Intelligent detection method and device for power distribution cabinet fault based on deep learning

Abstract

The present invention discloses a method and device for intelligent detection of distribution cabinet faults based on deep learning, and relates to the technical field of distribution cabinet fault detection. The method includes: identifying M key lines according to the distribution cabinet distribution line information; obtaining each connector and connection mode of each key line; configuring the parameters of the connection monitoring module according to the connection mode, performing connection monitoring on each connector, obtaining N corresponding connection monitoring data sets on each key line, inputting them into the line fault probability model, obtaining M connection fault probabilities, and generating fault detection results, and sending them to the distribution cabinet fault management center. The present invention solves the technical problem that the existing distribution cabinet fault detection is mostly aimed at electrical equipment and components, and ignores the potential fault hazards of line connectors, and achieves the technical effect of timely discovering potential fault hazards and improving the stability and reliability of the power system by real-time monitoring of the connector status on the key lines.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution cabinet fault detection, and in particular to a power distribution cabinet fault intelligent detection method and device based on deep learning. Background Art

[0002] In industrial production, the stability and safety of power supply are crucial to the smooth progress of the production process. As a key link in power supply, the failure of the distribution cabinet may lead to production interruption, equipment damage and even casualties. Therefore, fault detection of the distribution cabinet is an important measure to ensure the safety of industrial production. Existing distribution cabinet fault detection mainly focuses on fault detection of electrical equipment and components, ignoring the detection of line connector performance, which has potential fault hazards caused by connectors. Summary of the invention

[0003] The present application provides a method and device for intelligent detection of distribution cabinet faults based on deep learning, which is used to solve the technical problem that the existing distribution cabinet fault detection mainly targets electrical equipment and components, etc., but ignores the potential fault hazards of line connectors.

[0004] The first aspect of the present application provides a deep learning-based intelligent detection method for distribution cabinet faults, the method comprising: interacting with a distribution cabinet fault management center to obtain distribution cabinet distribution line information; analyzing the distribution cabinet distribution line information to identify M key lines; obtaining each connector of each of the M key lines, and identifying the connection mode of each connector; configuring the parameters of a connection monitoring module according to the connection mode, performing connection monitoring on each connector on each key line according to the connection monitoring module, and outputting N connection monitoring data sets corresponding to each key line, wherein N is the number of connectors; inputting each connection monitoring data set into a line fault probability model, wherein the line fault probability model is obtained by fitting a random fault probability layer and a concentrated fault probability layer; according to the line fault probability model, outputting M connection fault probabilities corresponding to the M key lines; generating a fault detection result according to the M connection failure probabilities, and sending the fault detection result to the distribution cabinet fault management center.

[0005] A second aspect of the present application provides a deep learning-based intelligent detection device for distribution cabinet faults, the device comprising: a distribution line information acquisition module, the distribution line information acquisition module is used to interact with a distribution cabinet fault management center to obtain distribution cabinet distribution line information;

[0006] A key line identification module, the key line identification module is used to analyze the distribution line information of the distribution cabinet and identify M key lines; a connector connection mode identification module, the connector connection mode identification module is used to obtain each connector of each key line in the M key lines and identify the connection mode of each connector; a connection monitoring data acquisition module, the connection monitoring data acquisition module is used to configure the parameters of the connection monitoring module according to the connection mode, perform connection monitoring on each connector on each key line according to the connection monitoring module, and output N connection monitoring data sets corresponding to each key line, wherein N is the number of connectors; a line fault analysis module, the line fault analysis module is used to input each connection monitoring data set into a line fault probability model, wherein the line fault probability model is obtained by fitting a random fault probability layer and a concentrated fault probability layer; a connection fault probability output module, the connection fault probability output module is used to output M connection fault probabilities corresponding to the M key lines according to the line fault probability model; a distribution cabinet fault management module, the distribution cabinet fault management module is used to generate a fault detection result according to the M connection fault probabilities, and send the fault detection result to the distribution cabinet fault management center.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] The deep learning-based intelligent detection method for distribution cabinet faults provided in the present application relates to the technical field of distribution cabinet fault detection. Through the distribution line information of the distribution cabinet, M key lines are identified, and the connectors and connection methods of each key line are obtained. The connection of each connector is monitored to obtain the corresponding N connection monitoring data sets, which are input into the line fault probability model to obtain M connection failure probabilities and generate fault detection results. The technical problem that the existing distribution cabinet fault detection mostly targets electrical equipment and components and ignores the potential fault hazards of line connectors is solved, and the technical effect of timely discovering potential fault hazards and improving the stability and reliability of the power system is achieved by real-time monitoring of the connector status on the key lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic diagram of a flow chart of a method for intelligent detection of power distribution cabinet faults based on deep learning provided in an embodiment of the present application;

