Intelligent robot inspection system in power distribution and transformation intelligent auxiliary system

Abstract

The invention provides an intelligent robot inspection system in a power distribution and transformation intelligent auxiliary system, and relates to the technical field of intelligent sensing and control, and the system comprises a real-time operation collection module which is used for planning a global path of an intelligent inspection robot, generating an initial inspection path, carrying out real-time operation collection, and obtaining a multi-modal detection data set; the fault feature recognition module is used for performing fault feature recognition, generating an abnormal feature set and constructing a dynamic risk map; the simulation maintenance feedback module is used for carrying out simulation maintenance feedback according to the inspection correction path in combination with the equipment maintenance suggestion and generating a maintenance feedback result; and the inspection control module is used for generating an optimized inspection path to perform inspection control on the intelligent inspection robot. According to the invention, the technical problem that the routing inspection efficiency and the fault diagnosis accuracy are influenced because routing inspection path planning often depends on a preset environment map and the intelligent routing inspection robot cannot automatically adjust the routing inspection path according to the change of the equipment state in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent perception and control, and particularly to an intelligent robot patrol inspection system in a power distribution and transformation intelligent auxiliary system. Background Art

[0002] The power distribution and transformation intelligent auxiliary system is an intelligent auxiliary system applied to the power distribution and transformation links in the power system. Its core lies in achieving comprehensive intelligent monitoring and management of power equipment, operating environment, and safety protection by integrating advanced technologies such as computer software and hardware, artificial intelligence, and the Internet of Things.

[0003] Currently, intelligent robot patrol inspection systems have been widely used in the management of power equipment. Intelligent robots achieve intelligent perception and control by carrying different types of sensors. However, due to the accuracy of the sensors themselves and interference from environmental factors, the intelligent perception and control in the existing technology result in insufficient perception accuracy, making the data collected by the sensors may contain noise or errors, thereby affecting the accurate determination of the equipment status and the accuracy of fault diagnosis. Moreover, the path planning of intelligent patrol inspection robots is usually based on the geographical distribution of equipment and patrol inspection requirements, lacking flexibility and unable to be adjusted dynamically in real time according to changes in the equipment status, resulting in high-risk equipment not receiving sufficient attention, which increases the risk of faults not being detected in time and delays fault repair, thereby affecting the operation safety of the power grid. Summary of the Invention

[0004] This application provides an intelligent robot patrol inspection system in a power distribution and transformation intelligent auxiliary system, aiming to solve the technical problem that the patrol inspection path planning in the existing technology often relies on a preset environmental map, and the intelligent patrol inspection robot cannot automatically adjust the patrol inspection path according to changes in the equipment status, resulting in the patrol inspection unable to efficiently cover the fault equipment area, thereby affecting the patrol inspection efficiency and the accuracy of fault diagnosis.

[0005] This application discloses an intelligent robot patrol inspection system in a power distribution and transformation intelligent auxiliary system. The system includes: a real-time operation acquisition module, which is used to plan the global path of the intelligent patrol inspection robot based on three-dimensional environmental data, generate an initial patrol inspection path, and perform real-time operation acquisition on the power distribution and transformation equipment in the power distribution and transformation intelligent auxiliary system to obtain a multi-modal detection data set; a fault feature recognition module, which is used to recognize the fault features of the multi-modal detection data set, generate an abnormal feature set, and construct a dynamic risk map according to the abnormal feature set. The dynamic risk map includes hierarchical alarm signals and equipment maintenance suggestions; a simulated maintenance feedback module, which is used to correct the initial patrol inspection path based on the hierarchical alarm signals, and perform simulated maintenance feedback according to the patrol inspection correction path combined with the equipment maintenance suggestions to generate a maintenance feedback result; a patrol inspection control module, which is used to generate an optimized patrol inspection path according to the maintenance feedback result and send it to the power distribution and transformation intelligent auxiliary system to perform patrol inspection control on the intelligent patrol inspection robot.

[0006] One or more technical solutions provided in this application have at least the following beneficial effects:

[0007] By using three-dimensional environmental data to plan the global path, the intelligent inspection robot can perform intelligent inspections in power transformation and distribution equipment according to the spatial distribution of the equipment and the environmental layout. Through the generation of a more accurate optimal inspection path, the inspection coverage and efficiency are improved. At the same time, through global path planning, the path repetition or empty running phenomenon of the robot during the inspection process is reduced, the energy consumption is lowered, and the inspection task can be efficiently completed. This process ensures that the robot can perceive and adjust the path planning in real time through the application of intelligent perception and control devices, ensuring the optimal execution of the inspection task; during the inspection process, by collecting and comprehensively analyzing multi-modal data in real time, the evaluation of the equipment status is made more comprehensive and accurate, and potential faults of the equipment can be identified more effectively. Through the automatic identification of fault characteristics, a risk map is generated to timely detect and handle potential problems of the equipment, avoiding the limitations and incompleteness of manual inspections; the dynamic risk map provides a hierarchical warning mechanism that can issue corresponding warnings for faults of different severity levels, helping operators to make corresponding responses at different fault levels, ensuring timely and accurate intervention. The construction of the dynamic risk map can improve the intelligent level of equipment management and significantly enhance the timeliness and accuracy of fault warnings; modifying the initial inspection path based on the hierarchical warning signal can not only adjust the inspection task priority according to the actual health status of the equipment, but also automatically adjust the inspection path according to the warning level, so as to ensure that high-risk equipment receives priority attention and optimize the inspection efficiency. The generation of simulated maintenance feedback enables the robot to be corrected according to equipment maintenance suggestions and simulate the effects of different maintenance measures. This process optimizes the equipment maintenance strategy, predicts and adjusts the inspection plan in advance, and improves the management efficiency of the equipment; finally, the optimized inspection path generated according to the maintenance feedback is used to precisely control the inspection of the equipment by the intelligent inspection robot. The optimized inspection path better conforms to the equipment status and maintenance requirements, helps to improve the accuracy and efficiency of the inspection, and also reduces the ineffective path of the robot operation; generally speaking, the application of intelligent perception and control devices ensures that the optimized path can be dynamically adjusted according to the real-time status of the equipment, making the inspection task more efficient and accurate, and greatly enhancing the adaptive ability of the robot's behavior during the inspection process.

[0008] The above description is only an overview of the technical solutions of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the specific embodiments of this application are specifically given below. Brief Description of the Drawings

[0009] Figure 1Schematic diagram of the intelligent robot inspection system in the power distribution and transformation intelligent auxiliary system provided by the embodiment of the present application.

[0010] Figure 2 Schematic diagram of the real-time operation acquisition module in the intelligent robot inspection system of the power distribution and transformation intelligent auxiliary system provided by the embodiment of the present application.

[0011] Explanation of reference numerals: Real-time operation acquisition module 10, fault feature recognition module 20, simulation maintenance feedback module 30, inspection control module 40, three-dimensional space coordinate extraction unit 11, thermal feature marking unit 12, path planning unit 13. Detailed implementation manners

[0012] The embodiment of the present application provides an intelligent robot inspection system in a power distribution and transformation intelligent auxiliary system, which solves the technical problems in the prior art that the inspection path planning often depends on a preset environmental map, and the intelligent inspection robot cannot automatically adjust the inspection path according to the change of the equipment state, resulting in the failure to efficiently cover the faulty equipment area during the inspection, thereby affecting the inspection efficiency and the accuracy of fault diagnosis.

