Frequency stop line intelligent diagnosis and treatment strategy generation system

Abstract

The invention discloses an intelligent diagnosis and treatment strategy generation system for a frequency stop line, which belongs to the technical field of line diagnosis and comprises a data acquisition module which is integrally provided with a high-frequency magnetic field sensor, a Rogowski coil galvanometer and an infrared thermometer and is used for acquiring line traveling wave signals, temperature and electromagnetic field data in real time; time-space information is synchronized through the three-channel radio frequency GPS module; the feature extraction module is used for performing wavelet packet decomposition and noise reduction on the traveling wave signal, calculating high-frequency signal information entropy based on a beat frequency attenuation principle and extracting a fault feature vector; the fault diagnosis module is used for inputting the feature vector into a deep neural network model through multi-criterion information fusion, and outputting a fault type and a positioning coordinate; and the strategy generation module is used for generating a governance strategy library. According to the invention, high-frequency transient characteristics can be captured, complex faults of high-resistance faults and broken line resonance are avoided, and the omission ratio is greatly reduced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of line diagnosis, and particularly relates to an intelligent diagnosis and treatment strategy generation system for a frequent outage line. BACKGROUND

[0002] With the continuous development of the power system, as the key infrastructure for power transmission, power transmission lines play a crucial role in ensuring the efficiency and safety of energy transmission. However, due to factors such as natural environment, human factors, and equipment aging, power transmission line faults occur frequently, which not only may cause large-scale power outages, but also may lead to serious economic losses and casualties. Therefore, how to effectively monitor the state of power transmission lines and timely detect and warn potential faults is of great significance to improving the safety and reliability of the power grid. Traditional power transmission line fault diagnosis methods mainly rely on manual inspection, which is not only costly and inefficient, but also difficult to achieve real-time monitoring of the entire power grid. With the development of big data technology, by collecting, integrating, and analyzing massive power transmission line operation data, fault patterns can be more accurately identified and predicted, providing a scientific basis for fault prevention. The automatic diagnosis of power transmission line fault states based on big data analysis is emerging in this context.

[0003] Currently, the main method for diagnosing cable faults is the traveling wave method, and common methods include impulse current experimentation (ICE) and time domain reflectometry (TDR). Although the ICE method can diagnose high-resistance cable faults, it requires high voltage to click through the fault, and its test accuracy is easily affected by external electromagnetic interference. The TDR method is a non-destructive testing method and is widely used in cable fault diagnosis due to its portable testing equipment. The TDR method locates cable faults by the time difference between incident and reflected pulse signals, but due to the low high-frequency signal component of the injected pulse signal, signal attenuation and dispersion effects during transmission, and the influence of electromagnetic interference on site, the test accuracy of the TDR method is relatively low. Nowadays, the emerging frequency domain reflectometry (FDR) method has more high-frequency components in the incident wave due to the use of sweep signals, making it more advantageous than traditional TDR methods in positioning. Although some scholars have used IFFT algorithms to explore cable fault positioning, due to the limitations of the algorithm, traditional methods (such as Fourier transform and time domain reflectometry) rely on a single sensor or base frequency signals, making it difficult to capture high-frequency transient characteristics, resulting in high missed detection rates for complex faults such as high-resistance faults and broken wire resonances. SUMMARY

[0004] The present application aims at overcoming the deficiencies of the prior art, and provides a frequency outage line intelligent diagnosis and treatment strategy generation system, which solves the problems in the background art.

[0005] The present application is achieved in that a frequency outage line intelligent diagnosis and treatment strategy generation system comprises: a data acquisition module, which is integrally provided with a high-frequency magnetic field sensor, a Rogowski coil current meter and an infrared temperature measuring instrument, is used for acquiring line traveling wave signals, temperature and electromagnetic field data in real time, and synchronizes time and space information through a three-channel radio frequency GPS module; a feature extraction module, which is used for wavelet packet decomposition noise reduction of the traveling wave signals, and calculates high-frequency signal information entropy based on the beat frequency attenuation principle to extract a fault feature vector; a fault diagnosis module, which inputs the feature vector into a deep neural network model through multi-criteria information fusion to output fault types and positioning coordinates; and a strategy generation module, which is used for matching historical fault cases based on a knowledge graph to generate a treatment strategy library.

