Power line fault identification intelligent snapshot system based on multi-source signal trigger mechanism

Abstract

The invention discloses an electric power line fault identification intelligent snapshot system based on a multi-source signal trigger mechanism, which belongs to the technical field of electric power line fault identification and comprises a multi-source analysis module, an acquisition end and a platform end. The multi-source analysis module is used for performing fault analysis on the power line and determining a multi-source monitoring item; the acquisition end comprises a monitoring module, a monitoring analysis module, a snapshot control module and a storage module; the monitoring module is used for monitoring a power line in real time to obtain line monitoring information of the power line; the monitoring analysis module is used for analyzing the received line monitoring information to obtain a corresponding snapshot analysis result; when the snapshot analysis result is snapshot, generating a snapshot instruction, and sending the snapshot instruction to a snapshot control module; the snapshot control module is used for controlling the high-speed camera to carry out snapshot to obtain a fault snapshot image when a snapshot instruction is received; and the storage module is used for storing the fault snapshot image and sending the stored fault snapshot image to the platform end.

Description

Technical Field

[0001] The invention belongs to the technical field of power line fault identification, and specifically relates to an intelligent capture system for power line fault identification based on a multi-source signal trigger mechanism. Background Art

[0002] Traditional power line fault monitoring relies on manual inspections and regular preventive tests, which are inefficient, difficult to detect hidden faults in a timely manner, and unable to capture the moment of failure. With the development of technology, although there are high-precision current and voltage sensors, fiber optic temperature sensors, intelligent algorithms, high-speed cameras, solid-state drives, and other single-signal monitoring devices, as well as various advanced sensing, signal processing, and image capture and storage technologies emerging, power line faults are complex, and single-signal monitoring is prone to misjudgment.

[0003] Based on this, the invention provides an intelligent capture system for power line fault identification based on a multi-source signal trigger mechanism. Summary of the Invention

[0004] To solve the problems existing in the above solutions, the invention provides an intelligent capture system for power line fault identification based on a multi-source signal trigger mechanism.

[0005] The object of the invention can be achieved by the following technical solutions:

[0006] An intelligent capture system for power line fault identification based on a multi-source signal trigger mechanism, comprising a multi-source analysis module, a collection end, and a platform end;

[0007] The multi-source analysis module is used to perform fault analysis on the power line and determine multi-source monitoring items.

[0008] Further, the method for determining multi-source monitoring items includes:

[0009] Establish a fault information table, which is used to count the potential faults of the power line and the fault phenomena corresponding to the potential faults;

[0010] Perform monitoring simulation on the fault phenomena according to the fault information table, obtain the monitoring simulation results of the fault phenomena, and determine the multi-source monitoring items according to the monitoring simulation results.

[0011] Further, the method for establishing the fault information table includes:

[0012] Obtain the historical fault data of the power line, perform fault identification on the historical fault data, obtain several historical line faults, and mark the de-duplicated historical line faults as potential faults;

[0013] Set the fault phenomena of the potential faults according to the historical fault data; and update the potential faults and fault phenomena according to the update of the historical fault data;

[0014] Set up a fault information table according to potential faults and fault phenomena.

[0015] Furthermore, the method for setting the fault phenomena of potential faults according to historical fault data includes:

[0016] Obtain the normal state data of the power line, set the normal phenomenon standard for each potential fault according to the normal state data, establish a difference recognition model based on the normal phenomenon standard, and the expression of the difference recognition model is:

[0017]

[0018] In the formula: U is the input data, the output data is the unit difference value CY(U), and the unit difference value is 1 or 0;

[0019] Classify the historical fault data according to the potential faults, obtain the fault classification data for each potential fault, extract the corresponding unit data from the fault classification data, and integrate the unit data corresponding to the potential faults after removing duplicates as the input data;

[0020] Analyze the input data through the difference recognition model to obtain the unit difference value of the unit data;

[0021] Set the fault phenomena according to the unit data with a unit difference value of 1.

[0022] Furthermore, the method for monitoring and simulating fault phenomena according to the fault information table includes:

[0023] Identify the fault phenomena in the fault information table, obtain various monitoring methods possessed by the fault phenomena in real time, and mark them as the primary selection methods;

[0024] Obtain the monitoring requirements of the user, evaluate the primary selection methods of the fault phenomena according to the monitoring requirements, evaluate whether the primary selection methods meet the monitoring requirements, and mark the primary selection methods that meet the monitoring requirements as the candidate methods;

[0025] Screen the candidate methods to obtain the multi-source monitoring methods for the fault phenomena, and set up multi-source monitoring items according to the multi-source monitoring methods.