[0013] After introducing the basic principle of the present application, the various non-limiting implementation manners of the present application will be specifically introduced below in conjunction with the accompanying drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0014] As Figure 1 shown, the embodiment of the present application provides an intelligent robot inspection system in a power distribution and transformation intelligent auxiliary system, and the system includes:

[0015] A real-time operation acquisition module 10, configured to plan the global path of the intelligent inspection robot based on three-dimensional environment data, generate an initial inspection path, perform real-time operation acquisition on the power distribution and transformation equipment in the power distribution and transformation intelligent auxiliary system, and obtain a multi-modal detection data set.

[0016] Generate three-dimensional point cloud data by scanning the substation with a lidar. The point cloud data contains the spatial positions and shape information of various objects in the environment, and extract the three-dimensional space coordinates of the power distribution and transformation equipment from the point cloud data. Combine the thermal distribution data obtained from the infrared thermal imaging sensor with the equipment spatial coordinate information extracted from the three-dimensional point cloud to generate a thermal-space fusion array, which is used to reflect the relationship between the thermal state of the equipment and its spatial position. Perform multi-target path planning on the power distribution and transformation intelligent auxiliary system based on the thermal-space fusion array to generate an initial inspection path.

[0017] Configure a visible light camera, an ultrasonic probe, and a partial discharge sensor on the intelligent inspection robot. Capture the real-time images of the power distribution and transformation equipment through the visible light camera, capture the ultrasonic signal data through the ultrasonic probe for structural health detection, and capture the partial discharge signals on the equipment through the partial discharge sensor for electrical detection. During the real-time operation of the acquisition process, use a hardware trigger signal to ensure the consistency of the acquisition timing of different sensors, so that there is a consistent time stamp during data fusion.

[0018] Perform preprocessing on the sensor data to obtain a multi-modal detection data set. Specifically, perform real-time distortion correction on the collected visible light images to ensure the geometric accuracy of the images, use the adaptive histogram equalization algorithm to perform illumination compensation on the images, enhance the contrast of the images, and process the image differences under different illumination conditions; perform band-pass filtering on the ultrasonic signals, select the effective echo signals within a specific frequency band (such as 0.5 - 5 MHz) for analysis, which helps to eliminate the noise signals outside the frequency band and enhance the recognition ability of the effective signals; perform amplitude-phase joint analysis on the partial discharge pulse signals, eliminate the electromagnetic interference pseudo-pulses, and extract the characteristics of the real partial discharge signals.

[0019] The fault feature recognition module 20 is used to recognize the fault features of the multi-modal detection data set, generate an abnormal feature set, and construct a dynamic risk map based on the abnormal feature set. The dynamic risk map includes graded warning signals and equipment maintenance suggestions.

[0020] Process the multi-modal detection data set, extract their respective features, and convert the features provided by each sensor into vectors. For example, the visible light image feature vector, the ultrasonic signal feature vector, and the partial discharge feature vector. Weightedly fuse these feature vectors to obtain a multi-dimensional abnormal feature set. This step weights the features of different sensors by using an attention mechanism, so as to more accurately reflect the fault features of different data sources.

[0021] Combine the historical equipment defect data, compare the currently collected abnormal features with the historical defect data through similarity analysis to identify potential fault modes, perform dynamic risk assessment based on the potential fault modes, generate graded warning signals, and the levels of the warning signals are divided into low risk, medium risk, high risk, etc. Give equipment maintenance suggestions based on the potential fault modes, including regular inspections, replacement of parts, reinforcement of structures, etc. Integrate the fault modes, warning signals, and equipment maintenance suggestions into a dynamic risk map, which can reflect the health status of the power distribution and transformation system in real time and help operators make faster and more accurate responses.

[0022] The simulation maintenance feedback module 30 is used to correct the initial inspection path based on the hierarchical alarm signal, and perform simulation maintenance feedback according to the inspection correction path combined with the equipment maintenance suggestions to generate a maintenance feedback result.

[0023] When the hierarchical alarm signal is triggered, real-time parsing is first performed. The hierarchical alarm signal reflects potential faults and risks during equipment operation. According to the hierarchical alarm signal, the target alarm equipment to be inspected is determined, and the three-dimensional coordinates and alarm risk levels of these equipment are extracted. The three-dimensional coordinates of the target alarm equipment are used as the nodes of the topological network, and the weights of the edges are calculated according to the alarm risk level and spatial distance to construct an alarm topological network. According to the alarm topological network, correction analysis is performed to generate a set of path correction factors. The set of path correction factors contains path parameters that must be adjusted to avoid risk areas, collisions, or high-risk areas. The initial inspection path is corrected according to the set of path correction factors to generate an inspection correction path.

[0024] After obtaining the inspection correction path, simulation maintenance feedback is performed in combination with the equipment maintenance suggestions. This means generating a simulated maintenance action plan based on the current health status of the equipment (through the analysis of alarm signals and fault modes), and generating a final simulated maintenance feedback result according to the simulated maintenance process. The simulated maintenance feedback result includes information such as the estimated maintenance time, required resources, and changes in equipment status. The simulated maintenance feedback result will be used as the basis for optimizing the inspection path to ensure that the inspection robot can perform inspections according to the optimal path and sequence, so as to minimize the equipment failure risk.

[0025] The inspection control module 40 is used to generate an optimized inspection path according to the maintenance feedback result and send it to the power distribution and transformation intelligent auxiliary system to control the inspection of the intelligent inspection robot.

[0026] By analyzing the maintenance feedback result, the current health status of the equipment and the maintenance suggestions are determined. Based on this, the inspection path is optimized. For example, for high-risk equipment, the inspection robot needs to increase the inspection frequency or change the inspection path. Specifically, a multi-objective evolutionary algorithm is used to optimize the inspection path. The optimization objectives include reducing the path length, avoiding high-risk areas, and shortening the inspection time.

[0027] After completing the path optimization, the optimized inspection path is sent to the power distribution and transformation intelligent auxiliary system. This path contains the corrected path sequence and new inspection points. The power distribution and transformation intelligent auxiliary system controls the inspection tasks of the intelligent inspection robot according to the optimized inspection path. The robot automatically performs inspections according to the optimized inspection path. This process ensures that the robot can perform inspection tasks along the optimal path, thereby improving the inspection efficiency and the monitoring effect of the equipment health status.

[0028] Furthermore, as Figure 2As shown, the real-time operation and acquisition module includes:

[0029] A three-dimensional space coordinate extraction unit 11, which is used to generate three-dimensional point cloud parameters by lidar scanning a substation and extract a set of three-dimensional space coordinates of power transformation and distribution equipment; a thermal feature annotation unit 12, which is used to fuse the thermal distribution data obtained by an infrared thermal imaging sensor with the set of three-dimensional space coordinates, perform thermal feature annotation on the three-dimensional point cloud parameters, and generate a thermal-space fusion array; a path planning unit 13, which is used to perform multi-objective path planning on the power transformation and distribution intelligent auxiliary system based on the thermal-space fusion array and generate the initial inspection path.

