Cable insulation on-line monitoring method based on tree topology structure

Abstract

The invention belongs to the technical field of cable insulation monitoring, and provides a cable insulation online monitoring method based on a tree topology structure, which comprises the following steps: constructing a tree structure comprising a trunk cable and a plurality of branch cables, the trunk cable serving as a root node of the tree structure, the branch cables serving as sub-nodes, and the branch cables serving as sub-nodes; each branch cable is connected with one or more cables as grandson nodes; installing a sensor on each cable node; all the sensors are connected to a central monitoring unit through a communication network to form a complete tree topology structure monitoring network; the central monitoring unit serves as the core of data processing and control, and adopts a convolutional neural network to process and analyze electrical parameter data collected from the sensors so as to identify the change trend of the insulation state of the cable. According to the method, diagnosis and early warning of extremely early hidden dangers of cable insulation faults are realized, and accidents such as line tripping and fire disasters caused by faults caused by insulation problems are avoided.

Description

Technical Field

[0001] The invention belongs to the technical field of cable insulation monitoring, and relates to an online cable insulation monitoring method based on a tree topology structure. Background Art

[0002] In the current distribution network construction process in my country, the tree topology has been increasingly used. As an important component of the distribution network, the cable has a strong adaptability to the laying environment, which largely ensures the safety of the workers after the cable is powered on. Therefore, it has been more and more widely used in the distribution network construction process. However, in the distribution network with a tree topology structure, most of the cables work in a humid environment with large changes. Cables working in a humid environment are prone to partial discharge due to extrusion, forming electrical trees caused by water trees, and eventually breaking through the cable insulation layer, and even forming single-phase grounding, forming arcs, and causing accidents, which seriously threaten the safety and reliability of the power supply of the distribution network.

[0003] At present, the online monitoring of cable insulation in my country mainly uses the preventive test method, which has two major defects: first, it is impossible to perform power outage detection on a specific cable, and it can only conduct centralized detection on cable defects in the entire distribution network, resulting in the formation of "rectification effect" and "cumulative effect" of cables that have not experienced degradation problems due to frequent power outages, causing additional insulation aging; second, the preventive test method can only be implemented under power outage conditions, and frequent detection will inevitably cause frequent power outages, which will greatly affect the safety of power grid operation; at the same time, there is currently no scientific and complete reference formula for the test voltage of the preventive test method, which has also brought many inconveniences to the practical application of the preventive test method.

[0004] Based on this research, a cable insulation online monitoring method is studied to realize the diagnosis and early warning of hidden dangers of cable insulation failure, and to avoid line tripping, fire and other incidents caused by failures due to insulation problems, which is of great significance. Summary of the invention

[0005] The present invention proposes an online cable insulation monitoring method based on a tree topology structure, which realizes the diagnosis and early warning of the hidden dangers of cable insulation failure.

[0006] The technical solution of the present invention is achieved in this way:

[0007] A cable insulation online monitoring method based on a tree topology structure is provided. A tree structure including a trunk cable and multiple branch cables is constructed, wherein the trunk cable is used as the root node of the tree structure, the branch cables are used as child nodes, and each branch cable is connected to one or more cables as grandchild nodes; a sensor is installed on each cable node to monitor the current, voltage and insulation resistance electrical parameters of the node in real time; all sensors are connected to a central monitoring unit through a communication network to form a complete tree topology structure monitoring network; the central monitoring unit is used as the core of data processing and control and is responsible for receiving and processing data from each child node and grandchild node; at the same time, the central monitoring unit uses a convolutional neural network to process and analyze the electrical parameter data collected from the sensor to identify the change trend of the cable insulation state.

[0008] Preferably, the sensors include a current sensor, a voltage sensor and an insulation resistance sensor, which are used to monitor the current, voltage and insulation resistance value of the cable segment in real time.

[0009] Preferably, each cable segment in the tree topology is also provided with identification information for identifying properties such as the location, length and type of the cable segment, so as to facilitate subsequent data analysis and processing.

