Fault monitoring system and method for transmission, transformation and power distribution cabinet based on artificial intelligence
Through the fault monitoring system based on artificial intelligence, data from the transmission and distribution cabinet are collected and analyzed in real time, and predicted models are built, which solves the problems of fault monitoring delay and misjudgment in the existing technology, and achieves timely early warning and accuracy of faults. Diagnosis improves the safety and reliability of the power system.
Patent Information
- Application Number
- CN202411057414.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
Smart Images

Figure CN118983940A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of fault monitoring, and in particular to a fault monitoring system and method for a power transmission and distribution cabinet based on artificial intelligence. Background Art
[0002] Transmission and distribution cabinet is a kind of equipment used in power system, mainly used for control and protection of power transmission, transformation and distribution. It is an indispensable and important part of power system, ensuring efficient and safe transmission of electric energy from power plant to user end. It is usually installed at user end or near distribution transformer, including low voltage circuit breaker, disconnector, fuse, protection relay and metering equipment, etc., used to distribute electric energy to different users and equipment, and provide necessary protection and control;
[0003] However, there are still areas for improvement in the existing technology for fault monitoring of power distribution cabinets, such as:
[0004] 1. When various faults occur during power transmission due to environmental factors or equipment failures, traditional fault monitoring methods usually require manual inspection and fault diagnosis, or rely solely on sensor information, lack comprehensive analysis, have a certain delay, and cannot respond and handle faults in real time, which is prone to missed detection or misjudgment, which may lead to further equipment damage or safety hazards;
[0005] 2. For the image data and operation data presented by visually identifiable faults, in the process of processing a large amount of operation data and image data, it is often impossible to fully explore the potential information and features in the data due to the massive and complex data, and it is difficult to discover hidden fault modes and abnormal behaviors. Usually, only fault detection and diagnosis can be performed, and the ability to predict future faults is weak. It is difficult to know possible faults in advance, thereby avoiding or reducing the occurrence of faults. Summary of the invention
[0006] 1. Technical issues to be resolved
[0007] In view of the above-mentioned shortcomings of the prior art, the present invention provides a fault monitoring system and method for a power transmission and distribution cabinet based on artificial intelligence, which can effectively solve the problems of the prior art.
[0008] (II) Technical solution
[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions:
[0010] The present invention discloses a fault monitoring system for a power transmission and distribution cabinet based on artificial intelligence, comprising:
[0011] The power distribution central control terminal is used as the central control terminal for each power distribution equipment to edit and send control instructions;
[0012] The operation collection module is used to collect the current operation data of several distribution cabinets according to a preset period and pre-process the acquired operation data;
[0013] The fluctuation identification module is used to capture abnormal fluctuations of the data collected by the operation collection module that exceed the fluctuation threshold, mark and extract the captured abnormal fluctuations, and form a data flow abnormality set corresponding to several distribution cabinet targets;
[0014] An image acquisition module is used to collect image data of the power distribution equipment and its associated equipment according to a preset period and pre-process the acquired operating data;
[0015] The image recognition module is used to analyze the image data collected by the image acquisition module, extract abnormal features through a machine learning algorithm, and classify the extracted abnormal features to form an image flow abnormality set;
[0016] A fault analysis module is used to obtain the definition data of the fault type, cause and associated equipment through data mining based on the data flow anomaly set and the image flow anomaly set in combination with the information in the historical database;
[0017] The power transmission behavior prediction module is used to obtain the current environment data, equipment data and definition data obtained by the fault analysis module, build a prediction model through a deep learning algorithm, make predictions based on the input data, and output the power transmission behavior of the specified distribution cabinet and its associated equipment in a specific period of time in the future;
[0018] The structure verification module is used to be triggered after the power transmission behavior prediction module is predicted, analyze the change data of the data flow anomaly set and the image flow anomaly set collected during the blank stage of the prediction operation of the power transmission behavior prediction module, verify whether such change data in the anomaly set is consistent with the change of the same data segment in the prediction result, and output the credibility coefficient of the prediction result for verification and evaluation;
[0019] The report sending module is used to integrate the fault analysis results and the credibility coefficient of the prediction results to generate a fault monitoring report and submit it to the power distribution central control end.
