Power equipment inspection method and system based on fault prediction

Abstract

The invention discloses a power equipment inspection method and system based on fault prediction, and relates to the technical field of power system monitoring and maintaining.The method comprises the steps that an intelligent sensor marked with current power equipment is arranged on the power equipment, and a corresponding sensor recognition device is set up on an inspection unmanned aerial vehicle; a primary inspection route is formulated by considering expert opinions, and key parameters are recognized; constructing a multi-dimensional-graph neural model to perform fault deviation prediction on the polled data, identifying a propagation path, and performing self-optimization; and formulating a secondary inspection route according to an optimized result, and performing secondary inspection again. According to the method, intelligent equipment monitoring, data analysis and dynamic adjustment strategies are combined, the fault prediction capacity of the power equipment is remarkably improved by implementing the efficient inspection method, the fault occurrence probability can be reduced, the resource configuration and use efficiency can be optimized, and finally safe and stable power supply and management are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system monitoring and maintenance, and particularly relates to a power equipment inspection method and system based on fault prediction. Background Art

[0002] In recent years, with the continuous expansion of the scale of power equipment, the problems of its reliability and safety have become increasingly prominent. Especially in the power industry, the normal operation of power equipment is crucial for the stability and economy of power supply. Therefore, the inspection and maintenance of power equipment have become an important link to ensure its efficient operation. Traditional inspection methods mainly rely on manual inspections to evaluate the status and health level of equipment through manual checks. However, this method not only has deficiencies in inspection efficiency but also is subject to the limitations of the professional level, experience, and subjective judgment of inspection personnel, resulting in defects in the early identification and warning of equipment failures. These traditional methods are often slow to respond to abnormal states of equipment and are difficult to detect potential fault hazards in a timely manner, which may lead to sudden equipment failures and outages, causing significant losses to the power system.

[0003] In response to the above problems, intelligent inspection technology has gradually become a development trend in power equipment maintenance. Currently, inspection schemes based on drones and intelligent sensors are gradually emerging. By using sensors carried by drones to regularly inspect power equipment, the inspection efficiency and the accuracy of data collection can be effectively improved. However, even with advanced intelligent inspection equipment, there are still some deficiencies in the existing technology. First, existing inspection schemes mainly focus on the immediate collection of data and often lack systematicness in the analysis of the collected data and fault prediction. Second, many intelligent inspection systems fail to fully utilize historical equipment data and operating parameters for fault trend prediction, resulting in the inability to take effective preventive measures in a timely manner when equipment failures occur. In addition, the comprehensive balance of cost and time during the inspection process is often overlooked, which may lead to low implementation efficiency of the inspection plan. Therefore, how to establish a comprehensive power equipment inspection method based on fault prediction has become the key to improving the stability and safety of power equipment.

[0004] The present invention proposes a power equipment inspection method based on fault prediction. The implementation of the method of the present invention not only improves the accuracy of early fault prediction but also helps to enhance the scientificity and rationality of inspection work, providing a new solution for the safe operation of power equipment. Summary of the Invention

[0005] In view of the above existing problems, the present invention predicts potential equipment failures by analyzing the equipment's historical records and current status, combined with the data obtained by sensors, so as to discover problems in a timely manner and reduce the risk of equipment failures; by constructing inspection routes and determining priorities, it realizes the optimal allocation of inspection resources to improve the efficiency and accuracy of inspections; by analyzing inspection data through a multi-dimensional graph neural model, extracting features, judging the health status of equipment, and making corresponding optimizations and adjustments, it improves the accuracy of fault identification and response; through early fault identification and optimized inspection strategies, it reduces the maintenance cost of equipment, extends the service life of equipment, and improves the overall stability and reliability of the power system.

[0006] To solve the above technical problems, a power equipment inspection method based on fault prediction is proposed, including,

[0007] Arrange intelligent sensors marked with the current power equipment on the power equipment, and build the corresponding sensor identification device on the inspection drone; consider expert opinions to formulate the initial inspection route and identify key parameters; construct a multi-dimensional graph neural model to predict fault deviations of the inspected data, identify the propagation path, and perform self-optimization; formulate the secondary inspection route according to the optimized results and conduct the secondary inspection again.