[0026] Furthermore, the method for screening the candidate methods includes:

[0027] Identify the potential faults corresponding to the candidate methods, obtain the weight coefficients of the potential faults; estimate the cost of the candidate methods, mark it as the candidate cost; analyze the candidate methods to obtain the compliance values of the candidate methods for the potential faults;

[0028] Mark potential faults as i, where i = 1, 2, ……, n and n is the number of potential faults; mark the compliance value of the candidate method for the potential fault as DB i , mark the weight coefficient of the potential fault as δ i ; mark the candidate cost of the candidate method as CB;

[0029] Calculate the screening value of the corresponding candidate method according to the screening formula, and the screening formula:

[0030]

[0031] In the formula: SQ is the screening value; b 1 , b 2 are both proportionality coefficients, and the value range is 0 < b 1 ≤1, 0 < b 2 ≤1;

[0032] Sort the candidate methods in descending order of the screening value to obtain the recommended list of the fault phenomenon, and the user determines the multi-source monitoring method of the fault phenomenon according to the recommended list.

[0033] The acquisition end includes a monitoring module, a monitoring and analysis module, a capture control module, and a storage module;

[0034] The monitoring module is used to monitor the power line in real time according to the preset multi-source monitoring items and multi-source monitoring methods, obtain the line monitoring information of the power line, and send the line monitoring information to the monitoring and analysis module;

[0035] The monitoring and analysis module is used to analyze the received line monitoring information to obtain the corresponding capture analysis result, and the capture analysis result includes capture and non-capture;

[0036] When the capture analysis result is capture, generate a capture instruction and send the capture instruction to the capture control module;

[0037] When the capture analysis result is non-capture, no corresponding processing is performed.

[0038] Further, the method for analyzing the line monitoring information includes:

[0039] Set the monitoring and analysis requirements, set the simulated monitoring data according to the monitoring and analysis requirements; set the target analysis method based on the monitoring and analysis requirements and the simulated monitoring data; set the corresponding monitoring and analysis model according to the target analysis method;

[0040] Analyze the line monitoring information through the preset monitoring and analysis model to obtain the line analysis result.

[0041] Further, the method for setting the target analysis method based on the monitoring analysis requirements and the simulated monitoring data includes:

[0042] Obtain a monitoring and analysis method that meets the monitoring and analysis requirements according to the monitoring and analysis requirements and the simulated monitoring data;

[0043] Identify the benchmark analysis method; compare the monitoring and analysis method with the benchmark analysis method to obtain the efficiency priority value and the accuracy priority value of the monitoring and analysis method

[0044] Calculate the priority evaluation value of the monitoring and analysis method according to the priority evaluation formula. The priority evaluation formula is:

[0045] PY = b 3 ×XL + b 4 ×QL;

[0046] In the formula: PY is the priority evaluation value; b 3 、b 4 are both proportionality coefficients, and the value range is 0 < b 3 ≤1, 0 < b 4 ≤1; XL is the efficiency priority value; QL is the accuracy priority value;

[0047] Eliminate the monitoring and analysis methods with a priority evaluation value less than 0, sort the remaining monitoring and analysis methods in descending order of the priority evaluation value to obtain a method sequence; the user determines the target analysis method according to the method sequence.

[0048] The capture control module is used to control the high-speed camera to capture images when receiving a capture instruction, obtain a fault capture image, and send the fault capture image to the storage module for storage.

[0049] The storage module is used to store the fault capture images and send the stored fault capture images to the platform side.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] It abandons the disadvantages of traditional single-signal monitoring and innovatively integrates multi-source signals such as line current, voltage, temperature, arc light, and sound in the surrounding environment. Taking short-circuit fault monitoring as an example, when a line short-circuit occurs suddenly, not only will the current rise sharply instantaneously, but the voltage will also drop correspondingly. At the same time, strong arc light, rapid temperature rise will be generated around the fault point, and there will also be special sounds of line discharge. With the multi-source signal acquisition module, these synchronous changes are captured sensitively and the information is summarized to the monitoring and analysis module. With the help of advanced machine learning algorithms such as built-in neural networks, learning and modeling are carried out on a large amount of historical data to accurately identify the fault feature combinations. This multi-pronged approach greatly reduces the misjudgment rate and significantly improves the accuracy rate, effectively ensuring the accuracy of fault determination and laying a solid foundation for subsequent emergency repair decisions.

[0052] The response speed from detecting a fault to starting the capture is extremely fast, and the entire process is completed in milliseconds. The high-speed data acquisition card "closely monitors" the line status in real time at an ultra-high sampling rate. Once fluctuations in multi-source signals conforming to the fault characteristics occur, the intelligent algorithm immediately performs rapid calculations and accurate judgments to trigger the capture. This module relies on a high-speed trigger circuit to wake up the high-speed camera in microseconds. Different from traditional interval or cyclic shooting, the device is online in real time all day long, always ready to capture the moment of the fault. Whether it is an instantaneous tripping caused by lightning, a flashover without tripping, or the moment when a tree accidentally touches the line, it can accurately freeze the frame and retain the most original and crucial fault images in the first time, racing against time for emergency repairs. At the same time, it can also provide conclusive evidence for accident liability determination such as lightning strikes without tripping, reducing unnecessary disputes.