[0030] The lidar uses laser beams to scan the environment and calculates the distance by measuring the time it takes for laser pulses to reflect back from the target object, thereby obtaining the position and shape data of the object. During the scanning process, the lidar generates three-dimensional point cloud parameters, and each point contains a point coordinate in space, usually in three dimensions of X, Y, and Z. These points form the three-dimensional coordinate information of all objects in the environment. According to the point cloud data, a set of three-dimensional space coordinates of power transformation and distribution equipment is extracted from a large number of points. These equipment include distribution transformers, circuit breakers, switches, lines, etc. The extracted set of three-dimensional space coordinates of the equipment is used for subsequent path planning and analysis.

[0031] The infrared thermal imaging sensor can capture the temperature distribution on the surface of the equipment. The sensor generates thermal distribution data by measuring the infrared radiation emitted by the surface of the object. Each point in the thermal distribution data represents the temperature at a certain position on the surface of the equipment, reflecting the temperature change on the surface of the equipment. In the power transformation and distribution system, the temperature change of the equipment can reveal the working state of the equipment. For example, overheating indicates equipment failure or excessive load.

[0032] Fuse the thermal distribution data with the set of three-dimensional space coordinates. Specifically, convert the thermal distribution data into coordinates in three-dimensional space so as to match the three-dimensional space coordinates of the equipment. The three-dimensional space coordinates of each equipment position correspond to a certain temperature value in the thermal distribution, forming a thermal-space data point. Map the thermal information of each equipment into the three-dimensional space of the equipment to form a thermal-space fusion array. This array reflects the position of the equipment in space and its temperature state.

[0033] Based on the thermal-space fusion array, multi-target path planning is carried out to ensure that the intelligent inspection robot can efficiently inspect all devices. Specifically, the first objective function is defined as maximizing the device detection coverage rate, the reachability weight coefficient of each device point is calculated, the second objective function is defined as the priority of the thermal anomaly area, the area detection urgency is set according to the temperature rise rate, the third objective function is defined as minimizing the path energy consumption, the steering angle energy consumption factor is calculated based on the robot motion model, and the weighted summation method is used to transform the multi-objective into a single-objective optimization problem. The constraint conditions include the obstacle avoidance radius and the minimum detection residence time of the device. An optimization algorithm, such as the A* algorithm, Dijkstra algorithm, genetic algorithm, etc., is used to calculate the optimal path. After the planning is completed, the initial inspection path is generated.

[0034] Furthermore, the thermal feature annotation unit includes:

[0035] The feature point matching channel is used to perform feature point matching based on the thermal distribution data and the three-dimensional space coordinate set to construct a coordinate system transformation matrix; the visualization layer generation channel is used to retrieve the dynamic pseudo-color mapping rule for setting the rated temperature threshold of the power distribution and transformation equipment in the power distribution and transformation intelligent auxiliary system, and generate a temperature gradient visualization layer according to the dynamic pseudo-color mapping rule; the anomaly analysis channel is used to traverse the temperature gradient visualization layer and combine the thermal distribution data for anomaly analysis to determine the temperature anomaly area, and extract the continuous thermal anomaly boundary contour according to the temperature anomaly area; the feature annotation channel is used to delimit the thermal anomaly area according to the continuous thermal anomaly boundary contour, perform thermal feature annotation on the three-dimensional point cloud parameters based on the thermal anomaly area, and generate the thermal-space fusion array.

[0036] Feature points are extracted from the thermal distribution data and the three-dimensional space coordinate data. Feature points in the three-dimensional space coordinate data are points with obvious features in the device or environment (such as the surface, corner, edge, etc. of the device), and in the thermal distribution data, they are points with significant temperature changes (such as positions with rapid temperature changes). Through spatial mapping methods, such as the least squares method, RANSAC algorithm, etc., the feature points in the thermal distribution data are matched with the feature points in the three-dimensional space coordinate data. This process ensures that the two-dimensional temperature map of the thermal data can be correctly mapped to a specific position in the three-dimensional space.

[0037] Based on the matched feature points, a coordinate system transformation matrix is constructed. This matrix is used to transform the thermal distribution data from a two-dimensional plane to a three-dimensional space, so that the thermal data can be accurately fused with the three-dimensional space coordinate set of the device. Using this transformation matrix, the thermal map captured by the thermal imaging sensor (usually two-dimensional data) is converted into a thermal-space fusion array aligned with the lidar point cloud data (usually three-dimensional data), which provides accurate data support for subsequent analysis and path planning.

[0038] Power transformation and distribution equipment usually has a set rated operating temperature threshold. Equipment with a temperature exceeding this threshold may pose a risk of failure. Therefore, first obtain the rated temperature thresholds of the equipment from the intelligent auxiliary system for power transformation and distribution. These temperature thresholds are the maximum safe operating temperatures specified during the equipment design.

[0039] Pseudo-color mapping is to map temperature data to color values, enabling temperature changes to be intuitively displayed through color changes. Different temperature ranges can be mapped to different colors. For example, red represents high temperature and blue represents low temperature. The dynamic pseudo-color mapping rule automatically adjusts the color mapping according to the real-time temperature data and temperature thresholds of the equipment. For example, when the temperature of the equipment approaches or exceeds the threshold, the equipment is displayed in a more prominent color for easy identification by inspection personnel. Based on the dynamic pseudo-color mapping rule, the temperature data of the equipment is converted into a temperature gradient visualization layer, and the temperature changes in the layer can be intuitively shown through color gradients. In this way, inspection personnel can quickly identify temperature anomaly areas.

[0040] In the generated temperature gradient visualization layer, traverse each area, mainly focusing on the anomalies of temperature values. The definition of an anomaly is that the temperature exceeds the rated temperature threshold of the equipment. By comparing the temperature of each equipment point with its rated temperature threshold, the temperature anomaly areas are determined. Extract the boundary contours of the temperature anomaly areas. For example, use the Canny edge detection algorithm, Sobel operator, or region-growing-based method to extract the boundaries of the anomaly areas and generate continuous thermal anomaly boundary contours. These contours identify the specific locations and shapes of the temperature anomalies and provide a reference for subsequent analysis.

[0041] According to the extracted continuous thermal anomaly boundary contours, delimit these contour areas as thermal anomaly areas. Thermal anomaly areas pay more attention to anomalies in heat energy transfer or distribution. It not only considers temperature changes but also deeply analyzes the source of heat energy, the heat transfer path, and the distribution among different areas. Anomalous phenomena are discovered from the perspective of thermodynamics. According to the delimited thermal anomaly areas, label the thermal information (such as temperature, temperature change rate, etc.) of these areas into the corresponding 3D point cloud data. This labeling helps subsequent analysis to determine which equipment has overheating problems and classify them according to the severity of the problems. After labeling, the 3D point cloud data and thermal data will be fused into a new data structure, namely the thermal-space fusion array. This array contains the 3D spatial coordinates and temperature information of the equipment and indicates which areas belong to the temperature anomaly areas.

[0042] Furthermore, the fault feature recognition module includes:

[0043] A feature extraction unit for performing cross - domain feature extraction on the multi - modal detection data set to generate multiple feature vectors; a weighted fusion unit for performing weighted fusion based on the multiple feature vectors using an attention mechanism to construct a multi - dimensional abnormal feature set; a similarity analysis unit for retrieving the historical defect database, migrating the multi - dimensional abnormal feature set to the historical defect database for similarity analysis, and identifying faults according to data similarity to determine potential fault modes; a dynamic risk assessment unit for performing dynamic risk assessment based on the potential fault modes to construct a hierarchical warning signal; a device maintenance analysis unit for performing device maintenance analysis based on the potential fault modes to formulate device maintenance suggestions; an association integration unit for associating and integrating the hierarchical warning signal with the device maintenance suggestions to construct a dynamic risk map.