[0008] As a preferred solution of the power equipment inspection method based on fault prediction described in the present invention, wherein: the formulation of the initial inspection route includes, when formulating the initial inspection route, considering expert opinions, including the historical fault records, current operating status, and environmental factors of the equipment, conducting basic modeling for the initial inspection, and determining the inspection cost:

[0009]

[0010] Where C is the comprehensive inspection cost, d i is the distance from the i-th equipment to the next equipment, ρ i is the important function factor of the i-th equipment, K is the control coefficient for adjusting the inspection strategy, n is the total number of equipment traversed in the inspection, and i is the variable index;

[0011] The initial inspection route is obtained as R = min(C + γ·t 1 ), where R is the initial inspection path, t 1 is the estimated initial inspection time, and γ is the weight factor for balancing cost and time; the drone equipped with the sensor identification device conducts the initial inspection according to the initial inspection route and collects key parameters in real time.

[0012] As a preferred solution of a power equipment inspection method based on fault prediction according to the present invention, wherein: the identification of key parameters includes that when inspecting according to the inspection route, the key parameters collected include electrical parameters, temperature parameters, vibration parameters, and unconventional parameters;

[0013] The electrical parameters include voltage change, current change, and harmonic distortion;

[0014] The temperature parameters include the surface temperature of the equipment and the temperature of the insulating material;

[0015] The vibration parameters include vibration spectrum analysis and mechanical wear index;

[0016] The unconventional parameters include gas leakage detection value and surface crack monitoring;

[0017] Taking the obtained data as new data input, comparing the mean and variance of the new data with the historical data, and making a preliminary evaluation of whether there is a collection deviation. When the newly input data is outside the mean range of μ±2σ, it is considered that there is a collection deviation, and data smoothing processing is performed to convert it to within the mean range for further output.

[0018] As a preferred solution of a power equipment inspection method based on fault prediction according to the present invention, wherein: the multi-dimensional graph neural model includes that in the process of using the multi-dimensional graph neural model to predict the fault deviation of the inspection data, feature extraction is performed on the inspected data to form a node feature matrix H and define an edge weight matrix W. After capturing the spatial feature H' in the data through an enhanced graph convolution mechanism, the fault deviation is predicted to obtain a fault deviation prediction value P fault , comparing the deviation degree according to the result of the deviation prediction with the set threshold:

[0019]

[0020] Among them, Δ represents the change degree of the prediction deviation relative to the threshold, and the fault state is judged; TH Δ represents the set critical threshold for the occurrence of a fault, and ∈ represents a small constant to prevent the denominator from being zero;

[0021] When Δ>TH Δs , it means that the power equipment has a serious deviation, and the propagation path is further identified and improved;

[0022] When TH Δm <Δ≤TH Δs , it means that the power equipment has a slight deviation, and self-optimization is performed;

[0023] When Δ≤TH Δm , it means that the power equipment has no deviation, and it is considered that the fault has not occurred and the current state is maintained;

[0024] Among them, TH Δs and TH Δm are the thresholds for severe and mild biases respectively.

[0025] As a preferred solution of a power equipment inspection method based on fault prediction according to the present invention, wherein: the identification of the propagation path includes, for the case of severe bias, by analyzing the adjacency matrix to identify the fault propagation path:

[0026]

[0027] Among them, V spread represents the feature representation of the impact of fault propagation through the network, represents the adjacency matrix between nodes, r is the number of propagation steps, indicating the number of graph convolutional layers performed through the adjacency matrix ; β is the adjustment weight of the impact of the propagation feature in the final feature;

[0028] After identifying the propagation path, an improvement strategy is adopted for response processing:

[0029] H revised = H' - δ·V spread

[0030] Among them, H revised represents the feature after being corrected by fault propagation, and δ is the response adjustment coefficient for controlling the propagation impact.

[0031] As a preferred solution of a power equipment inspection method based on fault prediction according to the present invention, wherein: the self-optimization includes, for the case of mild bias, applying a feedback adjustment mechanism for optimization:

[0032] H adjusted = H + θ·ΔH

[0033] Among them, H adjusted is the node feature matrix after feedback adjustment, θ is the adjustment coefficient, and ΔH is the adjustment value of the mild deviation.

