A distribution network fault location system and measurement device based on K-means
By using the K-mean algorithm and GPS/Beidou positioning system to process the fault traveling wave signals in the 10kV distribution network, a fault traveling wave feature library is established, and the rapid and accurate positioning and type identification of distribution network faults is achieved, and fault location in the existing technology is solved. Problems of long time and insufficient accuracy.
Patent Information
- Application Number
- CN202210423428.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-21
- Publication Date
- 2024-12-10
- Estimated Expiration
- 2042-04-21
AI Technical Summary
The existing 10kV distribution network fault location technology is difficult to locate fault points quickly and accurately, resulting in a long time to find faults and affecting power supply reliability.
Using a distribution network fault positioning system based on K mean, fault traveling wave signals are collected and processed through components such as current transformer, A/D converter, locked frequency multiplication circuit, FPGA programming processor, DSP signal processor, etc., and combined with GPS Positioning module and Beidou positioning system realize fault positioning and type identification, and use the K-mean algorithm to establish a fault traveling wave feature library for judgment.
It realizes rapid and accurate positioning of distribution network faults, eliminates positioning errors caused by terminal clocks and sampling frequency, improves the accuracy of fault type identification, and solves the problem of insufficient positioning accuracy in the prior art.
Smart Images
Figure CN115184719B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of digital power technology, and in particular to a distribution network fault location system based on K-means and a measuring device thereof. Background Art
[0002] With the rapid development of China's economy, the power grid business has grown rapidly, and the scale of 10kV distribution network has continued to expand. China's 10kV distribution network is widely distributed geographically, with complex line grid structure, long lines and many branches. Affected by factors such as operating environment, power load, equipment aging, tree and bamboo hidden dangers, external force damage, etc., power outages are frequent and fault location is difficult. After a distribution network failure occurs, the operation and inspection personnel often need to find the fault within a few kilometers, and the fault search takes several hours or even days, which is not conducive to timely repair and rapid restoration of reliable power supply. Therefore, we need a distribution network fault location device to quickly locate the fault.
[0003] For example, the invention application with application number: CN111983374A proposes a distribution network fault location device and its location method, which includes a data processing unit, a collection and transmission unit and a box. The data processing unit includes a synchronous clock unit, a traveling wave recording unit and a cloud service platform, which are used for data storage, processing and calculation. The synchronous clock unit adopts a high-precision synchronous clock; the traveling wave recording unit adopts a single-chip microcomputer module to generate a traveling wave arrival time message and send it to the cloud service platform; the cloud service platform adopts a virtual host for background location calculation and fault information sharing; the fault location platform based on the cloud service platform realizes background location calculation and convenient fault information sharing. In the case of Internet, the operation status can be checked at any time and the fault data can be shared, eliminating the inconvenience of fixed IP handling procedures and being vulnerable to virus attacks, shortening the time for finding distribution faults, improving the efficiency of power outage repairs due to distribution faults, and ensuring power supply reliability.
[0004] However, the application provides a device for recording fault traveling waves, but does not propose a feasible technical solution on how to process the fault traveling waves, how to locate the fault point through the fault traveling waves, and how to determine the fault type through the fault traveling waves.
[0005] Therefore, it is necessary to provide a distribution network fault location system and a measurement device based on K-means to solve one of the above technical problems. Summary of the invention
[0006] To solve one of the above technical problems, the present invention proposes a measuring device for locating distribution network faults, including a current transformer, an A / D converter, a phase-locked frequency multiplication circuit, a signal conditioning circuit, an FPGA programming processor, a DSP signal processor, a clock crystal oscillator, a data buffer, a data storage device, a GPS positioning module, a serial communication module, a remote communication module and a power supply module.
[0007] Specifically, the current transformer is used to collect current signals in the distribution network line and is electrically connected to the input pole of the signal conditioning circuit; the output pole of the signal conditioning circuit is electrically connected to the input pole of the phase-locked frequency multiplication circuit and the input pole of the A / D converter respectively; the A / D converter is provided with a CONVST pin and is electrically connected to the output pole of the phase-locked frequency multiplication circuit; the output pole of the A / D converter is electrically connected to the FPGA programming processor; the FPGA programming processor is also electrically connected to a DSP signal processor, a clock crystal oscillator, a data buffer, a data storage device, a GPS positioning module, a serial communication module, a remote communication module and a power supply module; the DSP signal processor is also electrically connected to a clock crystal oscillator, a data buffer and a data storage device.
