Power distribution network operation and maintenance inspection device and method based on artificial intelligence

Abstract

The invention relates to the technical field of power grid operation and maintenance, in particular to a power distribution network operation and maintenance inspection device and method based on artificial intelligence, and the device comprises a sound data acquisition module, a thermal imaging monitoring module, a data fusion diagnosis module and a path optimization and scheduling module. According to the invention, by integrating the acoustic sensor and the thermal imaging camera, the comprehensive state monitoring of the power grid equipment is realized, the accuracy of fault diagnosis and the foresight of operation are improved, and the potential fault point can be quickly positioned by integrating sound wave and temperature data, so that the fault early warning is more accurate, and the working efficiency is improved. The power grid safety risk caused by improper fault processing is effectively reduced, the stability and reliability of the power grid are improved, through real-time data analysis, continuous health assessment can be carried out on the power grid, a maintenance plan and resource allocation are optimized, the patrol is more efficient through automatic path optimization and a scheduling module, and the safety of the power grid is improved. The key area is ensured to be checked preferentially and thoroughly, and the overall operation efficiency of the power grid is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid operation and maintenance, and in particular to an artificial intelligence-based distribution network operation and maintenance inspection device and method. Background Art

[0002] The field of power grid operation and maintenance technology focuses on maintaining and improving the reliability, safety and efficiency of the power system. It includes multiple aspects such as monitoring, maintenance, upgrading and management of the power grid. By using advanced monitoring systems, such as smart sensors, automated control systems and data analysis tools, failures can be predicted and prevented to ensure the continuity and safety of power supply. Power grid operation and maintenance also involves disaster response plans, asset management, and the renewal or replacement of aging infrastructure, helping the operation and maintenance team to effectively manage daily operations and respond to emergencies, thereby optimizing power grid performance and reducing power outages.

[0003] Among them, the distribution network operation and maintenance inspection device is a technology specifically used in the field of power grid operation and maintenance. It usually includes drones or robots, equipped with high-definition cameras, infrared scanners and other sensors, and is used to inspect power transmission lines and facilities. The main purpose is to monitor the physical state of the power grid in real time, detect and diagnose potential problems such as line aging, damage or foreign object interference. By providing accurate real-time data, it can help power grid operation and maintenance personnel to promptly discover and solve problems, reduce the risk of power outages, and improve the overall operation efficiency and reliability of the power grid.

[0004] Existing grid operation and maintenance technologies show great limitations in terms of rapid fault location and real-time response. Despite covering multiple aspects of monitoring and management, traditional methods still rely on manual inspections and reactive processing, which often leads to untimely responses when dealing with line aging or emergencies. For example, for line damage or equipment failure, traditional monitoring can only diagnose after the failure occurs, missing the best time for prevention and early intervention. In addition, existing technologies are also insufficient in flexibility and adaptability. In particular, in the face of extreme weather or environmental changes, conventional asset management and disaster response plans are difficult to adjust in real time, affecting the operational efficiency and safety of the power grid. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings existing in the prior art and to propose a distribution network operation and maintenance inspection device and method based on artificial intelligence.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: A distribution network operation and maintenance inspection device based on artificial intelligence comprises:

[0007] The sound data acquisition module installs acoustic sensors in the power equipment of the distribution network to collect operating noise, extracts the frequency characteristics of sound waves, analyzes the differences in sound patterns, compares the patterns in the normal database, detects potential anomalies based on frequency changes, and obtains acoustic anomaly indicators;

[0008] The thermal imaging monitoring module uses a thermal imaging camera to scan the power lines and power equipment of the distribution network, record the surface temperature distribution data of the equipment, compare it with the normal temperature standard, identify the abnormal temperature hotspots, and judge whether the hotspot area corresponds to the abnormal frequency in the sound data based on the abnormal temperature hotspot identification results combined with the acoustic abnormality index, and mark the actual abnormal area to obtain the thermal imaging abnormal position record;

[0009] The data fusion diagnosis module combines the acoustic abnormality index and the thermal imaging abnormality position record to perform difference analysis on the collected data, determine the performance degradation damage mode of the power equipment, determine the matching degree of the result analysis with the known fault record according to the mode, determine the cause of the fault and divide the risk area, and issue a fault warning signal;

[0010] The path optimization and scheduling module receives the fault warning signal, marks the risk area as a key inspection area, replans the inspection route of the inspection device, gives priority to covering all key inspection areas, and formulates and outputs an inspection path optimization plan.

[0011] As a further solution of the present invention, the steps of obtaining the acoustic abnormality index are:

[0012] Deploy acoustic sensors on power equipment in the distribution network to collect sound signals during operation, convert them into electronic data, and generate raw sound signal data sets;

[0013] The original sound signal data set is subjected to Fourier transform to convert the time domain signal into the frequency domain signal using the formula:

[0014] ;

[0015] Calculate the frequency domain signal value of the sound , generate a frequency feature data set, where Representative time The signal strength value at is a frequency variable;

[0016] Based on the frequency feature data set, a comparative analysis is performed with the standard sound patterns in the normal database to identify frequency changes and abnormal patterns and obtain acoustic anomaly indicators.

