A system and method for remotely monitoring the power supply state of a ring main unit

Abstract

The application discloses a system and method for remotely monitoring power supply states of ring network cabinets, relates to the technical field of ring network cabinet data monitoring, and comprises the following steps: taking ring network cabinet power supply characteristic data sets as input to perform forward propagation calculation, obtaining health scores and failure probabilities of the power supply states of the ring network cabinets, using a dynamic adjustment mechanism to dynamically calibrate the health scores and the failure probabilities, obtaining the calibrated health scores and the failure probabilities, adopting a weighted decision algorithm on the calibrated health scores and the failure probabilities, obtaining comprehensive health evaluation results of the power supply states of the ring network cabinets, and generating risk levels and monitoring reports according to the comprehensive health evaluation results of the power supply states of the ring network cabinets and pushing the risk levels and the monitoring reports to operation and maintenance terminals. Through data analysis, the power supply state changes of the ring network cabinets under different time scales are captured, the monitoring reports are directly pushed to mobile terminals of operation and maintenance personnel, and remote monitoring is realized.

Description

Technical Field

[0001] This invention relates to power monitoring and remote control, and more specifically to the field of ring main unit power supply, and more specifically to a system and method for remotely monitoring the power supply status of ring main units. Background Technology

[0002] With the continuous advancement of power monitoring technology, ring main units (RNBs), as key components in distribution networks, are crucial for ensuring power quality due to their reliability and stability. In recent years, the development of IoT, big data analytics, and artificial intelligence technologies has improved the technology for remotely monitoring the power supply status of RNBs. By deploying advanced sensors on RNBs, real-time monitoring and data acquisition of their operating parameters can be achieved. Furthermore, leveraging low-power wide-area network (LPWAN) communication protocols, this collected data can be efficiently and securely transmitted to a remote control center for further processing and analysis, significantly enhancing the automation level and intelligent management capabilities of power units.

[0003] However, existing technical solutions for remotely monitoring the power supply status of ring main units still have room for improvement. First, in terms of data analysis, previous methods typically only use a single data processing approach, which cannot effectively integrate information from multiple sources and is difficult to comprehensively reflect the actual operating status of the ring main unit. Second, in terms of model building, most existing methods use basic classification models that lack adaptability and generalization ability. When faced with different operating conditions and new equipment, it is often necessary to retrain the model, which leads to greater monitoring difficulties. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a method for remotely monitoring the power supply status of ring main units to solve the problems of limited power supply monitoring data and high difficulty in status monitoring.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a method for remotely monitoring the power supply status of a ring main unit, comprising,

[0008] The power supply parameters of the ring main unit are collected in real time, preprocessed, and the power supply data of the ring main unit is obtained and transmitted to the remote control center.

[0009] A multimodal feature matrix is ​​constructed based on the power supply data of the ring main unit to obtain the power supply feature dataset of the ring main unit;

[0010] By applying the transfer learning mechanism, a residual connection structure is added to the intermediate layer of the basic classification model to obtain a fault prediction model. The power supply feature dataset of the ring main unit is used as input for forward propagation calculation to obtain the health score and fault probability of the power supply status of the ring main unit.

[0011] A dynamic adjustment mechanism is used to dynamically calibrate the health score and failure probability to obtain the calibrated health score and failure probability.

[0012] A weighted decision algorithm is used to calculate the comprehensive health assessment result of the ring main unit's power supply status by applying a weighted decision algorithm to the calibrated health score and fault probability.

[0013] Based on the comprehensive health assessment results of the ring main unit's power supply status, a risk level and monitoring report are generated and pushed to the operation and maintenance terminal.

[0014] As a preferred embodiment of the remote monitoring method for the power supply status of a ring main unit according to the present invention, the real-time acquisition of the ring main unit's power supply parameters, preprocessing to obtain the ring main unit's power supply data, and transmission to the remote control center specifically involves...

[0015] Sensors are installed at the power supply location of the ring main unit to collect voltage, current and power parameters from the ring main unit in real time.

[0016] Voltage, current, and power parameters are sent to edge computing nodes for noise reduction and normalization to obtain the power supply data of the ring main unit.