[0011] Figure 2 A schematic diagram of a process for identifying M key circuits in a power distribution cabinet fault intelligent detection method based on deep learning provided in an embodiment of the present application;

[0012] Figure 3 A schematic diagram of the structure of a power distribution cabinet fault intelligent detection device based on deep learning provided in an embodiment of the present application.

[0013] Explanation of the reference numerals: distribution line information acquisition module 11, key line identification module 12, connector connection mode identification module 13, connection monitoring data acquisition module 14, line fault analysis module 15, connection fault probability output module 16, distribution cabinet fault management module 17. DETAILED DESCRIPTION

[0014] The present application provides a deep learning-based intelligent detection method for distribution cabinet faults, which is used to solve the technical problem that the existing distribution cabinet fault detection mainly targets electrical equipment and components, but ignores the potential fault hazards of line connectors.

[0015] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0016] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices.

[0017] Embodiment 1

[0018] like Figure 1 As shown, the present application provides a distribution cabinet fault intelligent detection method based on deep learning, the method comprising:

[0019] P10: Interactive distribution cabinet fault management center, obtains distribution cabinet distribution line information;

[0020] Optionally, a network communication protocol is used to establish a connection with a distribution cabinet fault management center, and the distribution cabinet fault management center is interacted with through network API, database query, etc. to extract the distribution line information of the target distribution cabinet, wherein the distribution cabinet fault management center is a centralized fault management system of the target distribution cabinet, responsible for collecting, storing and distributing fault information related to the distribution cabinet, and performing fault management according to the fault detection results. In addition, the distribution cabinet distribution line information may include the line topology, electrical parameters (such as voltage, current, power, etc.), equipment connection relationship, line status (such as normal, faulty, under maintenance), etc., which can be used as a basic reference for subsequent key line identification, connector connection method identification, and fault detection.

[0021] P20: Analyze the distribution line information of the power distribution cabinet and identify M key lines;

[0022] Further, such as Figure 2 As shown, step P20 of the embodiment of the present application also includes:

[0023] P21: Interactive distribution cabinet fault management center, obtains line function type;

[0024] P22: Clustering the distribution line information of the power distribution cabinet according to the line function type, and outputting the line clustering result, wherein each line clustering result corresponds to a line function type;

[0025] P23: Perform key line analysis on each line clustering result and screen M key lines.

[0026] It should be understood that by analyzing the distribution line information of the power distribution cabinet, M key lines are identified from the complex distribution line information of the power distribution cabinet, that is, lines that are crucial to the stability and safety of power system operation.

[0027] Specifically, information is exchanged with the distribution cabinet fault management center again to obtain the functional type of each line, such as power supply line, backup line, special-purpose line, etc. Further, a clustering algorithm suitable for the current data characteristics and requirements is selected, such as K-means, hierarchical clustering, etc., and the line functional type is used as the clustering basis to cluster the distribution line information of the distribution cabinet, and the lines of the same specific functional type are clustered into one category, that is, the lines are grouped according to the similarity of the functional type, so that the lines in the same group have a high functional similarity, and the line clustering results are obtained, and each line clustering result corresponds to a line functional type.

[0028] Furthermore, a key line analysis is performed on each line clustering result, that is, key lines are extracted for each line type. By evaluating the importance of each line in the power system, line load, and fault impact range, M key lines are selected as key monitoring objects, providing an important basis for subsequent monitoring and fault detection.

[0029] Furthermore, step P23 of the embodiment of the present application also includes:

[0030] P23-1: Perform criticality assessment on the corresponding line in each line clustering result and output the criticality assessment result, where the criticality assessment is to assess the impact of line failure on system safety;

[0031] P23-2: M key paths are selected according to the criticality evaluation result, wherein the M key paths include a first key path with the highest criticality in each path clustering result, and key paths with a criticality greater than or equal to a preset criticality in all path clustering results.