[0017] As a further solution of the present invention, the step of identifying the abnormal temperature hotspot is:

[0018] Deploy thermal imaging cameras to scan the power lines and power equipment of the distribution network, collect real-time temperature readings of each part of the power equipment, and generate an initial temperature data set for the equipment surface;

[0019] The initial surface temperature data set of the equipment is compared point by point with the normal temperature standard, using the formula:

[0020] ;

[0021] Calculate the coordinates of each point Temperature deviation , and obtain the temperature deviation data set, where is the measured temperature value, is the normal temperature standard value of the coordinate;

[0022] Analyze the temperature deviation data set to identify deviation values ​​exceeding a set threshold The area is determined and marked as a temperature anomaly hotspot area to obtain temperature anomaly hotspot data.

[0023] As a further solution of the present invention, the steps for obtaining the abnormal position record of thermal imaging are:

[0024] Synchronize the temperature anomaly hotspot data and the acoustic anomaly index, perform spatial and temporal alignment, confirm the consistency of the timestamp and position coordinates of each temperature anomaly point, and generate a synchronized multi-source data set;

[0025] Based on the synchronized multi-source data set, the abnormal frequency is extracted and compared with the temperature abnormal point, using the formula:

[0026] ;

[0027] Calculate the correlation between acoustic anomalies and temperature anomalies , evaluate the matching degree and obtain the abnormal matching data set, where Representing coordinates Acoustic anomaly indicators on is the temperature anomaly index at the same coordinate;

[0028] The anomaly matching data set is analyzed to determine all areas where acoustic and temperature anomalies match, and are marked as actual anomaly areas. The actual position of each anomaly area is located, and a thermal imaging anomaly position record is generated.

[0029] As a further solution of the present invention, the steps for determining the performance degradation damage mode of the power equipment are:

[0030] Using the acoustic anomaly indicators and the thermal imaging anomaly location records, perform data integration, synchronize timestamps and geographic information tags, establish time-space associations, and obtain a comprehensive data set;

[0031] The composite data set was analyzed to calculate the standard deviation and the range of deviation between data points using the formula:

[0032] ;

[0033] Calculate variance analysis values , generate abnormal difference analysis results, where, represents the value of each data point in the synthetic dataset, represents the average value of the comprehensive data set, represents the total number of data points, represents the standard deviation of the data point values, Represents the deviation range value of the data point value;

[0034] The abnormal difference analysis results are analyzed, and the probability of occurrence of equipment damage modes is analyzed in combination with the historical failure data and performance parameters of the equipment to obtain the damage mode data of the power equipment performance degradation.

[0035] As a further solution of the present invention, the step of issuing the fault warning signal is:

[0036] Based on the power equipment performance degradation damage mode data, the data is compared with known fault records, fault instances with similar characteristics are screened, and a matching result list is generated;

[0037] Analyze the matching result list, analyze the relationship between the matched fault mode and the current operating parameters, reveal the inherent connection between the faults under different operating conditions, determine the fault causes, and generate a fault cause analysis report;

[0038] Based on the fault cause analysis report, the fault data is combined with the geographic coordinates, the affected area is visualized, the area is divided into risk levels, and a fault warning signal is issued.

[0039] As a further solution of the present invention, the steps for obtaining the patrol path optimization solution are:

[0040] Receive the fault warning signal, extract fault location information, mark the risk area as a key inspection area, and generate a fault location marking result;

[0041] Based on the fault location marking results, dynamic path planning is performed to ensure that all key inspection areas are covered using the formula:

[0042] ;

[0043] Calculate the total length of the inspection path , generate the optimal inspection route, where represents the number of points in the path, Representative The coordinates of the points, Representative The coordinates of the points;

[0044] According to the inspection planning route, the route is visualized to maximize the inspection efficiency, and the inspection path optimization plan is formulated and output in combination with environmental factors.

[0045] A distribution network operation and maintenance inspection method based on artificial intelligence includes the following steps:

[0046] S1: Deploy acoustic sensors on power equipment in the distribution network to collect sound signals during operation, perform Fourier transform, convert time domain signals into frequency domain signals, compare and analyze them with standard sound patterns, identify frequency changes and abnormal patterns, and obtain acoustic abnormality indicators;

[0047] S2: Deploy thermal imaging cameras to scan the power lines and power equipment of the distribution network, collect real-time temperature readings of each part of the power equipment, compare them point by point with the normal temperature standard, identify and mark the areas where the deviation value exceeds the set threshold, and obtain abnormal temperature hotspot data;

[0048] S3: Synchronize the temperature anomaly hotspot data and the acoustic anomaly index, perform spatial and temporal alignment, extract the anomaly frequency, and compare it with the temperature anomaly point, determine all areas where acoustic and temperature anomalies match, and generate thermal imaging anomaly position records;

[0049] S4: using the acoustic anomaly index and the thermal imaging anomaly position record, synchronizing the timestamp and geographic information tag, calculating the standard deviation and deviation range between data points and performing difference analysis, combining the historical equipment failure data and performance parameters, analyzing the probability of occurrence of equipment damage mode, and obtaining the power equipment performance degradation damage mode data;

[0050] S5: Based on the performance degradation damage mode data of the power equipment, the data is compared with known fault records to reveal the inherent connection between the faults under different operating conditions, determine the cause of the fault, visualize the affected area, and issue a fault warning signal;

[0051] S6: Receive the fault warning signal, extract the fault location information, mark the key inspection area, perform dynamic inspection route planning, visualize the route, and formulate and output the inspection route optimization plan in combination with environmental factors.