[0017] As a preferred embodiment of the remote monitoring method for the power supply status of a ring main unit according to the present invention, the step of constructing a multimodal feature matrix based on the power supply data of the ring main unit to obtain a power supply feature dataset of the ring main unit specifically involves...

[0018] Set N time windows, and use the N time windows sequentially to perform statistical analysis on the power supply data of the ring main unit to obtain the characteristics of the power supply data of the ring main unit;

[0019] The power supply data features of the ring main unit are combined into a power supply feature vector and arranged in order to form a multi-row, multi-column power supply feature matrix;

[0020] The power supply feature matrix is ​​standardized to obtain the power supply feature dataset of the ring main unit.

[0021] As a preferred embodiment of the remote monitoring method for the power supply status of a ring main unit according to the present invention, the application of transfer learning mechanism adds a residual connection structure to the intermediate layer of the basic classification model to obtain a fault prediction model, and uses the ring main unit power supply feature dataset as input for forward propagation calculation to obtain the health score and fault probability of the ring main unit power supply status. Specifically,

[0022] The collected historical power supply characteristics dataset for ring main units is divided into a training set and a validation set.

[0023] A basic classification model consisting of an input layer, hidden layers, and an output layer is built using a multilayer perceptron architecture, and all network weights are initialized using the Xavier-Uniform method.

[0024] The historical power supply feature dataset of the ring main unit is input into the basic classification model for training. After training, the transfer learning mechanism is applied to add a residual connection structure to the intermediate layer of the basic classification model to obtain the fault prediction model.

[0025] The power supply feature dataset of the ring main unit is converted into a format suitable for the input of the fault prediction model, and then passed to the input layer of the fault prediction model for linear transformation to obtain the health score and fault probability of the power supply status of the ring main unit.

[0026] As a preferred embodiment of the remote monitoring method for the power supply status of a ring main unit according to the present invention, the method of using a dynamic adjustment mechanism to dynamically calibrate the health score and fault probability to obtain the calibrated health score and fault probability specifically involves:

[0027] Collect historical health scores, form a sliding time window, and calculate the local mean and standard deviation;

[0028] The health score output by the fault prediction model is compared with the local mean in the sliding window to obtain the absolute deviation, and the correction intensity is adjusted according to the absolute deviation using a nonlinear feedback mechanism.

[0029] The health score output by the fault prediction model is calibrated using the adjusted correction strength to obtain the calibrated health score.

[0030] A linkage adjustment strategy is established, and the failure probability is adjusted in a linkage manner based on the calibrated health score to obtain the calibrated health score and failure probability.

[0031] As a preferred embodiment of the remote monitoring method for the power supply status of a ring main unit according to the present invention, the weighted decision algorithm is used to obtain a comprehensive health assessment result of the power supply status of the ring main unit based on the calibrated health score and fault probability. Specifically,

[0032] The calibrated health score and fault probability are weighted and used as input variables within a fuzzy integral framework. Through fuzzy measurement, an integrated calculation is performed to obtain the comprehensive health assessment score of the ring main unit's power supply status.

[0033] As a preferred embodiment of the remote monitoring method for the power supply status of a ring main unit according to the present invention, the step of generating a risk level and monitoring report based on the comprehensive health assessment results of the power supply status of the ring main unit and pushing it to the operation and maintenance terminal specifically involves:

[0034] The risk level range is preset to form K risk levels. The comprehensive health assessment results of the power supply status of the ring main unit are substituted into the preset risk level classification rules to obtain the risk level of the power supply status of the ring main unit.

[0035] Design a monitoring report template in PDF format, and use a report generation tool to fill in the ring main unit power supply data, the health score and failure probability of the ring main unit power supply status, the comprehensive health assessment results of the ring main unit power supply status, and the risk level of the ring main unit power supply status into the monitoring report, and push it to the mobile terminal of the operation and maintenance personnel.

[0036] Secondly, this invention provides a system for remotely monitoring the power supply status of a ring main unit, comprising,

[0037] The data acquisition module collects the power supply parameters of the ring main unit in real time, performs preprocessing, obtains the power supply data of the ring main unit, and transmits it to the remote control center.

