Method and platform for evaluating, regulating and controlling state of ring main unit

Abstract

The invention provides a state evaluation, regulation and control method and platform for a ring main unit, and relates to the technical field of intelligent power systems, and the method comprises the steps: predicting and obtaining the SF6 gas component of a target ring main unit at a current time node based on an inspection log and an SF6 gas component record; according to the predicted SF6 gas data, the SF6 gas concentration, the air temperature in the cabinet and the air humidity in the cabinet, insulation fault probability analysis is carried out in combination with the electric power internet, and the predicted insulation fault probability is obtained; and if the predicted insulation fault probability exceeds an expected probability threshold, generating a risk early warning signal, and performing dynamic regulation and control according to a predetermined scheme. The technical problems that a traditional insulation performance evaluation method usually depends on a single physical quantity for judgment, the insulation state of equipment cannot be comprehensively known, and the accuracy and reliability of insulation performance evaluation are low can be solved, the accuracy and reliability of ring main unit insulation performance evaluation can be improved, and the reliability of ring main unit insulation performance evaluation is improved. And potential insulation fault threats can be found in time.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent power systems, and in particular to a method and platform for status evaluation and control of ring main units. Background Art

[0002] Ring main units (RMUs) are crucial equipment in power systems, primarily responsible for power distribution and protection, ensuring stable power transmission under varying loads. The reliability and safety of RMUs are crucial to power system operation, and one of their key performance indicators is insulation performance. The condition of the insulation system directly impacts equipment safety and the stable operation of the power system.

[0003] Traditional methods for evaluating the insulation performance of ring main units usually rely on a single physical quantity, such as insulation resistance testing and gas pressure testing. These evaluation methods have certain limitations and cannot fully understand the insulation status of the equipment, resulting in insufficient accuracy and reliability in insulation performance evaluation. Summary of the Invention

[0004] The purpose of this application is to provide a ring network cabinet status evaluation and control method and platform to solve the technical problems that the traditional ring network cabinet insulation performance evaluation method usually relies on a single physical quantity for judgment, cannot fully understand the insulation status of the equipment, and has low accuracy and reliability of insulation performance evaluation.

[0005] In view of the above problems, the present application provides a ring main unit status assessment and control method and platform.

[0006] In the first aspect, the present application provides a method for status assessment and control of a ring network cabinet, which is implemented through a status assessment and control platform for a ring network cabinet, including: fixed-point monitoring to obtain the SF6 gas concentration, air temperature and air humidity in the target ring network cabinet, and transmit them to a cloud platform; in the cloud platform, query the inspection log to obtain the inspection data of the target ring network cabinet at the most recent inspection time node, wherein the inspection data includes SF6 gas composition records; based on the inspection log and SF6 gas composition records, predict and obtain the SF6 gas composition of the target ring network cabinet at the current time node to obtain predicted SF6 gas data; based on the predicted SF6 gas data, SF6 gas concentration, air temperature and air humidity in the cabinet, combine the power Internet to perform insulation fault probability analysis to obtain predicted insulation fault probability; if the predicted insulation failure probability exceeds the expected probability threshold, a risk warning signal is generated, and the target ring network cabinet is dynamically controlled according to a predetermined plan.