[0052] Compared with the prior art, the advantages and positive effects of the present invention are:

[0053] In the present invention, by integrating acoustic sensors and thermal imaging cameras, comprehensive status monitoring of power grid equipment is achieved, the accuracy of fault diagnosis and the foresight of operation are improved, and the sound wave and temperature data can be integrated to quickly locate potential fault points, making fault warnings more accurate, effectively reducing the power grid safety risks caused by improper fault handling, and improving the stability and reliability of the power grid. Through real-time analysis of data, continuous health assessment of the power grid can be carried out, maintenance plans and resource allocation can be optimized, and automated path optimization and scheduling modules make inspections more efficient, ensuring that key areas are given priority and thorough inspections, further improving the overall operating efficiency of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a flow chart of the device of the present invention;

[0055] Figure 2 The following is a flowchart of obtaining the acoustic anomaly index of the present invention;

[0056] Figure 3 This is a flow chart for identifying abnormal temperature hot spots of the present invention;

[0057] Figure 4 The flowchart of obtaining the abnormal position record of thermal imaging of the present invention;

[0058] Figure 5 A flow chart for determining the performance degradation damage mode of the power equipment of the present invention;

[0059] Figure 6 This is a flow chart of issuing a fault warning signal of the present invention;

[0060] Figure 7 This is a flow chart for obtaining the patrol path optimization solution of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0062] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0063] See also Figure 1 , a distribution network operation and maintenance inspection device based on artificial intelligence includes:

[0064] The sound data acquisition module installs acoustic sensors in the power equipment of the distribution network to collect operating noise, extracts the frequency characteristics of sound waves, analyzes the differences in sound patterns, compares the patterns in the normal database, detects potential anomalies based on frequency changes, and obtains acoustic anomaly indicators;

[0065] The thermal imaging monitoring module uses a thermal imaging camera to scan the power lines and power equipment of the distribution network, record the surface temperature distribution data of the equipment, compare it with the normal temperature standard, identify the abnormal temperature hotspots, and judge whether the hotspot area corresponds to the abnormal frequency in the sound data based on the abnormal temperature hotspot identification results combined with the acoustic abnormality index, and mark the actual abnormal area to obtain the thermal imaging abnormal position record;

[0066] The data fusion diagnosis module combines acoustic anomaly indicators and thermal imaging anomaly location records to perform differential analysis on the collected data, determine the performance degradation and damage mode of the power equipment, determine the matching degree of the result analysis with the known fault records based on the mode, determine the cause of the fault and divide the risk area, and issue a fault warning signal;

[0067] The path optimization and scheduling module receives fault warning signals, marks risk areas as key inspection areas, replans the inspection routes of the inspection devices, gives priority to covering all key inspection areas, and formulates and outputs inspection path optimization plans.

[0068] Acoustic anomaly indicators include frequency characteristics, pattern difference analysis results and anomaly detection records; thermal imaging anomaly location records include temperature distribution images, abnormal hotspots and abnormal area positioning results; fault warning signals include fault type identification results, matching degree analysis records and risk area division results; patrol path optimization plans include route replanning results, risk area priorities and path coverage strategies.

[0069] See also Figure 2 , the steps to obtain acoustic anomaly indicators are:

[0070] Deploy acoustic sensors on power equipment in the distribution network to collect sound signals during operation, convert them into electronic data, and generate raw sound signal data sets;

[0071] The installation of acoustic sensors in power equipment in the distribution network is, first of all, the key is to choose the right sensor. The sensor should be able to cover the low-frequency and high-frequency ranges that cannot be captured by the human ear to ensure that all relevant sound wave signals can be recorded; secondly, determine the best installation location of the sensor, which is usually located in the sound source activity area of ​​the power equipment, such as the cooling system of the transformer or near the operating mechanism of the circuit breaker. The choice of installation location depends on the previous sound field distribution test, which involves testing the sound intensity and quality at different locations to determine where the sound of the equipment can be most effectively captured when it is running; next, fix and adjust the sensor to ensure that its sensitivity to sound and response time meet the monitoring requirements, including adjusting its direction and sensitivity settings to maximize the accuracy and integrity of the received signal; finally, conduct field tests to verify the performance of the sensor and confirm that it can accurately capture sound data in the actual operating environment. The data is transmitted via wired or wireless means for subsequent analysis and processing to generate the original sound signal data set.