[0044] Based on the analysis of the multi - modal detection data set, a visible light image data set, an ultrasonic signal data set, and a partial discharge data set are obtained. Data of different modalities have different feature spaces. For cross - domain feature extraction, specifically, a convolutional neural network is used to extract spatial features from images, such as shape, color, texture, etc., to generate image feature vectors; the ultrasonic signal is subjected to spectral analysis, for example, through Fourier transform or wavelet transform, to extract the frequency components, energy distribution, etc. of the signal, and generate ultrasonic signal feature vectors; based on the time - domain or frequency - domain features of the partial discharge signal, features such as the amplitude, frequency, and pulse cluster of the partial discharge signal are extracted to generate partial discharge signal feature vectors. These feature vectors represent the state of the device in different modalities and will provide support for subsequent fault identification and risk analysis.

[0045] The attention mechanism can assign weights according to the importance of different features in a task. In the feature fusion of multi - modal data, the attention mechanism helps to automatically determine which modal data features are more important, thus strengthening the attention to important features. Multiple feature vectors are input into the attention mechanism, and the attention mechanism assigns weights to each feature vector. These weights are adjusted based on the importance of the features in the overall decision. Features with higher weights will have a greater impact on the final decision. Through weighted fusion, multiple feature vectors are combined into a multi - dimensional abnormal feature set, which represents the comprehensive information of different modal data in potential fault detection.

[0046] Retrieve the historical defect database. The historical defect database contains the fault modes, fault types, fault occurrence conditions, and repair records of past devices. By querying this database, historical fault modes similar to the current device state can be found.

[0047] Migrate the currently generated multi-dimensional anomaly feature set to the historical defect database for comparison. The migration is carried out by processing the features such as standardization and normalization so that they can match the features in the historical database, and calculate the similarity between these features and the historical data. For example, the cosine similarity metric method is used to calculate the similarity between the current feature vector and the historical data vector. The larger the value, the higher the feature similarity. High similarity means that the current device faces a fault similar to a certain fault mode in the historical data. According to the results of the similarity analysis, identify the potential fault modes of the current device, such as overheating, structural damage, electrical discharge, etc.

[0048] Conduct a risk assessment on the potential fault modes to determine the likelihood of the fault occurring and the impact on the system after the fault occurs. Specifically, according to factors such as historical data, device health status, and working environment, evaluate the probability of a certain fault mode occurring. The higher the probability, the greater the risk. Analyze the degree of impact on the device or the entire power transformation and distribution system when the fault occurs. For example, some faults cause the system to stop operating completely, while others only have a local impact or even no obvious impact. Quantify the probability and impact of the fault occurrence. For example, use a risk matrix to evaluate different fault modes and calculate the risk value of the potential fault mode. According to the results of the dynamic risk assessment, generate corresponding warning signals for the potential fault modes and classify them according to different risk value thresholds to construct a hierarchical warning signal.

[0049] According to the results of the dynamic risk assessment of the potential fault modes, formulate a targeted device maintenance plan to prevent the occurrence of faults or reduce the impact of faults on the system.

[0050] Exemplarily, for fault modes with a relatively low risk value, they usually have little impact on the device or system, generate low-risk warnings, and it is recommended to conduct regular inspections or monitoring. The maintenance of regular inspections includes daily maintenance such as equipment cleaning, lubrication, and temperature monitoring; for fault modes with a relatively high risk value, partial repair or maintenance is required, generate medium-risk warnings, and it is recommended to conduct preventive maintenance. This type of maintenance focuses on prevention by detecting and replacing some components or systems to reduce the likelihood of faults occurring; for fault modes with a very high risk value, they may cause equipment damage or system shutdown, generate high-risk warnings, and it is recommended to handle them immediately, including shutting down the equipment, emergency repair, or replacing key components, etc.

[0051] Associate the hierarchical warning signals with the corresponding device maintenance suggestions. The level (low, medium, high risk) of the warning signal determines the priority and handling method of the maintenance suggestions. The dynamic risk map is a visualization tool used to display the real-time health status, fault risks, and maintenance requirements of various parts of the device, system, or network. The map presents the warning signals and maintenance suggestions in a graphical way to help staff quickly identify the devices that need to be focused on.

[0052] Furthermore, the feature extraction unit includes:

[0053] A parsing channel for parsing based on the multi-modal detection data set to obtain a visible light image data set, an ultrasonic signal data set, and a partial discharge data set; a dual-channel convolution analysis channel for performing dual-channel convolution analysis based on the visible light image data set to obtain visible light image features; a wavelet packet energy entropy analysis channel for performing wavelet packet energy entropy analysis based on the ultrasonic signal data set to obtain ultrasonic signal features; a pulse cluster clustering channel for performing pulse cluster clustering based on the partial discharge data set to extract partial discharge features; a feature vector construction channel for analyzing the temporal evolution trends of the visible light image features, the ultrasonic signal features, and the partial discharge features to construct visible light image feature vectors, ultrasonic signal feature vectors, and partial discharge feature vectors; and a data integration channel for integrating the visible light image feature vectors, the ultrasonic signal feature vectors, and the partial discharge feature vectors to obtain the multiple feature vectors.

[0054] During the intelligent inspection process, a visible light image data set, an ultrasonic signal data set, and a partial discharge data set are collected through a visible light camera, an ultrasonic probe, and a partial discharge sensor configured on the intelligent inspection robot.

[0055] The dual-channel convolutional network is a deep learning method that processes input data from two different sources simultaneously through two channels. In visible light image processing, the dual-channel convolutional network can extract features of the image from different processing levels. Among them, the first channel extracts surface defect texture features, and the second channel extracts structural deformation geometric features. Each channel extracts image features through a series of convolutional operations. The features extracted from the two different channels are fused to form visible light image features.

[0056] Wavelet packet transform is a time-frequency analysis method that can better capture the detailed changes of a signal by decomposing the signal into different frequency bands (high-frequency and low-frequency components). It has stronger frequency decomposition ability than traditional wavelet transform and can extract more frequency components of the signal. Energy entropy is an index used to measure the uncertainty or complexity of a signal. In ultrasonic signal analysis, energy entropy reflects the complexity of the signal and the distribution of information. A signal with a higher energy entropy indicates the existence of complex physical phenomena such as cracks and corrosion, while a lower energy entropy indicates that the signal is relatively simple and regular.

[0057] The ultrasonic signal is decomposed into multiple frequency bands using wavelet packet transform. These frequency bands represent the components of the signal in different frequency ranges. Through wavelet packet decomposition, the high-frequency and low-frequency components of the ultrasonic signal are captured for detecting internal cracks or damages. The energy of the signal in each frequency band is calculated to obtain the energy values of each frequency band. By calculating the energy distribution of different frequency bands, the energy spectrum of the signal is obtained. Based on the energy spectrum, the entropy value of each frequency band is calculated. The larger the entropy value, the more complex the signal in that frequency band. After wavelet packet energy entropy analysis, the characteristics of the ultrasonic signal obtained reflect the complexity and frequency components of the signal.