[0006] The working principle is that electromagnetic, temperature and time and space data of the line are cooperatively collected by multiple sensors, faults are diagnosed through signal processing and artificial intelligence models, and finally a treatment strategy is generated by matching a knowledge base. High-frequency transient characteristics are captured, complex faults such as high-resistance faults and broken-line resonances are avoided, and the missed detection rate is greatly reduced. The data acquisition module uses a high-frequency magnetic field sensor (flexible Rogowski coil), a Rogowski coil current meter and an infrared temperature measuring instrument to synchronously collect time and space coordinates through a three-channel radio frequency GPS, and to acquire traveling wave signals, temperature and electromagnetic field data in real time. The feature extraction module performs wavelet packet decomposition noise reduction on the traveling wave signals, calculates high-frequency signal information entropy based on the beat frequency attenuation principle, and extracts a fault feature vector. The fault diagnosis module inputs the feature vector into a deep neural network (including a CNN branch and an LSTM branch), and outputs fault types and positioning coordinates through multi-layer information fusion. The strategy generation module matches historical cases in the knowledge graph to generate a strategy library containing treatment measures.

[0007] Comprehensiveness: multi-sensor combination covers electromagnetic, temperature, time and space multi-dimensional data, improving fault perception accuracy. Automation: full-process automation from data acquisition to strategy generation, reducing manual intervention. Real-time: GPS synchronization ensures accurate time and space data, supporting rapid fault positioning.

[0008] Further, the high-frequency magnetic field sensor adopts a flexible Rogowski coil with a frequency response range of 1 kHz-10 MHz, which is used to capture the transient magnetic field changes of high-impedance faults; the infrared temperature measuring instrument is arranged at the conductor joint, which is used to monitor the temperature abnormality coefficient in real time to early warn the poor contact fault. The working principle is: the transient magnetic field changes are captured by the high-frequency sensor, and the infrared temperature measuring instrument monitors the temperature abnormality of the conductor joint. Flexible Rogowski coil: frequency response range of 1 kHz-10 MHz, closely adheres to the conductor to capture the transient magnetic field of high-frequency faults. Infrared temperature measuring instrument: installed at the conductor joint, dynamically calculates the temperature abnormality coefficient (such as temperature rise ΔT). High sensitivity: wide frequency response (1 kHz-10 MHz) covers various types of high-frequency fault transient signals. Precise early warning: temperature monitoring at the joint can detect the hidden high-impedance fault caused by poor contact in advance.

[0009] Further, the fault diagnosis module includes: a convolutional neural network branch for identifying the frequency domain impedance spectrum mutation characteristics of the traveling wave signal; a long short-term memory network branch for analyzing the phase continuity anomaly in the time series. In use, the CNN extracts the frequency domain mutation characteristics, the LSTM identifies the time sequence phase anomaly, and the dual-branch fusion improves the diagnosis reliability. The CNN branch analyzes the frequency domain impedance spectrum mutation characteristics of the traveling wave signal, including the resonance frequency anomaly. The LSTM branch detects the phase continuity anomaly in the time series, including the phase angle jump. The spatial features (CNN) and time features (LSTM) are fused to avoid misjudgment of a single model and are suitable for complex fault broken resonant, phase intermittent.

[0010] Further, the strategy generation module accesses the power grid dispatching system to push the treatment instructions to the fault section breaker in real time to trigger the resonance suppression circuit or the reclosing operation.

[0011] A frequency outage line intelligent diagnosis and treatment strategy generation method, which includes: collecting line multi-source data through a distributed sensor network and marking the space-time stamp; wavelet packet decomposition is performed on the traveling wave signal to extract the frequency domain impedance spectrum and phase mutation point, and a diagnosis function is constructed; the differential principle is used to fuse the high-frequency signal information entropy and the temperature abnormality coefficient to locate the fault interval; a strategy library is called to dynamically generate a treatment scheme, and feedback is fed back to the line control terminal.

[0012] Further, the construction of the diagnosis function is specifically:

[0013] Based on the traveling wave equation, Calculate the propagation constant, and combine the wave impedance Correct the delay error, and the delay correction formula is,

[0014]

[0015] Wherein, R0 is the line resistance per unit length; L0 is the line inductance per unit length; G0 is the line conductance per unit length; C0 is the line capacitance per unit length; ω is the angular frequency, ω = 2πf, f is the frequency of the traveling wave; γ is the propagation constant, the real part α represents the attenuation coefficient, and the imaginary part β represents the phase constant; Z is the wave impedance, reflecting the characteristic impedance of the traveling wave on the transmission line; Δx is the distance between the fault point and the measurement point; Δt is the time error.