[0072] Perform Fourier transform on the original sound signal data set and convert the time domain signal into the frequency domain signal using the formula:

[0073] ;

[0074] Calculate the frequency domain signal value of the sound , generate a frequency feature data set, where Representative time The signal strength value at is a frequency variable;

[0075] Consider a specific sound signal, e.g. , where the frequency Hz, which means a cosine wave with a fundamental frequency of 100 Hz. Applying the Fourier transform formula to this signal, we can calculate to The integral of this time interval of seconds, because the sound signal is periodic, calculating one cycle can get the spectrum representing the entire signal.

[0076] Will Substitute into the formula and integrate:

[0077]

[0078] Select during calculation Hz to observe the response at the fundamental frequency of the signal:

[0079]

[0080] Using Euler's formula , the above expression can be simplified and the real part calculated:

[0081]

[0082]

[0083]

[0084] The integral calculation yields:

[0085]

[0086] The results show that at a frequency of 100 Hz, the amplitude of the spectrum is 0.5, indicating that this frequency is a significant component in the signal. The results verify the existence of the signal at the fundamental frequency. For acoustic analysis, it shows that the frequency feature data set can accurately reflect the acoustic characteristics of the equipment and can be used to detect whether the equipment is operating normally or whether there are potential faults.

[0087] Based on the frequency feature data set, a comparative analysis is performed with the standard sound patterns in the normal database to identify frequency changes and abnormal patterns and obtain acoustic anomaly indicators;

[0088] Comparing and analyzing the collected frequency feature data set with the sound patterns stored in the normal database is a detailed process. First, the frequency data needs to be transmitted from the sensor collection system to the main data processing center. Next, the frequency patterns of known normal operating sounds are extracted from the normal database. The patterns are derived from long-term monitoring and historical data analysis and are used as a benchmark for comparison; then data matching is performed, such as cross-correlation analysis, to compare the actual frequency data with the standard patterns in the database one by one to detect frequency deviations and anomalies; for each abnormal frequency point detected, its manifestation in the sound data, such as amplitude, duration and frequency of occurrence, is further analyzed to evaluate whether the anomaly is related to potential equipment failure or performance degradation; finally, an acoustic anomaly indicator is generated, reflecting the current health status of the equipment and possible maintenance needs.

[0089] See also Figure 3 , the steps for identifying abnormal temperature hot spots are:

[0090] Deploy thermal imaging cameras to scan the power lines and power equipment of the distribution network, collect real-time temperature readings of each part of the power equipment, and generate an initial temperature data set for the equipment surface;

[0091] Before collecting data on the surface temperature of the equipment, it is first necessary to configure the scanning parameters of the thermal imaging camera according to the specific environment and operating status of the power lines and power equipment, including adjusting the camera's focal length, sensitivity and scanning speed to ensure that various temperature anomalies in the operation of the equipment can be captured. In addition, by setting different scanning areas and paths, every corner of the power equipment can be covered so as not to miss any potential hot spots. Subsequently, the camera will automatically start scanning at the set time interval and transmit the captured thermal image data in real time. The data includes the coordinates of each measurement point and the corresponding temperature value. The temperature data is sent to the central processing unit through the real-time data transmission system to generate a preliminary equipment surface temperature data set as the basis for subsequent analysis.

[0092] The initial surface temperature data set of the equipment is compared point by point with the normal temperature standard, using the formula:

[0093] ;

[0094] Calculate the coordinates of each point Temperature deviation , and obtain the temperature deviation data set, where is the measured temperature value, is the normal temperature standard value of the coordinate;

[0095] Indicates the actual measured temperature, assuming it is 350 degrees Celsius, is the normal standard temperature of the location, assuming it is 300 degrees Celsius, and is calculated by substituting it into the formula:

[0096]

[0097] The calculation results show that the temperature deviation is 50 degrees Celsius, indicating that the temperature at this point is higher than normal and requires further analysis and attention to obtain a temperature deviation data set.

[0098] Analyze the temperature deviation data set and identify deviation values ​​exceeding the set threshold The area is determined and marked as a temperature abnormality hotspot area, and the temperature abnormality hotspot data is obtained;

[0099] After collecting the temperature deviation data, further data analysis is performed to identify abnormal temperature hotspots. In this process, the data clustering analysis method is used to divide the temperature data into multiple categories through the clustering algorithm. The categories are distinguished according to the size of the temperature deviation. Each category represents a temperature condition. For those temperature deviations exceeding the set threshold, The system identifies areas where potential faults or risks exist and marks them as abnormal. The thresholds set here are determined based on historical data and equipment operation standards. All areas marked as abnormal are then summarized to form a list of all abnormal temperature hotspots, which is convenient for the maintenance team to conduct subsequent inspections and repairs.