[0038] The feature construction module constructs a multimodal feature matrix based on the power supply data of the ring main unit, thereby obtaining the power supply feature dataset of the ring main unit;

[0039] The power supply status module applies a transfer learning mechanism to add a residual connection structure to the intermediate layer of the basic classification model to obtain a fault prediction model. The power supply feature dataset of the ring main unit is used as input for forward propagation calculation to obtain the health score and fault probability of the power supply status of the ring main unit.

[0040] The calibration module uses a dynamic adjustment mechanism to dynamically calibrate the health score and fault probability, obtaining the calibrated health score and fault probability. A weighted decision algorithm is then applied to the calibrated health score and fault probability to obtain a comprehensive health assessment result of the ring main unit's power supply status.

[0041] The report generation module generates risk levels and monitoring reports based on the comprehensive health assessment results of the ring main unit's power supply status and pushes them to the operation and maintenance terminal.

[0042] Thirdly, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the method for remotely monitoring the power supply status of a ring main unit as described in the first aspect of the present invention.

[0043] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the method for remotely monitoring the power supply status of a ring main unit as described in the first aspect of the present invention.

[0044] The beneficial effects of this invention are as follows: By collecting and efficiently processing the power supply parameters of the ring main unit in real time, and securely transmitting this data to a remote control center for further analysis, it ensures that the information obtained from the source is accurate and useful. Based on the collected data, by constructing a multimodal feature matrix, this method can capture the power supply status changes of the ring main unit at different time scales. At the same time, the risk level and monitoring report generated based on the comprehensive health assessment results are directly pushed to the mobile terminals of maintenance personnel. This not only realizes remote monitoring, but also improves the response efficiency to potential problems. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 A flowchart for a method to remotely monitor the power supply status of a ring main unit.

[0047] Figure 2 A flowchart for generating a monitoring report.

[0048] Figure 3 A flowchart for constructing the feature matrix.

[0049] Figure 4 This is a flowchart of a fault prediction model. Detailed Implementation

[0050] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0051] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0052] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0053] Reference Figures 1-4As one embodiment of the present invention, this embodiment provides a method for remotely monitoring the power supply status of a ring main unit, comprising the following steps:

[0054] S1. Real-time acquisition of ring main unit power supply parameters, preprocessing to obtain ring main unit power supply data, and transmission to the remote control center. Specifically,

[0055] Based on the design drawings of the ring main unit, determine the power supply nodes that need to be monitored, namely the main incoming line, the outgoing line, and the high-voltage and low-voltage sides of the transformer.

[0056] Voltage sensors, current sensors, temperature sensors, and smart meters are simultaneously installed at the main incoming and outgoing terminals of the ring main unit, as well as on the high-voltage and low-voltage sides of the transformer. The sampling rates of the voltage sensors, current sensors, temperature sensors, and smart meters are then set. For example, the sampling rate can be set to once per second.

[0057] Connect the voltage sensor, current sensor, and smart meter to the data acquisition interface of the edge computing node using standard cables. Configure the IP address, subnet mask, and gateway of the edge computing node according to the network environment, and enable the TCP protocol and set the polling interval.

[0058] After initial configuration of the voltage sensor, current sensor, smart energy meter, and edge computing node, the voltage sensor, current sensor, and smart power meter are activated to collect real-time power supply parameters (voltage, current, temperature, and power) of the ring main unit. These parameters are transmitted to the edge computing node via a standard cable. A denoising algorithm on the edge computing node is used to denoise the parameters using a sliding window midpoint filtering method. The denoised parameters are then normalized to ensure they fall within the [0,1] range. The normalized parameters are then encapsulated and named "Ring Main Unit Power Supply Data." This data is then encrypted using AES-256 and transmitted via a TLS 1.3 protocol-secured transmission channel from the edge computing node to the message middleware at the remote control center.

