelligence deep learning.

In order to accurately locate the interphase short-circuit fault and the single-phase earth fault, the system at least comprises three parts:

1. the station internal grounding judgment and line selection equipment;

2. an out-of-station line fault indicator;

3. and a wireless communication section.

The in-station fault judging and line selecting device is installed in a transformer substation and has the main functions of monitoring zero-sequence voltage and zero-sequence current of a neutral point of a transformer in real time, starting zero-sequence fault diagnosis when the zero-sequence voltage and the zero-sequence current are detected to exceed set values, identifying whether the single-phase earth fault is stable or not, then identifying a specific fault line and a fault phase, recalling fault indicator current of the fault phase through a wireless communication front-end processor, positioning the position of the specific fault point according to the principle that the current before and after the fault indicator breaks down, and outputting signals to monitoring equipment.

The fault indicator on the outside line (also called in line, the same below) is a main body for detecting load current and fault current, is directly hung on the line, obtains working current through line induction power taking and a backup battery, and can acquire transient current on the line by a current transformer coil. The fault indicator is provided with a short-distance wireless communication module and can upload real-time data, duration data and fault data.

The wireless communication part can be divided into short-distance wireless equipment and VPN 4G equipment, short-distance wireless modules are arranged on a communication front-end processor in the station and a line fault indicator outside the station, and the communication front-end processor in the station has the function of establishing a wireless link with the line fault indicator through the wireless modules and centralizing data; the wireless module on the out-of-station fault indicator functions as a wireless repeater and transmits local data; meanwhile, data interaction can be carried out on site through the VPN 4G module and a remote main station.

The short-distance wireless transmission module and the VPN 4G wireless module can be used for organically combining a fault indicator on a line with ground fault line selection equipment, relay protection equipment, a background main station and the like at a transformer substation end, and comprehensively judging the short-circuit fault and the ground fault of the line.

The whole system is divided from functional modules and comprises the following modules:

1) a positioning decision device: this is the core of the overall system and is used to complete the localization of the entire fault.

2) A data concentrator: and the positioning decision device is responsible for communication between each wireless communication module and the positioning decision device.

3) Ground fault detection (comprehensive diagnosis) device: and the system is responsible for judging the single-phase grounding branch of the system.

4) background/GIS: and the fault positioning device is responsible for displaying and storing fault positioning results. The partial device is a device originally arranged on the station side.

5) Line fault indicator: and judging and positioning faults, and providing real-time running conditions of line current for all background services.

Wherein, the fault indicator in the circuit is 3 per group, monitors ABC three-phase of circuit respectively. Each fault indicator comprises a short-distance wireless module, the transmission distance is 2-5km (open area), the fault indicators can be cascaded with one another to form a network, and data can be relayed back to a fault line selection device when a fault line selection call is received.

The networking mode of the fault indicators in the line is shown in fig. 3, a wireless cascade networking mode is adopted among the fault indicators, and a wireless cascade networking mode is also adopted between the fault indicators and the communication front-end processor.

According to the intelligent distribution network fault diagnosis and positioning integrated system in the technical scheme, the interphase short-circuit fault is mainly completed by the fault indicator hung on the line, when the fault indicator detects that the line current has sudden change and meets corresponding criteria, the line disconnection or interphase short-circuit fault can be judged, and then the upper computer of the fault diagnosis system collects data of each fault indicator and completes fault positioning. And for the ground fault, the upper computer of the comprehensive fault diagnosis system installed in the transformer substation is matched with a fault indicator on the line to complete positioning.

The interphase short circuit detection principle is as follows: a fault indicator hanging on a line confirms a short circuit fault by sensing an abrupt current flowing through the line.

According to the characteristics of short circuit, the fault indicator can judge the fault by measuring the current sudden change and the duration time in the line by an electromagnetic induction method. Therefore, the device is a fault detection device which is adaptive to load current change and only related to short-circuit current component in fault. The main criteria are as follows:

1. a positive sudden change in current in the line is greater than the set value (Iset) and continues for a period of time (default value is 2 cycles 40 ms);

2. and 2S after the current mutation, the circuit breaker is detected to be tripped, and the line current is reduced to 0 (the time can be set, the setting can be cancelled according to the actual requirement, and no matter whether the reclosing is successful, an event is reported).