[0020] Furthermore, the power distribution central control terminal is interactively connected to a data repository via a wireless network, and the data repository is interactively connected to the fault analysis module via a wireless network. The data repository is used to store historical fault data and equipment data. When the fault analysis module and the power transmission behavior prediction module are triggered, the data repository calls the historical database to perform matching analysis of the fault type, cause and related equipment content as a reference for fault diagnosis and prediction.
[0021] Furthermore, the specific operation logic of the fluctuation identification module includes the following steps:
[0022] Step 31: Dynamically set standard fluctuation thresholds based on historical data or custom data;
[0023] Step 32: receiving the collected data submitted by the running collection module, removing noise, filling missing values and performing standardization processing to obtain processed data;
[0024] Step 33: Compare the processed data with the standard fluctuation threshold to determine whether it exceeds the set threshold range;
[0025] Step 34: When it is detected that the data exceeds the threshold, it is marked as abnormal fluctuation, and the data points marked as abnormal fluctuations and their timestamps and fluctuation amplitudes are extracted and integrated as a data flow anomaly set.
[0026] Furthermore, the specific operation logic of the image recognition module includes the following steps:
[0027] Step 51: performing data cleaning, image enhancement and image normalization processing on the image data collected by the image acquisition module;
[0028] Step 52: Use an edge detection algorithm to segment different objects or regions in the image, and extract the device surface area and connection point area of the monitoring target;
[0029] Step 53: extracting geometric features, color features and texture features from the segmented image, and selecting and reducing the dimension of the extracted features through a feature selection algorithm;
[0030] Step 54: Use a classification algorithm to classify the image, detect abnormal features, mark the detected abnormal features, and form an image flow abnormality set.
[0031] Furthermore, when performing data mining, the fault analysis module fuses the data stream anomaly set and the image stream anomaly set of the fluctuation recognition module and the image recognition module into a comprehensive data set, and extracts key features, including: voltage, current, temperature, vibration frequency and visual features in the image. Through the analysis of key features, the association rules between the fault data in the comprehensive data set are mined to obtain frequently occurring fault combinations, and the current fault type and cause are marked in combination with the historical fault type, cause and related equipment data.
[0032] Furthermore, the prediction model constructed by the transmission behavior prediction module uses a long short-term memory network, divides historical data into a training set and a verification set, uses the training set to train the model, uses the verification set to verify the trained model, and evaluates the performance of the model. When the evaluated model is running, it receives input of current environmental data and equipment data to generate transmission behavior prediction results for a specific period in the future. The output prediction results include: future operating status, fault time point and fault type.
[0033] Furthermore, the structure verification module calculates the consistency index in the process of verifying whether the change data in the abnormal set is consistent with the change of the same data segment in the prediction result, and its expression is:
[0034]
[0035] Among them, CI represents the consistency index, ΔA represents the anomaly set change data, ΔP represents the prediction result change data, and |ΔA-ΔP| represents the difference between the anomaly set change data and the prediction result change data.
[0036] Furthermore, the power distribution central control terminal is interactively connected with the operation acquisition module, the image acquisition module and the report sending module through a wireless network, the operation acquisition module is interactively connected with the fluctuation recognition module through a wireless network, the image acquisition module is interactively connected with the image recognition module through a wireless network, the fault analysis module is interactively connected with the fluctuation recognition module, the image recognition module and the power transmission behavior prediction module through a wireless network, the power transmission behavior prediction module is interactively connected with the structure verification module through a wireless network, and the structure verification module is interactively connected with the report sending module through a wireless network.
[0037] The fault monitoring method of the power transmission and distribution cabinet based on artificial intelligence includes the following steps:
[0038] Step 1: Install an image collector and a data collector inside the power distribution cabinet to obtain current data in real time;
[0039] Step 2: Analyze the collected data to capture abnormal fluctuations that exceed the preset fluctuation threshold;
[0040] Step 3: Record the abnormal fluctuations that exceed the fluctuation threshold to form a data flow anomaly set;
[0041] Step 4: Collect image data of the power distribution equipment and image data of its associated equipment;
[0042] Step 5: Identify and analyze abnormal features of the collected image data, and record the identified abnormal features of the image to form an abnormal set of image streams;
[0043] Step 6: By matching the historical database, analyze the abnormal data in the data stream abnormal set and the image stream abnormal set to obtain the fault type, cause and related equipment content;
[0044] Step 7: Combine current environmental data and equipment data to implement link transmission behavior prediction based on machine learning and data mining technology to predict fault conditions in a specific period of time in the future;
[0045] Step 8: After the prediction is completed, by analyzing the changes in the data stream anomaly set and the image stream anomaly set collected during the blank phase of the prediction operation, verify whether the actual result is consistent with the changes in the data segment in the prediction result, and generate a fault monitoring report.