[0034] As a preferred solution of a power equipment inspection method based on fault prediction according to the present invention, wherein: the secondary inspection includes, according to the equipment bias situation in the fault prediction, if the severe bias fault probability of the current equipment exceeds 5%, then mark the current biased equipment as a high-priority inspection object, when all high-priority objects are subject to secondary inspection, and mark the current inspection area as a high-priority area until the fault probability of the high-priority inspection objects in the high-priority area is lower than 2%;

[0035] If the failure probability of the current device with a serious deviation is within the range of 2% to 5%, the current device with a deviation is marked as an inspection object with medium priority, and the current inspection area is marked as an area with medium priority. During the second inspection, a partial random inspection of some areas with medium priority is carried out within a defined scope;

[0036] If the failure probability of the current device with a serious deviation is less than 2%, the current device with a deviation is marked as an inspection object with low priority. Then, during the second inspection, a random inspection is carried out on all inspection objects with low priority;

[0037] During the inspection of the same area, when there are more than 3 devices with potential failures, re-evaluate the inspection frequency of the entire area and raise the inspection level of the current area.

[0038] Another object of the present invention is to provide a power equipment inspection system based on fault prediction. The present invention aims to solve the problem of early warning of possible faults during the operation of power equipment. Through intelligent sensors and drone inspection technologies, combined with a multi-dimensional graph neural network model for in-depth analysis of data, real-time monitoring of key parameters and prediction of fault deviations are realized. The system optimizes the primary and secondary inspection routes, improves the inspection efficiency and accuracy, ensures the safe and stable operation of power equipment, and reduces the economic losses and safety hazards caused by the occurrence of faults.

[0039] As a preferred solution of a power equipment inspection system based on fault prediction according to the present invention, it is characterized in that it includes a primary inspection module, a data acquisition module, a fault prediction module, and a secondary inspection module;

[0040] The primary inspection module formulates the primary inspection route according to expert opinions and equipment status, plans the inspection activities, transfers the primary inspection route and equipment information to the data acquisition module, and provides the primary inspection records, including the key parameters collected, to the fault prediction module;

[0041] The data acquisition module receives the power equipment information and inspection route to be inspected from the primary inspection module, and feeds back the collected key parameter data to the fault prediction module for analysis;

[0042] The fault prediction module receives the new data and preliminary evaluation results from the data acquisition module, and outputs the results of fault prediction to the secondary inspection module, including the deviation status and priority marking of the equipment;

[0043] The secondary inspection module obtains the equipment deviation situation and priority marking from the fault prediction module, formulates the secondary inspection route, and feeds back the results of the secondary inspection to the data acquisition module.

[0044] A computer device includes a memory and a processor. The memory stores a computer program. It is characterized in that when the processor executes the computer program, the steps of the power equipment inspection method based on fault prediction are realized.

[0045] A computer-readable storage medium stores a computer program. It is characterized in that when the computer program is executed by a processor, the steps of the power equipment inspection method based on fault prediction are realized.

[0046] Advantages of the present invention: By arranging intelligent sensors marked with the current power equipment on the power equipment and building an identification device on the inspection unmanned aerial vehicle, accurate monitoring of the real-time state of the equipment can be realized. In this way, the inspection unmanned aerial vehicle can quickly obtain the real-time data of the power equipment, including electrical parameters, temperature parameters, vibration parameters, etc., thereby improving the data collection efficiency. The fast response ability brought by real-time monitoring enables potential faults to be identified as early as possible, thus reducing equipment damage and downtime.

[0047] Considering expert opinions and the historical fault records of the equipment, an initial inspection route is formulated to effectively identify key parameters and implement a targeted maintenance strategy. By quantifying the cost and time of inspection and comprehensively evaluating the optimal inspection route, not only manpower and time costs are saved, but also it is ensured that important equipment is inspected first. Optimize resource allocation, minimize inspection costs and improve the fault discovery rate at the same time.