[0016] Further, the fault interval comprises:

[0017] The high-frequency magnetic field scanning is started for the 5km section before and after the fault point, and the broken line resonance and single-phase grounding fault are distinguished.

[0018] When the temperature anomaly coefficient is greater than 2.5, it is determined that the fault is caused by the oxidation of the contact point.

[0019] Further, when the traveling wave frequency f is greater than 1MHz, the influence of the conductance G0 is ignored, that is, G0 = 0, and the temperature compensation coefficient K is increased T :

[0020]

[0021] Wherein, ΔT is the temperature rise of the conductor, which is used for high resistance fault diagnosis;

[0022] For the broken line resonance fault, the resonance detection function is set:

[0023]

[0024] When H(ω) > 0.85, it is determined that the fault is a broken line resonance fault.

[0025] A data fusion unit is set between the feature extraction module and the fault diagnosis module, the data fusion unit inputs: high-frequency information entropy S f , S f from the feature extraction module, temperature anomaly coefficient δ T , δ T from the infrared thermometer; output: fault confidence P = 0.7S f + 0.3δ T , when P ≥ 0.9, the strategy generation module is triggered.

[0026] The working principle is: through the simplified line model (ignoring the conductance G i ) and the temperature compensation mechanism, the high resistance fault diagnosis is optimized, and the temperature and electromagnetic data are fused to improve the confidence. The implementation process model is simplified: when the traveling wave frequency is greater than 1Hz, the conductance G i= 0, reduce the computational complexity. Increase the temperature compensation coefficient: K_T = 1-0.0023ΔT (ΔT is the temperature rise of the wire). Resonance fault detection, resonance detection function H(ω)>0.83 to determine the broken wire fault. Data fusion unit input: high frequency information entropy S r (feature extraction module) + temperature anomaly coefficient δ r (thermometer); output: fault confidence P = 0.75S r + 0.3δ r (P≥0.9 trigger strategy generation). Simplified model reduces the power consumption, suitable for edge device deployment, temperature compensation eliminates environmental interference, improves high resistance fault recognition rate, multi-source data fusion avoids single sensor false alarm.

[0027] Based on the distance and temperature threshold, the fault interval is divided, and the fault type is distinguished by high frequency scanning. The fault interval scanning starts the high frequency magnetic field scanning within 5 kilometers before and after the fault point. The broken wire fault spectrum characteristics are abnormal, including impedance spectrum mutation, and the temperature anomaly coefficient of the hidden high resistance fault is >2.5 (caused by contact point oxidation). Fast positioning 5km interval scanning reduces the scope of investigation and saves repair time. Accurate classification temperature threshold + high frequency scanning distinguishes broken wire and high resistance fault, reduces misjudgment.

[0028] A computer readable storage medium stores a computer program, the program is executed by a processor to realize the steps of the above method.

[0029] An electronic device includes a processor, a memory, and a communication interface, the memory stores a computer readable storage medium.

[0030] The above method is solidified in a computer program / equipment, realizing software and hardware cooperative control, the storage medium stores the computer program for executing the above method, the electronic device integrates the processor (running algorithm), the memory (storing program), and the communication interface (data transmission). Through the storage medium, it can be quickly deployed to different hardware devices, and the special electronic device meets the real-time demand of power inspection and low delay processing. From feature extraction (wavelet entropy) to double branch AI model to knowledge graph matching, the whole process is autonomously decided, realizing intelligent diagnosis closed loop. Edge computing optimizes the model to simplify temperature compensation and adapt to low-power field devices.