[0053] By carefully arranging various types of sensors, multi-dimensional "perspective" of the power line is achieved. At the electrical level, current and voltage sensors are in charge, the physical state is monitored by temperature and arc sensors, and the acoustic characteristics are captured by sound sensors. Based on this, operation and maintenance personnel can not only observe the fault appearance from the captured images, but also dig deep into the root causes by combining data such as temperature, voltage, and sound, and even give early warnings. The wide-angle lens equipped on the high-speed camera shows its prowess, with a wide shooting range, which can not only show the macroscopic view of the line layout around the fault area, but also focus on the details of the fault point, providing comprehensive data support for fault diagnosis and emergency repairs. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0055] Figure 1 It is a block diagram of the principle of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] The following will clearly and completely describe the technical solutions of the present invention in combination with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0057] As Figure 1 shown, a power line fault recognition intelligent capture system based on a multi-source signal trigger mechanism includes a multi-source analysis module, a collection end, and a platform end;

[0058] The multi-source analysis module is used to perform fault analysis on power lines and determine multi-source monitoring items, such as multi-source monitoring items like current, voltage, temperature, arc light, sound, etc. Subsequently, the monitoring module in the acquisition end is set according to the multi-source monitoring items.

[0059] In one embodiment, the multi-source monitoring items can be set based on existing methods, such as summarizing various existing single-signal monitoring items, relying on manual experience, etc.

[0060] In one embodiment, the method for determining multi-source monitoring items includes:

[0061] Analyze the historical fault data of the power line to determine the potential faults of the power line. Combine the historical fault data and other existing materials to count the fault phenomena that each potential fault has when it occurs, that is, compared with the normal state, what different phenomena there are, and mark them as fault phenomena; set up a fault information table according to the potential faults and fault phenomena, and dynamically update the fault information table according to subsequent fault changes;

[0062] Conduct fault phenomenon monitoring simulation according to the fault information table to understand whether the fault phenomenon meets the user's monitoring requirements, that is, whether the fault phenomenon can be monitored under the monitoring requirements, and obtain the monitoring simulation results of each fault phenomenon; determine the multi-source monitoring items according to the monitoring simulation results of each fault phenomenon.

[0063] In one embodiment, the method for setting fault phenomena includes:

[0064] Obtain the normal state data of the power line, set the normal phenomenon standard for each potential fault according to the normal state data, set training data according to the normal phenomenon standard and historical fault data, and the training data is the normal phenomenon standard, fault data, and fault phenomena corresponding to the potential fault. Establish a difference recognition model according to the training data. The expression of the difference recognition model is:

[0065]

[0066] In the formula: U is the input data, which is composed of unit data corresponding to the corresponding potential fault, that is, the unit data set; the output data is the unit difference value CY(U), and the unit difference value is 1 or 0;

[0067] Classify the historical fault data according to the potential faults to obtain the fault classification data corresponding to each potential fault. Extract the corresponding unit data from the fault classification data according to the line phenomena, that is, extract the phenomenon data in the fault classification data as the unit data; de-duplicate and integrate the unit data corresponding to the potential faults into the input data;

[0068] Analyze the input data through the difference recognition model to obtain the unit difference value of the corresponding unit data;

[0069] Set the fault phenomenon according to the unit data with a unit difference value of 1.

[0070] In one embodiment, the method for monitoring and simulating the fault phenomenon according to the fault information table includes:

[0071] Real-time identify the fault phenomenon in the fault information table, determine various monitoring methods for monitoring the fault phenomenon based on existing monitoring technologies, and perform corresponding current, voltage, temperature, arc light, sound, etc. monitoring through Hall effect current sensors, capacitive voltage dividers, fiber optic temperature sensors, ultraviolet photosensitive sensors, sound sensors, etc.; mark the obtained monitoring methods as preliminary selection methods.

[0072] Obtain the monitoring requirements of the user, mainly for relevant requirements such as monitoring efficiency and anti-interference ability. Since subsequent application monitoring needs to be carried out in the power line, it is difficult to work properly without meeting the background monitoring requirements. Especially for subsequent timely analysis and capture, there are also restrictive requirements for monitoring efficiency and data processing efficiency. To cope with instantaneous faults such as lightning strikes causing tripping or not tripping, and tree branches touching the line, when a lightning strike occurs, the fault from the start to causing the line to trip is completed within milliseconds or even microseconds. Therefore, there are high efficiency requirements; therefore, set the corresponding monitoring requirements according to the actual needs of the user.