[0048] A method for predicting fault deviation of inspection data using a multi-dimensional graph neural model can accurately capture the spatial features in the data and judge the fault state of the equipment through feature extraction and enhanced graph convolution mechanism. This step not only realizes the quantitative evaluation of fault deviation, but also can identify the propagation path of faults and perform self-optimization. It improves the accuracy of prediction and the adaptive ability of the system, and can quickly adjust the inspection strategy when a fault occurs to reduce potential losses.

[0049] According to several different biased fault probabilities, a secondary inspection strategy is formulated to ensure that the equipment in high-priority areas is inspected in a timely manner. This strategy of differential management based on the analysis results of data makes the use of resources more efficient. While ensuring the safe operation of important equipment, it reduces the risk of downtime caused by equipment failures and improves the flexibility and efficiency of the overall inspection operation. Description of the Drawings

[0050] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0051] Figure 1 This is the overall flowchart of a power equipment inspection method based on fault prediction provided by an embodiment of the present invention.

[0052] Figure 2 This is the system scheme module diagram of a power equipment inspection system based on fault prediction provided by an embodiment of the present invention. Detailed implementation manners

[0053] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will describe the detailed implementation manners of the present invention with reference to the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0054] In the following description, many specific details are set forth to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0055] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation manner of the present invention. The "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it an embodiment that is mutually exclusive with other embodiments alone or selectively.

[0056] The present invention is described in detail in combination with schematic diagrams. When detailing the embodiments of the present invention, for the convenience of explanation, the cross-sectional views showing the device structure will be enlarged locally out of the general proportion, and the schematic diagrams are only examples and should not limit the protection scope of the present invention herein. In addition, in actual production, three-dimensional spatial dimensions including length, width, and depth should be included.

[0057] Meanwhile, in the description of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "upper, lower, inner, and outer" is based on the orientation or positional relationship shown in the drawings. It is 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. Therefore, it should not be construed as a limitation to the present invention. In addition, the terms "first, second, or third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0058] Unless otherwise clearly defined and limited in the present invention, the terms "installation, connection, and coupling" shall be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can also be a mechanical connection, an electrical connection, or a direct connection, and can also be indirectly connected through an intermediate medium, or can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0059] Example 1, referring to Figure 1 , which is the first embodiment of the present invention. This embodiment provides a power equipment inspection method based on fault prediction, including:

[0060] S1: Arrange intelligent sensors marked with the current power equipment on the power equipment, and build the corresponding sensor identification device on the inspection unmanned aerial vehicle.

[0061] S2: Consider expert opinions to formulate the initial inspection route and identify key parameters.

[0062] Furthermore, when formulating the initial inspection route, consider expert opinions, including the historical fault records, current operating status, and environmental factors of the equipment, conduct basic modeling for the initial inspection, and determine the inspection cost:

[0063]

[0064] Among them, C is the comprehensive inspection cost, d i is the distance from the i-th device to the next device, ρ i is the important function factor of the i-th device, K is the control coefficient for adjusting the inspection strategy, n is the total number of devices traversed in the inspection, and i is the variable index;

[0065] The initial inspection route obtained is R = min(C + γ·t 1 ), where R is the initial inspection path, t 1 is the estimated initial inspection time, and γ is the weight factor for balancing cost and time; the unmanned aerial vehicle equipped with the sensor identification device conducts the initial inspection according to the initial inspection route and collects key parameters in real time.

[0066] It should be noted that the key parameters collected include electrical parameters, temperature parameters, vibration parameters, and unconventional parameters;

[0067] The electrical parameters include voltage change, current change, and harmonic distortion;

[0068] The temperature parameters include the surface temperature of the equipment and the temperature of the insulating material;

[0069] The vibration parameters include vibration spectrum analysis and mechanical wear index;

[0070] The unconventional parameters include gas leakage detection value and surface crack monitoring;

[0071] Taking the obtained data as new data input, comparing the means and variances of the new data and the historical data, and making a preliminary assessment of whether there is a collection deviation. When the newly input data is outside the mean range of μ±2σ, it is considered that there is a collection deviation, and data smoothing processing is performed to convert it to within the mean range for further output.