[0031] The application has the beneficial effects that: through multi-sensor cooperative collection of electromagnetic, temperature and space-time data of the line, signal processing and artificial intelligence model diagnosis of faults, and finally matching of a knowledge base to generate a treatment strategy. High-frequency transient characteristics are captured, complex faults such as high-resistance faults and broken-line resonances are avoided, and the missed detection rate is greatly reduced. The data acquisition module uses a high-frequency magnetic field sensor (flexible Rogowski coil), a Rogowski coil current meter and an infrared thermometer to collect traveling wave signals, temperature and electromagnetic field data in real time through three-channel radio frequency GPS synchronous space-time coordinates. The feature extraction module performs wavelet packet decomposition and noise reduction on the traveling wave signals, calculates the information entropy of the high-frequency signals based on the principle of beat frequency attenuation, and extracts the fault feature vector. The fault diagnosis module inputs the feature vector into a deep neural network (including a CNN branch and an LSTM branch), outputs the fault type and positioning coordinates through multi-layer information fusion, simplifies the model to reduce the power consumption, is suitable for edge device deployment, eliminates environmental interference through temperature compensation, improves the high-resistance fault recognition rate, and avoids false alarms of a single sensor through multi-source data fusion. The fault interval is divided based on the distance and temperature threshold, and the fault type is distinguished through high-frequency scanning. The 5-kilometer interval before and after the fault point is scanned to start high-frequency magnetic field scanning, the spectrum characteristics of the broken-line fault are abnormal, including impedance spectrum mutation, and the temperature anomaly coefficient of the hidden high-resistance fault is greater than 2.5 (caused by contact point oxidation). The 5-kilometer interval scanning reduces the search range and saves repair time, and the temperature threshold and high-frequency scanning are used to accurately classify the broken-line and high-resistance faults to reduce misjudgment. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 is a system schematic diagram of the application. DETAILED DESCRIPTION

[0033] The application will be further described in detail below with reference to the accompanying drawings, and it should be pointed out that the following is only for more clearly illustrating and explaining the application.

[0034] Example 1

[0035] As shown in Figure 1 , the embodiment discloses a frequency suspension line intelligent diagnosis and treatment strategy generation system, which comprises: a data acquisition module: the data acquisition module is integrally provided with a high-frequency magnetic field sensor, a Rogowski coil current meter and an infrared thermometer, and is used for collecting line traveling wave signals, temperature and electromagnetic field data in real time, and synchronizing space-time information through a three-channel radio frequency GPS module; a feature extraction module: used for wavelet packet decomposition and noise reduction on the traveling wave signals, and calculating the information entropy of the high-frequency signals based on the principle of beat frequency attenuation, and extracting a fault feature vector; a fault diagnosis module: through multi-criterion information fusion, inputting the feature vector into a deep neural network model, and outputting a fault type and positioning coordinates; a strategy generation module: used for matching historical fault cases based on a knowledge graph to generate a treatment strategy library.

[0036] The working principle is: through the cooperation of multiple sensors to collect electromagnetic, temperature and space-time data of the line, the fault is diagnosed through signal processing and artificial intelligence model, and finally the knowledge base is matched to generate governance strategy. Realize the capture of high-frequency transient characteristics, avoid complex faults such as high-resistance faults and broken-line resonances, and greatly reduce the missed detection rate. The data acquisition module uses a high-frequency magnetic field sensor (flexible Rogowski coil), a Rogowski coil current meter, and an infrared thermometer to collect traveling wave signals, temperature, and electromagnetic field data in real time through a three-channel RF GPS synchronous space-time coordinate. The feature extraction module performs wavelet packet decomposition and noise reduction on the traveling wave signal, calculates the high-frequency signal information entropy based on the beat attenuation principle, and extracts the fault feature vector. The fault diagnosis module inputs the feature vector into a deep neural network (including a CNN branch and an LSTM branch), and outputs the fault type and positioning coordinates through multi-layer information fusion. The strategy generation module matches historical cases in the knowledge graph to generate a strategy library containing governance measures.

[0037] The combination of multiple sensors covers electromagnetic, temperature, and space-time multi-dimensional data, improving fault perception accuracy and comprehensiveness. From data acquisition to strategy generation, the whole process is automated, reducing human intervention and improving automation. GPS synchronization ensures accurate space-time data and supports rapid fault location.

[0038] Embodiment 2

[0039] As shown in Figure 1 The embodiment discloses an intelligent diagnosis and governance strategy generation system for frequency suspension lines, which includes: a data acquisition module: the data acquisition module is integrally provided with a high-frequency magnetic field sensor, a Rogowski coil current meter and an infrared thermometer, which is used for real-time acquisition of line traveling wave signals, temperature and electromagnetic field data, and synchronization of space-time information through a three-channel RF GPS module; a feature extraction module: for wavelet packet decomposition and noise reduction of traveling wave signals, and based on the beat attenuation principle to calculate the high-frequency signal information entropy and extract the fault feature vector; a fault diagnosis module: through multi-criteria information fusion, the feature vector is input into a deep neural network model, and the fault type and positioning coordinates are output; a strategy generation module: for matching historical fault cases based on a knowledge graph to generate a governance strategy library.