According to the above two conditions, the short circuit criterion can be set as follows:

It≥Iset；

Ins＝0。

in the above 2 criteria: it is the sudden change current starting value, Ins is the measured line current n seconds after the current sudden change. At this point, the circuit breaker has tripped and the line current drops to 0.

In the short circuit judgment process, lightning stroke is mainly filtered by the duration of 40ms, and exciting current is mainly removed by the direct current component of the current and information communication between the background main station and the relay protection device. And if the current is exciting current, the system can not trip, and the power supply of the power grid is still normal after n seconds.

According to the technical scheme, after the equipment in the station diagnoses the fault, the differential change of current signals before and after the fault point in the line is monitored on line through the fault indicator, so that the fault is positioned.

1. The core criterion is as follows:

the core criterion is as follows: the method comprises the steps of providing a multidimensional (time, space and target) urban distribution network fault multidimensional intelligent studying, judging and positioning framework, constructing an urban distribution network running state analysis and evaluation model based on artificial intelligence deep learning and high latitude spatio-temporal data driving, and establishing a multi-system coordination control and application mechanism and a corresponding system of the urban distribution network. Multidimensional (time, space and target) urban distribution network fault multidimensional intelligent studying and judging and positioning service framework. Comprises the following steps:

and (3) time dimension: acquiring transverse data, calculating edges and controlling a sequence in a longitudinal hierarchical manner; short-period fault judgment and positioning, long-period state evaluation, trend analysis and auxiliary decision (including earth fault capacitor current compensation, overvoltage, system flashover and other problem treatment measures and schemes)

And (3) space dimension: judging faults in the station, confirming fault branches, and accurately positioning sections outside the station (lines); data acquisition and intelligent study and judgment in the station, and data support and disposal outside the station (line) are linked.

Target dimension: fault discrimination, branch confirmation, section positioning, auxiliary decision and optimization promotion.

The intelligent judging and positioning service framework comprises: from 'fault discrimination' to 'fault branch confirmation' and then to 'section positioning', three-stage integrated hierarchical and hierarchical progressive reduction of fault points is realized, and fault types and fault point section positions are accurately identified

2. The detection principle is as follows:

when the power grid works normally and stable earth faults occur, the difference of the front fault component and the rear fault component of the fault point of the system is large, and by utilizing the characteristic, when the system has the stable earth faults, the recording state information of the fault branch fault indicator is called, and the fault point is confirmed through comprehensive fault positioning and judgment.

3. The judging method comprises the following steps:

as shown in fig. 4, in the present technical solution, a ground fault diagnosis and line selection diagnostic device in a substation is used to monitor the zero sequence voltage of the neutral point of the transformer in real time, when it is detected that the variation value of the zero sequence voltage exceeds 15% (30% of the system grounded through the arc suppression coil), the zero sequence fault diagnosis is started to identify whether the zero sequence voltage is a single-phase ground fault, and if the zero sequence voltage is single-phase grounded, the ground fault is selected by the fault diagnostic device in the substation, and the ground fault is located on a certain phase of a certain line. And when the single-phase earth fault is judged to be a stable fault, the fault earth line selection device recalls real-time current data of the fault indicator on the fault phase line through the communication front-end processor, compares data difference before and after a fault point through comparing the acquired signals, and judges that two fault indicators with sudden change can judge the specific section with the fault.

The technical scheme of the invention can automatically and efficiently detect the section where the fault (line break, interphase short circuit and single-phase grounding) is located based on the fault indicator, the wireless communication technology and the ground fault line selection technology. The fault location software system of the fault line selection upper computer is matched with a fault indicator with a communication function on a line, and the indication information of the fault position and the fault time can be given on a geographic information system diagram of a monitoring center within a few minutes after the fault occurs, so that maintenance personnel can be helped to rush to the site quickly, the fault is eliminated, normal power supply is recovered, and the power supply reliability is greatly improved. The main characteristics of the method include:

1) the fault section is automatically positioned without manual intervention, and the work load of line patrol is reduced;

2) a current transformer coil is redesigned, and the detection precision is better than 5%;

3) all fault indicators on a line are combined into a network to run by utilizing a 433MHz wireless communication module, the accuracy of fault discrimination and fault status can be greatly improved from the global level through background software, and the phenomena of false alarm and missing alarm are avoided;

4) the comprehensive fault diagnosis system is used as an upper control part of the fault positioning system, the upper control part realizes the management of electronic parts of fault indicators, realizes the control of an injection signal source and realizes a final positioning algorithm of fault positioning. The device has single-phase ground connection, resonance and other fault diagnosis functions simultaneously.