[0008] A distribution network fault location system based on K-means has measuring devices installed at various locations in the distribution network, and a distribution network fault location master station electrically connected to each measuring device is also arranged at the remote end. The distribution network fault is located by fault traveling wave extraction, fault position location and fault type identification.
[0009] Specifically, fault traveling wave extraction: the fault traveling wave signal is collected through the measuring device, and the first wave signal of the fault traveling wave in the distribution network is continuously monitored and the measuring device clock is calibrated; when the first wave signal of the fault traveling wave is monitored, the first wave waveform data of the fault is recorded to complete the fault traveling wave extraction; it includes the measuring device clock calibration step, the zero-sequence current change sensing step and the fault traveling wave first wave signal extraction step.
[0010] Specifically, fault location positioning: record the time when the first wave signal of the fault traveling wave arrives at each measuring device and calculate the time difference, locate the fault position through the time difference, and perform position calibration to complete the fault location positioning; including the first wave arrival time recording step, the measurement terminal time matrix establishment step, the distribution network fault location positioning step and the distribution network fault location calibration step.
[0011] Specifically, fault type identification: receiving the first wave waveform data of the fault of each measuring device in the distribution network, and extracting the key features, learning the key features through the machine learning model, and identifying and judging the distribution network fault type to complete the fault type identification; including the steps of extracting the first wave waveform data of the fault, extracting the key features of the first wave of the fault and judging the distribution network fault type.
[0012] As a further solution, the measuring device is provided with a GPS positioning module, and is deployed through a Beidou positioning module, and can establish a connection with the Beidou timing system; the distribution network fault location master station regularly sends a timing instruction S to each measuring device a :
[0013]
[0014] Among them, n a is the number of measuring devices in the distribution network, and s is the timing instruction of different measuring devices; each measuring device completes the clock calibration step of the measuring device according to the corresponding timing instruction.
[0015] As a further solution, the measuring device is provided with a current transformer and deployed through a zero-sequence current transformer. The zero-sequence current change sensing step is to sense the three-phase current through the zero-sequence current transformer in the measuring device, and to determine whether the zero-sequence current is generated by calculating whether the sum of the current vectors of the three phases is zero; when the sum of the current vectors of the three phases is not zero, it is determined that the zero-sequence current is generated, and the zero-sequence current i is calculated. z value:
[0016]
[0017] Among them, i a 、i b 、i c , are the current vectors of the three-phase current in the distribution network; G a is the transformation ratio of the zero-sequence current transformer in the measuring device.
[0018] As a further solution, the fault traveling wave first wave signal extraction step samples the fault traveling wave through the analog-to-digital conversion chip in the measuring device, and performs transient recording after the fault traveling wave first wave signal is detected, that is:
[0019]
[0020] Among them, ω i is the frequency component of different fault traveling waves in the distribution network, n c is the number of components of the fault traveling wave frequency, u i is the current value corresponding to different fault traveling wave frequencies in the distribution network, △f is the fault traveling wave current threshold, when it is greater than the fault traveling wave current threshold, the fault traveling wave signal B a Record the first wave signal of transient fault traveling wave.
[0021] As a further solution, when the current value corresponding to the fault traveling wave is greater than the fault traveling wave current threshold, the measuring device records the current time to obtain the first wave arrival time point; the distribution network fault location master station executes the step of establishing the measurement terminal time matrix, that is, collecting the first wave arrival time points of each measuring device and compiling them into the measurement terminal time matrix T a :
[0022]
[0023] Where t is the first wave arrival time recorded by different measuring devices, n d is the number of measuring devices in the distribution network.
[0024] As a further solution, the distribution network fault location location step is performed by double-terminal traveling wave ranging positioning, that is, in the measurement terminal time matrix T a The earliest and second earliest first wave arrival time points are selected, and the measuring devices corresponding to the time points are used as the ranging double ends, and the distance y between the fault point and the ranging double ends is calculated. b and c :
[0025]
[0026] Among them, t b and t c are the earliest and second earliest arrival times of the first wave, respectively. a is the fault traveling wave transmission speed in the distribution network.