[0032] In a possible embodiment of the present application, a criticality assessment is performed for each line in each line clustering result, that is, the impact of a line failure on system safety is assessed. The historical fault operation and maintenance data can be collected, and a fault impact assessment model can be trained in combination with machine learning principles. The fault impact assessment model can be used to assess the impact of each line on system safety in combination with the electrical parameters of the line (such as voltage level, current capacity), connected equipment, and position in the system, and obtain a criticality assessment result, which includes the impact of each line in each line clustering result on the safety of the power system.

[0033] Furthermore, according to the criticality evaluation results, critical lines are screened. First, the line with the highest criticality in each line clustering result is extracted as the first critical line in the line clustering result. Second, in each line clustering result, the lines with a criticality evaluation result greater than or equal to a preset criticality threshold are selected, and together with the first critical line in each line clustering result, they form M critical lines. The preset criticality threshold can be set by professionals according to actual monitoring needs.

[0034] P30: Acquire each connector of each key line in the M key lines, and identify the connection mode of each connector;

[0035] Specifically, through the interactive distribution cabinet fault management center, equipment database, etc., the circuit diagram, equipment layout diagram or related database information of the distribution cabinet is accessed to determine the specific location of the connector on each of the M key circuits. The connector is a device used to connect different components or equipment in the circuit, such as cable connectors, sockets, switches, etc. Furthermore, for each connector, the connection method of each connector is identified through direct observation, tool detection, reference to technical documents, or image recognition using deep learning technology. The connection method includes but is not limited to screw connection, welding, plug-in, crimping, etc., which may vary depending on the type, purpose and manufacturer of the connector. By identifying the connection method of each connector, intelligent detection of the contact performance of the connector contact surface and fault identification can be performed for the distribution cabinet connector failure.

[0036] P40: configuring the parameters of the connection monitoring module according to the connection mode, performing connection monitoring on each connector on each key line according to the connection monitoring module, and outputting N connection monitoring data sets corresponding to each key line, where N is the number of connectors;

[0037] It should be understood that corresponding connection monitoring modules are set for different connection modes, and because different connection modes have different electrical characteristics and failure modes, it is necessary to configure corresponding parameters for each connection monitoring module according to different connection modes, such as monitoring frequency, threshold, data sampling rate, etc., to adjust the accuracy and sensitivity of each connection monitoring module. In addition, the connection monitoring module is used to perform connection monitoring on each connector on each key line to obtain N connection monitoring data sets corresponding to the N connectors on each key line. Among them, each connection monitoring data set contains monitoring data related to a specific connector, such as timestamp, collection value, status identification, etc., which can be used as input data for subsequent fault analysis and diagnosis.

[0038] P50: inputting each connection monitoring data set into a line fault probability model, wherein the line fault probability model is obtained by fitting a random fault probability layer and a concentrated fault probability layer;

[0039] Furthermore, step P50 of the embodiment of the present application also includes:

[0040] Fit the random fault probability layer and the concentrated fault probability layer to obtain the line fault probability model:

[0041] ;

[0042] in, , characterizes the probability density function of the exponential distribution corresponding to the random failure probability layer, Characterize the probability density function of the normal distribution corresponding to the concentrated fault probability layer, is the line fault probability model, Connect monitoring datasets;

[0043] Characterize the connection monitoring data corresponding to the i-th connector, Characterize the rate parameter of the failure of the i-th connector, is a natural constant, The weights corresponding to the fitting random failure probability layer;

[0044] is the mean failure probability of the i-th connector based on normal distribution, The discrete degree of failure of the i-th connector based on normal distribution, To fit the weights corresponding to the concentrated fault probability layer, is the number of connectors on each critical line.

[0045] It should be understood that each connection monitoring data set is input into the line fault probability model for fault analysis and fault probability assessment, and the line fault probability model is obtained by fitting the random fault probability layer and the concentrated fault probability layer, wherein the random fault probability layer uses the probability density function of the exponential distribution to fit the probability of random faults. Exponential distribution is often used to describe the time interval between events, and when the probability of an event is independent of the time that has passed, the exponential distribution is a suitable choice. In this scenario, the exponential distribution is used to simulate random, independent fault events. The concentrated fault probability layer uses the probability density function of the normal distribution to fit the probability of concentrated faults. The normal distribution is often used to describe random variables that are affected by multiple factors and have a certain concentration trend. In this scenario, the normal distribution may be used to simulate faults caused by some common factors (such as environmental factors, design defects, etc.).