5) On the basis of ground fault discrimination, an incoming main stream artificial intelligent algorithm library is innovatively provided, high latitude space-time data are established, different types of fault simulation are carried out, deep learning-based feature self-learning is realized, and therefore the accuracy of ground fault discrimination is greatly improved.

The fault indicator in the technical scheme is installed on outdoor equipment and is directly hung on a 10KV line.

In the technical scheme, each fault indicator is arranged in a hanging mode in a grading mode (a rod), each grade (rod) is divided into A, B, C three phases, and each fault indicator corresponds to one fault indicator. Two stages (poles) too far apart to communicate are cascaded by adding repeaters. The fault indicator adopts independent identification codes, can transmit the current number and harmonic data on a line to a main station in a wireless cascade networking mode, and obtains the type and the position of a fault by analyzing the data and contrasting the codes through the main station.

The mechanical structure of the fault indicator is of great concern to the overall reliability of the product and the performance of the fault indicating system.

The fault indicator in the technical scheme of the invention adopts a semi-self-locking type clamping suspension mechanism. The dangerous factors and the dismounting operation of the fault indicator on the installation site are considered; the structure and the mechanism are matched to complete a certain semi-self-locking function.

Because the installation occasion of the fault indicator is special, the fixing structure of the fault indicator is similar to a clip of a clamp, and the fault indicator is hung on a 10KV electric wire through the clip. Because the device is installed at high altitude, the buckle can be fixed temporarily after being opened, when the buckle is closed when the buckle touches an overhead wire, the fault indicator is hung on the wire. When will take off, down pull down fault indicator, then the buckle is opened, opens to the certain degree, then fixes temporarily, and fault indicator just can follow the electric wire and take off.

The fault indicator realizes non-contact safe installation and disassembly through the design of the semi-self-locking clamping mechanism, and is convenient and flexible to operate.

As shown in fig. 5a to 5c, when the fault indicator is raised, the half self-locking clamping mechanism of the fault indicator is pressed down by the transmission line, and then the half self-locking clamping mechanism of the fault indicator is automatically closed, so that the fault indicator is installed. The fault indication system can then function properly through its configuration.

The fault indicator is assembled and disassembled as shown in fig. 6a to 6b, when the fault indicator is to be removed, the fault indicator is pulled down, the semi-self-locking clamping mechanism is separated and locked to be opened, and then the fault indicator can be removed.

The fault indicator is battery powered and the design of the battery compartment is very important to the product in view of the use of the fault indicator.

To facilitate installation and replacement of the battery; meanwhile, the fault indicator in the technical scheme of the invention adopts the structural form of the detachable battery compartment and the standby battery compartment to facilitate the replacement and expansion of the battery compartment, so that the convenience in taking and putting the fault indicator is ensured.

As shown in fig. 7a and 7b, the fault indicator and the battery compartment below the fault indicator are detachable, and the structure uses the rotary bayonet structure between the bayonet type electric bulb and the electric lamp holder for reference, after the battery compartment is rotated to one side, the battery compartment can be clamped at the lower end of the fault indicator, and after the battery compartment is rotated to the other side, the battery compartment can be taken down smoothly.

The expansion of the battery compartment can be realized by replacing the battery compartments with different capacities.

The fault indicator in the technical scheme of the invention realizes the detection and monitoring functions through the realization of the following functions:

1) getting electricity on line:

the fault indicator is hung on a 10KV high-voltage line to work, when the current in the line is small, the battery supplies power, and when the current in the line is large, the circuit is directly powered to work in order to prolong the power supply time of the battery. This improves the service life of the fault indicator.

2) Networking, collecting and uploading data:

the short-range wireless communication used by the fault indicators is limited in distance, so it is not possible to have each fault indicator directly connected to the concentrator. But at the same time to enable one concentrator to be connected with more fault indicators in order to save costs. Appropriate networking considerations need to be taken into account between the fault indicators. Finally, a communication relay mode is considered, and each time communication is carried out, one group of fault indicators are used as a unit to complete one communication. Of course, the number of fault indicators in a group cannot be too large, but the sum of the three phases is sufficient for a concentrator. Therefore, the system construction cost is saved, the networking function of short-range wireless communication is realized, and in addition, the power consumption of each fault indicator is similar in the communication aspect, and the service life difference of the fault indicators can not be too large.