[0058] Partial discharge refers to the phenomenon of small electrical discharges inside or on the surface of electrical equipment. It usually occurs in electrical insulation materials and can reflect the electrical health status of the equipment. The partial discharge signal often increases before the equipment fails. Pulse cluster clustering is a method for analyzing partial discharge signals. It clusters the pulses in the partial discharge signal according to a certain time interval or similarity. Clustering can reveal the nature, occurrence frequency, and area of the discharge event. By using clustering algorithms such as K-means clustering, DBSCAN clustering, etc., three-dimensional statistics of the discharge times per second, average amplitude, and skewness of the phase distribution are extracted, and similar discharge events are classified into the same cluster, thereby extracting the characteristic pattern of the discharge as the partial discharge characteristic.

[0059] Time-series evolution analysis refers to the analysis of the trend of equipment status data (including visible light images, ultrasonic signals, partial discharge signals) over time. This analysis can reveal the dynamic changes in the equipment's health status and help identify potential fault patterns. Specifically, by analyzing consecutive frames of visible light images, the trend of changes on the equipment surface is identified. The change trends include crack propagation, surface wear, etc. During the analysis, techniques such as image differencing method and optical flow method are used to track the movement and changes of objects between consecutive frames, and a feature vector of the visible light image is constructed to reflect the changes in the equipment's appearance; ultrasonic signals are used to detect internal structural problems of the equipment, such as cracks, corrosion, etc. Over time, the characteristics of the signal such as frequency, amplitude, and phase will change. Through time-series analysis, a feature vector of the ultrasonic signal is obtained to provide quantitative information on the internal health status of the equipment; partial discharge signals are important early signals of electrical equipment failures. As the equipment ages or is damaged, the frequency, amplitude, and pulse pattern of partial discharge will change. Through time-series analysis, a partial discharge feature vector is obtained to reflect the changes in the electrical state of the equipment.

[0060] The feature vectors of visible light images, ultrasonic signals, and partial discharge signals are integrated to obtain multiple feature vectors, which simultaneously reflect the health status of the equipment in different dimensions (visual, structural, electrical).

[0061] Furthermore, the simulation maintenance feedback module includes:

[0062] A real-time parsing unit for performing real-time parsing based on the hierarchical alarm signal to determine the target alarm device, and extracting three-dimensional alarm coordinates and alarm risk levels based on the target alarm device; A correction analysis unit for constructing an alarm topology network according to the three-dimensional alarm coordinates and the alarm risk level, and performing correction analysis according to the alarm topology network to generate a set of path correction factors; A correction unit for correcting the initial inspection path based on the set of path correction factors to generate an initial inspection correction path; A correction path determination unit for performing collision detection according to the initial inspection correction path, judging the completion progress index of the intelligent robot inspection task according to the detection result, and determining the inspection correction path according to the completion progress index.

[0063] Perform real-time parsing on the hierarchical alarm signal to determine which device has a problem, and identify the target alarm device that needs to be processed first through parsing. The target alarm device is usually a high-risk device, and their failures may have a serious impact on the system. Determine the location of the target alarm device according to the three-dimensional spatial coordinates of the device (data from lidar scanning), and extract the three-dimensional alarm coordinates, which are used for subsequent path planning and correction; at the same time, extract the alarm risk level of the target alarm device for path correction and priority sorting.

[0064] Use the three-dimensional alarm coordinates and alarm risk level of the target alarm device to construct an alarm topology network. In this network, each node represents a device, and the edges between devices represent their spatial or functional relationships. The weight of a node is determined by its risk level. For example, the node weight of a high-risk device is larger. After constructing the alarm topology network, perform path correction analysis. By calculating the distances, risk levels, and priorities of alarm signals between different devices, analyze which devices need to be inspected first and which devices can be skipped temporarily. Based on the analysis of the topology network, generate a set of path correction factors, which represent the priorities between devices, the adjusted inspection order, and the paths that need to avoid high-risk areas.

[0065] According to the generated set of path correction factors, correct the initial inspection path, including arranging high-risk devices at the front of the path first; avoiding areas where high risks or failures have been detected to ensure that the robot does not enter potentially dangerous areas; optimizing the path to reduce unnecessary repeated inspections and improve efficiency. After path correction, generate an initial inspection correction path, which can more effectively respond to the health status and failure risks of devices.

[0066] Perform collision detection according to the initial inspection correction path. The collision detection includes detecting whether the robot will collide with obstacles in the environment (such as equipment, walls, facilities, etc.), and detecting whether the robot will collide due to narrow space or equipment configuration problems during its movement. Use simulation technology to virtually run the correction path and detect potential collision risks.

[0067] Based on the results of the collision detection, judge the completion progress of the inspection task. The progress index is a value reflecting the completion degree of the robot inspection task, which is statistically calculated based on the equipment that has completed the inspection and the inspection tasks that have not been completed. For example, if there are obstacles or equipment failures on the path, the progress index decreases; if the robot successfully completes the inspection without encountering obstacles, the progress index increases. According to the calculated completion progress index, further optimize the inspection path. For example, if the progress index is low, adjust the path or add additional inspection points to ensure that the task can be successfully completed. After the optimization is completed, determine the inspection correction path to improve the inspection efficiency.

[0068] Furthermore, the correction analysis unit includes:

[0069] An edge weight parameter acquisition channel for calculating node space distance data based on the three-dimensional alarm coordinates as topological nodes, and using the alarm risk level combined with the node space distance data as edge weight parameters; an alarm topology network construction channel for constructing an alarm topology network based on the topological nodes and the edge weight parameters; a calculation channel for calculating the initial inspection path according to the alarm topology network to obtain a correction cost matrix, where the correction cost matrix contains path correction weights; a descending order arrangement channel for performing multi-objective evaluation based on the path correction weights, arranging them in descending order according to the evaluation results, and generating a path correction priority list; a simulation channel for determining the path correction factor set according to the path correction priority list combined with the movement trajectory simulation of the intelligent robot.

[0070] In the alarm topology network, the position of each device is represented by three-dimensional alarm coordinates, serving as a node of the topology network. Calculate the space distance between topological nodes, and the Euclidean distance is used to calculate the space distance. When constructing the topology network, the connection (i.e., edge) between each pair of devices will calculate the weight according to its space distance and alarm risk level. The edge weight combines the physical distance between devices and the urgency of equipment failures. Among them, the shorter the distance, the smaller the weight of the connection, indicating the relative closeness between devices; the higher the risk level (such as high-risk alarm), the greater the weight, indicating that the device needs to be processed first.

[0071] Using the calculated edge weight parameters and topological nodes, an alarm topology network is constructed. This network consists of several nodes (alarm devices) and the edges between them (representing the relationships or connections between devices). The weight of each edge is jointly determined by the spatial distance between devices and the alarm risk level.

[0072] The initial inspection path does not consider the risk level and spatial distance. Based on the alarm topology network, the correction cost of the initial inspection path is calculated. The correction cost represents the path cost from one device to another, including the weighted influence of spatial distance and risk level. After calculation, a correction cost matrix is obtained, which represents the correction cost of each path.

[0073] Multi-objective evaluation is carried out. The optimization objectives include: minimizing the path length, that is, optimizing the total length of the inspection path to reduce the distance and time of the robot's movement; minimizing the risk, that is, preferentially inspecting high-risk devices to ensure the timely monitoring of the health status of the devices; avoiding collisions, that is, preventing the robot from colliding with obstacles or other devices during the inspection process. According to these objectives, the comprehensive score of each path is calculated by combining the path correction weight. According to the results of the multi-objective evaluation, all the corrected paths are sorted in descending order to generate a path correction priority list. Among them, the higher the priority of the path, the more important the path is during the inspection process. The high-priority paths are selected for inspection first, and the low-priority paths are postponed or skipped, especially when the path contains low-risk devices or has passed through other inspection paths.