[0046] The line fault probability model is:

[0047] ;

[0048] in, , characterizes the probability density function of the exponential distribution corresponding to the random failure probability layer, Characterize the probability density function of the normal distribution corresponding to the concentrated fault probability layer, is the line fault probability model, Connect monitoring datasets; Characterize the connection monitoring data corresponding to the i-th connector, Characterize the rate parameter of the failure of the i-th connector, is a natural constant, The weights corresponding to the fitting random failure probability layer; is the mean failure probability of the i-th connector based on normal distribution, The discrete degree of failure of the i-th connector based on normal distribution, To fit the weights corresponding to the concentrated fault probability layer, is the number of connectors on each critical line.

[0049] By inputting the various connection monitoring data sets into the line fault probability model, after being processed by the random fault probability layer and the concentrated fault probability layer, a comprehensive line fault probability value is output in combination with their respective weights. The line fault probability value will reflect the possibility of failure of the entire line and take into account two different types of failure modes: random and concentrated.

[0050] Furthermore, step P50 of the embodiment of the present application also includes:

[0051] P51: Among them, the connection monitoring data set includes contact resistance monitoring data, contact temperature monitoring data, contact displacement monitoring data and contact vibration monitoring data;

[0052] P52: Input each connection monitoring data set into the line fault probability model, identify the fault probability of each connection monitoring data set according to the random fault probability layer and the concentrated fault probability layer of the line fault probability model, and output the line random fault probability and the line concentrated fault probability;

[0053] P53: Output the connector failure probability on each key line based on the random failure probability of the line and the concentrated failure probability of the line.

[0054] Specifically, the connection monitoring data set includes contact resistance monitoring data, contact temperature monitoring data, contact displacement monitoring data and contact vibration monitoring data, wherein the contact resistance monitoring data can reflect the resistance value of the connector contact surface, and high resistance may mean poor contact or corrosion. Excessively high contact temperature monitoring data may indicate overload, poor contact or impending failure. The contact displacement monitoring data can be used to detect slight movements of the connector under mechanical or thermal stress. The contact vibration monitoring data may reflect situations where the connector is loose, damaged or improperly installed.

[0055] Furthermore, each connection monitoring data set is input into the line fault probability model, and the fault probability of each connection monitoring data set is identified according to the random fault probability layer and the concentrated fault probability layer of the line fault probability model. According to the random fault probability layer, the faults caused by random factors, such as material aging, accidental impact, etc., are analyzed, and the possibility of random faults is evaluated by the corresponding exponential distribution probability density function, and the line random fault probability is output. At the same time, according to the concentrated fault probability layer, the faults caused by common factors, such as environmental changes, design defects, etc., are analyzed, and the possibility of concentrated occurrence is evaluated by the corresponding normal distribution probability density function, and the line concentrated fault probability is output.

[0056] Furthermore, by integrating the random failure probability of the line and the concentrated failure probability of the line, the connector failure probability on each key line can be calculated through weighted average, extraction of maximum probability and other methods, that is, the comprehensive failure probability assessment value, which can more comprehensively and accurately reflect the possibility of connector failure.

[0057] Furthermore, the embodiment of the present application further includes step P50a, and step P50a further includes:

[0058] P51a: Get the P connector positions corresponding to each key line;

[0059] P52a: Obtaining the distance between two adjacent connectors according to the positions of the P connectors;

[0060] P53a: Perform fault probability impact analysis based on the distance between two adjacent connectors and configure the distance penalty factor;

[0061] P54a: Performing feedback optimization on the line fault probability model based on the distance penalty factor.