[0027] As a further solution, the distribution network fault location calibration step further pairs the remaining measuring devices, performs double-end traveling wave ranging positioning, obtains the fault point location and performs mean calculation to obtain the fault point correction location r a :
[0028]
[0029] Among them, y i The fault location for different measuring devices; n g is the number of measuring devices in the distribution network.
[0030] As a further solution, the first-wave waveform data of the fault is obtained through the first-wave waveform data extraction step, and the first-wave waveform data of the fault is used for model training; the machine learning model in the first-wave key feature extraction step of the fault is further deployed through the K-mean algorithm, and data classification is performed through clustering values, and the clustering value D of the K-mean algorithm is k for:
[0031]
[0032] Among them, n k is the number of K-mean cluster centers, w i is a sample set of traveling wave data for different distribution network faults, u i is the square error of traveling wave data for different distribution network faults.
[0033] As a further solution, the key features of the first wave of each fault are clustered using the K-mean algorithm to obtain the distribution network fault traveling wave feature library; when performing the distribution network fault type judgment step, the current fault first wave waveform data is classified and compared with the distribution network fault traveling wave feature library. If the similarity exceeds the clustering threshold, the distribution network fault type corresponding to the clustering point is output to complete the distribution network fault type judgment.
[0034] Compared with the related art, the distribution network fault location system and its measuring device based on K-means provided by the present invention have the following beneficial effects:
[0035] The present invention proposes a distribution network fault location system based on K-means and its measuring device, which starts the distribution network fault traveling wave measuring device through zero-sequence current change induction, and realizes the extraction of the arrival time and waveform of the first wave of the distribution network fault traveling wave. Then, the mean calculation method is used to calibrate the distribution network fault position to eliminate the influence of the positioning error caused by the terminal clock and sampling frequency. On this basis, the K-mean algorithm is used to establish a distribution network fault traveling wave feature library. By comparing the distribution network fault traveling wave with the value of the feature library, the type of distribution network fault is determined. It solves the problem that only the distribution network short-circuit fault section can be located, the positioning accuracy is several kilometers, and the distribution network fault cannot be accurately located. The double-end partial discharge location method uses the time point when the partial discharge waveform arrives at the device to locate the fault, but is limited by the device clock error and the partial discharge sampling frequency, and the positioning accuracy of the double-end partial discharge location method is not high. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 A flow chart of a preferred embodiment of a distribution network fault location system based on K-means provided by the present invention;
[0037] Figure 2 A schematic structural diagram of a preferred embodiment of a measuring device for locating distribution network faults provided by the present invention. DETAILED DESCRIPTION
[0038] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.
[0039] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.
[0040] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present disclosure.
[0041] like Figure 1 to Figure 2 As shown, the present invention provides a measuring device for locating distribution network faults, including a housing and a mainboard, on which a current transformer, an A / D converter, a phase-locked frequency multiplication circuit, a signal conditioning circuit, an FPGA programming processor, a DSP signal processor, a clock crystal oscillator, a data buffer, a data storage device, a GPS positioning module, a serial communication module, a remote communication module and a power supply module are arranged.
[0042] Specifically, the current transformer is used to sample the current signal in the distribution network line and is electrically connected to the input pole of the signal conditioning circuit; the output pole of the signal conditioning circuit is electrically connected to the input pole of the phase-locked frequency multiplication circuit and the input pole of the A / D converter respectively; the A / D converter is provided with a CONVST pin and is electrically connected to the output pole of the phase-locked frequency multiplication circuit; the output pole of the A / D converter is electrically connected to the FPGA programming processor; the FPGA programming processor is also electrically connected to a DSP signal processor, a clock crystal oscillator, a data buffer, a data storage device, a GPS positioning module, a serial communication module, a remote communication module and a power supply module; the DSP signal processor is also electrically connected to a clock crystal oscillator, a data buffer and a data storage device.
[0043] Specifically, the DSP signal processor is deployed through a chip of model TMS320; the FPGA programming processor is deployed through a chip of model EP1K30, and a dedicated configuration device EPC2 is set externally; the A / D converter is deployed through a chip of model AD7656.
[0044] As a further solution, a watchdog chip is also provided on the main board. The watchdog chip is deployed through a chip model SP706 and is electrically connected to the DSP signal processor. A key module and a display module are also installed on the shell. The key module and the display module are electrically connected to the FPGA programming processor respectively. A buzzer alarm is also provided on the main board.