[0010] Preferably, the convolutional neural network model includes an input layer, multiple convolutional layers, a pooling layer, a fully connected layer and an output layer, wherein the input layer receives data from a sensor, the convolutional layer is used to extract features of the data, the pooling layer is used to reduce the dimension of the data, the fully connected layer is used to integrate features and perform classification or regression, and the output layer outputs the judgment result of the cable insulation status.

[0011] Preferably, the convolutional neural network model also includes using historical insulation monitoring data and known insulation status labels to train the convolutional neural network model to optimize the weight and bias parameters of the model and improve the accuracy of insulation status identification.

[0012] Preferably, the convolutional neural network model also includes inputting the insulation monitoring data collected in real time into a trained convolutional neural network model, the model outputs a judgment result of the cable insulation status, and generates an insulation status assessment report based on the judgment result.

[0013] Preferably, the insulation status assessment report includes information such as the grade of the cable insulation status, degradation rate, estimated remaining life, and possible fault locations and causes.

[0014] Preferably, when the central monitoring unit identifies that the cable insulation state is abnormal, an alarm signal is triggered to remind the operation and maintenance personnel to deal with it in time, and record relevant information of the abnormal event, including the time, location and cause of the abnormality.

[0015] Preferably, the data received by the central monitoring unit is pre-processed, including data cleaning and standardization, to remove noise and outliers.

[0016] Preferably, each of the child nodes and grandchild nodes includes a battery module and a wireless communicator, and the battery module provides working power for the sensor and the wireless communicator.

[0017] The working principle and beneficial effects of the present invention are:

[0018] 1. The cable insulation online monitoring method based on the tree topology structure provided by the present invention has multiple advantages such as high efficiency, accuracy, scalability, flexibility, easy maintenance and cost reduction. These advantages make the method have broad application prospects and important practical value in the field of cable insulation monitoring.

[0019] 2. The method provided by the present invention adopts a tree topology structure, in which the trunk cable is used as the root node, the branch cable is used as the child node, and each branch cable can be connected to multiple grandchild nodes. This structure enables the monitoring system to efficiently cover the entire cable network, ensuring that each cable node can be monitored in real time; the sensors installed on each cable node can monitor electrical parameters such as current, voltage and insulation resistance in real time, which are important indicators for evaluating the insulation status of the cable; the central monitoring unit uses a convolutional neural network to process and analyze the collected electrical parameter data, which can accurately identify the changing trend of the cable insulation status. This intelligent analysis method improves the accuracy and reliability of monitoring; the tree topology structure provides high scalability, allowing new cable nodes to be easily added to the existing network, which enables the monitoring system to adapt to the changing needs of the cable network; at the same time, the structure also provides flexibility, allowing the layout and configuration of the monitoring network to be adjusted according to actual needs; the tree topology structure makes the monitoring system easy to maintain. If a cable node fails, it can be quickly located and repaired without affecting the operation of the entire monitoring network; in addition, by real-time monitoring and analysis of electrical parameter data, the central monitoring unit can promptly detect potential safety hazards and take corresponding preventive measures to avoid accidents; compared with traditional periodic power outage preventive tests, this method can monitor the insulation status of the cable online in real time, significantly reducing the power outage time of electrical equipment, thereby reducing operation and maintenance costs; at the same time, this method can also improve monitoring efficiency, allowing operation and maintenance personnel to discover problems more quickly and take corresponding treatment measures to ensure the stable operation of the cable network; by real-time monitoring of the insulation status of the cable, the method can promptly detect and deal with potential insulation defects and safety hazards, thereby effectively preventing the occurrence of cable failures; this not only improves the safety of the cable network, but also ensures the reliability and stability of power supply, which is of great significance for maintaining the normal operation of the power system. DETAILED DESCRIPTION

[0020] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 creative work are within the scope of protection of the present invention.