[0040] The working principle is: through the cooperation of multiple sensors to collect electromagnetic, temperature and space-time data of the line, through signal processing and artificial intelligence model diagnosis, finally match the knowledge base to generate governance strategy. It can capture high-frequency transient characteristics, avoid complex faults such as high-impedance faults and broken-line resonance, and greatly reduce the missed detection rate. The data acquisition module uses high-frequency magnetic field sensors (flexible Rogowski coil), Rogowski coil current meter, and infrared thermometer to collect traveling wave signal, temperature and electromagnetic field data in real time through three-channel radio frequency GPS synchronous space-time coordinates. The feature extraction module performs wavelet packet decomposition on the traveling wave signal to reduce noise, calculates the information entropy of the high-frequency signal based on the principle of beat frequency attenuation, and extracts the fault feature vector. The fault diagnosis module inputs the feature vector into the deep neural network (including CNN branch and LSTM branch), and outputs the fault type and positioning coordinates through multi-layer information fusion. The strategy generation module matches the historical cases in the knowledge graph to generate a strategy library containing governance measures.

[0041] The combination of multiple sensors covers electromagnetic, temperature, and space-time multi-dimensional data, improving fault perception accuracy and comprehensiveness. The whole process from data acquisition to strategy generation is fully automated, reducing human intervention and improving automation. GPS synchronization ensures accurate space-time data and supports rapid fault location.

[0042] To better effect, the high-frequency magnetic field sensor uses a flexible Rogowski coil with a frequency response range of 1 kHz-10 MHz, which is used to capture the transient magnetic field changes of high-impedance faults. The infrared thermometer is installed at the conductor joint to monitor the temperature anomaly coefficient in real time to warn of poor contact faults. The working principle is: through the high-frequency sensor to capture the transient magnetic field changes, and the infrared thermometer to monitor the temperature anomaly of the conductor joint. Flexible Rogowski coil: frequency response range 1 kHz-10 MHz, closely attached to the conductor to capture high-frequency fault transient magnetic field. Infrared thermometer: installed at the conductor joint, dynamically calculates the temperature anomaly coefficient (such as temperature rise ΔT). High sensitivity: wide frequency response (1 kHz-10 MHz) covers various high-frequency fault transient signals. Accurate early warning: temperature monitoring at the joint can detect hidden high-impedance faults caused by poor contact in advance.

[0043] To better effect, the fault diagnosis module includes: a convolutional neural network branch for identifying frequency domain impedance spectrum mutation characteristics of the traveling wave signal; a long short-term memory network branch for analyzing phase continuity anomalies in time series. In use, CNN extracts frequency domain mutation characteristics, LSTM identifies time series phase anomalies, and dual-branch fusion improves diagnostic reliability. The CNN branch analyzes the frequency domain impedance spectrum mutation characteristics of the traveling wave signal, including resonance frequency anomalies. The LSTM branch detects phase continuity anomalies in time series, including phase angle jumps. The fusion of spatial features (CNN) and time features (LSTM) avoids misjudgment by a single model and is suitable for complex fault broken-line resonance and phase discontinuity.

[0044] For better effect, the strategy generation module accesses the power grid dispatching system to push the treatment instructions to the fault section breaker in real time to trigger the resonance suppression circuit or the reclosing operation.

[0045] Embodiment 3

[0046] As Figure 1 shown, the embodiment discloses a frequent outage line intelligent diagnosis and treatment strategy generation system, which comprises a data acquisition module, a feature extraction module, a fault diagnosis module and a strategy generation module. The data acquisition module is integrally provided with a high-frequency magnetic field sensor, a Rogowski coil ammeter and an infrared thermometer, is used for acquiring line traveling wave signals, temperature and electromagnetic field data in real time, and synchronizes space-time information through a three-channel radio frequency GPS module. The feature extraction module is used for wavelet packet decomposition and noise reduction of the traveling wave signals, and extracts a fault feature vector based on the beat frequency attenuation principle. The fault diagnosis module inputs the feature vector into a deep neural network model through multi-criteria information fusion, and outputs a fault type and positioning coordinates. The strategy generation module is used for matching historical fault cases based on a knowledge graph, and generating a treatment strategy library.