[0045] As a further solution, the housing is deployed with a signal shielding shell and seals the mainboard inside.
[0046] As a further solution, the current transformer is deployed via a zero-crossing current transformer and samples the zero-crossing current signal in the distribution network line.
[0047] As a further solution, the serial communication module is deployed through a USB serial communication module of model CH340T; the remote communication module is deployed through a GPRS remote communication module; the GPS positioning module is deployed through a Beidou positioning module; wherein the GPRS remote communication module and the Beidou positioning module are deployed through a positioning and communication two-in-one module of model G500-GL GSM.
[0048] As a further solution, an environmental parameter sensor is also provided on the main board, and the environmental parameter sensor includes a temperature sensor, a humidity sensor, an image collector and an infrared sensor; the temperature sensor, the humidity sensor, the image collector and the infrared sensor are electrically connected to the FPGA programming processor respectively.
[0049] As a further solution, the power supply module is deployed through a battery management circuit and an energy storage battery.
[0050] As a further solution, the power supply module is also provided with an AC power supply circuit, which includes a filter and rectifier circuit, a switch control circuit, a DC / DC converter and an overcurrent protection circuit; the input pole of the filter and rectifier circuit is electrically connected to the current transformer, and the output pole is electrically connected to the overcurrent protection circuit; the output pole of the overcurrent protection circuit is electrically connected to the input pole of the DC / DC converter, and the switch control circuit is connected between the energy storage battery and the output pole of the DC / DC converter, and receives a switch control signal from the battery management circuit.
[0051] As a further solution, the signal conditioning circuit is implemented by a signal conditioning circuit module of model LMH6505.
[0052] A distribution network fault location system based on K-means has measuring devices installed at various locations in the distribution network, and a distribution network fault location master station electrically connected to each measuring device is also arranged at the remote end. The distribution network fault is located by fault traveling wave extraction, fault position location and fault type identification.
[0053] Specifically, fault traveling wave extraction: the fault traveling wave signal is collected through the measuring device, and the first wave signal of the fault traveling wave in the distribution network is continuously monitored and the measuring device clock is calibrated; when the first wave signal of the fault traveling wave is monitored, the first wave waveform data of the fault is recorded to complete the fault traveling wave extraction; it includes the measuring device clock calibration step, the zero-sequence current change sensing step and the fault traveling wave first wave signal extraction step.
[0054] Specifically, fault location positioning: record the time when the first wave signal of the fault traveling wave arrives at each measuring device and calculate the time difference, locate the fault position through the time difference, and perform position calibration to complete the fault location positioning; including the first wave arrival time recording step, the measurement terminal time matrix establishment step, the distribution network fault location positioning step and the distribution network fault location calibration step.
[0055] Specifically, fault type identification: receiving the first wave waveform data of the fault of each measuring device in the distribution network, and extracting the key features, learning the key features through the machine learning model, and identifying and judging the distribution network fault type to complete the fault type identification; including the steps of extracting the first wave waveform data of the fault, extracting the key features of the first wave of the fault and judging the distribution network fault type.
[0056] It should be noted that: in the fault traveling wave extraction link, first, the Beidou technology is used to calibrate the clock of the distribution network fault traveling wave measuring device to eliminate the impact of time asynchrony on the distribution network fault identification. Secondly, the distribution network fault traveling wave measuring device senses the change of the zero-sequence current of the distribution network. When the zero-sequence current appears in the distribution network, the distribution network fault traveling wave recording device is started. Finally, the distribution network fault traveling wave measuring device extracts the first wave data of the fault traveling wave. In the fault location link, first, the distribution network fault traveling wave measuring device records the time when the first wave of the fault arrives at the device, and then establishes the time matrix of the first wave arriving at each measurement terminal. On this basis, the time when the traveling wave arrives at the two nearest measurement devices is determined, and the fault point is located according to the time difference between the fault traveling wave arriving at the above two measurement devices. Finally, the mean calculation method is used to calibrate the distribution network fault location. In terms of fault type identification, first, the distribution network fault traveling wave waveform data is extracted. Then, the key change features of the first wave of the fault are extracted, and a typical fault feature library is established through the K-mean algorithm. Finally, the fault traveling wave is compared with the value of the feature library to determine the type of distribution network fault.