[0059] By accurately collecting power supply data from ring main units, efficiently processing it through edge computing nodes, and reliably receiving it from the remote control center, a complete and closed-loop mechanism for collecting and preprocessing ring main unit power supply data is established. This improves the integrity and reliability of the collected ring main unit power supply data, solves the problems of low data quality, large response delay, and insufficient security of ring main unit data, and lays the foundation for subsequent feature extraction of ring main unit data.

[0060] S2. Construct a multimodal feature matrix based on the ring main unit power supply data to obtain the ring main unit power supply feature dataset, specifically...

[0061] Set the number of time windows (e.g., 5), where each time window refers to a specific time period used for feature extraction (e.g., 30 seconds, 1 minute, 5 minutes, 15 minutes, 30 minutes). Each time window corresponds to a ring main unit power supply data point, used to capture the changing patterns of ring main unit power supply data at different time frequencies.

[0062] Perform statistical analysis on the power supply data of the ring main unit for each time window, specifically as follows:

[0063] Mean: Calculate the average values ​​of voltage, current, temperature, and power parameters of the ring main unit power supply data;

[0064] Maximum and minimum value records: Find the maximum and minimum values ​​of voltage, current, temperature and power parameters in the ring main unit power supply data;

[0065] Frequency domain analysis: The main frequency components and amplitudes of the ring main unit power supply data are extracted using Fast Fourier Transform.

[0066] Skewness and Kurtosis Description: Statistical tools are used to calculate the skewness (measures asymmetry) and kurtosis (measures sharpness) of the power supply data distribution of the ring main unit to describe the distribution characteristics of the power supply data of the ring main unit;

[0067] Each time window extracts the above four features, and there are five time windows in total, generating a total of 20 features. These 20 features are combined into a power supply feature vector of dimension 20. The power supply feature vectors are stacked in chronological order to form a two-dimensional matrix (i.e., the feature matrix). Each row of the feature matrix corresponds to a power supply feature vector at a specific time point, where the time point refers to the power supply feature vector formed based on the data analysis results of the above time windows at a specific moment. Each column corresponds to a type of feature (such as the average voltage and the skewness of the ring main unit power supply data distribution).

[0068] Perform Min-Max standardization on each power supply feature vector in the feature matrix to obtain a standardized feature matrix. Export the standardized feature matrix as a structured file, such as train.csv, to obtain the power supply feature dataset of the ring main unit.

[0069] This feature matrix construction method not only effectively captures the power supply status changes of ring main units at different time scales, but also improves the efficiency and accuracy of subsequent modeling and analysis through power supply feature vectorization and matrix organization, and provides a high-quality data foundation for intelligent monitoring, health assessment and fault prediction.

[0070] S3. A residual connection structure is added to the intermediate layer of the basic classification model using a transfer learning mechanism to obtain a fault prediction model. The ring main unit power supply feature dataset is then used as input for forward propagation calculation to obtain the health score and fault probability of the ring main unit power supply status. Specifically...

[0071] Obtain historical power supply characteristic datasets of ring main units from the remote control center over a period of time, such as historical power supply characteristic datasets of ring main units from 6 months ago.

[0072] The basic classification model network structure is set up with four layers: an input layer, hidden layer 1, hidden layer 2, and an output layer. The input layer receives the power supply feature dataset from the ring main unit. Hidden layer 1 is a fully connected layer with 128 neurons and the ReLU activation function. Hidden layer 2 is also a fully connected layer with 64 neurons and the ReLU activation function. The output layer has two nodes, outputting the health score and the fault probability, respectively. A joint loss function composed of mean squared error loss and cross-entropy loss is incorporated into the basic classification model to simultaneously optimize both the health score and the fault probability.

[0073] The Xavier-Uniform method was used to initialize all weights of the base classification model, ensuring that the base classification model was trained from scratch. Adam was selected as the optimizer for the base classification model (learning rate set to 0.001, momentum set to 0.9). The historical ring main unit power supply feature dataset was divided into training and validation sets in an 8:2 ratio, and the input batch size of the training set was set to 64 to facilitate the convergence speed of the base classification model.

