A safe power distribution cabinet based on intelligent monitoring

Abstract

The present application provides a safe power distribution cabinet based on intelligent monitoring, which relates to the field of power distribution cabinet monitoring technology, including: a fault sample collection module, which generates a fault sample set; a data classification module, which classifies data and generates multiple component fault sample sets; a model construction module, which trains and constructs multiple component fault identification models; a state monitoring module, which monitors the operating state and obtains multiple real-time operating state data of multiple components; a first fault identification module, which outputs multiple first fault identification results; a second fault identification module, which performs acoustic signal monitoring; and a fault control decision module, which performs fault processing. This application can solve the problem that the existing technology is insufficient in real-time state monitoring capability, has a single fault diagnosis method, and cannot fully utilize various fault indication information, resulting in the inability to detect potential faults in a timely manner, further affecting the safety performance of the power distribution cabinet and the accuracy of fault identification, reducing false alarms and missed alarms, and improving the efficiency and safety of fault processing.

Description

Technical Field

[0001] The present application relates to the technical field of power distribution cabinet monitoring, and in particular to a safe power distribution cabinet based on intelligent monitoring. Background Art

[0002] Safety distribution cabinet is a kind of electrical equipment used to distribute, control, protect and monitor electrical energy. Safety distribution cabinets play a vital role in industrial, commercial and residential fields. However, due to the complexity of the power system and human operating errors, power accidents occur from time to time, posing a serious threat to people's lives and property. In order to improve the safety of distribution cabinets and reduce the incidence of power accidents, safety distribution cabinets based on intelligent monitoring came into being. Traditional distribution cabinets are often composed of air switches, AC contactors, and leakage protectors. In seasons with high power consumption or affected by other factors, the leakage protection facilities of the distribution cabinets often trip, affecting the reliability of the low-voltage distribution network. Especially in summer when the temperature is high, the distribution cabinet often trips due to excessive load, affecting the continuous power consumption.

[0003] At present, the existing distribution cabinet fault diagnosis system has certain limitations in terms of real-time performance, accuracy, fault handling strategy, etc. Problems such as incomplete fault sample collection, inaccurate fault type and location identification, unintelligent fault control strategy, insufficient real-time status monitoring capability, single multi-modal fault diagnosis method, and lack of fault confidence assessment still exist.

[0004] In summary, the existing technology is insufficient in real-time status monitoring capability and has a single fault diagnosis method, which cannot fully utilize various fault indication information, resulting in the inability to detect potential faults in a timely manner, further affecting the safety performance of the distribution cabinet and the accuracy of fault identification. Summary of the invention

[0005] The purpose of this application is to provide a safe distribution cabinet based on intelligent monitoring, so as to solve the problems in the prior art such as insufficient real-time status monitoring capability, single fault diagnosis method, and inability to fully utilize various fault indication information, resulting in the inability to timely detect potential faults, further affecting the safety performance of the distribution cabinet and the accuracy of fault identification, reducing false alarms and missed alarms, and improving the efficiency and safety of fault handling.

[0006] In view of the above problems, the present application provides a safe power distribution cabinet based on intelligent monitoring.