[0072] S3: Construct a multi-dimensional graph neural model to predict fault deviations for the data after inspection, identify the propagation path, and perform self-optimization.

[0073] It should be noted that in the process of using the multi-dimensional graph neural model to predict fault deviations for the inspection data, feature extraction is performed on the data after inspection to form a node feature matrix H and define the weight matrix W of the edges, and the spatial features in the data are captured through an enhanced graph convolution mechanism:

[0074]

[0075] Among them, H' represents the node feature matrix after graph convolution operation, N(i) is the neighbor set of node i, including all nodes directly connected to node i; i and j are variable indices, H j is the feature vector of node j, H i is the feature vector of node i, W 0 represents the learned weight matrix, which is used to perform a linear transformation on the features of neighbor nodes; b i 、b j represent the degrees of nodes i and j, indicating the number of edges connected to nodes i and j respectively; W 1 represents the weight matrix of the feature linear transformation of the self-node i, and ReLU is the activation function;

[0076] Perform the prediction of fault deviation:

[0077]

[0078] Among them, P faultis the predicted value of the fault deviation, and α represents the adjustment parameter for the predicted result of the final control output. represents the additional information generated by processing the node feature H'; k f represents the weight parameter, which determines the influence of the part increased according to the Sigmoid function in the final prediction; e -H' represents the exponential quantity of the entire H' matrix, enhancing the model's response to different features;

[0079] Compare the degree of deviation between the result of the deviation prediction and the set threshold:

[0080]

[0081] Among them, Δ represents the change degree of the prediction deviation relative to the threshold, and is used to judge the fault state; TH Δ represents the set critical threshold for the occurrence of a fault, and ∈ represents a small constant to prevent the denominator from being zero;

[0082] When Δ > TH Δs it indicates that the power equipment has a serious deviation, and further identify the propagation path and make improvements;

[0083] When TH Δm < Δ ≤ TH Δs it indicates that the power equipment has a slight deviation, and then perform self-optimization;

[0084] When Δ ≤ TH Δm it indicates that the power equipment has no deviation, and it is considered that the fault has not occurred and the current state is maintained;

[0085] Among them, TH Δs and TH Δm are the thresholds for serious and slight deviations respectively.

[0086] Specifically, for the case of serious deviation, identify the fault propagation path by analyzing the adjacency matrix :

[0087]

[0088] Among them, V spread represents the feature representation of the influence of fault propagation through the network, represents the adjacency matrix between nodes, r is the number of propagation steps, indicating the number of graph convolution layers performed through the adjacency matrix ; β is the adjustment weight of the influence of the propagation feature in the final feature;

[0089] After identifying the propagation path, adopt an improvement strategy for response processing:

[0090] H revised = H' - δ·V spread

[0091] Among them, H revised represents the feature after fault propagation correction, and δ is the response adjustment coefficient for controlling the propagation impact.

[0092] In addition, for slightly biased situations, a feedback adjustment mechanism is applied for optimization:

[0093] H adjusted = H + θ·ΔH

[0094] Among them, H adjusted is the node feature matrix after feedback adjustment, θ is the adjustment coefficient, and ΔH is the adjustment value for slight deviation.

[0095] S4: Formulate a secondary inspection route based on the optimized result and conduct a secondary inspection again.

[0096] Furthermore, according to the equipment bias situation in the fault prediction, if the severe bias fault probability of the current equipment exceeds 5%, mark the current biased equipment as a high-priority inspection object. When conducting a secondary inspection on all high-priority objects, mark the current inspection area as a high-priority area until the fault probability of the high-priority inspection objects in the high-priority area is lower than 2%;

[0097] If the severe bias fault probability of the current equipment is within the range of 2% to 5%, mark the current biased equipment as a medium-priority inspection object and mark the current inspection area as a medium-priority area. During the secondary inspection, conduct a range sampling inspection on some medium-priority areas;

[0098] If the severe bias fault probability of the current equipment is lower than 2%, mark the current biased equipment as a low-priority inspection object. Then, during the secondary inspection, conduct a sampling inspection on all low-priority inspection objects;

[0099] During the inspection in the same area, when there are more than 3 devices with potential faults, re-evaluate the inspection frequency of the entire area and upgrade the inspection level of the current area.