[0074] The training set is input into the basic classification model for training, with 50 training rounds. After each round, the training results of the basic classification model are validated using a validation set. Training is terminated after three consecutive rounds of accurate validation to prevent overfitting. After the basic classification model is trained, the parameters of the input layer and hidden layer 1 are kept unchanged. Then, a residual connection structure is inserted into hidden layer 2. The output of the residual connection structure is added to the output of hidden layer 2 and then fed into the output layer to enhance the basic classification model's ability to learn deep features, finally yielding the fault prediction model. The ring main unit power supply feature dataset is converted into a format suitable for the fault prediction model input and passed to the input layer of the fault prediction model. The ring main unit power supply feature dataset undergoes linear transformation and nonlinear activation in the fault prediction model network architecture to obtain the health score and fault probability of the ring main unit power supply status. The fault prediction model constructed using the above methods not only inherits the training stability of the basic classification model, but also enhances the basic classification model's ability to express complex features through residual connections. Combined with the transfer learning mechanism, it effectively improves the generalization performance and adaptability of the basic classification model. It exhibits stronger transfer capabilities and prediction accuracy when facing different working conditions and new equipment scenarios, providing solid data support and technical assurance for subsequent uncertainty assessment, dynamic calibration, and comprehensive health assessment.

[0075] S4. A dynamic adjustment mechanism is used to dynamically calibrate the health score and failure probability, resulting in the calibrated health score and failure probability. Specifically:

[0076] Choose an appropriate time period as the length of the sliding window (e.g., the past 7 days). Obtain the health scores and failure probabilities for the past 7 days from the remote control center, ensuring these health scores are accurate and arranged chronologically. Calculate local statistics (local mean and standard deviation), where the local mean represents the typical performance level of the ring main unit's backup power supply status during this period. The standard deviation reflects the fluctuation of the health scores during this period.

[0077] Based on the health score output by the fault prediction model, the absolute deviation between the health score output by the fault prediction model and the local mean is calculated. Specifically, the local mean is subtracted from the health score output by the fault prediction model. If the result is negative, the absolute value is taken. This absolute value is the degree of deviation of the current health score from the historical average level.

[0078] A set of rules is established for adjusting the correction strength of the nonlinear feedback function based on the absolute deviation. Specifically, a smaller adjustment magnitude is used when the absolute deviation is small, and a larger adjustment magnitude is used when the absolute deviation is large. Simultaneously, the standard deviation is introduced as an adjustment factor to enhance robustness. This can be expressed by the formula:

[0079]

[0080] Where f represents the correction strength coefficient, β represents the amplification factor of the standard deviation, ε represents the minimum constant, γ represents the nonlinear exponent, and ΔH represents the absolute deviation;

[0081] This adjustment method can be segmented or continuously changing, with the aim of responding more sensitively to abnormal situations.

[0082] The health score output by the fault prediction model is adjusted using a correction strength coefficient. This adjustment can be achieved by adding or subtracting a certain percentage of the correction value from the original health score, making the final health score closer to the true state. Specifically,

[0083] k = V + α·f;

[0084] Where k represents the calibrated health score, V represents the health score output by the fault prediction model, α represents the offset correction value, and f represents the correction strength coefficient.

[0085] Based on collected historical health scores and corresponding failure probabilities, regression analysis is used to identify the relationship parameters (e.g., 0.8 and 0.1) between health scores and failure frequency. Specifically, these relationship parameters are typically coefficients in a linear model, designed to accurately reflect the impact of changes in health scores on failure probability. Furthermore, health scores and failure probabilities are negatively correlated; that is, the lower the health score, the higher the failure probability. Therefore, the corrected failure probability is expressed by a linear model as follows:

[0086] P = 0.8·(1-k) + 0.1;

[0087] Where P represents the corrected failure probability, and k represents the calibrated health score;

[0088] By dynamically adjusting prediction results based on historical operational data, the health score more closely reflects the actual power supply status of the ring main unit, while improving the accuracy and robustness of fault probability estimation. This mechanism effectively compensates for the lag and bias that may exist in a single prediction model, thus providing a more reliable data foundation for subsequent comprehensive health assessments and risk level classifications.