[0074] After determining the path correction priority list, the motion trajectory simulation of the intelligent robot is combined to simulate the movement of the robot along the corrected path to identify potential problems in the actual execution process, such as collisions and route adjustments. According to the simulation results, a set of path correction factors is determined. The correction factors include the inspection priority of the device, the obstacle avoidance route, the health status of the device, etc. Each correction factor represents the part of the path that needs to be adjusted during the inspection process.

[0075] Furthermore, the corrected path determination unit includes:

[0076] A collision detection channel for performing multi-modal collision detection according to the initial inspection corrected path and obtaining the spatial distribution data of obstacles in real time; a progress evaluation channel for introducing the intelligent robot inspection task for progress evaluation, obtaining the completion progress index of the intelligent robot inspection task, and calculating the remaining path risk assessment value based on the completion progress index; a local correction channel for performing local correction according to the spatial distribution data of obstacles in combination with the remaining path risk assessment value to generate a candidate set of obstacle avoidance paths; an update channel for updating the initial inspection corrected path according to the candidate set of obstacle avoidance paths to determine the inspection corrected path.

[0077] During the robot's inspection along the corrected path according to the initial inspection route, real-time images of power transformation and distribution equipment are captured by a visible light camera, ultrasonic signal data is captured by an ultrasonic probe for structural health detection, and partial discharge signals on the equipment are captured by a partial discharge sensor for electrical detection. These data provide real-time updates for collision detection, enabling the robot to dynamically adjust its path. Based on the sensor data, spatial distribution data of obstacles in the environment is generated. This data structure is an obstacle map with spatial coordinates, reflecting the type, location, and size of the obstacles.

[0078] Evaluate the progress of the task based on the actual inspection status of the robot. The progress index is a digital indicator representing the percentage of progress of the robot in the inspection task. For example, if the robot has completed 50% of the inspection path, the progress index is 0.5. Based on the current progress index, evaluate the risk of the remaining path. Among them, if the risk level of the remaining equipment is high, the risk assessment value of the remaining path will be high.

[0079] Based on the spatial distribution data of obstacles and the risk assessment value of the remaining path, make local corrections to the inspection path. The goal is to avoid obstacles and ensure the smooth completion of the inspection task in high-risk areas. Specifically, in places with obstacles, use path planning algorithms such as the A* algorithm to find a new path to bypass the obstacles or high-risk areas; if the path passes through a high-risk area, avoid entering this area through path correction or re-plan the path to reduce the likelihood of failures. According to the local corrections, generate a candidate set of obstacle avoidance paths. These candidate paths represent different detour methods or optimized paths.

[0080] After generating the candidate set of obstacle avoidance paths, select the optimal path according to the evaluation results and update this path as the corrected inspection path. The new corrected inspection path features obstacle avoidance and low risk. This path will be provided to the intelligent inspection robot so that it can smoothly complete the inspection task according to the new path.

[0081] Furthermore, the inspection control module includes:

[0082] A status change analysis unit is used to perform a status change analysis on the power distribution and transformation equipment of the power distribution and transformation intelligent auxiliary system based on the maintenance feedback result to determine the equipment status improvement parameters; a secondary detection analysis unit is used to traverse the power distribution and transformation equipment of the power distribution and transformation intelligent auxiliary system according to the equipment status improvement parameters to perform secondary detection analysis to determine the secondary detection priority coefficient; a health prediction unit is used to retrieve historical cruise data and combine it with the secondary detection priority coefficient to perform a health prediction on the power distribution and transformation equipment of the power distribution and transformation intelligent auxiliary system to generate an equipment health prediction score; a multi-objective evolution unit is used to perform multi-objective evolution on the inspection correction path according to the equipment health prediction score to generate the optimized inspection path; an inspection control unit is used to send the optimized inspection path to the power distribution and transformation intelligent auxiliary system to perform inspection control on the intelligent inspection robot to form a closed-loop control loop.

[0083] Based on the maintenance feedback result, perform a change analysis on the status of the power distribution and transformation equipment. The purpose of the analysis is to identify whether the status of the equipment has improved after maintenance, whether it has returned to the normal operating state, or whether there are still potential problems. For example, if the equipment returns to normal operation after repair, or if the equipment has not fully recovered after maintenance, or if the health status of the equipment fluctuates, according to the analysis result, determine the equipment status improvement parameters. The equipment status improvement parameters are indicators used to quantify the change in the equipment status.

[0084] After the initial maintenance, perform secondary detection analysis on the equipment based on the equipment status improvement parameters. The secondary detection aims to further confirm whether the equipment has returned to the ideal state or to check whether new problems have occurred. Traverse all the power distribution and transformation equipment in the power distribution and transformation intelligent auxiliary system and determine which equipment needs to be subjected to secondary detection according to the equipment status change parameters (such as temperature, efficiency, failure rate, etc.). For example, if the equipment status improvement is not obvious or the failure recurs, mark it as a high-priority equipment for secondary detection; if the equipment status has improved well, then secondary detection is not required, or a routine inspection is carried out. To optimize resource and time allocation, assign a secondary detection priority coefficient to each equipment according to factors such as the risk level, failure mode, and maintenance history of the equipment, sort the equipment by priority, and high-priority equipment will be given priority for detection, while low-priority equipment can be inspected later.

[0085] Historical cruise data refers to the equipment operation data collected by intelligent inspection robots during past inspection tasks. This data includes information such as the operation status of equipment, inspection paths, maintenance records, equipment failure history, and environmental data. Based on the historical cruise data and combined with the secondary detection priority coefficient, health prediction is carried out on the power distribution and transformation intelligent auxiliary system's power distribution and transformation equipment. Specifically, machine learning models based on historical data and current equipment status, such as regression analysis, time series analysis, or neural networks, are used to perform health prediction on each device. By adjusting the focus of health prediction according to the secondary detection priority of the device, it is ensured that the health status of high-priority devices is more accurately evaluated. By analyzing the historical operation data, failure records, and environmental factors of the device, the health status of the device in the future period is predicted, and a device health prediction score is generated. The device health prediction score is a quantitative score based on the above prediction results, indicating the current and future health status of the device.

[0086] Based on the device health prediction score, multi-objective path optimization is carried out on the inspection correction path. The optimization objectives include minimizing inspection time, minimizing risk, and optimizing resources (reducing unnecessary repeated inspections, saving the working time and energy of the robot), etc. Evolutionary algorithms, such as genetic algorithms and particle swarm optimization algorithms, are used to optimize the inspection path. These algorithms find the optimal solution by iteratively integrating different objectives. The evolutionary algorithm adjusts the inspection path according to the device health prediction score, giving priority to arranging devices with low health scores and higher risks, and optimizing the path on this basis. After multi-objective evolution, an optimized inspection path is generated. This path achieves an optimal balance in terms of time, risk, resources, etc. The optimized path will be used in subsequent inspection tasks to ensure the efficiency of the robot inspection task and the health management of the equipment.