[0073] Evaluate each preliminary selection method according to the monitoring requirements, judge whether the preliminary selection method meets the monitoring requirements, and mark the preliminary selection method that meets the monitoring requirements as the candidate method.

[0074] Screen the candidate methods to determine the multi-source monitoring method for the corresponding fault phenomenon, and set the multi-source monitoring items according to the multi-source monitoring method.

[0075] In one embodiment, screening the candidate methods can be based on existing screening methods, such as screening according to parameters such as cost, accuracy, and efficiency; or using an intelligent model for screening.

[0076] In one embodiment, the method for screening the candidate methods includes:

[0077] Identify the potential faults monitored by the candidate methods, that is, the data monitored by the candidate methods are used to identify which potential faults. Since some fault phenomena are common to several potential faults, there may be multiple corresponding potential faults. Moreover, even for different candidate methods corresponding to the same fault phenomenon, the content of the monitored data may vary due to differences in the candidate methods. If some candidate methods may monitor additional data in other aspects, they may also be applicable to other potential faults. Determine whether it can correspond to the potential fault according to the user's monitoring requirements, usually in the alternatives of the corresponding potential fault; set the weight coefficient for each potential fault, which can be set according to the importance of the potential fault, fault loss, etc., usually set according to the user's needs;

[0078] Estimate the cost of each candidate method, marked as the candidate cost;

[0079] Based on neural networks such as CNN or DNN, establish a standard evaluation model, and establish a corresponding training set through manual means for training. The training set includes input data and output data. The input data is the candidate method, the user's monitoring requirements, potential faults, and weight coefficients. The output data is the compliance degree of the candidate method with respect to the potential fault, marked as the compliance value, which is mainly calculated comprehensively according to the degree of exceeding each requirement in the monitoring requirements. For example, set the coefficients of the corresponding requirement items in advance and then perform the calculation; analyze through the successfully trained standard evaluation model to obtain the compliance value of the candidate method with respect to the corresponding potential fault;

[0080] Mark the potential faults as i, where i = 1, 2,..., n, and n is the number of potential faults;

[0081] Mark the compliance value of the candidate method with respect to the potential fault as DB i , when it does not belong to the potential fault corresponding to the candidate method, its compliance value is 0; mark the weight coefficient of the potential fault as δ i ; mark the candidate cost of the candidate method as CB;

[0082] Calculate the screening value of the corresponding candidate method according to the screening formula. The screening formula:

[0083]

[0084] In the formula: SQ is the screening value; b 1 , b 2 are both proportionality coefficients, and the value range is 0 < b 1 ≤1, 0 < b 2 ≤1;

[0085] Sort the candidate methods in descending order of the screening value to obtain the recommended list of the corresponding fault phenomena, and the user determines the multi-source monitoring method of the corresponding fault phenomenon according to the recommended list.

[0086] The acquisition end is used to be subsequently installed in a power line, generally installed on a transmission tower, and includes a monitoring module, a monitoring and analysis module, a capture control module, and a storage module;

[0087] The monitoring module is used to perform real-time monitoring on the power line according to preset multi-source monitoring items and multi-source monitoring methods, obtain line monitoring information of the power line, and send the line monitoring information to the monitoring and analysis module.

[0088] Exemplarily, the multi-source monitoring items include current, voltage, temperature, arc light, sound, etc.; the corresponding monitoring methods are to monitor through a high-precision Hall effect current sensor, a capacitive voltage divider voltage sensor, an optical fiber temperature sensor, an ultraviolet photosensitive sensor, and a sound sensor;

[0089] Current sensor: A high-precision Hall effect current sensor is adopted. Based on the Hall effect principle, when there is a current I passing through the power line, a magnetic field B is applied perpendicular to the current direction. The carriers deflect under the action of the Lorentz force, and a Hall voltage VH will be generated on both sides of the Hall element of the sensor. The relationship satisfies the formula:

[0090] V H =k H ·I·B;

[0091] Among them, kH is the Hall coefficient, which is related to factors such as the material and size of the Hall element. By reasonably designing the magnetic field intensity and selecting a suitable Hall element, the line current can be accurately measured. The measurement range is from milliamperes to thousands of amperes, and the output is a standard analog electrical signal (0 - 5V or 4 - 20mA). This signal can sensitively reflect the subtle changes in the current and provide a key basis for subsequent fault judgment. The sensor is installed on the phase wire and neutral wire of the power line near the branch point, joint and other fault-prone parts, and is closely attached through a special fixture to ensure accurate measurement.