[0100] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

[0101] Embodiment 2, the second embodiment of the present invention, which is different from the previous two embodiments in that:

[0102] If the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0103] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a predefined sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or used in combination with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

[0104] More specific examples (non-exhaustive list) of computer-readable media include the following: electrical connection parts with one or more wirings (electronic devices), portable computer disk cartridges (magnetic devices), random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memories), optical fiber devices, and portable compact disc read-only memories (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, then editing, interpreting, or otherwise processing it as necessary, and then storing it in a computer memory.

[0105] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits with logic gate circuits for implementing logical functions on data signals, application specific integrated circuits with suitable combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0106] Embodiment 3, referring to Figure 2 , is the third embodiment of the present invention. This embodiment provides a power equipment inspection system based on fault prediction, including a primary inspection module 10, a data acquisition module 20, a fault prediction module 30, and a secondary inspection module 40;

[0107] The primary inspection module 10 formulates a primary inspection route according to expert opinions and equipment status, plans inspection activities, transmits the primary inspection route and equipment information to the data acquisition module 20, and provides the primary inspection record, including the collected key parameters, to the fault prediction module 30.

[0108] The data acquisition module 20 receives the power equipment information and inspection route to be inspected from the primary inspection module 10, and feeds back the collected key parameter data to the fault prediction module 30 for analysis.

[0109] The fault prediction module 30 receives new data and preliminary evaluation results from the data acquisition module 20, and outputs the fault prediction results, including the deviation status and priority mark of the equipment, to the secondary inspection module 40.

[0110] The secondary inspection module 40 obtains the equipment deviation situation and priority mark from the fault prediction module 30, formulates a secondary inspection route, and feeds back the results of the secondary inspection to the data acquisition module 20.

[0111] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for inspecting electric power equipment based on fault prediction, characterized in that: include, Arrange smart sensors marked with current power equipment on the power equipment, and install corresponding sensor identification devices on the inspection drones; Consider expert opinions to develop the initial inspection route and identify key parameters; Construct a multi-dimensional graph neural model to predict fault deviations in the data after inspection, identify the propagation path, and perform self-optimization; According to the optimized results, a secondary inspection route is formulated and a secondary inspection is carried out again.

2. The method for inspecting electric power equipment based on fault prediction according to claim 1, characterized in that: The formulation of the initial inspection route includes: when formulating the initial inspection route, taking into account expert opinions, including historical fault records of the equipment, current operating status, and environmental factors, conducting basic modeling for the initial inspection, and determining the inspection cost: Among them, C is the comprehensive cost of inspection, d i is the distance from the ith device to the next device, ρ i The importance function factor of the i-th device, K is the control coefficient for adjusting the inspection strategy, n is the total number of devices traversed during the inspection, and i is the variable index; The initial inspection route is R=min(C+γ·t1), where R is the initial inspection path, t1 is the estimated initial inspection time, and γ is the weight factor that balances cost and time. The UAV equipped with a sensor recognition device performs the initial inspection along the initial inspection route and collects key parameters in real time.

3. A method for inspecting electric power equipment based on fault prediction according to claim 2, characterized in that: The identification of key parameters includes, when performing inspection according to the inspection route, collecting key parameters including electrical parameters, temperature parameters, vibration parameters, and unconventional parameters; The electrical parameters include voltage variation, current variation, and harmonic distortion; The temperature parameters include equipment surface temperature and insulation material temperature; The vibration parameters include vibration spectrum analysis and mechanical wear index; The unconventional parameters include gas leakage detection value and surface crack monitoring; The acquired data is used as new data input, and the mean and variance of the new data are compared with those of the historical data to make a preliminary assessment of whether there is a collection bias. When the new input data is outside the mean range μ±2σ, it is considered that there is a collection bias, and the data is smoothed and converted to the mean range for further output.