[0047] The working principle is that the electromagnetic, temperature and space-time data of the line are cooperatively collected by multiple sensors, the fault is diagnosed through signal processing and an artificial intelligence model, and finally the treatment strategy is generated by matching a knowledge base. High-frequency transient characteristics are captured, complex faults such as high-resistance faults and broken-line resonance are avoided, and the missed detection rate is greatly reduced. The data acquisition module uses a high-frequency magnetic field sensor (flexible Rogowski coil), a Rogowski coil ammeter and an infrared thermometer, synchronizes space-time coordinates through a three-channel radio frequency GPS, and acquires traveling wave signals, temperature and electromagnetic field data in real time. The feature extraction module performs wavelet packet decomposition and noise reduction on the traveling wave signals, calculates the information entropy of the high-frequency signals based on the beat frequency attenuation principle, and extracts a fault feature vector. The fault diagnosis module inputs the feature vector into a deep neural network (including a CNN branch and an LSTM branch), and outputs a fault type and positioning coordinates through multi-layer information fusion. The strategy generation module matches historical cases in the knowledge graph, and generates a strategy library containing treatment measures.

[0048] The multi-sensor combination covers electromagnetic, temperature, space-time multi-dimensional data, improves fault perception accuracy, and improves comprehensiveness. The whole process from data acquisition to strategy generation is automated, reducing manual intervention and improving automation. GPS synchronization ensures accurate space-time data and supports rapid fault location.

[0049] To better effect, the high-frequency magnetic field sensor adopts a flexible Rogowski coil with a frequency response range of 1 kHz-10 MHz for capturing transient magnetic field changes of high-resistance faults; the infrared temperature measuring instrument is arranged at the conductor joint for real-time monitoring of temperature anomaly coefficients to prewarn poor contact faults. The working principle is: capturing transient magnetic field changes through the high-frequency sensor and monitoring temperature anomalies of the conductor joint through the infrared temperature measuring instrument. The flexible Rogowski coil: frequency response range of 1 kHz-10 MHz, closely adhering to the conductor to capture high-frequency fault transient magnetic fields. The infrared temperature measuring instrument: installed at the conductor joint, dynamically calculating temperature anomaly coefficients (such as temperature rise ΔT). High sensitivity: wide frequency response (1 kHz-10 MHz) covering various high-frequency fault transient signals. Precise prewarning: temperature monitoring at the joint can discover hidden high-resistance faults caused by poor contact in advance.

[0050] To better effect, the fault diagnosis module includes: a convolutional neural network branch for identifying frequency domain impedance spectrum mutation characteristics of the traveling wave signal; a long short-term memory network branch for analyzing phase continuity anomalies in the time series. In use, the CNN extracts frequency domain mutation characteristics, the LSTM identifies time sequence phase anomalies, and the dual-branch fusion improves the diagnosis reliability. The CNN branch analyzes the frequency domain impedance spectrum mutation characteristics of the traveling wave signal, including resonance frequency anomalies. The LSTM branch detects phase continuity anomalies in the time series, including phase angle jumps. The spatial features (CNN) and the time features (LSTM) are fused to avoid misjudgment of a single model and are suitable for complex fault broken-line resonance and phase discontinuity.

[0051] To better effect, the strategy generation module accesses the power grid dispatching system for real-time pushing of treatment instructions to the fault section breaker to trigger the resonance suppression circuit or the reclosing operation.

[0052] A frequency outage line intelligent diagnosis and treatment strategy generation method, which includes: collecting line multi-source data through a distributed sensor network and marking space-time stamps; performing wavelet packet decomposition on the traveling wave signal, extracting frequency domain impedance spectrum and phase mutation points, and constructing a diagnosis function; fusing high-frequency signal information entropy and temperature anomaly coefficients based on the differential principle to locate the fault interval; calling a strategy library to dynamically generate a treatment scheme and feeding back to the line control terminal.

[0053] To better effect, the construction of the diagnosis function is specifically:

[0054] Based on the traveling wave wave equation, Calculate the propagation constant, combine the wave impedance Correct the delay error, and the delay correction formula is,

[0055]

[0056] Wherein, R0 is the line resistance per unit length; L0 is the line inductance per unit length; G0 is the line conductance per unit length; C0 is the line capacitance per unit length; ω is the angular frequency, ω = 2πf, f is the frequency of the traveling wave; γ is the propagation constant, the real part α represents the attenuation coefficient, and the imaginary part β represents the phase constant; Z is the wave impedance, reflecting the characteristic impedance of the traveling wave on the transmission line; Δx is the distance between the fault point and the measurement point; Δt is the time error.