[0100] See also Figure 4 ,The steps for obtaining the thermal imaging abnormal position record are:

[0101] Synchronize temperature anomaly hotspot data and acoustic anomaly indicators, perform spatial and temporal alignment, confirm the consistency of timestamp and location coordinates of each temperature anomaly point, and generate a synchronized multi-source data set;

[0102] When synchronizing acoustic anomaly indicators and temperature anomaly data, we first extract the timestamps and spatial coordinates of the two sensor data through an automated data processing script, and then use the time series analysis in the database management system to ensure the consistency of the acoustic data and temperature data on the time axis. The process includes correcting time deviations and data lags. At the same time, spatial alignment is performed to ensure that the geocoding of each piece of data matches accurately, and the data is calibrated through the geographic information system (GIS). Finally, the data that has been synchronized in time and space will be integrated into a unified multi-source data platform to provide more accurate data support for identifying potential anomalies.

[0103] Based on the synchronous multi-source data set, the abnormal frequency is extracted and compared with the temperature abnormal points, using the formula:

[0104] ;

[0105] Calculate the correlation between acoustic anomalies and temperature anomalies , evaluate the matching degree and obtain the abnormal matching data set, where Representing coordinates Acoustic anomaly indicators on is the temperature anomaly index at the same coordinate;

[0106] Assuming a specific location Acoustic anomaly indicators is 0.75, and the corresponding temperature anomaly index is 0.45. Substitute the value into the formula to calculate:

[0107]

[0108] Calculation results It indicates that the correlation between acoustic anomalies and temperature anomalies at this location is low, and the value 0.3 indicates a small difference, indicating that although there are certain differences, such differences are not significant, indicating that the sources of the anomalies are the same or related to each other.

[0109] Analyze the anomaly matching data set, determine all areas where acoustic and temperature anomalies match, mark them as actual anomaly areas, locate the actual position of each anomaly area, and generate thermal imaging anomaly position records;

[0110] After confirming the correlation between acoustic anomalies and temperature anomalies, the next task is to convert these data into practical and actionable information. First, all matching points identified as anomalies will be entered into the geographic information system. This system is equipped with an automated geographic tagging tool that can accurately mark the location of each anomaly point on the map; then, using the analysis function of GIS, the distribution pattern and potential correlation of the anomaly area are further analyzed to provide specific navigation information for the maintenance team. Finally, a detailed anomaly location report is generated, which not only lists the specific coordinates of all anomaly areas, but also includes the intensity of the related anomalies and the possible impact range.

[0111] See also Figure 5 , the steps to determine the damage mode of power equipment performance degradation are:

[0112] Using acoustic anomaly indicators and thermal imaging anomaly location records, data integration is performed to synchronize timestamps and geographic information tags, establish temporal and spatial associations, and obtain a comprehensive data set;

[0113] The data integration process starts with collecting the raw data output by the sensors, which are collected in real time by sensors distributed in key parts of the equipment. The sensors can capture sound wave fluctuations ranging from weak to strong. At the same time, the thermal imaging camera captures the thermal image of the equipment at preset time intervals. The image is saved with high resolution and marked with timestamps. Data integration is performed in a central processing unit that runs custom software to automatically align acoustic and thermal map data based on timestamps and geotags to ensure perfect synchronization of the two types of data. The integrated data is further analyzed through complex data fusion algorithms, including data cleaning and preliminary anomaly identification, to eliminate environmental noise and thermal changes not related to equipment failures, providing an accurate and clean data set for subsequent in-depth analysis.

[0114] The comprehensive data set was analyzed to calculate the standard deviation and the range of deviation between data points using the formula:

[0115] ;

[0116] Calculate variance analysis values , generate abnormal difference analysis results, where, represents the value of each data point in the synthetic dataset, represents the average value of the comprehensive data set, represents the total number of data points, represents the standard deviation of the data point values, Represents the deviation range value of the data point value;

[0117] Assumptions Represents the value of each data point in the comprehensive data set, such as {15, 20, 22, 21, 19}. Calculate the average of these data , by adding up all the data values ​​and dividing by the number of data points:

[0118]

[0119] Total number of data points is 5.

[0120] Standard Deviation It is a statistical indicator to measure the degree of dispersion of data. First, the square of the difference between each data point and the mean is calculated, then the average is calculated, and finally the square root is taken:

[0121]

[0122] Deviation range is the difference between the maximum and minimum values ​​in the data:

[0123]

[0124] Substitute these values ​​into the original formula to calculate:

[0125]

[0126]

[0127]

[0128] The results show that the difference analysis value in the comprehensive data set is 2.378, indicating that there is a large degree of anomaly in the data set. The normalization of the data and the calculation of the normalized deviation range provide a quantitative indicator for evaluating the degree of anomaly of acoustic and thermal imaging data under different environmental conditions. The high value indicates that there are some abnormal events or signs of equipment performance degradation in the collected data, and further analysis is needed on the specific impact of these anomalies on the performance of power equipment in order to take appropriate maintenance or repair measures.