[0092] Voltage sensor: The high-voltage capacitor C1 and the low-voltage capacitor C2 are connected in series for voltage division, and the line voltage signal is accurately obtained according to the following voltage division formula, and a digital pulse signal is output. It is insulated and connected between the cross arm of the tower and the phase wire to stably monitor the voltage situation.

[0093] Among them,

[0094] Optical fiber temperature sensor: It is closely wound around wires, insulators, etc. Relying on the photothermal characteristics, according to the relationship between the optical transmission loss and temperature function L = f(T), it accurately senses the temperature fluctuation, converts the optical signal into an electrical signal output, with an accuracy of up to ±0.5°C, and gives a real-time warning of overheating hazards.

[0095] Ultraviolet light-sensitive sensor: A photosensitive element sensitive to ultraviolet light is used. When an arc occurs due to a power line fault, the intensity of ultraviolet light in the arc will increase sharply. Let the ultraviolet light intensity be IUV, and the electrical signal IUV output by the sensor is approximately proportional to the ultraviolet light intensity, that is:

[0096] V UV = k UV ·I UV ;

[0097] where kUV is the proportionality coefficient, which is determined by the material and characteristics of the sensor. The sensor is installed near insulators, at line joints and other areas where arcs are likely to occur, fixed on the accessory brackets of the pole tower, and the angle is adjusted so that it faces the direction where arcs may appear to ensure maximum capture of arc signals.

[0098] Sound sensor: A piezoelectric sound sensor is selected. Based on the piezoelectric effect, when an external sound causes the piezoelectric crystal inside the sensor to vibrate, an electrical signal related to the sound intensity, frequency, etc. will be generated. Let the sound pressure be P and the generated electric charge be Q, and the two satisfy the piezoelectric equation:

[0099] Q = d·P;

[0100] where d is the piezoelectric constant, which depends on the characteristics of the piezoelectric material. This sensor has the characteristics of high sensitivity, wide frequency response range, etc., and can effectively capture various sounds in the environment around the power line. It is installed on the cross arm of the pole tower near the power line or suspended at a certain distance below the wire, and shock absorption, wind prevention and other measures are taken to avoid false signals caused by factors such as pole tower vibration or wind blowing. Its output electrical signal is preliminarily amplified by a preamplifier circuit and then sent to the subsequent processing link.

[0101] Electric field sensor: Based on the charge induction principle, lightning will generate a strong electric field change in the surrounding space. When the electric field intensity E changes, the induction electrode in the sensor will generate induced charges. By measuring the current or voltage generated by the induced charges, the change of the electric field intensity can be reflected. The relationship between its output signal and the electric field intensity satisfies the formula:

[0102] V = k·E;

[0103] where V is the voltage signal output by the sensor and k is the proportionality constant related to the sensor. The sensor can monitor the change of the electric field before lightning strikes and can achieve early warning. It has high sensitivity and high response speed, and can quickly and accurately detect the change of the electric field intensity. At the same time, it has good anti-interference ability and can work stably in a complex environment. It is installed in an open area near the power line and can be fixed on the pole tower through a bracket to ensure that it can effectively sense the surrounding electric field change.

[0104] Due to the different signal strengths and types, they first enter the preamplifier. The weak signal of the sound sensor is amplified 500-1000 times, and other sensor signals are amplified according to their own characteristics. A low-noise operational amplifier chip (OP07) is used to ensure the signal purity and increase the amplitude for subsequent fine processing.

[0105] Next, the filter group starts to operate. The current signal is filtered through a low-pass filter (transfer function H(s) = 1 / (1+sRC), with a cutoff frequency of 10kHz) to remove high-frequency noise; the voltage signal is filtered through a band-pass filter (50Hz-5kHz frequency band) to filter out key frequency components; the temperature, arc light, and sound signals are also filtered through adaptive filters to remove irrelevant interference, ensuring the quality of the signals input to subsequent links.

[0106] The signals that have been pre-amplified and filtered and optimized quickly flow to the high-speed data acquisition card. The acquisition card samples at a high speed of 10MS / s and synchronously collects the signals of each sensor through multiple channels. The analog signals are converted into digital signals according to the following quantization formula (12-16-bit sampling accuracy) and stored in a 1MB-10MB first-in-first-out (FIFO) high-speed cache area to ensure that the data is complete and real-time for immediate algorithm call.

[0107]

[0108] Where D represents the quantized digital value; Va is the analog input voltage; Vr is the reference voltage; and n is the number of bits. The quantized digital value D is determined by comparing Va with Vr and according to the quantization level 2n. In simple terms, this is the process of converting a continuous analog signal into a discrete digital signal. After that, the data is normalized and effective features are extracted to complete data preprocessing and obtain line monitoring information.