[0007] The present application provides a safe power distribution cabinet based on intelligent monitoring, wherein the safe power distribution cabinet includes: a fault sample collection module, which is used to collect historical fault records of the target power distribution cabinet within the service interval to generate a fault sample set, wherein any fault sample includes fault status data, fault type, fault control mode and fault location; a data classification module, which is used to classify the fault sample set based on the fault location to generate multiple component fault sample sets; a model construction module, which is used to train and construct multiple component fault identification models with the fault status data and fault type in the multiple component fault sample sets, and set the startup strategy of the fault control decision model in combination with the fault control mode; a state monitoring module, which is used to classify the fault sample set based on the fault location to generate multiple component fault sample sets; a model construction module, which is used to train and construct multiple component fault identification models with the fault status data and fault type in the multiple component fault sample sets, and set the startup strategy of the fault control decision model in combination with the fault control mode; a state monitoring module, which is used to classify the fault sample set based on the fault status data and fault type in the multiple component fault sample sets The target power distribution cabinet is used to perform operating status monitoring to obtain multiple real-time operating status data of multiple components; a first fault identification module is used to input the multiple real-time operating status data into the multiple component fault identification models, and output multiple first fault identification results; a second fault identification module is used to monitor the acoustic signals of the multiple components through a voiceprint sensor, generate multiple chromaticity diagrams for fault analysis, and output multiple second fault identification results; a fault control decision module is used to perform fault confidence verification on the multiple first fault identification results and the multiple second fault identification results, generate multiple confident fault identification results, and start the fault control decision model through the startup strategy control, generate a fault control decision, and perform fault processing.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] A fault sample collection module is used to collect historical fault records of a target distribution cabinet within a service interval to generate a fault sample set, wherein any fault sample includes fault status data, fault type, fault control mode and fault location; a data classification module is used to classify the fault sample set based on the fault location to generate multiple component fault sample sets; a model construction module is used to train and construct multiple component fault identification models with the fault status data and fault type in the multiple component fault sample sets, and set the startup strategy of the fault control decision model in combination with the fault control mode; a status monitoring module is used to monitor the operating status of the target distribution cabinet to obtain multiple real-time operating status data of multiple components; a first fault identification module is used to input the multiple real-time operating status data into the multiple component fault identification A model is provided to output multiple first fault identification results; a second fault identification module is used to monitor the acoustic signals of the multiple components through a voiceprint sensor, generate multiple chromaticity diagrams for fault analysis, and output multiple second fault identification results; a fault control decision module is used to perform fault confidence verification on the multiple first fault identification results and the multiple second fault identification results, generate multiple confident fault identification results, and start the fault control decision model through the startup strategy control, generate a fault control decision, and perform fault processing, which effectively solves the problem that the prior art is unable to fully utilize various fault indication information due to insufficient real-time status monitoring capabilities and a single fault diagnosis method, resulting in the inability to timely detect potential faults, further affecting the safety performance of the distribution cabinet and the accuracy of fault identification, reducing false alarms and missed alarms, and improving the efficiency and safety of fault processing.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0012] Figure 1 This is a schematic diagram of the structure of a safety distribution cabinet based on intelligent monitoring for this application;

[0013] Figure 2 This is a flow chart of generating maintenance signals in a safety distribution cabinet based on intelligent monitoring according to the present application.

[0014] Description of reference numerals:

[0015] Fault sample collection module 11, data classification module 12, model construction module 13, state monitoring module 14, first fault identification module 15, second fault identification module 16, fault control decision module 17. DETAILED DESCRIPTION

[0016] The present application provides a safe power distribution cabinet based on intelligent monitoring, which solves the problems in the prior art such as insufficient real-time status monitoring capability, single fault diagnosis method, and inability to fully utilize various fault indication information, resulting in the inability to timely detect potential faults, further affecting the safety performance of the power distribution cabinet and the accuracy of fault identification, reducing false alarms and missed alarms, and improving the efficiency and safety of fault handling.

[0017] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0018] This application provides a safe power distribution cabinet based on intelligent monitoring. Figure 1 , the safety distribution cabinet comprises:

[0019] The fault sample collection module 11 is used to collect historical fault records of the target distribution cabinet within the service range and generate a fault sample set, wherein any fault sample includes fault status data, fault type, fault control mode and fault location.

[0020] Specifically, the service interval is from the starting node where the target distribution cabinet begins to serve to the current node, and the records in this interval contain all fault records after use. First, use the sensors deployed on the target distribution cabinet to monitor key parameters such as voltage, current, temperature, and humidity in real time. At the same time, record the details of each fault, including fault status data, such as abnormal fluctuations in voltage and current, fault types such as short circuit, overload, etc., fault control modes such as automatic reclosing, manual switching, etc., and fault locations such as circuit breakers and cables. The characteristic data, fault type, fault control mode, and fault location corresponding to each fault event are integrated together to form a fault sample. Then all fault samples are combined to generate a fault sample set.

[0021] The data classification module 12 is used to perform data classification on the fault sample set based on the fault location to generate multiple component fault sample sets.