[0129] Analyze the abnormal difference analysis results, combine the equipment historical fault data and performance parameters, analyze the probability of equipment damage mode, and obtain the power equipment performance degradation damage mode data;

[0130] After obtaining the data processed by the differential analysis, damage mode prediction is performed by combining the historical failure records and current performance parameters of the equipment. This process involves comparing the current analysis results with the historical failure modes stored in the database. The specific operation is to filter out historical failure data through advanced query statements, which include information such as failure time, failure type and its duration. Then, the data is combined with machine learning technology to build a prediction model to estimate the probability of different failure modes. In the model building process, the first thing to do is feature selection to ensure that the data input to the model are key variables that affect the failure mode. Next, statistical methods are used to evaluate the weights of different features, such as decision tree analysis, which can reveal which parameters can best explain the changes in failure modes. Finally, these models are applied to perform real-time data analysis to predict possible future failures and evaluate their possible impacts, so that measures can be taken in advance to avoid or mitigate the impact of failures.

[0131] See also Figure 6 , the steps for issuing fault warning signals are:

[0132] Based on the power equipment performance degradation damage mode data, it is compared with known fault records, and fault instances with similar characteristics are screened to generate a matching result list;

[0133] After calling the power equipment performance degradation damage mode data, the feature vector of each historical fault record is first extracted, including the fault type, frequency of occurrence, affected system components, and environmental conditions. At the same time, similar feature vectors are extracted from the sensor data of the current equipment. The features include data collected from acoustic and thermal imaging sensors, which reflect the operating status of the equipment and any initial signs of abnormalities. The feature matching process is completed by calculating the Euclidean distance between the historical fault feature vector and the current fault feature vector. Each matching score is based on the inverse of the distance, and the higher the score, the higher the similarity. The system automatically arranges the matching results according to the score and displays the most relevant fault records on the user interface for further analysis.

[0134] Analyze the matching result list, analyze the relationship between the matched fault mode and the current operating parameters, reveal the inherent connection between the faults under different operating conditions, determine the fault causes, and generate a fault cause analysis report;

[0135] After obtaining a list of matching results, further fault cause analysis is performed by combining historical data of the fault. This analysis does not rely solely on a single fault instance, but clusters all fault instances with similar characteristics to form statistically significant fault categories. For example, if a certain type of fault frequently occurs under specific temperature and humidity conditions, the condition will be identified as a key factor in the fault triggering. An in-depth analysis of each type of fault is performed, using data visualization tools such as scatter plots and heat maps to display the conditions and frequency of the fault occurrence, helping technicians identify which operating parameters need to be adjusted to prevent future failures. In addition, by comparing the occurrence patterns of faults in different time periods, the trend of equipment performance degradation can be identified, providing decision support for equipment maintenance and updates.

[0136] Based on the fault cause analysis report, the fault data is combined with the geographic coordinates to visualize the affected area, classify the area into risk levels, and issue a fault warning signal;

[0137] Using a geographic information system (GIS) to divide and visualize risk areas, first extract the geographic location data related to the fault from the database, which indicates the exact location of each fault. Then conduct a risk assessment, analyze the fault history and severity of each location, and automatically classify the risk level of each location according to the set standards. The calculation of the risk level takes into account the frequency of the fault, the severity of the impact, and the difficulty of fault recovery. After that, the risk data is converted into a visual layer and superimposed on the map. Each risk level is represented by a different color, allowing the maintenance team to see the high-risk areas at a glance. In addition, when the data in the risk area is updated or changed, the GIS system automatically updates the view and sends alarms to the monitoring center and mobile terminals in real time through the network, so that the response team can respond.

[0138] See also Figure 7 , the steps to obtain the patrol path optimization solution are:

[0139] Receive fault warning signals, extract fault location information, mark risk areas as key inspection areas, and generate fault location marking results;

[0140] Receiving fault warning signals is a highly automated process involving real-time data monitoring and analysis. First, the fault location information is extracted, each risk area is recorded, and the data is transmitted to the central processing unit in real time. The received data is processed immediately. For the identified risk areas, they are marked as key inspection areas. At the same time, the system will immediately mark them and evaluate their potential risk levels. This process ensures the accuracy of monitoring and the timeliness of response, and generates fault location marking results.

[0141] Based on the fault location marking results, dynamic path planning is performed to ensure that all key inspection areas are covered using the formula:

[0142] ;

[0143] Calculate the total length of the inspection path , generate the optimal inspection route, where represents the number of points in the path, Representative The coordinates of the points, Representative The coordinates of the points;

[0144] Consider a practical example where there are four inspection points with coordinates (2, 3), (5, 7), (8, 2), and (10, 5).

[0145] First, calculate the distance from the first point to the second point:

[0146]

[0147] Next, calculate the distance from the second point to the third point:

[0148]

[0149] Then, calculate the distance from the third point to the fourth point:

[0150]

[0151] Add these distances together to get the total route length:

[0152]

[0153] The calculated result 14.44 units represents the total inspection path length, which is the physical distance to ensure that the inspection robot or personnel covers all key areas. By optimizing this path length, the efficiency of the inspection process and the completeness of coverage can be ensured, resource consumption can be reduced, and the timeliness of the inspection can be improved. The value is extremely critical in actual inspection route planning because it is directly related to the economic cost and time management of the inspection task.