[0021] Example 1

[0022] A cable insulation online monitoring method based on a tree topology structure, comprising:

[0023] Construct a tree structure including a trunk cable and multiple branch cables, wherein the trunk cable serves as a root node of the tree structure, the branch cables serve as child nodes, and each branch cable is connected to one or more cables as grandchild nodes;

[0024] A sensor is installed on each cable node to monitor the current, voltage and insulation resistance electrical parameters of the node in real time; each child node and grandchild node includes a battery module and a wireless communicator, the battery module provides working power for the sensor and the wireless communicator, and the sensor includes a current sensor, a voltage sensor and an insulation resistance sensor, which are used to monitor the current, voltage and insulation resistance value of the cable segment in real time;

[0025] All sensors are connected to the central monitoring unit through a communication network to form a complete tree-type topology monitoring network. As the core of data processing and control, the central monitoring unit is responsible for receiving and processing data from each child node and grandchild node. At the same time, the central monitoring unit uses a convolutional neural network to process and analyze the electrical parameter data collected from the sensors to identify the changing trend of the cable insulation status.

[0026] The specific steps are as follows:

[0027] Uniquely number each cable segment and sensor to identify the location, length and type attributes of the cable segment;

[0028] Cluster the nodes where the sensors are located according to their geographical locations;

[0029] The sensor node sends and receives information to the adjacent sensor nodes in the same cluster in parallel through cables; the information is uploaded to the central monitoring unit step by step through the tree topology structure, and finally the aggregated information is sent to the central server;

[0030] Based on the received information, the central server performs preprocessing, including data cleaning and standardization, to remove noise and outliers. Then, it uses a convolutional neural network to process and analyze the electrical parameter data collected from the sensors to identify the changing trend of the cable insulation status, thereby monitoring the status of the cable insulation.

[0031] When the central monitoring unit identifies an abnormality in the cable insulation status, it triggers an alarm signal to remind the operation and maintenance personnel to deal with it in time and record relevant information of the abnormal event, including the time, location and cause of the abnormality.

[0032] The convolutional neural network model includes an input layer, multiple convolutional layers, a pooling layer, a fully connected layer and an output layer, wherein the input layer receives data from a sensor, the convolutional layer is used to extract features of the data, the pooling layer is used to reduce the dimension of the data, the fully connected layer is used to integrate features and perform classification or regression, and the output layer outputs a judgment result of the cable insulation status; the convolutional neural network model also includes training the convolutional neural network model using historical insulation monitoring data and known insulation status labels to optimize the weight and bias parameters of the model and improve the accuracy of insulation status recognition; at the same time, the insulation monitoring data collected in real time is input into the trained convolutional neural network model, the model outputs the judgment result of the cable insulation status, and generates an insulation status assessment report based on the judgment result; the insulation status assessment report includes information such as the grade of the cable insulation status, degradation rate, expected remaining life, and possible fault locations and causes.

[0033] Among them: The establishment of the convolutional neural network database includes the following steps:

[0034] Collect data: Collect a large number of samples from historical insulation monitoring data, including data under various insulation conditions.

[0035] Label data: Label the collected data with known insulation status labels to ensure that each sample has a corresponding insulation status.

[0036] Data storage: Store the labeled data in the database to facilitate subsequent training and testing.

[0037] The construction of the convolutional neural network model includes the following steps:

[0038] Input layer: Receives data from the sensor and sets the size of the input layer according to the dimension of the data.

[0039] Convolutional layers: Multiple convolutional layers are used to extract features of the data, each of which is usually followed by an activation function (such as ReLU) and possible batch normalization layers.

[0040] Pooling layer: Pooling layers are inserted between convolutional layers to reduce the dimension of the data and reduce the amount of computation.

[0041] Fully connected layer: After the convolutional layer and the pooling layer, a fully connected layer is used to integrate features and perform classification or regression.

[0042] Output layer: Set the number of neurons in the output layer according to the task requirements (such as binary classification) and use an appropriate activation function (such as softmax).

[0043] Convolutional neural network model training and optimization includes the following steps:

[0044] The model uses the training data in the database to train the model and optimize the weight and bias parameters of the model.

[0045] Use validation data to monitor the performance of the model and prevent overfitting.