[0089] S5. A weighted decision algorithm is used to calculate the calibrated health score and fault probability to obtain a comprehensive health assessment result of the ring main unit's power supply status. Specifically,

[0090] Based on the calibrated health score and failure probability, the calibrated health score and failure probability are normalized so that the values ​​are all within the range [0,1] to ensure data consistency and comparability. The calibrated health score and failure probability are labeled as input variables x1 and x2, respectively, forming an input variable set X = {x1, x2}. Fuzzy measures are applied to assign initial weights (e.g., 0.6 and 0.4) to input variables x1 and x2.

[0091] Compare the numerical values ​​of input variable x1 and input variable x2. Based on the comparison result, sort the two variables in ascending order. For example, if x1 > x2, the order is x2, x1; otherwise, it is x1, x2. Record the sorted order of the input variables and their corresponding original names.

[0092] Based on the fuzzy measure and the sorted input variables, the fuzzy integral value is calculated. The fuzzy integral value comprehensively considers all input variables and their interactions, providing a comprehensive indicator of the system's health. According to the sorted order of the input variables, the contribution of each input variable to the final score is calculated sequentially, expressed as:

[0093] C = x + zx·z l ;

[0094] Where C represents the final score, x represents the input variable with the smallest value, and z represents the input variable with the largest value. l This represents the weight of the input variable with the largest value, where l represents the weight.

[0095] The final score is a numerical result obtained through fuzzy measurement, input variable ranking, and fuzzy integral calculation. This value comprehensively reflects the overall health status of the ring main unit during power supply, i.e., the comprehensive health assessment result of the ring main unit's power supply status. Contribution refers to the degree of influence of each input variable on the final score during its formation. The entire assessment process not only considers the contribution of each input variable itself but also introduces a fuzzy measurement mechanism to quantify the interaction between input variables, making the final comprehensive health assessment result closer to the actual operating status of the ring main unit. This assessment method, which integrates subjective experience and objective data, demonstrates stronger adaptability and robustness in the face of complex and ever-changing power equipment operating environments.

[0096] S6. Based on the comprehensive health assessment results of the ring main unit's power supply status, generate a risk level and monitoring report and push it to the operation and maintenance terminal. Specifically,

[0097] Based on the comprehensive health assessment results of the ring main unit's power supply status, three risk levels and threshold ranges are set, with each level's threshold range corresponding to a specific risk level, named "Low -0.2 to 0.4", "Medium -0.4 to 0.6", and "High 0.6 to 0.8". Since the final score is obtained through fuzzy integration of the comprehensive health assessment result, the risk level is determined by which range the score falls into. For example, if the comprehensive health assessment score is 0.7, the risk level of the ring main unit's power supply status is "High".

[0098] Open your PDF editing software and create a new blank PDF document. Add a title at the top: "Ring Main Unit Power Supply Status Monitoring Report". Create a table to display key information such as the ring main unit number, power supply data, health score and failure probability, comprehensive health assessment results, and risk level. Ensure each section has a clear label. Save the designed PDF template as a reusable file named "Ring Main Unit Power Supply Status Monitoring Report Template.pdf".

[0099] Using a report generation tool, the ring main unit number, power supply data, health score and failure probability, comprehensive health assessment results, and risk level were populated into a PDF template according to a preset data format. The completed PDF file was carefully checked to ensure that all data was accurate and formatted correctly. The populated PDF file was saved, named "Ring Main Unit Power Supply Status Monitoring Report_Date.pdf," and transmitted to the maintenance personnel's mobile terminal via email address for remote monitoring. By creating a standardized PDF monitoring report template and automatically populating the data using a report generation tool, work efficiency was improved, and the consistency of report format and data accuracy were ensured. This automated process reduced the possibility of human error and made information transmission more efficient and transparent. Finally, these monitoring reports were sent directly to the maintenance terminal via email, achieving remote monitoring and real-time management, greatly improving the convenience and timeliness of maintenance work.