3) And (3) fault judgment:

the accurate judgment of the fault is the core of the fault indicator. In order to ensure that the fault signal can be collected, in addition to AD sampling, whether the positive half wave and the negative half wave of the signal exceed a certain value or not is judged through the external interruption of the main chip, so that the fault signal can be captured more quickly.

4) Low power consumption techniques:

since the fault indicator is battery powered, the power consumption of the circuit directly determines the service life of the fault indicator.

5) Normally working for a long time:

since the fault indicator is a non-stop operation, it needs to be used for several years, or has a problem but can be quickly recovered to a normal state.

The fault indicator in the technical scheme of the invention adopts a power supply scheme that three power supply modes of a battery, an electric storage element and on-line power taking coexist, and when the current of a power grid is large enough to realize on-line power taking, the power supply device preferentially adopts mutual inductance power taking work and stores energy for the electric storage element, so that the consumption of the battery is saved, and the service life of the battery is prolonged.

Structurally, the fault indicator adds a backup battery compartment design. Under the condition that the volume of the fault indicator is basically unchanged, a spare battery bin is designed, and a spare battery can be added according to actual conditions so as to prevent the service life of the fault indicator product from being influenced by the fact that the original battery is not enough. This ensures that the sufficiency of the power supply is ensured.

On the basis, the power supply switching circuit and the working mode are optimized, the main control chip is awakened when the current of the power grid is large, and the main control chip enters a dormant state when the current of the power grid is small. And a power-taking mutual inductor rapid saturation function and an overcurrent protection circuit are added in the circuit, so that the safe and stable operation of the equipment is ensured.

In the aspect of data communication of the fault indicator, a current load curve is reported at regular time every day, and communication is only generated when a fault occurs at other moments, so that the control of the power consumption of the wireless communication module is ensured.

The in-station fault discrimination line selection device is matched with the out-station line fault indicator unit, and the discrimination and positioning are progressively researched and carried out in a grading and layering manner through three-stage integration of fault discrimination, fault branch confirmation and zone positioning.

The specific implementation way and mode are as follows:

1) the whole-process real-time wave recording technology comprises the following steps:

all zero sequence loop signals are recorded in the whole process by adopting a real-time wave recording technology utilizing full-time domain, full-frequency band and full-type monitoring, and comprehensive fault online diagnosis is carried out.

2) Fault feature library technique:

and a unique fault waveform feature library is established, and the accurate identification of the fault is realized by comparing the field fault recording and broadcasting data with the fault feature library. Meanwhile, the waveform with specificity can be added into the original feature library so as to enrich the feature library.

3) The line selection algorithm based on transient and steady state and artificial intelligence deep learning comprises the following steps:

based on high sampling rate and full-synchronous real-time wave recording, the system calculates whether the zero-sequence voltage of each bus of the system is out of limit in real time, and if the zero-sequence voltage is out of limit, whether the voltage is abnormal caused by faults such as PT disconnection and the like is judged firstly to prevent misjudgment of line selection. After confirming that the PT input signal is normal, judging that the power grid is grounded in a single phase, and immediately starting to judge the grounding branch.

For any single-phase earth fault, the system selects the earth branch by adopting an appropriate earth route selection algorithm aiming at the signal according to the characteristics of the earth signal. The grounding line selection algorithm of the system comprises the following steps: transient current method, transient energy method, steady state current method, steady state energy method and artificial intelligence deep learning method: the judgment on special and high-resistance grounding faults is effectively increased through deep learning of the wave recording data.

4) Calculating and monitoring the dynamic characteristics of the arc suppression coil:

establish corresponding mathematical model to arc suppression coil, establish the relation of record ripples data and arc suppression coil dynamic characteristic index through mathematical model to calculate arc suppression coil dynamic characteristic, include: the method comprises the steps of real-time online analysis of action time, dynamic response speed and tracking sensitivity, dynamic monitoring of capacitance current detection precision and arc suppression coil adjustment range, identification capability of a grounding state and a series resonance state and the like, and calculation and monitoring of dynamic characteristics of the arc suppression coil also ensure the accuracy of line selection in an arc suppression grounding system.

5) And (3) overvoltage monitoring:

monitoring the overvoltage phenomenon of the power grid, recording overvoltage wave recording data, diagnosing the type and the area of the fault of the overvoltage phenomenon exceeding 10 seconds on line, and recording the overvoltage phenomenon less than 10 seconds. And analyzing the insulation coordination of each device of the transformer substation according to the actual overvoltage condition of the transformer substation.