[0057] For better effect, the fault interval comprises:

[0058] The high-frequency magnetic field scanning is started for the 5km section before and after the fault point, and the broken line resonance and single-phase grounding fault are distinguished;

[0059] When the temperature anomaly coefficient is greater than 2.5, it is determined that the fault is caused by the oxidation of the contact point.

[0060] For better effect, when the traveling wave frequency f is greater than 1MHz, the influence of the conductance G0 is ignored, that is, G0 = 0, and the temperature compensation coefficient K is increased T :

[0061]

[0062] Wherein, ΔT is the temperature rise of the conductor, which is used for high resistance fault diagnosis;

[0063] For the broken line resonance fault, the resonance detection function is set:

[0064]

[0065] When H(ω) is greater than 0.85, it is determined that the fault is a broken line resonance fault;

[0066] A data fusion unit is set between the feature extraction module and the fault diagnosis module, the data fusion unit inputs: high-frequency information entropy S f , S f from the feature extraction module, temperature anomaly coefficient δ T , δ T from the infrared thermometer; output: fault confidence P = 0.7S f + 0.3δ T , when P is greater than or equal to 0.9, the strategy generation module is triggered.

[0067] The working principle is: through the simplified line model (ignoring the conductance G i ) and the temperature compensation mechanism, the high resistance fault diagnosis is optimized, and the temperature and electromagnetic data are fused to improve the confidence. The implementation process model is simplified: when the traveling wave frequency is greater than 1Hz, the conductance G i=0, reducing computational complexity. Adding temperature compensation coefficient: K_T=1-0.0023ΔT (ΔT is the conductor temperature rise). Resonance fault detection: when the resonance detection function H(ω)>0.83, a wire break fault is determined. Data fusion unit input: high-frequency information entropy S r (Feature extraction module) + temperature anomaly coefficient δr (thermometer); Output: Fault confidence P = 0.75S r +0.3δ r (Policy generation is triggered when P ≥ 0.9.) The simplified model reduces computing power consumption and is suitable for edge device deployment. Temperature compensation eliminates environmental interference, improves high-resistance fault recognition, and multi-source data fusion avoids false alarms from a single sensor.

[0068] Fault intervals are divided based on distance and temperature thresholds, and high-frequency scanning is used to distinguish fault types. Fault interval scanning initiates a high-frequency magnetic field scan within 5 kilometers before and after the fault point. Spectral abnormalities of broken wire faults include sudden changes in the impedance spectrum and temperature anomalies greater than 2.5 for latent high-resistance faults (caused by contact oxidation). Rapid location within the 5-kilometer interval scan narrows the investigation scope, saving repair time. Accurate classification of temperature thresholds combined with high-frequency scanning distinguishes between broken wire faults and high-resistance faults, reducing misdiagnosis.

[0069] A computer-readable storage medium stores a computer program, which implements the steps of the above method when executed by a processor.

[0070] An electronic device includes a processor, a memory, and a communication interface, wherein the memory stores a computer-readable storage medium.

[0071] The above-mentioned method is solidified in a computer program / device to achieve coordinated control of software and hardware. The storage medium stores the computer program that executes the above-mentioned method. The electronic device integrates a processor (to run the algorithm), a memory (to store the program), and a communication interface (for data transmission). It is quickly deployed to different hardware devices through the storage medium. The dedicated electronic device meets the real-time requirements of power inspections with low latency processing. From feature extraction (wavelet entropy) → dual-branch AI model → knowledge graph matching, autonomous decision-making is made throughout the entire process to achieve an intelligent diagnostic closed loop. The edge computing optimization model simplifies temperature compensation and is suitable for low-computing field equipment.

[0072] The above are only preferred specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes based on the technical solutions and concepts of the present invention within the technical scope disclosed by the present invention, and they should be covered by the scope of protection of the present invention.