[0100] This embodiment also provides a system for remotely monitoring the power supply status of a ring main unit, including:

[0101] The data acquisition module collects the power supply parameters of the ring main unit in real time, performs preprocessing, obtains the power supply data of the ring main unit, and transmits it to the remote control center.

[0102] The feature construction module constructs a multimodal feature matrix based on the power supply data of the ring main unit, thereby obtaining the power supply feature dataset of the ring main unit;

[0103] The power supply status module applies a transfer learning mechanism to add a residual connection structure to the intermediate layer of the basic classification model to obtain a fault prediction model. The power supply feature dataset of the ring main unit is used as input for forward propagation calculation to obtain the health score and fault probability of the power supply status of the ring main unit.

[0104] The calibration module uses a dynamic adjustment mechanism to dynamically calibrate the health score and fault probability, obtaining the calibrated health score and fault probability. A weighted decision algorithm is then applied to the calibrated health score and fault probability to obtain a comprehensive health assessment result of the ring main unit's power supply status.

[0105] The report generation module generates risk levels and monitoring reports based on the comprehensive health assessment results of the ring main unit's power supply status and pushes them to the operation and maintenance terminal.

[0106] This embodiment also provides a computer device applicable to the method of remotely monitoring the power supply status of a ring main unit, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method of remotely monitoring the power supply status of a ring main unit as proposed in the above embodiment.

[0107] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0108] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the method for remotely monitoring the power supply status of a ring main unit as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0109] In summary, this invention ensures the accuracy and usefulness of information obtained from the source by: real-time acquisition and efficient processing of power supply parameters of the ring main unit, and secure transmission of this data to a remote control center for further analysis. Based on the collected data, by constructing a multimodal feature matrix, this method can capture the power supply status changes of the ring main unit at different time scales. At the same time, the risk level and monitoring report generated based on the comprehensive health assessment results are directly pushed to the mobile terminals of maintenance personnel. This not only realizes remote monitoring but also improves the response efficiency to potential problems.

[0110] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for remotely monitoring the power supply status of a ring main unit, characterized in that: include, S1. Real-time acquisition of ring main unit power supply parameters, preprocessing to obtain ring main unit power supply data, and transmission to remote control center; S2. Construct a multimodal feature matrix based on the power supply data of the ring main unit to obtain the power supply feature dataset of the ring main unit; S3. A residual connection structure is added to the intermediate layer of the basic classification model using a transfer learning mechanism to obtain a fault prediction model. The ring main unit power supply feature dataset is then used as input for forward propagation calculation to obtain the health score and fault probability of the ring main unit power supply status. Specifically, The collected historical power supply characteristics dataset for ring main units is divided into a training set and a validation set. A basic classification model consisting of an input layer, hidden layer 1, hidden layer 2, and output layer is built using a multilayer perceptron architecture, and all network weights are initialized using the Xavier-Uniform method. The historical power supply feature dataset of the ring main unit is input into the basic classification model for training. After training, the transfer learning mechanism is applied to add a residual connection structure to the intermediate layer of the basic classification model to obtain the fault prediction model. A residual connection structure is inserted into hidden layer 2. The output of the residual connection structure is added to the output of hidden layer 2 and then enters the output layer. The power supply feature dataset of the ring main unit is converted into a format suitable for the input of the fault prediction model and passed to the input layer of the fault prediction model for linear transformation to obtain the health score and fault probability of the power supply status of the ring main unit. S4. A dynamic adjustment mechanism is used to dynamically calibrate the health score and failure probability, resulting in calibrated health scores and failure probabilities; specifically, this involves... Collect historical health scores, form a sliding time window, and calculate the local mean and standard deviation; The health score output by the fault prediction model is compared with the local mean in the sliding window to obtain the absolute deviation, and the correction intensity is adjusted according to the absolute deviation using a nonlinear feedback mechanism. The health score output by the fault prediction model is calibrated using the adjusted correction strength to obtain the calibrated health score. Establish a linkage adjustment strategy and adjust the failure probability according to the calibrated health score to obtain the calibrated health score and failure probability. S5. A weighted decision algorithm is used to evaluate the calibrated health score and fault probability to obtain a comprehensive health assessment result of the ring main unit's power supply status. This includes: based on the calibrated health score and failure probability, normalizing the calibrated health score and failure probability so that the values ​​are all within the range of [0,1], labeling the calibrated health score and failure probability as input variables x1 and x2 respectively to form the input variable set X={x1,x2}, and applying fuzzy measure to set initial weights for input variables x1 and x2; Compare the values ​​of input variable x1 and input variable x2, and sort the two variables in ascending order based on the comparison results; Based on the sorted order of the input variables, the contribution of each input variable to the final score is calculated sequentially, and expressed as follows: ； in, This indicates the final score. This represents the input variable with the smallest value. This represents the input variable with the largest value. The weight represents the input variable with the largest value. Indicates weight; S6. Based on the comprehensive health assessment results of the ring main unit's power supply status, generate risk level and monitoring report and push them to the operation and maintenance terminal.