[0062] Specifically, based on design documents or installation records, the specific position of the connector on each critical line is obtained to obtain the corresponding P connector positions. Further, based on the P connector positions, the distance between two adjacent connectors is calculated, and the distance information can reflect the distribution and density of the connectors on the critical line. Since the distance between the connectors may affect the probability of failure, for example, if the distance between the two connectors is too close, they may be affected by similar environmental factors, thereby increasing the risk of simultaneous failure. On the contrary, if the distance is far, the probability of being affected by the same environmental factors is small, and the failure risk is relatively independent. Therefore, the impact analysis of the failure probability is performed based on the distance between the two adjacent connectors, and a distance penalty factor is configured for each distance range to reflect the increase or decrease of the connector failure probability within the distance range. For example, for connectors that are too close, a larger penalty factor can be set to increase the evaluation value of their failure probability. Further, the distance penalty factor is used as feedback information to perform feedback optimization on the line fault probability model to improve the evaluation accuracy and reliability of the line fault probability model.

[0063] P60: outputting M connection failure probabilities corresponding to the M key lines according to the line failure probability model;

[0064] Optionally, line fault analysis is performed by inputting the connection monitoring data set of M key lines into the line fault probability model. The line fault probability model identifies and calculates the failure probability of the connector on each key line based on the input connection monitoring data set and the algorithm and logic within the model, and outputs M connection failure probabilities corresponding to the M key lines, thereby providing a basis for subsequent fault prevention, early warning and maintenance decisions.

[0065] P70: Generate a fault detection result according to the M connection failure probabilities, and send the fault detection result to a distribution cabinet fault management center.

[0066] Furthermore, step P70 of the embodiment of the present application further includes:

[0067] P71: After performing a criticality assessment on the corresponding route in each route clustering result, the criticality assessment result is called to generate M criticality indicators;

[0068] P72: Calculate the M connection failure probabilities based on the M key indicators and output the fault detection result.

[0069] It should be understood that a fault detection result is generated based on the M connection failure probabilities, and the fault detection result is sent to the distribution cabinet fault management center for fault management. Specifically, first, a criticality assessment is performed on multiple key lines in each line clustering result, including assessing the position of the line in the distribution system, the importance of the load carried, connectivity with other key facilities, etc., and a key indicator is generated for each line, that is, an indicator reflecting the importance or criticality of the line in the system. Furthermore, the M connection failure probabilities are calculated based on the M key indicators, and the M key indicators can be used as weights to perform weighted or adjusted priority sorting on the M connection failure probabilities, thereby more accurately reflecting the severity and criticality of the fault, so that the distribution cabinet fault management center can give priority to faults that have a greater impact on the system according to the fault detection results, improve the accuracy and effectiveness of fault handling, and improve the safety of system operation.

[0070] In summary, the embodiments of the present application have at least the following technical effects:

[0071] This application identifies M key lines through the distribution line information of the distribution cabinet, and obtains the connectors and connection methods of each key line, configures the parameters of the connection monitoring module according to the connection method, performs connection monitoring on each connector, obtains the corresponding N connection monitoring data sets, inputs them into the line fault probability model, obtains M connection failure probabilities, and generates fault detection results, which are sent to the distribution cabinet fault management center.

[0072] The technical effect of real-time monitoring of the connector status on key lines, timely detection of potential fault hazards, and improvement of the stability and reliability of the power system has been achieved.

[0073] Embodiment 2

[0074] Based on the same inventive concept as the intelligent detection method for power distribution cabinet fault based on deep learning in the aforementioned embodiment, Figure 3 As shown, the present application provides a power distribution cabinet fault intelligent detection device based on deep learning, and the device and method embodiments in the present application embodiments are based on the same inventive concept. Among them, the device includes:

[0075] A distribution line information acquisition module 11, wherein the distribution line information acquisition module 11 is used for interacting with the distribution cabinet fault management center to acquire the distribution line information of the distribution cabinet;

[0076] A key line identification module 12, the key line identification module 12 is used to analyze the distribution line information of the power distribution cabinet and identify M key lines;

[0077] A connector connection mode identification module 13, the connector connection mode identification module 13 is used to obtain each connector of each key line in the M key lines, and identify the connection mode of each connector;

[0078] A connection monitoring data acquisition module 14, the connection monitoring data acquisition module 14 is used to configure the parameters of the connection monitoring module according to the connection mode, perform connection monitoring on each connector on each key line according to the connection monitoring module, and output N connection monitoring data sets corresponding to each key line, where N is the number of connectors;

[0079] A line fault analysis module 15, the line fault analysis module 15 is used to input each connection monitoring data set into a line fault probability model, wherein the line fault probability model is obtained by fitting a random fault probability layer and a concentrated fault probability layer;

[0080] A connection failure probability output module 16, wherein the connection failure probability output module 16 is used to output M connection failure probabilities corresponding to the M key lines according to the line failure probability model;

[0081] The distribution cabinet fault management module 17 is used to generate a fault detection result according to the M connection failure probabilities, and send the fault detection result to the distribution cabinet fault management center.