[0022] Specifically, the fault location information is used as the classification basis to divide the fault sample set into multiple subsets. Each subset contains all fault samples related to a specific component. For example, the fault sample set can be divided into a circuit breaker fault sample set, a transformer fault sample set, a cable fault sample set, etc.

[0023] The model building module 13 is used to train and build multiple component fault identification models based on the fault state data and fault types in the multiple component fault sample sets, and to set a startup strategy for the fault control decision model in combination with the fault regulation mode.

[0024] Specifically, each component fault sample set contains fault status data and fault type labels. The fault status data includes the measured values ​​of parameters such as voltage, current, temperature, and vibration, while the fault type label indicates the type of fault corresponding to each fault sample, such as short circuit, overload, poor contact, etc. Machine learning algorithms such as random forest, support vector machine, neural network, etc. are used to train each component fault sample set. The relationship between fault status data and fault type is learned to build a model that can identify specific component faults. After the construction is completed, the trained model is verified through cross-validation to evaluate its performance indicators such as accuracy and recall rate. According to the verification results, the model parameters are adjusted to improve the fault identification ability of the model. The fault control modal data in the fault sample set is analyzed to identify the control measures for different fault types when they occur, such as automatic reclosing, manual switching, protection device action, etc. According to the fault control modal analysis results, the corresponding startup strategy is set for each fault identification model. For example, temperature anomalies can be eliminated through control of heat dissipation devices; if electric sparks or other phenomena that may cause fire occur, the fire extinguishing device is activated to extinguish the fire and the power is cut off to generate a fault signal. For faults that cannot be handled, it can be determined based on the severity whether to directly cut off the power to generate a fault signal to alert relevant personnel to perform control and processing.

[0025] The status monitoring module 14 is used to monitor the operating status of the target power distribution cabinet and obtain multiple real-time operating status data of multiple components.

[0026] Specifically, sensors configured on the target power distribution cabinet, including voltage sensors, current sensors, temperature sensors, humidity sensors, vibration sensors, etc., are used to monitor different types of operating status data. The sensors collect data on parameters such as voltage, current, temperature, humidity, and vibration in real time.

[0027] The first fault identification module 15 is used to input the multiple real-time operating status data into the multiple component fault identification models and output multiple first fault identification results.

[0028] Specifically, before the real-time operating status data is input into the fault identification model, the data is preprocessed to ensure that the quality and format of the data are consistent with the data used in model training. Preprocessing includes denoising, normalization, feature extraction, etc. The component fault identification model is loaded, and the preprocessed real-time operating status data is input into the corresponding fault identification model. The fault identification model of each component will process the data related to the component. The model calculates and analyzes the input data and outputs the fault identification results.

[0029] The second fault identification module 16 is used to monitor the acoustic signals of the multiple components through the voiceprint sensor, generate multiple chromaticity diagrams for fault analysis, and output multiple second fault identification results.

[0030] Specifically, a voiceprint sensor is a sensor that can monitor and analyze the sound of equipment operation. It can help diagnose faults by capturing the sound characteristics of the equipment in normal and abnormal states. Chromagram is a technology that converts sound signals into visual representations, which can help technicians identify sound patterns and analyze faults more intuitively. Use voiceprint sensors to monitor the sound of multiple components of the distribution cabinet. The sensor captures the sound generated when the equipment is running, including noise generated by mechanical vibration, electrical discharge, airflow, etc. The sound data captured by the voiceprint sensor is converted into digital signals and recorded. These data include characteristics such as the amplitude, frequency, rhythm and timbre of the sound. The collected sound data is preprocessed to remove irrelevant noise and interference and enhance the sound characteristics related to the fault. Preprocessing includes steps such as filtering, noise reduction, and feature enhancement. A chromagram is generated using the preprocessed sound data. Chromagram is a technology that maps the frequency components of sound to color space, using a spectrogram or a three-dimensional spectrogram. In the chromagram, different colors represent different frequency components, and the brightness and saturation of the color may indicate the intensity and duration of the sound. By analyzing the generated chromaticity diagram, technicians can identify changes in sound patterns. For example, abnormal vibration or discharge sounds may show specific color patterns on the chromaticity diagram. Based on the analysis results of the chromaticity diagram, a second fault identification result is output. These results include information such as the presence of a fault, the type of fault, and the severity of the fault.