2. The method for remotely monitoring the power supply status of a ring main unit as described in claim 1, characterized in that: The real-time acquisition of ring main unit power supply parameters is preprocessed to obtain ring main unit power supply data, which is then transmitted to the remote control center. Specifically, Sensors are installed at the power supply location of the ring main unit to collect voltage, current and power parameters from the ring main unit in real time. Voltage, current, and power parameters are sent to edge computing nodes for noise reduction and normalization to obtain the power supply data of the ring main unit.

3. The method for remotely monitoring the power supply status of a ring main unit as described in claim 2, characterized in that: The process involves constructing a multimodal feature matrix based on the power supply data of the ring main unit to obtain a power supply feature dataset for the ring main unit. Set N time windows, and use the N time windows sequentially to perform statistical analysis on the power supply data of the ring main unit to obtain the characteristics of the power supply data of the ring main unit; The power supply data features of the ring main unit are combined into a power supply feature vector and arranged in order to form a multi-row, multi-column power supply feature matrix; The power supply feature matrix is ​​standardized to obtain the power supply feature dataset of the ring main unit.

4. The method for remotely monitoring the power supply status of a ring main unit as described in claim 1, characterized in that: Based on the comprehensive health assessment results of the ring main unit's power supply status, a risk level and monitoring report are generated and pushed to the operation and maintenance terminal. Specifically, The risk level range is preset to form K risk levels. The comprehensive health assessment results of the power supply status of the ring main unit are substituted into the preset risk level classification rules to obtain the risk level of the power supply status of the ring main unit. Design a monitoring report template in PDF format, and use a report generation tool to fill in the ring main unit power supply data, the health score and failure probability of the ring main unit power supply status, the comprehensive health assessment results of the ring main unit power supply status, and the risk level of the ring main unit power supply status into the monitoring report, and push it to the mobile terminal of the operation and maintenance personnel.

5. A system for remotely monitoring the power supply status of a ring main unit, based on the method for remotely monitoring the power supply status of a ring main unit as described in any one of claims 1 to 4, characterized in that: include, The data acquisition module collects the power supply parameters of the ring main unit in real time, performs preprocessing, obtains the power supply data of the ring main unit, and transmits it to the remote control center. The feature construction module constructs a multimodal feature matrix based on the power supply data of the ring main unit, thereby obtaining the power supply feature dataset of the ring main unit. The power supply status module applies a transfer learning mechanism to add a residual connection structure to the intermediate layer of the basic classification model to obtain a fault prediction model. The power supply feature dataset of the ring main unit is used as input for forward propagation calculation to obtain the health score and fault probability of the power supply status of the ring main unit. The calibration module uses a dynamic adjustment mechanism to dynamically calibrate the health score and fault probability, obtaining the calibrated health score and fault probability. A weighted decision algorithm is then applied to the calibrated health score and fault probability to obtain a comprehensive health assessment result of the ring main unit's power supply status. The report generation module generates risk levels and monitoring reports based on the comprehensive health assessment results of the ring main unit's power supply status and pushes them to the operation and maintenance terminal.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the method for remotely monitoring the power supply status of a ring main unit as described in any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the method for remotely monitoring the power supply status of the ring main unit as described in any one of claims 1 to 4.

Images