6) The advancement of software and hardware systems:

the whole system adopts a nanosecond and distributed multi-module control hardware structure of DSP + ARM + FPGA and a layered software architecture, so that the system is open in functional mode and easy to realize multiple functions.

7) Selecting tripping and alternate cutting tripping and the like:

the tripping of a branch outlet can be realized, and the tripping and selective tripping functions of a branch wheel can be set by matching with the operation requirement; and the system can realize the acceleration function after the fault by matching with the line protection and the reclosing, thereby ensuring the safe and stable operation of the system.

8) Engineering wiring and polarity error correction:

the wave recording data can be analyzed through background wave recording data off-line analysis software, the conditions of wrong connection and reverse connection of signals can be found through the analysis of the wave recording data, and the adjustment can be carried out through the software.

9) The method has the following functions of background analysis and management:

the device is accessed to a background analysis workstation through an IP network, and the background analysis workstation is used for uniformly managing and analyzing wave recording data of each device and monitoring the operation of each device. The background analysis workstation may provide a variety of analysis tools as desired.

10) High reliability and easy expansion:

the system adopts a centralized and distributed optical fiber communication mode, the hardware is modularized, and the system data acquisition extension unit and the core controller can be connected through optical fibers, so that reliable electrical isolation is realized, and the safety of system operation is improved.

The technical scheme of the invention is based on a fault indicator, a wireless communication technology and a ground fault line selection technology, an in-station fault discrimination line selection device is matched with an out-station line fault indicator unit, and is integrated, graded, layered and developed for judgment and positioning through three stages of fault discrimination, fault branch confirmation and section positioning, a fault positioning software system in a fault line selection upper computer is matched with a fault indicator on a line, and indication information of fault position and fault time is given on a geographic information system diagram of a monitoring center, so that accurate fault positioning service is provided when a 10kV distribution network line has ground fault, short circuit and line break fault, meanwhile, blind area-free monitoring on a network is realized, maintenance personnel can be helped to quickly arrive at a site, faults are eliminated, normal power supply is recovered, and power supply reliability is greatly improved.

The invention can be widely applied to the field of monitoring and positioning of the operation faults of the power distribution network.

Claims (10)

1. The utility model provides an intelligent distribution network fault diagnosis location integrated system under artificial intelligence degree of depth study which characterized by:

the intelligent distribution network fault diagnosis and positioning integrated system comprises an in-station fault judgment and line selection device, a fault indicator on an out-station line and a wireless communication part;

the in-station fault judging and line selecting device is arranged in a transformer substation and used for monitoring zero sequence voltage and zero sequence current of a neutral point of a transformer in real time, when the zero sequence voltage and the zero sequence current are detected to exceed a set value, zero sequence fault diagnosis is started, whether the single-phase earth fault is stable or not is identified, then a specific fault line and a fault phase are identified, meanwhile, fault indicator current of the fault phase is recalled through a wireless communication front-end processor, the position of the specific fault point is positioned according to the principle that the current magnitude of the fault indicator before and after the fault point changes suddenly, and a signal is output to monitoring equipment;

the fault indicator on the outside line is a main body for detecting load current and fault current, is directly hung on the line, obtains working current through line induction power taking and a backup battery, and collects transient current on the line through a current transformer coil; the fault indicator is provided with a short-distance wireless communication module and can upload real-time data, historical data and fault data;

the wireless communication part comprises short-distance wireless equipment and VPN 4G equipment, short-distance wireless modules are arranged on a communication front-end processor in the station and a line fault indicator outside the station, and the function of the communication front-end processor in the station establishes a wireless link with the line fault indicator through the wireless modules to concentrate data; the wireless module on the out-of-station fault indicator functions as a wireless repeater and transmits local data; meanwhile, data interaction is carried out on the on-site data through the VPN 4G module and a remote master station;

a short-distance wireless transmission module and a VPN 4G wireless module are used for organically combining a fault indicator on a line with ground fault line selection equipment, relay protection equipment, a background main station and the like at a transformer substation end, and comprehensively judging short-circuit faults and ground faults of the line.