[0046] Example 2

[0047] A cable insulation online monitoring method based on a tree topology structure, comprising:

[0048] Construct a tree structure including a trunk cable and multiple branch cables, wherein the trunk cable serves as the root node (central node) of the tree structure, the branch cables serve as child nodes, and each branch cable is connected to one or more cables as grandchild nodes;

[0049] A sensor is installed on each cable node to monitor the current, voltage and insulation resistance electrical parameters of the node in real time; each child node and grandchild node includes a battery module and a wireless communicator, the battery module provides working power for the sensor and the wireless communicator, and the sensor includes a current sensor, a voltage sensor and an insulation resistance sensor, which are used to monitor the current, voltage and insulation resistance value of the cable segment in real time;

[0050] All sensors are connected to the central monitoring unit through the communication network to form a complete tree-type topology monitoring network. The central monitoring unit, as the core of data processing and control, is responsible for receiving and processing data from each child node and grandchild node. At the same time, the central monitoring unit uses a convolutional neural network to process and analyze the electrical parameter data collected from the sensors to identify the changing trend of the cable insulation status.

[0051] The central monitoring unit sets up a cloud server to store and manage large amounts of monitoring data. The cloud server and the central node exchange data through a secure communication protocol to ensure data security and privacy protection. The cloud server also supports remote access, allowing operation and maintenance personnel to view real-time monitoring data and historical records through mobile devices or computer terminals. The cloud server provides data analysis tools to help operation and maintenance personnel perform trend analysis and fault prediction.

[0052] The convolutional neural network model is improved by using an improved AlexNet model. The convolutional neural network database is established in the same manner as in Example 1.

[0053] The construction of the convolutional neural network model includes the following steps:

[0054] The model is improved based on AlexNet, retaining its basic convolutional layer, pooling layer, fully connected layer and output layer structure.

[0055] Increase the number or depth of convolutional layers to extract more levels of features.

[0056] Use a larger convolution kernel or a smaller stride to capture more details.

[0057] A dropout layer is introduced to reduce overfitting.

[0058] Convolutional neural network model training and optimization includes the following steps:

[0059] Use the improved AlexNet model for training, and also use the training data and validation data in the database.

[0060] Try different hyperparameters like learning rate, optimizer, batch size, etc. to find the best training configuration.

[0061] Example 3

[0062] A cable insulation online monitoring method based on a tree topology structure, comprising:

[0063] Construct a tree structure including a trunk cable and multiple branch cables, wherein the trunk cable serves as a root node of the tree structure, the branch cables serve as child nodes, and each branch cable is connected to one or more cables as grandchild nodes;

[0064] A sensor is installed on each cable node to monitor the current, voltage and insulation resistance electrical parameters of the node in real time; each child node and grandchild node includes a battery module and a wireless communicator, the battery module provides working power for the sensor and the wireless communicator, and the sensor includes a current sensor, a voltage sensor and an insulation resistance sensor, which are used to monitor the current, voltage and insulation resistance value of the cable segment in real time;

[0065] All sensors are connected to the central monitoring unit through the communication network to form a complete tree-type topology monitoring network; the central monitoring unit, as the core of data processing and control, is responsible for receiving and processing data from each child node and grandchild node; at the same time, the central monitoring unit uses convolutional neural networks to process and analyze the electrical parameter data collected from the sensors to identify the changing trend of the cable insulation status; the central monitoring unit will also provide specific maintenance suggestions based on the analysis results of the convolutional neural network to help operation and maintenance personnel take timely measures to prevent failures. The communication between child nodes and grandchild nodes is achieved through self-organizing network technology to improve the flexibility and reliability of the system; self-organizing network technology enables nodes to automatically discover and connect to nearby nodes to form a dynamic network topology; supports multiple communication protocols, including but not limited to Wi-Fi, Zigbee, LoRa, etc., to adapt to different application scenarios and environmental conditions; the choice of communication protocol can be flexibly configured according to actual needs to ensure the stability and efficiency of data transmission.

[0066] The convolutional neural network model combines the CNN model of the feedback mechanism with the dynamic update of the database. The convolutional neural network database is established in the same manner as in Example 1; a dynamic update mechanism is introduced to allow the model to continuously collect new insulation monitoring data during real-time operation, and automatically mark (or manually review and mark) and add it to the database.