[0087] The optimized inspection path is sent to the power distribution and transformation intelligent auxiliary system, and the power distribution and transformation intelligent auxiliary system controls the movement of the intelligent inspection robot according to this path to independently execute the inspection task. During the inspection process, the robot continuously collects the operation data and feedback information of the equipment based on sensors (including lidar, cameras, ultrasonic detectors, etc.), and adjusts the inspection path in real time accordingly. For example, if the health status of a certain device changes significantly or a new warning signal appears, the inspection path is dynamically adjusted based on the real-time data to ensure that the robot can respond to the health changes of the equipment in real time. In this way, a closed-loop control loop is formed to ensure the dynamic optimization of the inspection task. This closed-loop control process enables the robot to make flexible adjustments according to the continuously changing health status of the equipment, ensuring continuous monitoring and timely maintenance of the equipment's health.

[0088] Furthermore, the optimized inspection path is sent to the power distribution and transformation intelligent auxiliary system for inspection control of the intelligent inspection robot. The inspection control unit includes:

[0089] The inspection control channel is used to send the optimized inspection path to the power distribution and transformation intelligent auxiliary system for inspecting and controlling the intelligent inspection robot, and collect and obtain the mechanical vibration feedback data of the intelligent inspection robot; the twin simulation channel is used to perform twin simulation on the intelligent inspection robot according to the mechanical vibration feedback data to obtain the equipment state evolution simulation parameters; the closed-loop control channel is used to start the online replanning mechanism for closed-loop control and construct the closed-loop control loop when the error between the equipment state improvement parameters and the equipment state evolution simulation parameters exceeds a preset threshold.

[0090] Send the optimized inspection path to the power distribution and transformation intelligent auxiliary system, and the system controls the intelligent inspection robot according to the path instruction to ensure that the robot moves along the optimized path to cover all the equipment that needs to be inspected. During the inspection, the intelligent inspection robot continuously monitors its own mechanical state through built-in vibration sensors such as accelerometers and gyroscopes. These sensors can capture the vibration information of the robot during the inspection task, especially the vibration modes of mechanical components (such as motors and gears).

[0091] Twin simulation refers to the virtual modeling and simulation of the robot and the equipment, and the state of the equipment is simulated through real-time data to predict the health status of the equipment in the future for a period of time. According to the mechanical vibration feedback data, simulate the inspection process of the robot in the virtual model, and predict the possible state changes of the equipment during the inspection process through the action trajectory, vibration amplitude and frequency changes of the robot, including the health status of the equipment, fault evolution and possible performance degradation. The output equipment state evolution simulation parameters during the simulation process include the predicted equipment health status, potential faults, etc. in the simulation.

[0092] Compare the equipment state improvement parameters with the equipment state evolution simulation parameters predicted in the twin simulation. When the error exceeds the preset threshold, it means that the state of the equipment has not changed as expected and further monitoring or intervention is required. At this time, start the online replanning mechanism, and through this mechanism, adjust the inspection path, priority, task allocation, etc. in real time to ensure that the inspection task of the robot continues to be effective and potential problems of the equipment can be discovered and handled in time. After the replanning mechanism is started, a closed-loop control loop is formed. Through real-time data collection, simulation feedback and path optimization, the inspection plan can be dynamically adjusted according to the health status of the equipment and the actual progress of the inspection task to ensure the continuous optimization and efficient execution of the inspection task.

[0093] In summary, the intelligent robot inspection system in the power distribution and transformation intelligent auxiliary system provided by the embodiments of the present application has the following technical effects:

[0094] By utilizing three-dimensional environmental data to plan the global path, the intelligent inspection robot can conduct intelligent inspections in power distribution and transformation equipment according to the spatial distribution of equipment and the environmental layout. Through the generation of a more accurate optimal inspection path, the inspection coverage rate and efficiency are improved. At the same time, through global path planning, the path repetition or empty running phenomenon of the robot during the inspection process is reduced, the energy consumption is lowered, and the inspection task can be efficiently completed. This process ensures that the robot can sense and adjust the path planning in real time through the application of intelligent perception and control devices, ensuring the optimal execution of the inspection task; during the inspection process, by collecting and comprehensively analyzing multi-modal data in real time, the evaluation of the equipment status is made more comprehensive and accurate, and potential faults of the equipment can be identified more effectively. Through the automatic identification of fault characteristics, a risk map is generated to timely detect and handle potential problems of the equipment, avoiding the limitations and incompleteness of manual inspections; the dynamic risk map provides a hierarchical warning mechanism, which can issue corresponding warnings for faults of different severity levels, helping operators to make corresponding responses at different fault levels to ensure timely and accurate intervention. The construction of the dynamic risk map can improve the intelligent level of equipment management and significantly enhance the timeliness and accuracy of fault warnings; based on the hierarchical warning signal to correct the initial inspection path, not only can the priority of the inspection task be adjusted according to the actual health status of the equipment, but also the inspection path can be automatically adjusted according to the warning level, so as to ensure that high-risk equipment receives priority attention and optimize the inspection efficiency. The generation of the simulated maintenance feedback enables the robot to be corrected according to the equipment maintenance suggestions and simulate the effects of different maintenance measures. This process optimizes the equipment maintenance strategy, predicts and adjusts the inspection plan in advance, and improves the management efficiency of the equipment; finally, according to the optimized inspection path generated by the maintenance feedback, the intelligent inspection robot conducts precise inspection control on the equipment. The optimized inspection path better conforms to the equipment status and maintenance requirements, helps to improve the accuracy and efficiency of the inspection, and at the same time reduces the invalid path of the robot operation; generally speaking, the application of intelligent perception and control devices ensures that the optimized path can be dynamically adjusted according to the real-time status of the equipment, making the inspection task more efficient and accurate, and greatly enhancing the adaptive ability of the robot's behavior during the inspection process.

[0095] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The intelligent robot inspection system in the distribution and transformation intelligent auxiliary system is characterized by: The system comprises: The real-time operation acquisition module is used to plan the global path of the intelligent inspection robot based on the three-dimensional environmental data, generate the initial inspection path, and perform real-time operation acquisition of the distribution and transformation equipment in the distribution and transformation intelligent auxiliary system to obtain a multi-modal detection data set; A fault feature recognition module is used to perform fault feature recognition on the multimodal detection data set, generate an abnormal feature set, and construct a dynamic risk map according to the abnormal feature set, wherein the dynamic risk map includes a graded alarm signal and equipment maintenance suggestions; A simulated maintenance feedback module, used to correct the initial inspection path based on the graded alarm signal, perform simulated maintenance feedback according to the inspection correction path combined with the equipment maintenance suggestion, and generate a maintenance feedback result; The inspection control module is used to generate an optimized inspection path according to the maintenance feedback result and send it to the distribution and transformation intelligent auxiliary system to control the intelligent inspection robot for inspection.

2. The intelligent robot inspection system in the intelligent auxiliary system for power distribution and transformation as claimed in claim 1, characterized in that: The real-time operation acquisition module includes: A three-dimensional space coordinate extraction unit is used to generate three-dimensional point cloud parameters by scanning the power distribution station through a laser radar, and extract the three-dimensional space coordinate set of the power distribution and transformation equipment; A thermal feature annotation unit is used to fuse the thermal distribution data acquired by the infrared thermal imaging sensor with the three-dimensional space coordinate set, perform thermal feature annotation on the three-dimensional point cloud parameters, and generate a thermal-space fusion array; A path planning unit is used to perform multi-objective path planning on the distribution and transformation intelligent auxiliary system based on the thermal-spatial fusion array to generate the initial inspection path.