[0057] As a further solution, the measuring device is provided with a GPS positioning module, and is deployed through a Beidou positioning module, and can establish a connection with the Beidou timing system; the distribution network fault location master station regularly sends a timing instruction S to each measuring device a :
[0058]
[0059] Among them, n a is the number of measuring devices in the distribution network, and s is the timing instruction of different measuring devices; each measuring device completes the clock calibration step of the measuring device according to the corresponding timing instruction.
[0060] It should be noted that: the 10kV distribution network has many T nodes, and the distribution network fault traveling wave propagates in many directions and has a large attenuation. Therefore, the deployment principle of the 10kV distribution network fault traveling wave measuring device is set as follows: a set of measuring devices is deployed at both ends of the trunk line, and a set of measuring devices is deployed every 4 kilometers. For branch lines greater than 3 kilometers, a set of measuring devices is deployed at both ends. For branch lines less than 3 kilometers, a set of measuring devices is deployed at the head end. After completing the distribution network fault traveling wave measuring device, the spacing between the devices is maintained in the system. Due to the large number of 10kV distribution network fault traveling wave positioning measuring devices, in order to ensure the accurate positioning of the distribution network fault, the clocks between the measuring devices must be synchronized to eliminate the impact of time asynchrony on the distribution network fault identification.
[0061] As a further solution, the measuring device is provided with a current transformer and deployed through a zero-sequence current transformer. The zero-sequence current change sensing step is to sense the three-phase current through the zero-sequence current transformer in the measuring device, and to determine whether the zero-sequence current is generated by calculating whether the sum of the current vectors of the three phases is zero; when the sum of the current vectors of the three phases is not zero, it is determined that the zero-sequence current is generated, and the zero-sequence current i is calculated. z value:
[0062]
[0063] Among them, i a 、i b 、i c , are the current vectors of the three-phase current in the distribution network; G a is the transformation ratio of the zero-sequence current transformer in the measuring device.
[0064] It should be noted that China's 10kV distribution network operates in a neutral point ungrounded mode. In this mode, the sum of the current vectors of the three phases A, B, and C is zero. When a fault occurs in the distribution network, the vector sum of the three phases A, B, and C of the line is not zero, that is, zero-sequence current is generated. In order to reduce the operating power consumption of the distribution network fault traveling wave measurement device, under the normal operating conditions of the distribution network, the distribution network fault traveling wave measurement device is in standby mode. When a fault occurs on the distribution line and a zero-sequence current is generated, the distribution network fault traveling wave measurement device starts and performs fault recording detection.
[0065] As a further solution, the fault traveling wave first wave signal extraction step samples the fault traveling wave through the analog-to-digital conversion chip in the measuring device, and performs transient recording after the fault traveling wave first wave signal is detected, that is:
[0066]
[0067] Among them, ω i is the frequency component of different fault traveling waves in the distribution network, n cis the number of components of the fault traveling wave frequency, u i is the current value corresponding to different fault traveling wave frequencies in the distribution network, △f is the fault traveling wave current threshold, when it is greater than the fault traveling wave current threshold, the transient fault traveling wave first wave signal of the fault traveling wave signal Ba is recorded.
[0068] It should be noted that when a distribution network fault occurs, the generated distribution network fault wave will propagate along the distribution line at a propagation speed of 300m / us
[12] . All distribution network fault wave measurement devices on the 10kV distribution network will receive the arrival time of the first wave of the distribution network fault wave. When the distribution network fault current threshold is greater than △f in Formula 3, the arrival time of the first wave of the distribution network fault wave is recorded.
[0069] As a further solution, when the current value corresponding to the fault traveling wave is greater than the fault traveling wave current threshold, the measuring device records the current time to obtain the first wave arrival time point; the distribution network fault location master station executes the step of establishing the measurement terminal time matrix, that is, collecting the first wave arrival time points of each measuring device and compiling them into the measurement terminal time matrix T a :
[0070]
[0071] Where t is the first wave arrival time recorded by different measuring devices, n d is the number of measuring devices in the distribution network.
[0072] As a further solution, the distribution network fault location location step is performed by double-terminal traveling wave ranging positioning, that is: in the measurement terminal time matrix T a The earliest and second earliest first wave arrival time points are selected, and the measuring devices corresponding to the time points are used as the ranging double ends, and the distance y between the fault point and the ranging double ends is calculated. b and c :
[0073]
[0074] Among them, t b and t c are the earliest and second earliest arrival times of the first wave, respectively. a is the fault traveling wave transmission speed in the distribution network.