[0067] Convolutional neural network model construction:

[0068] Build a basic CNN model and introduce a feedback mechanism to adjust the size and number of convolution kernels.

[0069] Based on entropy theory, a semantic error information entropy measurement evaluation index is established to evaluate the uncertain detection results of insulators in real time.

[0070] According to the evaluation results of the semantic error information entropy measurement index, the size and number of convolution kernels are dynamically adjusted through the constructed feedback mechanism to refine the extracted features.

[0071] Convolutional neural network model training and optimization:

[0072] Use the data from the initial database for training.

[0073] During real-time operation, new data is continuously collected and the database is updated.

[0074] Retrain models periodically or on demand to adapt to new data distributions and insulation state changes.

[0075] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A cable insulation online monitoring method based on a tree topology structure, characterized in that: Construct a tree structure including a trunk cable and multiple branch cables, wherein the trunk cable serves as a root node of the tree structure, the branch cables serve as child nodes, and each branch cable is connected to one or more cables as grandchild nodes; Sensors are installed on each cable node to monitor the current, voltage and insulation resistance electrical parameters of the node in real time. All sensors are connected to the central monitoring unit through the communication network to form a complete tree-type topology monitoring network. The central monitoring unit, as the core of data processing and control, is responsible for receiving and processing data from each child node and grandchild node. At the same time, the central monitoring unit uses a convolutional neural network to process and analyze the electrical parameter data collected from the sensors to identify the changing trend of the cable insulation status.

2. The cable insulation online monitoring method based on a tree topology structure according to claim 1 is characterized in that: The sensors include a current sensor, a voltage sensor and an insulation resistance sensor, which are used to monitor the current, voltage and insulation resistance value of the cable segment in real time.

3. The cable insulation online monitoring method based on a tree topology structure according to claim 1 is characterized in that: Each cable segment in the tree topology is also provided with identification information for identifying the location, length and type attributes of the cable segment.

4. The cable insulation online monitoring method based on a tree topology structure according to claim 1 is characterized in that: The convolutional neural network model includes an input layer, multiple convolutional layers, a pooling layer, a fully connected layer and an output layer, wherein the input layer receives data from a sensor, the convolutional layer is used to extract features of the data, the pooling layer is used to reduce the dimension of the data, the fully connected layer is used to integrate features and perform classification or regression, and the output layer outputs the judgment result of the cable insulation status.

5. The cable insulation online monitoring method based on a tree topology structure according to claim 4 is characterized in that: The convolutional neural network model also includes using historical insulation monitoring data and known insulation status labels to train the convolutional neural network model to optimize the weight and bias parameters of the model and improve the accuracy of insulation status identification.

6. The cable insulation online monitoring method based on a tree topology structure according to claim 4 is characterized in that: The convolutional neural network model also includes inputting the insulation monitoring data collected in real time into the trained convolutional neural network model, the model outputs the judgment result of the cable insulation status, and generates an insulation situation assessment report based on the judgment result.

7. The cable insulation online monitoring method based on a tree topology structure according to claim 6 is characterized in that: The insulation status assessment report includes information such as the grade of the cable insulation status, the rate of degradation, the estimated remaining life, and possible fault locations and causes.

8. The cable insulation online monitoring method based on a tree topology structure according to claim 1 is characterized in that: When the central monitoring unit identifies an abnormality in the cable insulation status, it triggers an alarm signal to remind the operation and maintenance personnel to deal with it in time and record relevant information of the abnormal event, including the time, location and cause of the abnormality.

9. The cable insulation online monitoring method based on a tree topology structure according to claim 1 is characterized in that: The data received by the central monitoring unit undergoes preprocessing, including data cleaning and standardization, to remove noise and outliers.

10. The cable insulation online monitoring method based on a tree topology structure according to claim 1, characterized in that: Each of the child nodes and grandchild nodes includes a battery module and a wireless communicator, and the battery module provides working power for the sensor and the wireless communicator.