[0154] According to the inspection planning route, the route is visualized to maximize the inspection efficiency. In combination with environmental factors, an optimization plan for the inspection route is formulated and output;

[0155] Based on the planned inspection route, a detailed inspection route optimization plan is formulated. The process starts with route planning. Through GIS (geographic information system) technology, the optimized route is visualized. This representation not only includes the direction and length of the route, but also the coordinates and turning angles of each turning point. Through simulation operation tests, the impact of various environmental factors on inspection efficiency, such as terrain, weather conditions, etc., is evaluated. At the same time, the system dynamically adjusts the route through algorithms to adapt to emergencies, such as temporary obstacles or the expansion of the fault area. Each optimized route will undergo a cost-benefit analysis to ensure that the path is not only safe but also economical.

[0156] A distribution network operation and maintenance inspection method based on artificial intelligence includes the following steps:

[0157] S1: Deploy acoustic sensors on power equipment in the distribution network to collect sound signals during operation, perform Fourier transform, convert time domain signals into frequency domain signals, compare and analyze them with standard sound patterns, identify frequency changes and abnormal patterns, and obtain acoustic abnormality indicators;

[0158] S2: Deploy thermal imaging cameras to scan the power lines and power equipment of the distribution network, collect real-time temperature readings of each part of the power equipment, compare them point by point with the normal temperature standard, identify and mark the areas where the deviation value exceeds the set threshold, and obtain abnormal temperature hotspot data;

[0159] S3: Synchronize temperature anomaly hotspot data and acoustic anomaly indicators, perform spatial and temporal alignment, extract anomaly frequencies, and compare them with temperature anomaly points to determine all areas where acoustic and temperature anomalies match, and generate thermal imaging anomaly location records;

[0160] S4: Using acoustic anomaly indicators and thermal imaging anomaly location records, synchronized timestamps and geographic information tags, calculate the standard deviation and deviation range between data points and perform difference analysis, combine historical equipment failure data and performance parameters, analyze the probability of equipment damage mode occurrence, and obtain power equipment performance degradation damage mode data;

[0161] S5: Based on the data of power equipment performance degradation and damage mode, compare it with known fault records to reveal the inherent connection of faults under different operating conditions, determine the cause of the fault, visualize the affected area, and issue a fault warning signal;

[0162] S6: Receive fault warning signals, extract fault location information, mark key inspection areas, perform dynamic inspection route planning, visualize the route, and formulate and output inspection route optimization plans based on environmental factors.

[0163] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A distribution network operation and maintenance inspection device based on artificial intelligence, characterized in that: The device comprises: The sound data acquisition module installs acoustic sensors in the power equipment of the distribution network to collect operating noise, extracts the frequency characteristics of sound waves, analyzes the differences in sound patterns, compares the patterns in the normal database, detects potential anomalies based on frequency changes, and obtains acoustic anomaly indicators; The thermal imaging monitoring module uses a thermal imaging camera to scan the power lines and power equipment of the distribution network, record the surface temperature distribution data of the equipment, compare it with the normal temperature standard, identify the abnormal temperature hotspots, and judge whether the hotspot area corresponds to the abnormal frequency in the sound data based on the abnormal temperature hotspot identification results combined with the acoustic abnormality index, and mark the actual abnormal area to obtain the thermal imaging abnormal position record; The data fusion diagnosis module combines the acoustic abnormality index and the thermal imaging abnormality position record to perform difference analysis on the collected data, determine the performance degradation damage mode of the power equipment, determine the matching degree of the result analysis with the known fault record according to the mode, determine the cause of the fault and divide the risk area, and issue a fault warning signal; The path optimization and scheduling module receives the fault warning signal, marks the risk area as a key inspection area, replans the inspection route of the inspection device, gives priority to covering all key inspection areas, and formulates and outputs an inspection path optimization plan.

2. The distribution network operation and maintenance inspection device based on artificial intelligence according to claim 1 is characterized in that: The steps for obtaining the acoustic anomaly index are as follows: Deploy acoustic sensors on power equipment in the distribution network to collect sound signals during operation, convert them into electronic data, and generate raw sound signal data sets; The original sound signal data set is subjected to Fourier transform to convert the time domain signal into the frequency domain signal using the formula: ； Calculate the frequency domain signal value of the sound , generate a frequency feature data set, where Representative time The signal strength value at is a frequency variable; Based on the frequency feature data set, a comparative analysis is performed with the standard sound patterns in the normal database to identify frequency changes and abnormal patterns and obtain acoustic anomaly indicators.

3. The distribution network operation and maintenance inspection device based on artificial intelligence according to claim 2 is characterized in that: The steps of identifying the abnormal temperature hotspot are as follows: Deploy thermal imaging cameras to scan the power lines and power equipment of the distribution network, collect real-time temperature readings of each part of the power equipment, and generate an initial temperature data set for the equipment surface; The initial surface temperature data set of the equipment is compared point by point with the normal temperature standard, using the formula: ； Calculate the coordinates of each point Temperature deviation , and obtain the temperature deviation data set, where is the measured temperature value, is the normal temperature standard value of the coordinate; Analyze the temperature deviation data set to identify deviation values ​​exceeding a set threshold The area is determined and marked as a temperature anomaly hotspot area to obtain temperature anomaly hotspot data.