2. The intelligent distribution network fault diagnosis and location integrated system under artificial intelligence deep learning of claim 1, wherein the in-station fault judgment and line selection device comprises a location decision device, a data concentrator, a grounding detection comprehensive diagnosis device and a background/GIS;

the positioning decision device is used for positioning the whole fault;

the data concentrator is responsible for communication between each wireless communication module and the positioning decision device;

the grounding detection comprehensive diagnosis device is responsible for judging a single-phase grounding branch of the system;

and the background/GIS is responsible for displaying and storing fault positioning results.

3. The intelligent distribution network fault diagnosis and location integrated system under artificial intelligence deep learning of claim 1, wherein the fault indicator on the outside-station line is used for fault judgment and location, and provides real-time operation conditions of line current for all background services.

4. The intelligent distribution network fault diagnosis and location integrated system under artificial intelligence deep learning of claim 3, wherein each group of 3 fault indicators respectively monitors ABC three phases of a line; each fault indicator comprises a short-distance wireless module, the short-distance wireless modules can be mutually cascaded to form a network, and when a fault line selection call is received, data can be relayed and transmitted back to a fault line selection device;

the fault indicator is an external hanging fault indicator and is suspended in a grading or rod-dividing mode, each grade or rod is divided into A, B, C three-phase lines, and each phase line is correspondingly suspended with one fault indicator; each fault indicator adopts an independent identification code, the current number and harmonic data on the line are transmitted to the main station in a wireless cascade networking mode, and the main station analyzes the data and compares the codes to obtain the type and the position of the fault.

5. The intelligent distribution network fault diagnosis and positioning integrated system under artificial intelligence deep learning of claim 1, which is characterized in that the intelligent distribution network fault diagnosis and positioning integrated system is used for detecting interphase short circuit faults by fault indicators hung on a line, when the fault indicators detect that line current has sudden change and meet corresponding criteria, the line disconnection or the interphase short circuit faults can be judged, and then an upper computer of the fault diagnosis system collects data of each fault indicator and completes fault positioning; for the earth fault, differential changes of current signals before and after a fault point in a line are monitored on line through a fault indicator, and positioning is completed by matching an upper computer of a comprehensive fault diagnosis system installed in a transformer substation with the fault indicator on the line.

6. The intelligent distribution network fault diagnosis and location integrated system under artificial intelligence deep learning of claim 5, wherein the corresponding criteria include interphase short circuit criteria and grounding criteria;

the interphase short circuit criterion comprises the following steps:

1) the positive sudden change current in the line is larger than a set value Iset and lasts for a period of time, namely It is larger than or equal to Iset;

2) 2S after the current is suddenly changed, the circuit breaker is detected to be tripped, the line current is reduced to 0, and in other words, Ins is 0;

wherein It is an abrupt change current starting value, Ins is a line current measured n seconds after current abrupt change, and at the moment, the circuit breaker has tripped and the line current is reduced to 0;

the grounding criterion comprises the following steps:

the method comprises the steps that a line selection diagnosis device is judged by utilizing ground faults in a transformer substation, zero sequence voltage of a neutral point of a transformer is monitored in real time, when the fact that the variation value of the zero sequence voltage exceeds 15% -30% is detected, zero sequence fault diagnosis is started, and whether the zero sequence voltage is a single-phase ground fault or not is identified; if the single-phase grounding is achieved, grounding and line selection are carried out through a fault diagnosis device in the station, and the grounding fault is positioned on a certain phase of a certain line; and when the single-phase earth fault is judged to be a stable fault, the fault earth line selection device recalls real-time current data of the fault indicator on the fault phase line through the communication front-end processor, compares data difference before and after a fault point through comparing the acquired signals, and judges that two fault indicators with sudden change can judge the specific section with the fault.

7. The intelligent distribution network fault diagnosis and location integrated system under artificial intelligence deep learning of claim 6, wherein the default value of the duration is 2 power frequency cycles, namely 40 ms.

8. The intelligent distribution network fault diagnosis and positioning integrated system under the artificial intelligence deep learning of claim 6, characterized in that the intelligent distribution network fault diagnosis and positioning integrated system, when the system has a stable earth fault, calls the recording state information of the fault branch fault indicator, and determines the fault point through the fault comprehensive positioning judgment.