3. The intelligent robot inspection system in the intelligent auxiliary system for power distribution and transformation as claimed in claim 2, characterized in that: The thermal feature annotation unit comprises: A feature point matching channel, used to perform feature point matching based on the thermal distribution data and the three-dimensional space coordinate set to construct a coordinate system conversion matrix; A visualization layer generation channel is used to retrieve a dynamic pseudo-color mapping rule for setting a rated temperature threshold value of the power distribution and transformation equipment in the power distribution and transformation intelligent auxiliary system, and generate a temperature gradient visualization layer according to the dynamic pseudo-color mapping rule; An anomaly analysis channel, used to traverse the temperature gradient visualization layer and combine the thermal distribution data to perform anomaly analysis, determine the temperature anomaly area, and extract continuous thermal anomaly boundary contours based on the temperature anomaly area; A feature annotation channel is used to delineate a thermal anomaly region according to the continuous thermal anomaly boundary contour, perform thermal feature annotation on the three-dimensional point cloud parameters based on the thermal anomaly region, and generate the thermal-spatial fusion array.

4. The intelligent robot inspection system in the intelligent auxiliary system for distribution and transformation as claimed in claim 1, characterized in that: The fault feature identification module comprises: A feature extraction unit, used to perform cross-domain feature extraction on the multimodal detection data set to generate multiple feature vectors; A weighted fusion unit, configured to perform weighted fusion based on the plurality of feature vectors using an attention mechanism to construct a multi-dimensional abnormal feature set; A similarity analysis unit is used to retrieve a historical defect database, migrate the multi-dimensional abnormal feature set to the historical defect database for similarity analysis, identify faults based on data similarity, and determine potential fault modes; A dynamic risk assessment unit, used to perform dynamic risk assessment based on the potential failure mode and construct a graded alarm signal; An equipment maintenance analysis unit, used to perform equipment maintenance analysis based on the potential failure mode and formulate equipment maintenance recommendations; The correlation and integration unit is used to correlate and integrate the graded alarm signal with the equipment maintenance suggestion to construct a dynamic risk map.

5. The intelligent robot inspection system in the intelligent auxiliary system for power distribution and transformation as claimed in claim 4, characterized in that: The feature extraction unit comprises: An analysis channel, used for performing analysis based on the multimodal detection data set to obtain a visible light image data set, an ultrasonic signal data set, and a partial discharge data set; A dual-channel convolution analysis channel, used for performing dual-channel convolution analysis based on the visible light image data set to obtain visible light image features; A wavelet packet energy entropy analysis channel, used to perform wavelet packet energy entropy analysis based on the ultrasonic signal data set to obtain ultrasonic signal characteristics; A pulse clustering channel, used for performing pulse cluster clustering based on the partial discharge data set to extract partial discharge features; A feature vector construction channel is used to analyze the time-series evolution trend of the visible light image feature, the ultrasonic signal feature, and the partial discharge feature, and to construct a visible light image feature vector, an ultrasonic signal feature vector, and a partial discharge feature vector; A data integration channel is used to integrate the visible light image feature vector, the ultrasonic signal feature vector, and the partial discharge feature vector to obtain the multiple feature vectors.

6. The intelligent robot inspection system in the intelligent auxiliary system for distribution and transformation as claimed in claim 1, characterized in that: The simulated maintenance feedback module includes: A real-time analysis unit, configured to perform real-time analysis based on the graded alarm signal, determine a target alarm device, and extract a three-dimensional alarm coordinate and an alarm risk level based on the target alarm device; A correction analysis unit, configured to construct an alarm topology network according to the three-dimensional alarm coordinates and the alarm risk level, perform correction analysis based on the alarm topology network, and generate a set of path correction factors; A correction unit, configured to correct the initial inspection path based on the path correction factor set to generate an initial inspection correction path; The correction path determination unit is used to perform collision detection according to the initial inspection correction path, determine the completion progress index of the intelligent robot inspection task according to the detection result, and determine the inspection correction path according to the completion progress index.

7. The intelligent robot inspection system in the intelligent auxiliary system for power distribution and transformation as claimed in claim 6, characterized in that: The correction analysis unit comprises: An edge weight parameter acquisition channel is used to use the three-dimensional alarm coordinates as topological nodes, calculate node space distance data according to the topological nodes, and use the alarm risk level combined with the node space distance data as edge weight parameters; An alarm topology network construction channel, used to construct an alarm topology network based on the topology nodes and the edge weight parameters; A calculation channel, used for calculating the initial inspection path according to the alarm topology network to obtain a modified cost matrix, wherein the modified cost matrix includes a path modification weight; A descending order channel is used to perform multi-objective evaluation based on the path correction weights, perform descending order according to the evaluation results, and generate a path correction priority ranking list; The simulation channel is used to determine the path correction factor set according to the path correction priority list combined with the motion trajectory simulation of the intelligent robot.

8. The intelligent robot inspection system in the intelligent auxiliary system for power distribution and transformation as claimed in claim 6, characterized in that: The correction path determination unit comprises: A collision detection channel, used to implement multi-modal collision detection according to the initial inspection correction path and obtain obstacle spatial distribution data in real time; A progress evaluation channel is used to introduce the intelligent robot inspection task for progress evaluation, obtain a completion progress index of the intelligent robot inspection task, and calculate a remaining path risk evaluation value based on the completion progress index; A local correction channel, used to perform local correction according to the obstacle spatial distribution data combined with the remaining path risk assessment value to generate an obstacle avoidance path candidate set; An update channel is used to update the initial inspection and correction path according to the obstacle avoidance path candidate set to determine the inspection and correction path.

9. The intelligent robot inspection system in the intelligent auxiliary system for power distribution and transformation as claimed in claim 1, characterized in that: The inspection control module comprises: A state change analysis unit, configured to perform state change analysis on the power distribution and transformation equipment of the power distribution and transformation intelligent auxiliary system based on the maintenance feedback result, and determine equipment state improvement parameters; A secondary detection analysis unit, used to traverse the distribution and transformation equipment of the distribution and transformation intelligent auxiliary system according to the equipment status improvement parameter to perform secondary detection analysis and determine the secondary detection priority coefficient; The health prediction unit is used to retrieve historical cruise data and combine it with the secondary detection priority coefficient to make health predictions on the distribution and transformation equipment of the distribution and transformation intelligent auxiliary system, and generate equipment health prediction scores; A multi-objective evolution unit, used for performing multi-objective evolution on the inspection correction path according to the equipment health prediction score to generate the optimized inspection path; The inspection control unit is used to send the optimized inspection path to the distribution and transformation intelligent auxiliary system to perform inspection control on the intelligent inspection robot to form a closed-loop control circuit.

10. The intelligent robot inspection system in the intelligent auxiliary system for power distribution and transformation as claimed in claim 9, characterized in that: The inspection control unit comprises: The inspection control channel is used to send the optimized inspection path to the distribution and transformation intelligent auxiliary system to control the intelligent inspection robot and collect mechanical vibration feedback data of the intelligent inspection robot; A twin simulation channel, used to perform twin simulation on the intelligent inspection robot according to the mechanical vibration feedback data to obtain simulation parameters of equipment state evolution; The closed-loop control channel is used to start the online re-planning mechanism to perform closed-loop control and construct the closed-loop control loop when the error between the device state improvement parameter and the device state evolution simulation parameter exceeds a preset threshold.

Images