[0109] The monitoring and analysis module is used to analyze the received line monitoring information to obtain corresponding snapshot analysis results, where the snapshot analysis results include snapshot and no snapshot;

[0110] When the snapshot analysis result is a snapshot, a snapshot instruction is generated and sent to a snapshot control module;

[0111] When the snapshot analysis result is not to capture, no corresponding processing is performed.

[0112] The line monitoring information is analyzed to determine whether there is an abnormal fault. If there is an abnormal fault, the snapshot analysis result is snapshot, otherwise the snapshot analysis result is no snapshot.

[0113] In one embodiment, it is possible to determine whether a power line has an abnormal fault based on the prior art. For example, an in-built neural network algorithm is used. The input layer sets corresponding nodes according to the characteristics of the line monitoring information collected (such as current mutation rate, voltage sag amplitude, abnormal temperature rise value, arc light intensity, sound characteristics, etc.). After weighted calculation by the hidden layer (1 - 3 layers, the number of points is adjusted according to the complexity), the output layer outputs a fault - corresponding coding numerical table for fault judgment. During training, a large amount of historical fault and normal data are used as training data, and the weight thresholds are adjusted through backpropagation to build an accurate fault feature model. During operation, real - time feature values are input to determine faults in real - time, and the model is adaptively updated regularly based on new data.

[0114] In one embodiment, with the iteration of technology and the complex and changeable monitoring situation, it is necessary to continuously achieve efficient monitoring and analysis. Therefore, in this embodiment, the method for analyzing line monitoring information includes:

[0115] Set the monitoring and analysis requirements. The monitoring and analysis requirements are to analyze the line monitoring information to determine whether it has an abnormal fault, and other requirements such as analysis efficiency are added. Set simulated monitoring data according to the monitoring and analysis requirements. The simulated monitoring data is the preset line monitoring information and the corresponding capture and analysis results.

[0116] Determine the target analysis method according to the monitoring and analysis requirements and the simulated monitoring data. Set the corresponding monitoring and analysis model according to the target analysis method. The monitoring and analysis model is used to analyze the line monitoring to determine the capture and analysis results. That is, with the replacement of technology, the change of line monitoring information, etc., the target analysis method and the monitoring and analysis model are dynamically adjusted.

[0117] Analyze the line monitoring information through the preset monitoring and analysis model to obtain the line analysis result.

[0118] In one embodiment, the method for determining the target analysis method according to the monitoring and analysis requirements and the simulated monitoring data includes:

[0119] Determine the monitoring and analysis method that meets the monitoring and analysis requirements according to the monitoring and analysis requirements and the simulated monitoring data.

[0120] Mark the currently applied target analysis method as the reference analysis method. Compare the monitoring and analysis method with the reference analysis method to determine the efficiency priority value and the accuracy priority value of the monitoring and analysis method. The efficiency priority value is the efficiency exceeding the reference analysis method. The accuracy priority value is the correct rate exceeding the reference analysis method. For example, if they are 10 and 9 respectively, it is better by 10 - 9 = 1. If they are 9 and 10, it is better by 9 - 10 = - 1.

[0121] Calculate the priority evaluation value of the monitoring and analysis method according to the priority evaluation formula. The priority evaluation formula is:

[0122] PY = b 3 × XL + b 4 × QL;

[0123] Where: PY is the priority evaluation value; b 3 、b 4 are both proportionality coefficients, and the value range is 0 < b 3 ≤ 1, 0 < b 4 ≤ 1; XL is the efficiency priority value; QL is the precision priority value;

[0124] Eliminate the monitoring and analysis methods with a priority evaluation value less than 0, and sort the remaining monitoring and analysis methods in descending order of the priority evaluation value to obtain a method sequence;

[0125] The user determines the target analysis method according to the method sequence, that is, finally determined according to the user's needs.

[0126] The capture control module is used to control the high-speed camera to capture when receiving a capture instruction, obtain a fault capture image, and send the fault capture image to the storage module for storage.

[0127] Exemplarily, after receiving the capture instruction, a microsecond-level startup response mechanism is used for active fault capture. Through a carefully designed high-speed trigger circuit (ensuring that the trigger instruction transmission delay td ≤ 1 microsecond, realized by a high-speed logic chip and optimized wiring), a high-speed camera with a frame rate of 100 frames per second, a 120-degree wide-angle lens and optical image stabilization is instantly awakened, and the fault scene is recorded within 8 microseconds.

[0128] When the camera starts, the autofocus system uses phase detection autofocus (PDF). According to the image phase difference Δφ and the focus distance function d = f(Δφ), it accurately focuses on the key area of the circuit; automatic exposure monitors the light intensity I according to the built-in light sensor p , uses the center-weighted average photometry algorithm to intelligently adjust the aperture, shutter, and sensitivity to ensure that the picture is clear and bright, and completely capture the fault instant details such as electric arcs and line jitters.