[0082] Furthermore, the key line identification module 12 is also used to perform the following steps:

[0083] Interactive distribution cabinet fault management center, obtains line function type;

[0084] Clustering the distribution line information of the power distribution cabinet according to the line function type, and outputting the line clustering result, wherein each line clustering result corresponds to a line function type;

[0085] Perform key line analysis on each line clustering result and screen M key lines.

[0086] Furthermore, the key line identification module 12 is also used to perform the following steps:

[0087] Perform criticality assessment on the corresponding line in each line clustering result and output the criticality assessment result, wherein the criticality assessment is to assess the impact of line failure on system safety;

[0088] M key paths are screened according to the criticality evaluation result, wherein the M key paths include a first key path with the highest criticality in each path clustering result, and key paths with a criticality greater than or equal to a preset criticality in all path clustering results.

[0089] Furthermore, the line fault analysis module 15 is further configured to perform the following steps:

[0090] The connection monitoring data set includes contact resistance monitoring data, contact temperature monitoring data, contact displacement monitoring data and contact vibration monitoring data;

[0091] Inputting each connection monitoring data set into a line fault probability model, performing fault probability identification on each connection monitoring data set according to a random fault probability layer and a concentrated fault probability layer of the line fault probability model, and outputting a line random fault probability and a line concentrated fault probability;

[0092] The connector failure probability on each key line is output according to the line random failure probability and the line concentrated failure probability.

[0093] Furthermore, the line fault analysis module 15 is further configured to perform the following steps:

[0094] Fitting the random fault probability layer and the concentrated fault probability layer to obtain the line fault probability model includes:

[0095] ;

[0096] in, , characterizes the probability density function of the exponential distribution corresponding to the random failure probability layer, , characterizes the probability density function of the normal distribution corresponding to the concentrated fault probability layer, is the line fault probability model, To connect monitoring datasets;

[0097] Characterize the connection monitoring data corresponding to the i-th connector, Characterize the rate parameter of the failure of the i-th connector, is a natural constant, The weights corresponding to the fitting random failure probability layer;

[0098] is the mean failure probability of the i-th connector based on normal distribution, is the discrete degree of failure of the i-th connector based on normal distribution, To fit the weights corresponding to the concentrated fault probability layer, is the number of connectors on each critical line.

[0099] Furthermore, the line fault analysis module 15 is further configured to perform the following steps:

[0100] Get the corresponding information of each key line connector positions;

[0101] According to the Connector positions, get the distance between two adjacent connectors;

[0102] Perform fault probability impact analysis based on the distance between two adjacent connectors and configure the distance penalty factor;

[0103] Feedback optimization is performed on the line fault probability model based on the distance penalty factor.

[0104] Furthermore, the power distribution cabinet fault management module 17 is also used to perform the following steps:

[0105] After performing a criticality assessment on the corresponding route in each route clustering result, calling the criticality assessment result to generate M criticality indicators;

[0106] The M connection failure probabilities are calculated based on the M key indicators, and a fault detection result is output.

[0107] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0108] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0109] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A distribution cabinet fault intelligent detection method based on deep learning, characterized in that: The method comprises: Interactive distribution cabinet fault management center, obtains distribution cabinet distribution line information; Analyze the distribution line information of the power distribution cabinet and identify M key lines; Acquire each connector of each key line in the M key lines, and identify the connection mode of each connector; Configure the parameters of the connection monitoring module according to the connection mode, perform connection monitoring on each connector on each key line according to the connection monitoring module, and output N connection monitoring data sets corresponding to each key line, where N is the number of connectors; Inputting each connection monitoring data set into a line fault probability model, wherein the line fault probability model is obtained by fitting a random fault probability layer and a concentrated fault probability layer; Outputting M connection failure probabilities corresponding to the M key lines according to the line failure probability model; Generate a fault detection result according to the M connection failure probabilities, and send the fault detection result to a distribution cabinet fault management center.