9. The intelligent distribution network fault diagnosis and location integrated system under artificial intelligence deep learning according to claim 1, characterized in that the intelligent distribution network fault diagnosis and location integrated system establishes a multi-dimensional urban distribution network multi-system coordination control and application mechanism by constructing an urban distribution network operation state analysis and evaluation model based on artificial intelligence deep learning and high latitude spatio-temporal data driving;

the control and application mechanism of multi-dimensional urban distribution network multi-system coordination at least comprises:

and (3) time dimension: acquiring transverse data, calculating edges and controlling a sequence in a longitudinal hierarchical manner; short-period fault judgment and positioning, long-period state evaluation, trend analysis and auxiliary decision making; the assistant decision includes the measures and schemes for managing the problems of earth fault capacitance current compensation, overvoltage, system flashover and the like;

and (3) space dimension: judging faults in the station, confirming fault branches and accurately positioning the line sections outside the station; data acquisition and intelligent study and judgment in the station, and data support and disposal linkage outside the station;

target dimension: fault discrimination, branch confirmation, section positioning, auxiliary decision and optimization promotion;

the intelligent judging and positioning service framework comprises: from 'fault discrimination' to 'fault branch confirmation' and then to 'section positioning', three-stage integrated hierarchical and hierarchical progressive reduction of fault points is realized, and fault types and fault point section positions are accurately identified.

10. The intelligent distribution network fault diagnosis and location integrated system under artificial intelligence deep learning of claim 1, wherein the intelligent distribution network fault diagnosis and location integrated system realizes intelligent study and judgment and fault location by the following means:

1) recording all zero sequence loop signals in the whole process by using a real-time wave recording technology of full time domain, full frequency band and full type monitoring, and carrying out comprehensive fault on-line diagnosis;

2) establishing a unique fault waveform characteristic library, and comparing field fault recording and broadcasting data with the fault characteristic library to realize accurate identification of the fault; meanwhile, the waveform with specificity can be added into the original feature library so as to enrich the feature library;

3) based on high sampling rate and full-synchronous real-time wave recording, the system calculates whether the zero-sequence voltage of each bus of the system is out of limit in real time, if the zero-sequence voltage is out of limit, whether the voltage is abnormal caused by faults such as PT disconnection and the like is judged firstly, so as to prevent misjudgment of line selection; after confirming that the PT input signal is normal, judging that the power grid is grounded in a single phase, and immediately starting to judge a grounding branch;

for any single-phase earth fault, the system selects an earth branch by adopting a proper earth line selection algorithm aiming at the signal according to the characteristics of the earth signal; the grounding line selection algorithm of the system comprises the following steps: transient current method, transient energy method, steady state current method, steady state energy method and artificial intelligence deep learning method: the judgment on special and high-resistance grounding faults is effectively increased through deep learning of wave recording data;

4) establish corresponding mathematical model to arc suppression coil, establish the relation of record ripples data and arc suppression coil dynamic characteristic index through mathematical model to calculate arc suppression coil dynamic characteristic, include: the real-time online analysis of action time, dynamic response speed and tracking sensitivity, the dynamic monitoring of capacitance current detection precision and arc suppression coil adjusting range, the identification capability of grounding state and series resonance state and the like, and the calculation and monitoring of the dynamic characteristics of the arc suppression coil also ensure the accuracy of line selection in an arc suppression grounding system;

5) monitoring the overvoltage phenomenon of the power grid, recording overvoltage wave recording data, diagnosing the type and the fault area of the overvoltage phenomenon exceeding 10 seconds on line, and recording the overvoltage phenomenon less than 10 seconds; analyzing the insulation coordination of each device of the transformer substation according to the actual overvoltage condition of the transformer substation;

6) the tripping of a branch outlet can be realized, and the tripping and selective tripping functions of a branch wheel can be set by matching with the operation requirement; the system can realize the acceleration function after the fault by matching with the line protection and the reclosing, thereby ensuring the safe and stable operation of the system;

7) the wave recording data are analyzed through background wave recording data off-line analysis software, the conditions of wrong connection and reverse connection of signals can be found through the analysis of the wave recording data, and the adjustment can be carried out through the software;

8) the system adopts a centralized and distributed optical fiber communication mode, the hardware is modularized, and the system data acquisition expansion unit and the core controller can be connected through optical fibers, so that reliable electrical isolation is realized, and the safety of system operation is improved;

9) accessing a background analysis workstation through an IP network, uniformly managing and analyzing wave recording data of each device by the background analysis workstation, and monitoring the operation of each device;

10) the system adopts a centralized and distributed optical fiber communication mode, the hardware is modularized, and the system data acquisition extension unit and the core controller can be connected through optical fibers, so that reliable electrical isolation is realized, and the safety of system operation is improved.

Images