[0031] The fault control decision module 17 is used to perform fault confidence verification on the multiple first fault identification results and the multiple second fault identification results, generate multiple confident fault identification results, and start the fault control decision model through the startup strategy control, generate a fault control decision, and perform fault processing.

[0032] Specifically, the first fault identification result and the second fault identification result are comprehensively analyzed. The two identification results are compared and contrasted to see whether there are consistent or inconsistent fault diagnoses. For consistent fault diagnoses, the confidence level is increased; for inconsistent fault diagnoses, further analysis is performed to determine which result is more reliable. The confidence assessment combines historical fault data, sensor reliability, environmental conditions, etc. Based on the confidence assessment, the final confidence fault identification result is generated. These results should include the existence, type, location and confidence level of the fault. Based on the confidence fault identification result and the pre-set startup strategy, it is decided whether to start the fault control decision model. The startup strategy includes the severity of the fault, the confidence threshold, the availability of operation and maintenance personnel, etc. When the conditions of the startup strategy are met, the fault control decision model is triggered. The model will generate fault handling suggestions or automatically execute the fault handling process based on the fault identification results and the preset rules or algorithms. Fault handling is performed according to the suggestions or instructions output by the fault control decision model. Including adjusting equipment parameters, switching power paths, starting backup equipment, dispatching maintenance personnel, etc.

[0033] Further, if Figure 2 As shown, the safety power distribution cabinet further includes a frequency analysis module, and the frequency analysis module is used to:

[0034] Any fault sample also includes a fault time, and a fault frequency feature is identified based on the fault time in the multiple component fault sample sets; if any fault identification result among the multiple confident fault identification results shows that there is no fault, the next fault prediction time node is located based on the fault frequency feature; the time proximity coefficient between the current moment and the next fault prediction time node is calculated; if the time proximity coefficient is greater than or equal to a preset coefficient, a maintenance signal is generated.

[0035] Specifically, the failure time data in multiple component failure sample sets are analyzed to identify the frequency characteristics of the failure. This includes determining whether the failure tends to occur in certain specific time periods, such as peak power consumption periods, weather changes, etc., or whether there is a periodic failure mode. If any of the multiple confidence fault identification results shows no fault, but the historical fault frequency characteristics indicate a potential fault risk, the next fault prediction time node is located based on these frequency characteristics. By using time series analysis or machine learning algorithms, the time when the next failure may occur is predicted. The time proximity coefficient between the current moment and the next fault prediction time node is calculated. This coefficient is an indicator of the degree of proximity between two time points, which can be calculated based on factors such as time difference, fault frequency, and system operation status. If the calculated time proximity coefficient is greater than or equal to the preset coefficient threshold, a maintenance signal is generated. This signal instructs the operation and maintenance personnel or the automatic maintenance system to take preventive measures before the predicted failure occurs, such as arranging maintenance, replacing wearing parts, or performing system optimization.

[0036] Furthermore, the model building module 13 is also used for:

[0037] A first component fault sample set is extracted from the multiple component fault sample sets, and multiple first fault identification branches are constructed by training using the first component fault sample set, wherein the network structures of the multiple first fault identification branches are different; multiple prediction accuracies of the multiple first fault identification branches are obtained through testing; the multiple first fault identification branches are integrated and fused based on the multiple prediction accuracies to generate a first component fault identification model; and the first component fault identification model is added to the multiple component fault identification models.