[0075] As a further solution, the distribution network fault location calibration step further pairs the remaining measuring devices, performs double-end traveling wave ranging positioning, obtains the fault point location and performs mean calculation to obtain the fault point correction location r a :
[0076]
[0077] Among them, y i is the location of the fault point located by different measuring devices; ng is the number of measuring devices in the distribution network.
[0078] It should be noted that: in order to eliminate the distribution network fault location error caused by the distribution network fault traveling wave clock synchronization and sampling progress, the first wave recording points of all measuring devices under this 10kV distribution network line are used for average calculation, thereby reducing the distribution network fault traveling wave location error.
[0079] As a further solution, the first-wave waveform data of the fault is obtained through the first-wave waveform data extraction step, and the first-wave waveform data of the fault is used for model training; the machine learning model in the first-wave key feature extraction step of the fault is further deployed through the K-mean algorithm, and data classification is performed through clustering values, and the clustering value D of the K-mean algorithm is k for:
[0080]
[0081] Among them, n k is the number of K-mean cluster centers, w i is a sample set of traveling wave data for different distribution network faults, u i is the square error of traveling wave data for different distribution network faults.
[0082] As a further solution, the key features of the first wave of each fault are clustered using the K-mean algorithm to obtain the distribution network fault traveling wave feature library; when performing the distribution network fault type judgment step, the current fault first wave waveform data is classified and compared with the distribution network fault traveling wave feature library. If the similarity exceeds the clustering threshold, the distribution network fault type corresponding to the clustering point is output to complete the distribution network fault type judgment.
[0083] It should be noted that after locating the traveling wave of the distribution network fault, the first wave waveform data of the distribution network fault traveling wave is extracted. Because the traveling wave waveforms of different distribution network fault types present different states, the distribution network fault type is identified based on the traveling wave waveform of the distribution network fault. The K-mean algorithm is an unsupervised learning clustering method that divides the distribution network fault traveling wave data type into several categories. Each category is composed of a cluster center, and the traveling wave data with similar distribution network fault traveling wave characteristics are divided into this category.
[0084] The above are only embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A distribution network fault location system based on K-means, characterized in that: Measuring devices are installed at various locations in the distribution network, and a distribution network fault location master station electrically connected to each measuring device is also set at the remote end to locate the distribution network fault by extracting fault traveling waves, locating the fault position and identifying the fault type; Fault traveling wave extraction: The fault traveling wave signal is collected through the measuring device, and the first wave signal of the fault traveling wave in the distribution network is continuously monitored and the measuring device clock is calibrated; when the first wave signal of the fault traveling wave is monitored, the waveform data of the first wave of the fault traveling wave is recorded to complete the fault traveling wave extraction; including the measuring device clock calibration step, the zero-sequence current change sensing step and the fault traveling wave first wave signal extraction step; The fault traveling wave first wave signal extraction step samples the fault traveling wave through the analog-to-digital conversion chip in the measuring device, and performs transient recording after the fault traveling wave first wave signal is detected, that is: Among them, ωi is the component of different fault traveling wave frequencies in the distribution network, nc is the number of components of the fault traveling wave frequency, ui is the current value corresponding to different fault traveling wave frequencies in the distribution network, △f is the fault traveling wave current threshold, when it is greater than the fault traveling wave current threshold, the transient fault traveling wave first wave signal of the fault traveling wave signal Ba is recorded; Fault location: record the time when the first wave signal of the fault traveling wave arrives at each measuring device and calculate the time difference, locate the fault position by the time difference, and perform position calibration to complete the fault location; including the steps of recording the first wave arrival time, establishing the measurement terminal time matrix, locating the distribution network fault position and calibrating the distribution network fault position; The distribution network fault location calibration step also pairs the remaining measuring devices in pairs, performs double-end traveling wave ranging positioning, obtains the fault point location and performs mean value calculation to obtain the fault point correction location r a : Among them, y i The fault location for different measuring devices; n g is the number of measuring devices in the distribution network; Fault type identification: Receive the first wave waveform data of each measuring device in the distribution network, extract key features, learn the key features through the machine learning model, identify and judge the distribution network fault type, and complete the fault type identification; including the steps of extracting the first wave waveform data of the fault, extracting the key features of the first wave of the fault, and judging the distribution network fault type; The first-wave waveform data of the fault is obtained by extracting the first-wave waveform data of the fault, and the first-wave waveform data of the fault is used for model training; the machine learning model in the first-wave key feature extraction step of the fault is further deployed by the K-mean algorithm, and data classification is performed by clustering value, and the clustering value D of the K-mean algorithm is k for: Among them, n k is the number of K-mean cluster centers, w i is a sample set of traveling wave data for different distribution network faults, u i is the square error of the traveling wave data for different distribution network faults; The key features of the first wave of each fault are clustered using the K-mean algorithm to obtain the distribution network fault traveling wave feature library. When performing the distribution network fault type judgment step, the current fault first wave waveform data is classified and compared with the distribution network fault traveling wave feature library. If the similarity exceeds the clustering threshold, the distribution network fault type corresponding to the clustering point is output to complete the distribution network fault type judgment.