[0046] (III) Beneficial effects
[0047] Compared with the known prior art, the technical solution provided by the present invention has the following beneficial effects:
[0048] 1. By real-time monitoring of the operating data and image data of the distribution cabinet, and by analyzing the data flow anomaly set and the image flow anomaly set, it is possible to capture abnormal situations in time and warn of faults in advance. By using deep learning algorithms, effective fault prediction can be performed based on the input data to help maintenance personnel take measures in advance without manual intervention, thereby improving efficiency, reducing human errors, preventing and reducing the occurrence of distribution equipment failures, and improving the reliability and safety of the power system.
[0049] 2. By combining the operating data and image data of the distribution cabinet and using image analysis technology, abnormal features in the image data can be detected to further enhance the fault diagnosis capability. The operating data and image data can be fused into a comprehensive data set through data mining technology, and then the characteristics of different data sources can be comprehensively analyzed to improve the accuracy of fault diagnosis and prediction.
[0050] 3. Through the constructed prediction model, the system can predict the power transmission behavior of the specified distribution cabinet and its associated equipment in a specific period of time in the future, so as to formulate maintenance plans in advance and reduce the losses caused by failures. Through matching analysis with the historical database, the definition data of the fault type, cause and associated equipment can be obtained, providing a reference for fault diagnosis and prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0052] Figure 1 It is a schematic diagram of the architecture of the fault monitoring system of the power transmission and distribution cabinet based on artificial intelligence in the present invention;
[0053] Figure 2 It is a flow chart of a fault monitoring method of a power transmission and distribution cabinet based on artificial intelligence in the present invention;
[0054] Figure 3 It is a schematic diagram of the framework of the present invention.
[0055] The numbers in the figure represent: 1. Power distribution central control terminal; 2. Operation acquisition module; 3. Fluctuation recognition module; 4. Image acquisition module; 5. Image recognition module; 6. Fault analysis module; 7. Transmission behavior prediction module; 8. Structure verification module; 9. Report sending module; 10. Data storage repository. DETAILED DESCRIPTION
[0056] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0057] The present invention will be further described below in conjunction with the embodiments.
[0058] Example 1
[0059] The fault monitoring system of the power transmission and distribution cabinet based on artificial intelligence in this embodiment is as follows: Figure 1 and Figure 3 As shown, including:
[0060] The power distribution central control terminal 1 is used as the central control terminal of each power distribution equipment to edit and send control instructions; the power distribution central control terminal 1 is interactively connected to a data storage library 10 through a wireless network, and the data storage library 10 is interactively connected to the fault analysis module 6 through a wireless network. The data storage library 10 is used to store historical fault data and equipment data. When the fault analysis module 6 and the power transmission behavior prediction module 7 are triggered, the data storage library 10 calls the historical database to perform matching analysis of the fault type, cause and related equipment content as a reference for fault diagnosis and prediction;
[0061] Operation collection module 2 is used to collect the current operation data of several power distribution cabinets according to a preset period, and pre-process the acquired operation data to ensure the real-time and accuracy of the data;
[0062] The fluctuation identification module 3 is used to capture abnormal fluctuations of the data collected by the operation collection module 2 that exceed the fluctuation threshold, mark the captured abnormal fluctuations and extract them, and form a number of data flow abnormality sets corresponding to the distribution cabinet targets; it can accurately locate potential fault points and quickly generate data flow abnormality sets, which is helpful to find problems in time;
[0063] The specific operation logic of the fluctuation identification module 3 includes the following steps:
[0064] Step 31: Dynamically set standard fluctuation thresholds based on historical data or custom data;
[0065] Step 32: receiving the collected data submitted by the running collection module 2, removing noise, filling missing values and performing standardization processing to obtain processed data;
[0066] Step 33: Compare the processed data with the standard fluctuation threshold to determine whether it exceeds the set threshold range;
[0067] Step 34: When it is detected that the data exceeds the threshold, it is marked as abnormal fluctuation, and the data points marked as abnormal fluctuation and their timestamps and fluctuation amplitudes are extracted and integrated as a data flow abnormal set;
[0068] The image acquisition module 4 is used to collect image data of the power distribution equipment and its associated equipment according to a preset period, pre-process the acquired operation data, and provide intuitive equipment operation status information;