[0038] Specifically, fault samples related to the first component are extracted from multiple component fault sample sets to form a first component fault sample set. This sample set is used to train the fault identification model of the first component. Using the first component fault sample set, multiple first fault identification branches are trained and constructed. Each branch can adopt a different network structure or algorithm to capture different fault characteristics and patterns. For example, the first branch uses a convolutional neural network, while the other branches use a recurrent neural network or a support vector machine. Each first fault identification branch is tested to obtain their prediction accuracy on the test set. Evaluate through cross-validation. Based on the multiple prediction accuracies obtained from the test, multiple first fault identification branches are integrated and fused to generate a first component fault identification model. Integration can be performed by average fusion, weighted fusion, stacking or other advanced integration techniques. The trained first component fault identification model is added to the set of multiple component fault identification models.

[0039] Furthermore, the model building module 13 is also used for:

[0040] The fault control mode includes a fault recovery mode and a fault emergency mode; the multiple component fault sample sets are mapped and associated with each other according to the fault control mode, and a first mode-fault type mapping data set and a second mode-fault type mapping data set are constructed, wherein any first mode-fault type mapping data includes fault control parameters and fault status data, and any second mode-fault type mapping data includes fault control strategy and fault type; the startup strategy is set based on the fault type mapping association; a first fault control decision channel is constructed with the first mode-fault type mapping data set, and a second fault control decision channel is constructed with the second mode-fault type mapping data set; the fault control decision model is generated with the first fault control decision channel and the second fault control decision channel.

[0041] Specifically, fault type mapping association is performed on multiple component fault sample sets according to fault control modes, namely, fault recovery mode and fault emergency mode. The fault type of each fault sample is associated with the corresponding fault control mode. A first mode-fault type mapping data set and a second mode-fault type mapping data set are respectively constructed. The first mode-fault type mapping data set contains fault control parameters and fault state data, while the second mode-fault type mapping data set contains fault control strategy and fault type. Based on the fault type mapping association, the startup strategy of the fault control decision model is set. The startup strategy will determine when to activate the fault control decision model and how to adjust the response of the model according to different fault types and control modes. The first mode-fault type mapping data set is used to construct a first fault control decision channel, which is specifically for the fault recovery mode. At the same time, the second mode-fault type mapping data set is used to construct a second fault control decision channel, which is specifically for the fault emergency mode. Combine the first fault control decision channel and the second fault control decision channel to generate a fault control decision model. This model will be able to provide corresponding fault control strategies and parameter adjustment suggestions based on the current fault type and control mode.

[0042] Furthermore, the second fault identification module 16 is further configured to:

[0043] The acoustic signals of the multiple components are monitored by a voiceprint sensor to generate multiple voiceprint signal data; the multiple voiceprint signal data are short-time Fourier transformed to generate multiple voiceprint time-frequency diagrams; a chromaticity order-frequency mapping table is established to map and transform the multiple voiceprint time-frequency diagrams to generate multiple chromaticity spectra; the multiple chromaticity spectra are input into a fault chromaticity spectrum database for comparison to obtain the multiple second fault identification results.

[0044] Specifically, a voiceprint sensor is used to monitor the acoustic signals of multiple components of the power distribution cabinet. The sensor will record the acoustic wave signals generated when the equipment is running. These signals will vary due to different states of the equipment, such as normal, wear, and failure. The collected voiceprint signal data is subjected to short-time Fourier transform. Short-time Fourier transform can reveal the changes in the frequency components of the acoustic signal over time. Through short-time Fourier transform, multiple voiceprint time-frequency graphs are generated. These time-frequency graphs show the frequency distribution of the acoustic signal at different time points. A chromaticity order-frequency mapping table is established. This mapping table maps the frequency components of the acoustic signal to different color spaces, so that the frequency changes of the acoustic signal can be represented by color changes. Using the established mapping table, the voiceprint time-frequency graph is mapped and converted to generate multiple chromaticity spectra. The generated chromaticity spectra are input into the fault chromaticity spectrum database for comparison. The database stores standard chromaticity spectra under various fault conditions, and the current fault type can be identified by comparison. According to the results of the database comparison, multiple second fault identification results are obtained. These results include information such as the existence, type, and severity of the fault.