4. A method for inspecting electric power equipment based on fault prediction as claimed in claim 3, characterized in that: The multidimensional-graph neural model includes, in the process of using the multidimensional-graph neural model to predict the fault deviation of the inspection data, extracting features from the data after the inspection to form a node feature matrix H and a weight matrix W for defining edges, and predicting the fault deviation after capturing the spatial features H' in the data through an enhanced graph convolution mechanism to obtain a fault deviation prediction value P fault , compare the deviation degree based on the deviation prediction result and the set threshold: Among them, Δ represents the change degree of the prediction deviation relative to the threshold value, which determines the fault state; TH Δ It represents the critical threshold of the set fault occurrence, ∈ represents a small constant to prevent the denominator from being zero; When Δ>TH Δs When it is, it indicates that there is a serious bias in the power equipment, and further identification of the propagation path and improvement are made; When TH Δm <Δ≤TH Δs When , it means that the power equipment has a slight bias, and self-optimization is performed; When Δ≤TH Δm When , it means that the power equipment is not biased, and it is considered that no fault has occurred and the current state is maintained; Among them, TH Δs and TH Δm are the thresholds for severe and mild bias, respectively.

5. The method for inspecting electric power equipment based on fault prediction according to claim 4, characterized in that: The identification propagation pathways include, for severe biased cases, analyzing the adjacency matrix Identify fault propagation paths: Among them, V spread A feature representation that represents the effect of fault propagation through the network, Represents the adjacency matrix between nodes, r is the number of propagation steps, and represents the adjacency matrix The number of graph convolution layers performed; β is the adjustment weight of the influence of propagated features in the final features; After identifying the transmission path, adopt improved strategies to respond: H revised =H'-δ·V spread Among them, H revised represents the characteristic after fault propagation correction, and δ is the response adjustment coefficient to control the influence of propagation.

6. A method for inspecting electric power equipment based on fault prediction according to claim 5, characterized in that: The self-optimization includes applying a feedback adjustment mechanism to optimize for slightly biased situations: H adjusted =H+θ·ΔH Among them, H adjusted is the node feature matrix after feedback adjustment, θ is the adjustment coefficient, and ΔH is the adjustment value of slight deviation.

7. A method for inspecting electric power equipment based on fault prediction according to claim 6, characterized in that: The secondary inspection includes, according to the equipment bias in the fault prediction, if the fault probability of the severe bias of the current equipment exceeds 5%, marking the current biased equipment as a high-priority inspection object, performing secondary inspection on all high-priority equipment, and marking the current inspection area as a high-priority area, until the fault probability of the high-priority inspection objects in the high-priority area is less than 2%; If the probability of serious biased failure of the current equipment is within the range of 2% to 5%, the current biased equipment will be marked as a medium priority inspection object, and the current inspection area will be marked as a medium priority area. During the second inspection, a range of medium priority areas will be randomly inspected; If the probability of serious bias failure of the current device is less than 2%, the current biased device will be marked as a low-priority inspection object, and all low-priority inspection objects will be randomly inspected during the second inspection; During the inspection of the same area, if more than three devices have potential faults, the inspection frequency of the entire area will be re-evaluated and the inspection level of the current area will be increased.

8. A system using a power equipment inspection method based on fault prediction as claimed in any one of claims 1 to 7, characterized in that: It includes the initial inspection module, data collection module, fault prediction module and secondary inspection module; The initial inspection module formulates the initial inspection route according to the expert opinions and the equipment status, and plans the inspection activities, transmits the initial inspection route and equipment information to the data acquisition module, and provides the initial inspection record to the fault prediction module, including the key parameters collected; The data acquisition module receives the information of the power equipment to be inspected and the inspection route from the initial inspection module, and feeds back the collected key parameter data to the fault prediction module for analysis; The fault prediction module receives new data and preliminary evaluation results from the data acquisition module, and outputs the fault prediction results, including the bias status and priority mark of the equipment, to the secondary inspection module; The secondary inspection module obtains the equipment bias and priority mark from the fault prediction module, formulates a secondary inspection route, and feeds back the result of the secondary inspection to the data acquisition module.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of a power equipment inspection method based on fault prediction described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a method for inspecting electric power equipment based on fault prediction described in any one of claims 1 to 7 are implemented.

Images