[0069] The image recognition module 5 is used to analyze the image data collected by the image acquisition module 4, extract abnormal features through machine learning algorithms, and classify the extracted abnormal features to form an image flow abnormality set. The multimodal data fusion method enhances the comprehensiveness of fault detection;
[0070] The specific operation logic of the image recognition module 5 includes the following steps:
[0071] Step 51: performing data cleaning, image enhancement and image normalization processing on the image data collected by the image acquisition module 4;
[0072] Step 52: Use an edge detection algorithm to segment different objects or regions in the image, and extract the device surface area and connection point area of the monitoring target;
[0073] Step 53: extracting geometric features, color features and texture features from the segmented image, and selecting and reducing the dimension of the extracted features through a feature selection algorithm;
[0074] Step 54: using a classification algorithm to classify the image, detect abnormal features, mark the detected abnormal features, and form an image flow abnormality set;
[0075] Fault analysis module 6 is used to obtain the definition data of fault type, cause and associated equipment through data mining based on the data flow anomaly set and the image flow anomaly set, combined with the information in the historical database, to fully understand the cause and impact range of the fault, and improve the accuracy and comprehensiveness of fault diagnosis;
[0076] When performing data mining, the fault analysis module 6 fuses the data stream anomaly set and the image stream anomaly set of the fluctuation recognition module 3 and the image recognition module 5 into a comprehensive data set, extracts key features, including: voltage, current, temperature, vibration frequency and visual features in the image, mines the association rules between the fault data in the comprehensive data set through the analysis of key features, obtains frequently occurring fault combinations, and marks the current fault type and cause in combination with the historical fault type, cause and related equipment data;
[0077] The power transmission behavior prediction module 7 is used to obtain the current environment data, equipment data and definition data obtained by the fault analysis module 6, build a prediction model through a deep learning algorithm, make predictions based on the input data, and output the power transmission behavior of the specified distribution cabinet and its associated equipment in a specific period of time in the future;
[0078] The prediction model constructed by the transmission behavior prediction module 7 uses a long short-term memory network, divides historical data into a training set and a validation set, uses the training set to train the model, uses the validation set to validate the trained model, and evaluates the performance of the model. When the evaluated model is running, it receives the input of current environmental data and equipment data, generates the transmission behavior prediction results for a specific period in the future, and outputs the prediction results including: future operating status, fault time point and fault type; this makes the system more advantageous in processing time series data and the prediction results more accurate;
[0079] The structure verification module 8 is used to be triggered after the power transmission behavior prediction module 7 is predicted, analyze the change data of the data flow anomaly set and the image flow anomaly set collected during the prediction operation blank phase of the power transmission behavior prediction module 7, verify whether such change data in the anomaly set is consistent with the change of the same data segment in the prediction result, and output the credibility coefficient of the verification and evaluation prediction result; thereby helping to dynamically adjust and optimize the prediction model to ensure the reliability of the prediction result;
[0080] The report sending module 9 is used to integrate the fault analysis results and the credibility coefficients of the prediction results to generate a fault monitoring report, and submit it to the power distribution central control terminal 1. The integrated report helps decision makers to fully understand the operating status of the system and quickly make maintenance and repair decisions.
[0081] As a preferred implementation in this embodiment, Figure 1As shown, the power distribution central control terminal 1 is interactively connected with the operation acquisition module 2, the image acquisition module 4 and the report sending module 9 through a wireless network, the operation acquisition module 2 is interactively connected with the fluctuation recognition module 3 through a wireless network, the image acquisition module 4 is interactively connected with the image recognition module 5 through a wireless network, the fault analysis module 6 is interactively connected with the fluctuation recognition module 3, the image recognition module 5 and the power transmission behavior prediction module 7 through a wireless network, the power transmission behavior prediction module 7 is interactively connected with the structure verification module 8 through a wireless network, and the structure verification module 8 is interactively connected with the report sending module 9 through a wireless network.