[0045] Furthermore, the second fault identification module 16 is further configured to:

[0046] The fault chromaticity spectrum database includes multiple groups of fault chromaticity spectrum mapping data, wherein any group of fault chromaticity spectrum mapping data includes a fault type and a fault chromaticity spectrum; the multiple chromaticity spectra are input into the fault chromaticity spectrum database, and a chromaticity spectrum similarity comparison is performed with each group of fault chromaticity spectrum mapping data to obtain a fault type that meets a preset discrimination constraint, and generate the multiple second fault identification results.

[0047] Specifically, the fault chromaticity spectrum database contains multiple groups of fault chromaticity spectrum mapping data, each group of data includes a fault type and a corresponding fault chromaticity spectrum. These data are constructed through historical fault case analysis. Multiple chromaticity spectra generated in real time are input into the fault chromaticity spectrum database, and similarity comparison is performed with each group of fault chromaticity spectrum mapping data in the database. Similarity comparison can use various image processing and pattern recognition technologies, such as template matching, feature extraction, machine learning classifiers, etc. In the comparison process, preset discrimination constraints are applied to filter the comparison results. These constraints include similarity thresholds, specific conditions for spectrum matching, etc. According to the comparison results, the fault types that meet the preset discrimination constraints are obtained. These fault types are the possible fault types of the current device. Based on the identified fault types, multiple second fault identification results are generated. These results include information such as the existence, type, and possible causes of the fault.

[0048] Furthermore, the fault control decision module 17 is also used for:

[0049] Based on the multiple confident fault identification results, locate the faulty component and the fault type, and match the target fault control mode in combination with the startup strategy; if the target fault control mode is the fault recovery mode, call the real-time operating status data of the faulty component to input the first fault control decision channel of the fault control decision model, perform fault control parameter analysis, and output target fault control parameters; add the target fault control parameters into the fault control decision.

[0050] Specifically, according to the fault identification result, the specific component and type of the fault are determined through the electrical parameters, acoustic characteristics, fault frequency and other data of the fault. According to the fault component and type, combined with the preset startup strategy, the target fault control mode is determined. Including factors such as the severity of the fault, the scope of impact, and the processing priority. If the target fault control mode is the fault recovery mode, the real-time operating status data of the faulty component is analyzed. These data include parameters such as voltage, current, temperature, and vibration, which reflect the current operating status of the faulty component. The real-time operating status data of the faulty component is input into the first fault control decision channel of the fault control decision model. This channel is specifically used for the fault recovery mode, and can analyze according to the input data and output the target fault control parameters. The fault control decision model analyzes the input data, identifies the key parameters that cause the fault, and outputs the corresponding control parameters. These parameters include adjusting the voltage, current setting, replacing components, etc. The output target fault control parameters are added to the fault control decision.

[0051] Furthermore, the fault control decision module 17 is also used for:

[0052] If the target fault regulation mode is a fault emergency mode, the second fault control decision channel of the fault control decision model is started through the startup strategy control to generate an emergency control decision; and the emergency control decision is added to the fault control decision.

[0053] Specifically, according to the startup strategy, if it is determined that the fault needs to be handled immediately, the second fault control decision channel of the fault control decision model will be controlled to start. This channel is specially used to handle emergency situations and can quickly generate emergency control decisions. The fault control decision model generates emergency control decisions based on the urgency and impact range of the fault. These decisions include immediately cutting off power, starting the backup system, and evacuating personnel urgently. The generated emergency control decisions are added to the fault control decisions and are executed immediately.