[0129] In one embodiment, according to the development of capture technology, dynamic capture analysis is performed to realize the dynamic update of the capture device and system.

[0130] The storage module is used to store the fault capture images and send the stored fault capture images to the platform side.

[0131] If a large-capacity solid-state drive (SSD) is adopted, with sequential read and write speeds of up to 500 MB / s and above, it can quickly accept and store this multi-source data. According to strict storage rules, a multi-level directory structure is carefully built based on time, line number, and fault type, and each captured image is named in a standard format such as {date-line-faulttype-seq}.jpg, where date represents the date, line refers to the line number, faulttype is the fault type, and seq is the image sequence number, so as to ensure that the image data is filed in an orderly manner, providing great convenience for subsequent maintenance personnel to quickly and accurately retrieve.

[0132] In one embodiment, to ensure the continuous and efficient utilization of the SSD storage space, the system is built with an intelligent cleaning program to regularly perform data cleaning tasks. It operates according to two preset cleaning conditions: First, screening by time period, every 30 days, old image data beyond this period is cleaned; Second, based on the storage capacity threshold judgment, once the used space ratio of the SSD reaches 90%, the cleaning process is immediately started, accurately screening out expired or duplicate images, always reserving sufficient storage space for newly generated fault images, and ensuring the stable and efficient operation of the storage link.

[0133] At the moment when local storage is completed, important fault images and related detailed information are synchronously transmitted to the platform end by means of advanced wireless communication technologies such as 4G / 5G / APN. To ensure the security and reliability of data transmission, the widely recognized HTTPS encryption protocol is adopted throughout the process, effectively resisting external interference and potential data theft risks, and ensuring that key data arrives at the destination intact.

[0134] Considering the limited wireless transmission bandwidth and the large amount of data, intelligent compression processing is performed on the images at the sending end. The JPEG format compression algorithm is preferably selected. With its excellent compression performance, the original image data size of S o is greatly compressed to S c , and usually a compression ratio of up to 10:1 or even better can be achieved, that is, satisfying:

[0135]

[0136] This greatly reduces the network transmission burden and improves the transmission efficiency, enabling data to be delivered to the platform end as quickly as "lightning". To further ensure the accuracy and integrity of the platform end, the system has carefully deployed a CRC check mechanism. At the data sending end, a unique CRC check code is generated for each batch of transmitted data according to a specific algorithm and sent together with the data; when the platform end receives the data as the receiving end, it immediately recalculates the check code for the received data using the same algorithm and strictly compares it with the check code attached by the sending end. Once it is found that the two are inconsistent, the receiving end immediately sends a retransmission instruction to the sending end, requesting the retransmission of this batch of data. Through such a rigorous check process, it comprehensively ensures that the fault data received by the operation and maintenance personnel is accurate and error-free, and finally realizes the remote real-time monitoring and efficient management of power line faults, laying a solid foundation for line operation and maintenance and the safe and stable operation of the power grid.

[0137] The platform end is used by users and generally serves as an operation and maintenance center and a data center.

[0138] The above formulas are all calculated by removing the dimension and taking their numerical values. The formulas are obtained by collecting a large amount of data and performing software simulation to get a formula closest to the actual situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained through a large amount of data simulation.

[0139] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent capture system for power line fault identification based on a multi-source signal trigger mechanism, characterized in that: Includes multi-source analysis module, collection end and platform end; The multi-source analysis module is used to perform fault analysis on the power line and determine multi-source monitoring items; The acquisition end includes a monitoring module, a monitoring and analysis module, a capture control module and a storage module; The monitoring module is used to perform real-time monitoring of the power line according to the preset multi-source monitoring items and multi-source monitoring methods, obtain line monitoring information of the power line, and send the line monitoring information to the monitoring and analysis module; The monitoring and analysis module is used to analyze the received line monitoring information to obtain corresponding snapshot analysis results, wherein the snapshot analysis results include snapshot and non-snapshot; When the snapshot analysis result is a snapshot, a snapshot instruction is generated and sent to a snapshot control module; When the snapshot analysis result is no snapshot, no corresponding processing is performed; The capture control module is used to control the high-speed camera to capture when receiving the capture instruction, obtain the fault capture image, and send the fault capture image to the storage module for storage; The storage module is used to store the fault snapshot images and send the stored fault snapshot images to the platform end.

2. According to claim 1, a power line fault identification intelligent capture system based on a multi-source signal trigger mechanism is characterized in that: Methods for determining multi-source monitoring items include: Establishing a fault information table, wherein the fault information table is used to count potential faults of the power line and the fault phenomena corresponding to the potential faults; Perform fault phenomenon monitoring simulation according to the fault information table, obtain monitoring simulation results of the fault phenomenon, and determine multi-source monitoring items according to the monitoring simulation results.