[0082] Compared with existing technologies, the system can obtain the operating status in a timely and accurate manner, avoiding misjudgment caused by delayed or inaccurate data. By dynamically setting the standard fluctuation threshold and combining historical data with custom data, the system makes anomaly detection more flexible and accurate. It provides intuitive equipment operating status information through image data, enhancing the comprehensiveness of fault detection.
[0083] By fusing the data stream anomaly set and the image stream anomaly set, and combining the information in the historical database for data mining, we can fully understand the cause and impact range of the fault, and improve the accuracy and comprehensiveness of fault diagnosis. Multimodal data fusion can provide richer information and improve the effect of fault detection and diagnosis. By analyzing the change data of the data stream anomaly set and the image stream anomaly set, we can verify the credibility coefficient of the prediction result, dynamically adjust and optimize the prediction model, and ensure the reliability of the prediction result.
[0084] Example 2
[0085] In other aspects, this embodiment also provides a fault monitoring method for a power transmission and distribution cabinet based on artificial intelligence, such as Figure 2 As shown, the following steps are included:
[0086] Step 1: Install an image collector and a data collector inside the power distribution cabinet to obtain current data in real time to reflect the operating status and performance of the power distribution cabinet;
[0087] Step 2: Analyze the collected data to capture abnormal fluctuations that exceed the preset fluctuation threshold. These abnormal fluctuations indicate potential failures or abnormal conditions.
[0088] Step 3: Record abnormal fluctuations that exceed the fluctuation threshold to form a data flow anomaly set, and use the abnormal data for subsequent fault diagnosis and predictive analysis;
[0089] Step 4: Collect image data of the power distribution equipment and its associated equipment. Image data can provide more intuitive information to assist in fault identification and analysis.
[0090] Step 5: Identify and analyze abnormal features of the collected image data, and record the identified abnormal features of the image to form an abnormal set of image streams;
[0091] Step 6: By matching the historical database, analyze the abnormal data in the data stream abnormal set and the image stream abnormal set to obtain the fault type, cause and related equipment content;
[0092] Step 7: Combine current environmental data and equipment data to implement link transmission behavior prediction based on machine learning and data mining technology to predict fault conditions in a specific period of time in the future and take measures in advance to repair or prevent them;
[0093] Step 8: After the prediction is completed, by analyzing the changes in the data stream anomaly set and the image stream anomaly set collected during the blank phase of the prediction operation, verify whether the actual result is consistent with the changes in the data segment in the prediction result, and generate a fault monitoring report.
[0094] Compared with the prior art, this embodiment can monitor the operating status of the distribution cabinet in real time and respond quickly when an abnormality is detected, thereby improving the safety and stability of the distribution cabinet. By combining data stream and image stream data, the comprehensiveness and accuracy of fault detection are improved. Intelligent prediction is performed through a deep learning model, and the prediction results are verified, thereby improving the reliability of the prediction. The generated fault monitoring report is detailed and comprehensive, providing strong support for maintenance decisions. The historical database provides rich fault data and equipment data, enhancing the diagnosis and prediction capabilities of distribution cabinet faults.
[0095] Example 3
[0096] In this embodiment, a method for calculating the consistency index is provided. In the process of verifying whether the change data in the abnormal set is consistent with the change of the same data segment in the prediction result, the expression is:
[0097]
[0098] Among them, CI represents the consistency index, which indicates the degree of consistency between the abnormal set change data and the predicted result change data. The value range of the consistency index is between 0 and 1. The closer the value is to 1, the higher the consistency. ΔA represents the abnormal set change data, which indicates the change of the data collected in the abnormal set during the prediction period and the blank period. ΔP represents the predicted result change data, which indicates the data change of the specified distribution cabinet and its associated equipment in the predicted result within a specific period of time. |ΔA-ΔP| represents the difference between the abnormal set change data and the predicted result change data.
[0099] In summary, the system of the present invention that performs fault detection on the distribution cabinet through artificial intelligence technology can collect various current data of the distribution cabinet, capture abnormal fluctuations that exceed the fluctuation threshold, constitute a data flow anomaly set, collect image data of the distribution equipment and image data of its associated equipment, identify and analyze abnormal features of the obtained image data, constitute an image flow anomaly set, match the historical database for the data flow anomaly set and the image flow anomaly data set, analyze the fault type, cause and associated equipment content, implement the link transmission behavior prediction based on the current environmental data and equipment data, and after the prediction is completed, analyze the changes in the data flow anomaly set and the image flow anomaly set collected during the blank stage of the prediction operation to verify whether they conform to the changes in the data segment in the prediction result, input the credibility of the prediction result, and output a fault monitoring report.