4. The distribution network operation and maintenance inspection device based on artificial intelligence according to claim 3 is characterized in that: The steps for obtaining the thermal imaging abnormal position record are: Synchronize the temperature anomaly hotspot data and the acoustic anomaly index, perform spatial and temporal alignment, confirm the consistency of the timestamp and position coordinates of each temperature anomaly point, and generate a synchronized multi-source data set; Based on the synchronized multi-source data set, the abnormal frequency is extracted and compared with the temperature abnormal point, using the formula: ； Calculate the correlation between acoustic anomalies and temperature anomalies , evaluate the matching degree and obtain the abnormal matching data set, where Representing coordinates Acoustic anomaly indicators on is the temperature anomaly index at the same coordinate; The anomaly matching data set is analyzed to determine all areas where acoustic and temperature anomalies match, and are marked as actual anomaly areas. The actual position of each anomaly area is located, and a thermal imaging anomaly position record is generated.

5. The distribution network operation and maintenance inspection device based on artificial intelligence according to claim 4 is characterized in that: The steps for determining the damage mode of the power equipment performance degradation are as follows: Using the acoustic anomaly indicators and the thermal imaging anomaly location records, perform data integration, synchronize timestamps and geographic information tags, establish time-space associations, and obtain a comprehensive data set; The composite data set was analyzed to calculate the standard deviation and the range of deviation between data points using the formula: ； Calculate variance analysis values , generate abnormal difference analysis results, where, represents the value of each data point in the synthetic dataset, represents the average value of the comprehensive data set, represents the total number of data points, represents the standard deviation of the data point values, Represents the deviation range value of the data point value; The abnormal difference analysis results are analyzed, and the probability of occurrence of equipment damage modes is analyzed in combination with the historical failure data and performance parameters of the equipment to obtain the damage mode data of the power equipment performance degradation.

6. The distribution network operation and maintenance inspection device based on artificial intelligence according to claim 5 is characterized in that: The steps of issuing the fault warning signal are: Based on the power equipment performance degradation damage mode data, the data is compared with known fault records, fault instances with similar characteristics are screened, and a matching result list is generated; Analyze the matching result list, analyze the relationship between the matched fault mode and the current operating parameters, reveal the inherent connection between the faults under different operating conditions, determine the fault causes, and generate a fault cause analysis report; Based on the fault cause analysis report, the fault data is combined with the geographic coordinates, the affected area is visualized, the area is divided into risk levels, and a fault warning signal is issued.

7. The artificial intelligence-based distribution network operation and maintenance inspection device according to claim 6 is characterized in that: The steps for obtaining the patrol path optimization solution are as follows: Receive the fault warning signal, extract fault location information, mark the risk area as a key inspection area, and generate a fault location marking result; Based on the fault location marking results, dynamic path planning is performed to ensure that all key inspection areas are covered using the formula: ； Calculate the total length of the inspection path , generate the optimal inspection route, where represents the number of points in the path, Representative The coordinates of the points, Representative The coordinates of the points; According to the inspection planning route, the route is visualized to maximize the inspection efficiency, and the inspection path optimization plan is formulated and output in combination with environmental factors.

8. A distribution network operation and maintenance inspection method based on artificial intelligence, characterized in that: The device according to any one of claims 1 to 7 is implemented, comprising the following steps: Deploy acoustic sensors on power distribution network equipment to collect sound signals during operation, perform Fourier transform, convert time domain signals into frequency domain signals, compare and analyze them with standard sound patterns, identify frequency changes and abnormal patterns, and obtain acoustic anomaly indicators; Deploy thermal imaging cameras to scan the power lines and power equipment of the distribution network, collect real-time temperature readings of each part of the power equipment, compare them point by point with the normal temperature standard, identify and mark the areas where the deviation value exceeds the set threshold, and obtain abnormal temperature hotspot data; Synchronize the temperature anomaly hotspot data and the acoustic anomaly index, perform spatial and temporal alignment, extract the anomaly frequency, and compare it with the temperature anomaly point, determine all areas where acoustic and temperature anomalies match, and generate thermal imaging anomaly location records; Using the acoustic anomaly index and the thermal imaging anomaly position record, synchronizing the timestamp and geographic information tag, calculating the standard deviation and deviation range between data points and performing difference analysis, combining the historical equipment failure data and performance parameters, analyzing the probability of occurrence of equipment damage mode, and obtaining the power equipment performance degradation damage mode data; Based on the power equipment performance degradation damage mode data, the data is compared with known fault records to reveal the inherent connection between faults under different operating conditions, determine the cause of the fault, visualize the affected area, and issue a fault warning signal; Receive the fault warning signal, extract the fault location information, mark the key inspection area, perform dynamic inspection route planning, visualize the route, and formulate and output the inspection route optimization plan in combination with environmental factors.

Images