Claims

1. A system for intelligent diagnosis and management strategy generation of frequent outage lines, characterized by: include: Data acquisition module: The data acquisition module is integrated with a high-frequency magnetic field sensor, a Rogowski coil ammeter, and an infrared thermometer to collect line traveling wave signals, temperature, and electromagnetic field data in real time, and synchronize time and space information through a three-channel radio frequency GPS module; Feature extraction module: used to perform wavelet packet decomposition and noise reduction on traveling wave signals, calculate the information entropy of high-frequency signals based on the beat frequency attenuation principle, and extract fault feature vectors; Fault diagnosis module: Through multi-criteria information fusion, the feature vector is input into the deep neural network model to output the fault type and location coordinates; Policy generation module: used to match historical failure cases based on the knowledge graph and generate a governance policy library.

2. The system for intelligent diagnosis and management strategy generation of frequent outage lines according to claim 1 is characterized in that: The high-frequency magnetic field sensor uses a flexible Rogowski coil with a frequency response range of 1kHz-10MHz, which is used to capture transient magnetic field changes of high-resistance faults; the infrared thermometer is set at the wire joint to monitor the temperature anomaly coefficient in real time to warn of poor contact faults.

3. The system for intelligent diagnosis and management strategy generation of frequent outage lines according to claim 1 is characterized in that: The fault diagnosis module includes: The convolutional neural network branch is used to identify the frequency domain impedance spectrum mutation characteristics of the traveling wave signal; Long short-term memory network branch for analyzing phase-continuous anomalies in time series.

4. The system for intelligent diagnosis and management strategy generation of frequent outage lines according to claim 1 is characterized in that: The strategy generation module is connected to the power grid dispatching system to push management instructions to the fault section circuit breaker in real time to trigger the resonance suppression circuit or reclosing operation.

5. A method for intelligent diagnosis and management strategy generation of frequent outage lines, the method being used in a system for intelligent diagnosis and management strategy generation of frequent outage lines, characterized in that: include: Collect multi-source data from the line through a distributed sensor network and mark it with time and space stamps; Perform wavelet packet decomposition on the traveling wave signal, extract the frequency domain impedance spectrum and phase mutation points, and construct a diagnostic function; The differential principle is used to fuse the high-frequency signal information entropy and the temperature anomaly coefficient to locate the fault area; The policy library is called to dynamically generate governance solutions and feed them back to the line control terminal.

6. The method for generating intelligent diagnosis and management strategies for frequent outage lines according to claim 5, characterized in that: The construction of the diagnostic function is specifically as follows: Based on the traveling wave equation, Calculate the propagation constant and combine it with the wave impedance Correct the delay error, and the positioning delay correction formula is: Among them, R0 is the line resistance per unit length; L0 is the line inductance per unit length; G0 is the line inductance per unit length; C0 is the line capacitance per unit length; ω is the angular frequency, ω=2πf, where f is the traveling wave frequency; γ is the propagation constant, whose real part α represents the attenuation coefficient and whose imaginary part β represents the phase constant; Z is the wave impedance, which reflects the characteristic impedance of the traveling wave on the transmission line; Δx is the distance between the fault point and the measurement point; and Δt is the time error.

7. The method for generating intelligent diagnosis and management strategies for frequent outage lines according to claim 5, characterized in that: The fault interval includes: A high-frequency magnetic field scan is initiated within 5 km before and after the fault point to distinguish between line-break resonance and single-phase grounding faults. When the temperature anomaly coefficient is greater than 2.5, it is determined to be a hidden high-resistance fault caused by oxidation of the contact point.

8. The method for generating intelligent diagnosis and management strategies for frequent outage lines according to claim 6, characterized in that: When the traveling wave frequency f>1MHz, ignore the influence of conductivity G0, that is, set G0=0, and increase the temperature compensation coefficient K T : Where ΔT is the conductor temperature rise, which is used for high resistance fault diagnosis; For disconnection resonance fault, set the resonance detection function: When H(ω)>0.85, it is determined to be a line-break resonance fault; A data fusion unit is set between the feature extraction module and the fault diagnosis module. The data fusion unit inputs: high-frequency information entropy S f , S f From the feature extraction module, temperature anomaly coefficient δ T ,δ T From infrared thermometer; output: fault confidence P = 0.7S f +0.3δ T , when P≥0.9, the strategy generation module is triggered.

9. A computer-readable storage medium, characterized in that: A computer program is stored, and when the program is executed by a processor, the steps of the method according to any one of claims 5 to 8 are implemented.

10. An electronic device, characterized in that: The system comprises a processor, a memory and a communication interface, wherein the memory stores the medium according to claim 9.

Images