[0100] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. The fault monitoring system of power transmission and distribution cabinet based on artificial intelligence is characterized by: include: The power distribution central control terminal (1) is used as the central control terminal of each power distribution device to edit and send control instructions; An operation collection module (2) is used to collect current operation data of a number of power distribution cabinets according to a preset period and pre-process the acquired operation data; A fluctuation identification module (3) is used to capture abnormal fluctuations of the data collected by the operation collection module (2) that exceed the fluctuation threshold, mark the captured abnormal fluctuations and extract them, and form a data flow abnormality set corresponding to a number of distribution cabinet targets; An image acquisition module (4) is used to acquire image data of the power distribution equipment and its associated equipment according to a preset period, and pre-process the acquired operating data; An image recognition module (5) is used to analyze the image data collected by the image collection module (4), extract abnormal features through a machine learning algorithm, and classify the extracted abnormal features to form an image flow abnormality set; A fault analysis module (6) is used to obtain definition data of fault type, cause and associated equipment through data mining based on the data flow anomaly set and the image flow anomaly set in combination with information in the historical database; The power transmission behavior prediction module (7) is used to obtain the current environment data, equipment data and definition data obtained by the fault analysis module (6), build a prediction model through a deep learning algorithm, make predictions based on the input data, and output the power transmission behavior of the specified distribution cabinet and its associated equipment in a specific period of time in the future; A structure verification module (8) is used to be triggered after the power transmission behavior prediction module (7) is predicted, analyze the change data of the data flow anomaly set and the image flow anomaly set collected during the prediction operation blank phase of the power transmission behavior prediction module (7), verify whether such change data in the anomaly set is consistent with the change of the same data segment in the prediction result, and output the credibility coefficient of the prediction result for verification and evaluation; The report sending module (9) is used to integrate the fault analysis results and the credibility coefficients of the prediction results to generate a fault monitoring report and submit it to the power distribution central control terminal (1).
2. The fault monitoring system for power transmission and distribution cabinet based on artificial intelligence according to claim 1 is characterized in that: The power distribution central control terminal (1) is interactively connected to a data repository (10) via a wireless network. The data repository (10) is interactively connected to a fault analysis module (6) via a wireless network. The data repository (10) is used to store historical fault data and equipment data. When the fault analysis module (6) and the power transmission behavior prediction module (7) are triggered, the data repository (10) calls the historical database to perform matching analysis of the fault type, cause and associated equipment content as a reference for fault diagnosis and prediction.
3. The fault monitoring system for power transmission and distribution cabinet based on artificial intelligence according to claim 1 is characterized in that: The specific operation logic of the fluctuation identification module (3) includes the following steps: Step 31: Dynamically set standard fluctuation thresholds based on historical data or custom data; Step 32: receiving the collected data submitted by the running collection module (2), removing noise, filling missing values and performing standardization processing to obtain processed data; Step 33: Compare the processed data with the standard fluctuation threshold to determine whether it exceeds the set threshold range; Step 34: When it is detected that the data exceeds the threshold, it is marked as abnormal fluctuation, and the data points marked as abnormal fluctuations and their timestamps and fluctuation amplitudes are extracted and integrated as a data flow anomaly set.
4. The fault monitoring system for power transmission and distribution cabinet based on artificial intelligence according to claim 1 is characterized in that: The specific operation logic of the image recognition module (5) includes the following steps: Step 51: performing data cleaning, image enhancement and image normalization processing on the image data collected by the image acquisition module (4); Step 52: Use an edge detection algorithm to segment different objects or regions in the image, and extract the device surface area and connection point area of the monitoring target; Step 53: extracting geometric features, color features and texture features from the segmented image, and selecting and reducing the dimension of the extracted features through a feature selection algorithm; Step 54: Use a classification algorithm to classify the image, detect abnormal features, mark the detected abnormal features, and form an image flow abnormality set.