2. A distribution network fault location system based on K-means according to claim 1, characterized in that: The measuring device is provided with a GPS positioning module, and is deployed through a Beidou positioning module, and can establish a connection with the Beidou timing system; the distribution network fault positioning master station sends a timing instruction S to each measuring device at regular intervals. a : Among them, n a is the number of measuring devices in the distribution network, and s is the timing instruction of different measuring devices; each measuring device completes the clock calibration step of the measuring device according to the corresponding timing instruction.
3. A distribution network fault location system based on K-means according to claim 1, characterized in that: The measuring device is provided with a current transformer and deployed through a zero-sequence current transformer. The zero-sequence current change sensing step is to sense the three-phase current through the zero-sequence current transformer in the measuring device, and to determine whether the zero-sequence current is generated by calculating whether the sum of the current vectors of the three phases is zero; when the sum of the current vectors of the three phases is not zero, it is determined that the zero-sequence current is generated, and the zero-sequence current i is calculated. z value: Among them, i a 、i b 、i c , are the current vectors of the three-phase current in the distribution network; G a is the transformation ratio of the zero-sequence current transformer in the measuring device.
4. A distribution network fault location system based on K-means according to claim 1, characterized in that: When the current value corresponding to the fault traveling wave is greater than the fault traveling wave current threshold, the measuring device records the current time to obtain the first wave arrival time point; the distribution network fault location master station executes the step of establishing the measurement terminal time matrix, that is, collecting the first wave arrival time points of each measurement device and compiling them into the measurement terminal time matrix T a : Where t is the first wave arrival time recorded by different measuring devices, n d is the number of measuring devices in the distribution network.
5. A distribution network fault location system based on K-means according to claim 1, characterized in that: The distribution network fault location location step is performed by double-terminal traveling wave ranging positioning, that is, in the measurement terminal time matrix T a The earliest and second earliest first wave arrival time points are selected, and the measuring devices corresponding to the time points are used as the ranging double ends, and the distance y between the fault point and the ranging double ends is calculated. b and c : Among them, t b and t c are the earliest and second earliest arrival times of the first wave, respectively. a is the fault traveling wave transmission speed in the distribution network.
6. A distribution network fault location system based on K-means according to claim 1, characterized in that: The measuring device includes a current transformer, an A / D converter, a phase-locked frequency multiplication circuit, a signal conditioning circuit, an FPGA programming processor, a DSP signal processor, a clock crystal oscillator, a data buffer, a data storage device, a GPS positioning module, a serial communication module, a remote communication module and a power supply module; The current transformer is used to collect current signals in the distribution network line and is electrically connected to the input pole of the signal conditioning circuit; the output pole of the signal conditioning circuit is electrically connected to the input pole of the phase-locked frequency multiplication circuit and the input pole of the A / D converter respectively; the A / D converter is provided with a CONVST pin and is electrically connected to the output pole of the phase-locked frequency multiplication circuit; the output pole of the A / D converter is electrically connected to the FPGA programming processor; the FPGA programming processor is also electrically connected to a DSP signal processor, a clock crystal oscillator, a data buffer, a data storage device, a GPS positioning module, a serial communication module, a remote communication module and a power supply module; the DSP signal processor is also electrically connected to a clock crystal oscillator, a data buffer and a data storage device.