3. According to claim 2, a power line fault identification intelligent capture system based on a multi-source signal trigger mechanism is characterized in that: The method for establishing the fault information table includes: Acquire historical fault data of the power line, perform fault identification on the historical fault data, obtain a number of historical line faults, and mark the historical line faults after deduplication as potential faults; Setting the fault phenomenon of the potential fault according to the historical fault data; and updating the potential fault and the fault phenomenon according to the update of the historical fault data; Set up the fault information table according to potential faults and fault phenomena.

4. According to claim 3, a power line fault identification intelligent capture system based on a multi-source signal trigger mechanism is characterized in that: Methods for setting the fault phenomenon of potential faults based on historical fault data include: The normal state data of the power line is obtained, the normal phenomenon standard of each potential fault is set according to the normal state data, and a difference recognition model is established according to the normal phenomenon standard. The expression of the difference recognition model is: Where: U is the input data, the output data is the unit difference value CY(U), and the unit difference value is 1 or 0; Classifying the historical fault data according to the potential faults to obtain fault classification data of each potential fault, extracting corresponding unit data from the fault classification data, and removing duplicates from the unit data corresponding to the potential faults and integrating them into input data; Analyzing the input data through a difference recognition model to obtain a unit difference value of the unit data; Set the fault phenomenon based on the unit data with a unit difference value of 1.

5. According to claim 2, a power line fault identification intelligent capture system based on a multi-source signal trigger mechanism is characterized in that: The method of performing fault phenomenon monitoring simulation according to the fault information table includes: Identify the fault phenomenon in the fault information table, obtain various monitoring methods of the fault phenomenon in real time, and mark them as preliminary selection methods; Obtain the monitoring requirements of the user, evaluate the primary selection method of the fault phenomenon according to the monitoring requirements, evaluate whether the primary selection method meets the monitoring requirements, and mark the primary selection method that meets the monitoring requirements as the candidate method; Screen the candidate methods to obtain the multi-source monitoring method of the fault phenomenon, and set multi-source monitoring items according to the multi-source monitoring method.

6. The intelligent capture system for power line fault identification based on multi-source signal triggering mechanism according to claim 5 is characterized in that: The method for screening the candidate methods includes: Identify the potential faults corresponding to the candidate methods for monitoring, and obtain the weight coefficients of the potential faults; estimate the cost of the candidate methods, marked as the candidate cost; analyze the candidate methods to obtain the compliance values of the candidate methods for the potential faults; The potential fault is marked as i, i = 1, 2, ..., n, n is the number of potential faults; the standard value of the selected method for the potential fault is marked as DB i , the weight coefficient of potential fault is marked as δ i ; Mark the alternative cost of the alternative method as CB; Calculate the screening values of the corresponding candidate methods according to the screening formula. The screening formula is: In the formula: SQ is the screening value; b1 and b2 are both proportionality coefficients, and the value range is 0 < b1 ≤ 1, 0 < b2 ≤ 1; Sort the candidate methods in descending order of the screening values to obtain the recommendation list of the fault phenomenon, and the user determines the multi-source monitoring method of the fault phenomenon according to the recommendation list.

7. The intelligent capture system for power line fault identification based on multi-source signal triggering mechanism according to claim 1 is characterized in that: The method for analyzing the line monitoring information includes: Set the monitoring analysis requirements, and set the simulated monitoring data according to the monitoring analysis requirements; set the target analysis method based on the monitoring analysis requirements and the simulated monitoring data; set the corresponding monitoring analysis model according to the target analysis method; Analyze the line monitoring information through the preset monitoring analysis model to obtain the line analysis result.

8. The intelligent capture system for power line fault identification based on multi-source signal triggering mechanism according to claim 7 is characterized in that: The method for setting the target analysis method based on the monitoring analysis requirements and the simulated monitoring data includes: Obtain the monitoring analysis methods that meet the monitoring analysis requirements according to the monitoring analysis requirements and the simulated monitoring data; Identify the benchmark analysis method; compare the monitoring analysis method with the benchmark analysis method to obtain the efficiency priority value and the precision priority value of the monitoring analysis method Calculate the priority evaluation value of the monitoring analysis method according to the priority evaluation formula. The priority evaluation formula is: PY = b3 × XL + b4 × QL; In the formula: PY is the priority evaluation value; b3 and b4 are both proportionality coefficients, and the value range is 0 < b3 ≤ 1, 0 < b4 ≤ 1; XL is the efficiency priority value; QL is the precision priority value; Eliminate the monitoring analysis methods with the priority evaluation value less than 0, sort the remaining monitoring analysis methods in descending order of the priority evaluation value to obtain the method sequence; the user determines the target analysis method according to the method sequence.

Images