[0054] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0055] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A safe power distribution cabinet based on intelligent monitoring, characterized in that: The safety distribution cabinet comprises: The fault sample collection module is used to collect historical fault records of the target distribution cabinet within the service interval and generate a fault sample set, wherein any fault sample includes fault status data, fault type, fault control mode and fault location; A data classification module, used for performing data classification on the fault sample set based on the fault location to generate multiple component fault sample sets; A model building module, used to train and build multiple component fault identification models based on the fault state data and fault types in the multiple component fault sample sets, and to set a startup strategy for the fault control decision model in combination with the fault regulation mode; A status monitoring module is used to monitor the operating status of the target power distribution cabinet and obtain multiple real-time operating status data of multiple components; A first fault identification module, configured to input the plurality of real-time operating status data into the plurality of component fault identification models, and output a plurality of first fault identification results; A second fault identification module is used to monitor the acoustic signals of the multiple components through a voiceprint sensor, generate multiple chromaticity diagrams for fault analysis, and output multiple second fault identification results; A fault control decision module is used to perform fault confidence verification on the multiple first fault identification results and the multiple second fault identification results, generate multiple confident fault identification results, and start the fault control decision model through the start-up strategy control, generate a fault control decision, and perform fault processing; Wherein, the safety power distribution cabinet further includes a frequency analysis module, and the frequency analysis module is used to: Any fault sample further includes a fault time, and fault frequency characteristics are identified based on the fault times in the plurality of component fault sample sets; If any fault identification result among the multiple confident fault identification results shows no fault, locating the next fault prediction time node based on the fault frequency feature; Calculate the time proximity coefficient between the current moment and the next fault prediction time node; If the temporal proximity coefficient is greater than or equal to a preset coefficient, a maintenance signal is generated.

2. The safety distribution cabinet according to claim 1, characterized in that: The model building module is also used to: Extracting a first component fault sample set from the multiple component fault sample sets, using the first component fault sample set to train and construct multiple first fault identification branches, wherein the multiple first fault identification branches have different network structures; Testing and obtaining a plurality of prediction accuracies of the plurality of first fault identification branches; Integrate and fuse the multiple first fault identification branches based on the multiple prediction accuracies to generate a first component fault identification model; The first component fault identification model is added to the plurality of component fault identification models.

3. The safety distribution cabinet according to claim 2, characterized in that: The model building module is also used to: The fault control mode includes a fault recovery mode and a fault emergency mode; According to the fault control mode, the multiple component fault sample sets are associated with fault type mapping, and a first mode-fault type mapping data set and a second mode-fault type mapping data set are constructed, wherein any first mode-fault type mapping data includes fault control parameters and fault state data, and any second mode-fault type mapping data includes fault control strategy and fault type; Setting the startup strategy based on the fault type mapping association; constructing a first fault control decision channel using the first mode-fault type mapping data set, and constructing a second fault control decision channel using the second mode-fault type mapping data set; The fault control decision model is generated using the first fault control decision channel and the second fault control decision channel.

4. The safety distribution cabinet according to claim 1, characterized in that: The second fault identification module is also used for: Monitoring the acoustic signals of the multiple components by using a voiceprint sensor to generate multiple voiceprint signal data; Performing short-time Fourier transform on the multiple voiceprint signal data to generate multiple voiceprint time-frequency graphs; Establishing a chromaticity order-frequency mapping table, mapping and converting the multiple voiceprint time-frequency graphs, and generating multiple chromaticity spectra; The multiple chromaticity spectra are input into a fault chromaticity spectra database for comparison to obtain the multiple second fault identification results.

5. The safety distribution cabinet according to claim 4, characterized in that: The second fault identification module is also used for: The fault chromaticity spectrum database includes multiple groups of fault chromaticity spectrum mapping data, wherein any group of fault chromaticity spectrum mapping data includes a fault type and a fault chromaticity spectrum; The multiple chromaticity spectra are input into a fault chromaticity spectra database, and chromaticity spectra similarity comparison is performed with each group of fault chromaticity spectra mapping data to obtain fault types that meet preset discrimination constraints, and generate the multiple second fault identification results.

6. The safety distribution cabinet according to claim 3, characterized in that: The fault control decision module is also used for: Based on the multiple confident fault identification results, locate the faulty component and the fault type, and match the target fault control mode in combination with the startup strategy; If the target fault control mode is a fault recovery mode, the real-time operating status data of the fault component is called to input into the first fault control decision channel of the fault control decision model, and the fault control parameter analysis is performed to output the target fault control parameter; The target fault regulation parameter is added into the fault control decision.

7. The safety distribution cabinet according to claim 6, characterized in that: The fault control decision module is also used for: If the target fault regulation mode is a fault emergency mode, starting the second fault control decision channel of the fault control decision model through the start-up strategy control to generate an emergency control decision; The emergency control decision is added to the fault control decision.

Images