2. The method according to claim 1, characterized in that Analyze the distribution line information of the power distribution cabinet to identify M key lines, the method includes: Interactive distribution cabinet fault management center, obtains line function type; Clustering the distribution line information of the power distribution cabinet according to the line function type, and outputting the line clustering result, wherein each line clustering result corresponds to a line function type; Perform key line analysis on each line clustering result and screen M key lines.

3. The method according to claim 2, characterized in that Perform key line analysis on each line clustering result and select M key lines. The method includes: Perform criticality assessment on the corresponding line in each line clustering result and output the criticality assessment result, wherein the criticality assessment is to assess the impact of line failure on system safety; M key paths are screened according to the criticality evaluation result, wherein the M key paths include a first key path with the highest criticality in each path clustering result, and key paths with a criticality greater than or equal to a preset criticality in all path clustering results.

4. The method according to claim 1, characterized in that Each connection monitoring data set is input into the line fault probability model. include: The connection monitoring data set includes contact resistance monitoring data, contact temperature monitoring data, contact displacement monitoring data and contact vibration monitoring data; Input each connection monitoring data set into the line fault probability model, perform fault probability identification on each connection monitoring data set according to the random fault probability layer and the concentrated fault probability layer of the line fault probability model, and output the line random fault probability and the line concentrated fault probability; According to the random failure probability of the line and the concentrated failure probability of the line, the connection failure probability on each key line is output.

5. The method according to claim 1, characterized in that Fitting the random fault probability layer and the concentrated fault probability layer to obtain the line fault probability model includes: ； in, , characterizes the probability density function of the exponential distribution corresponding to the random failure probability layer, , characterizes the probability density function of the normal distribution corresponding to the concentrated fault probability layer, is the line fault probability model, To connect monitoring data sets; Characterize the connection monitoring data corresponding to the i-th connector, Characterize the rate parameter of the failure of the i-th connector, is a natural constant, The weights corresponding to the fitting random failure probability layer; is the mean failure probability of the i-th connector based on normal distribution, is the discrete degree of failure of the i-th connector based on normal distribution, To fit the weights corresponding to the concentrated fault probability layer, is the number of connectors on each critical line.

6. The method according to claim 5, characterized in that After obtaining the line fault probability model, the method further includes: Obtain the P connector positions corresponding to each key line; According to the positions of the P connectors, obtaining the distance between two adjacent connectors; Perform fault probability impact analysis based on the distance between two adjacent connectors and configure the distance penalty factor; Feedback optimization is performed on the line fault probability model based on the distance penalty factor.

7. The method according to claim 3, characterized in that Generating a fault detection result according to the M connection fault probabilities, the method includes: After performing a criticality assessment on the corresponding route in each route clustering result, calling the criticality assessment result to generate M criticality indicators; The M connection failure probabilities are calculated based on the M key indicators, and a fault detection result is output.

8. A power distribution cabinet fault intelligent detection device based on deep learning, characterized in that: The device comprises: A distribution line information acquisition module, which is used for interacting with the distribution cabinet fault management center to obtain distribution line information of the distribution cabinet; A key line identification module, the key line identification module is used to analyze the distribution line information of the power distribution cabinet and identify M key lines; A connector connection mode identification module, the connector connection mode identification module is used to obtain each connector of each key line in the M key lines, and identify the connection mode of each connector; A connection monitoring data acquisition module, the connection monitoring data acquisition module is used to configure the parameters of the connection monitoring module according to the connection mode, perform connection monitoring on each connector on each key line according to the connection monitoring module, and output N connection monitoring data sets corresponding to each key line, where N is the number of connectors; A line fault analysis module, the line fault analysis module is used to input each connection monitoring data set into a line fault probability model, wherein the line fault probability model is obtained by fitting a random fault probability layer and a concentrated fault probability layer; A connection failure probability output module, the connection failure probability output module is used to output M connection failure probabilities corresponding to the M key lines according to the line failure probability model; A distribution cabinet fault management module, the distribution cabinet fault management module is used to generate a fault detection result according to the M connection failure probabilities, and send the fault detection result to a distribution cabinet fault management center.

Images