5. The fault monitoring system for power transmission and distribution cabinet based on artificial intelligence according to claim 1 is characterized in that: When performing data mining, the fault analysis module (6) fuses the data flow anomaly set and the image flow anomaly set of the fluctuation recognition module (3) and the image recognition module (5) into a comprehensive data set, extracts key features, including: voltage, current, temperature, vibration frequency and visual features in the image, mines the association rules between the fault data in the comprehensive data set through analysis of the key features, obtains frequently occurring fault combinations, and marks the current fault type and cause in combination with the historical fault type, cause and associated equipment data.
6. The fault monitoring system for power transmission and distribution cabinet based on artificial intelligence according to claim 1 is characterized in that: The prediction model constructed by the power transmission behavior prediction module (7) uses a long short-term memory network, divides historical data into a training set and a verification set, uses the training set to train the model, uses the verification set to verify the trained model, evaluates the performance of the model, and receives the input of current environmental data and equipment data during operation of the evaluated model to generate a power transmission behavior prediction result for a specific period in the future, and outputs the prediction result including: future operating status, fault time point and fault type.
7. The fault monitoring system for power transmission and distribution cabinet based on artificial intelligence according to claim 1 is characterized in that: The structure verification module (8) calculates the consistency index in the process of verifying whether the change data in the abnormal set is consistent with the change of the same data segment in the prediction result, and its expression is: Among them, CI represents the consistency index, ΔA represents the anomaly set change data, ΔP represents the prediction result change data, and |ΔA-ΔP| represents the difference between the anomaly set change data and the prediction result change data.
8. The fault monitoring system for power transmission and distribution cabinet based on artificial intelligence according to claim 1 is characterized in that: The power distribution central control terminal (1) is interactively connected with the operation acquisition module (2), the image acquisition module (4) and the report sending module (9) via a wireless network; the operation acquisition module (2) is interactively connected with the fluctuation recognition module (3) via a wireless network; the image acquisition module (4) is interactively connected with the image recognition module (5) via a wireless network; the fault analysis module (6) is interactively connected with the fluctuation recognition module (3), the image recognition module (5) and the power transmission behavior prediction module (7) via a wireless network; the power transmission behavior prediction module (7) is interactively connected with the structure verification module (8) via a wireless network; and the structure verification module (8) is interactively connected with the report sending module (9) via a wireless network.
9. A fault monitoring method for a power transmission and distribution cabinet based on artificial intelligence, wherein the system is an implementation method of the fault monitoring system for a power transmission and distribution cabinet based on artificial intelligence according to any one of claims 1 to 8, characterized in that: The following steps are involved: Step 1: Install an image collector and a data collector inside the power distribution cabinet to obtain current data in real time; Step 2: Analyze the collected data to capture abnormal fluctuations that exceed the preset fluctuation threshold; Step 3: Record the abnormal fluctuations that exceed the fluctuation threshold to form a data flow anomaly set; Step 4: Collect image data of the power distribution equipment and image data of its associated equipment; Step 5: Identify and analyze abnormal features of the collected image data, and record the identified abnormal features of the image to form an abnormal set of image streams; Step 6: By matching the historical database, analyze the abnormal data in the data stream abnormal set and the image stream abnormal set to obtain the fault type, cause and related equipment content; Step 7: Combine current environmental data and equipment data to implement link transmission behavior prediction based on machine learning and data mining technology to predict fault conditions in a specific period of time in the future; Step 8: After the prediction is completed, by analyzing the changes in the data stream anomaly set and the image stream anomaly set collected during the blank phase of the prediction operation, verify whether the actual result is consistent with the changes in the data segment in the prediction result, and generate a fault monitoring report.
Citation Information
Patent Citations
Monitoring power distribution cabinet for automatically detecting and alarming, and line fault judgment method
CN110470939A
Equipment evaluation method and system based on LSTM (Long Short Term Memory)
CN114444607A
Intelligent equipment management method and system based on big data analysis
CN116862442A
Power distribution cabinet operation state monitoring management system based on big data
CN116914917A
Intelligent early warning system and early warning method for multifunctional power distribution cabinet
CN117477774A
Cited By
Self-diagnosis system of intelligent power distribution cabinet
CN119231761A
A self-diagnosis system for intelligent power distribution cabinet
CN119231761B
Electric inspection well safety early warning method and system
CN119599642A
Heterogeneous industrial equipment-oriented adaptive data fusion and decision optimization system
CN119887177A
Adaptive Data Fusion and Decision Optimization System for Heterogeneous Industrial Equipment
CN119887177B