[0007] In the second aspect, the present application also provides a ring network cabinet status assessment and control platform for executing a ring network cabinet status assessment and control method as described in the first aspect, including: a data monitoring module, which is used to monitor and obtain the SF6 gas concentration, air temperature and air humidity of the target ring network cabinet at a fixed point, and transmit them to the cloud platform; an inspection data acquisition module, which is used to query the inspection log in the cloud platform to obtain the inspection data of the target ring network cabinet at the most recent inspection time node, wherein the inspection data includes SF6 gas composition records; a gas composition prediction module, which is used to predict and obtain the SF6 gas composition of the target ring network cabinet at the current time node based on the inspection log and SF6 gas composition records, and obtain predicted SF6 gas data; an insulation fault probability analysis module, which is used to perform insulation fault probability analysis based on the predicted SF6 gas data, SF6 gas concentration, air temperature and air humidity in the cabinet, combined with the power Internet, to obtain predicted insulation fault probability; a dynamic control module, which is used to generate a risk warning signal if the predicted insulation failure probability exceeds the expected probability threshold, and dynamically control the target ring network cabinet according to a predetermined plan.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: Through fixed-point monitoring, the SF6 gas concentration, cabinet air temperature and cabinet air humidity of the target ring main unit are obtained and transmitted to the cloud platform; then, in the cloud platform, the inspection log is queried to obtain the inspection data of the target ring main unit at the most recent inspection time node, wherein the inspection data includes SF6 gas composition records; further based on the inspection log and SF6 gas composition records, the SF6 gas composition of the target ring main unit at the current time node is predicted to obtain predicted SF6 gas data; then, based on the predicted SF6 gas data, SF6 gas concentration, cabinet air temperature and cabinet air humidity, an insulation fault probability analysis is performed in combination with the power Internet to obtain a predicted insulation fault probability; if the predicted insulation fault probability exceeds the expected probability threshold, a risk warning signal is generated, and finally, the target ring main unit is dynamically regulated according to a predetermined plan; the above method can improve the accuracy and reliability of the insulation performance evaluation of the ring main unit, so that potential insulation fault threats can be discovered in a timely manner, and targeted means can be collected for dynamic regulation to ensure the power supply reliability and safety of the power system.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, which can be implemented in accordance with the contents of the description, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically listed below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without creative work.

[0011] Figure 1 This application provides a flow chart of a method for status evaluation and control of a ring main unit.

[0012] Figure 2 This application provides a structural diagram of a ring main unit status evaluation and control platform.

[0013] Description of reference numerals: Data monitoring module 11, inspection data acquisition module 12, gas composition prediction module 13, insulation fault probability analysis module 14, dynamic control module 15. DETAILED DESCRIPTION

[0014] This application provides a ring main unit (RMU) status assessment and control method and platform, addressing the technical issues of traditional RMU insulation performance assessment methods, which typically rely on a single physical quantity for judgment, fail to fully understand the insulation status of the equipment, and suffer from low accuracy and reliability of insulation performance assessment. This method improves the accuracy and reliability of RMU insulation performance assessment, enabling timely detection of potential insulation fault threats and the acquisition of targeted measures for dynamic control, ensuring the reliability and safety of power supply in the power system.

[0015] Below, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited to the example embodiments described herein. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. It should also be noted that, for the convenience of description, only the parts related to this application, rather than all of them, are shown in the accompanying drawings.

[0016] For example, see the attached Figure 1 The present application provides a ring main unit status evaluation and control method, which is applied to a ring main unit status evaluation and control platform, and specifically includes the following steps: Step 1: Obtain the SF6 gas concentration, air temperature, and air humidity of the target ring network cabinet through fixed-point monitoring and transmit them to the cloud platform.

[0017] Specifically, at a predetermined monitoring time node (e.g., every 6 hours), the target ring network cabinet is monitored through the gas concentration sensor, temperature sensor, and humidity sensor deployed in the target ring network cabinet to collect data and obtain the SF6 gas concentration, cabinet air temperature, and cabinet air humidity at the monitoring node. SF6 gas is widely used in the gas insulation system of the ring network cabinet. As the use time increases, SF6 gas will gradually be affected by factors such as temperature, pressure, and electrical load, and gas leakage, increased impurities, and increased acid value may occur, thereby affecting the insulation performance. Therefore, the monitoring Measuring the SF6 gas concentration is key to ensuring the normal operation of the ring main unit. The density of SF6 gas is inversely proportional to the temperature. When the temperature rises, the density of SF6 gas decreases, and the gas volume expands, resulting in an increase in gas pressure. Under high temperatures, the insulation performance of the gas may decrease. At the same time, the leakage rate of SF6 gas may increase, thereby affecting the insulation performance. When moisture enters the ring main unit, it will reduce the purity of SF6 gas, resulting in a decrease in insulation performance. The insulation capacity of SF6 gas depends on high purity. When the gas contains impurities or moisture, the discharge voltage will decrease, making the equipment more susceptible to discharge or breakdown.

[0018] The collected SF6 gas concentration, cabinet air temperature, and cabinet air humidity data are then communicated to the cloud platform. Through this method, the SF6 gas concentration, cabinet air temperature, and cabinet air humidity data of the ring main cabinet can be accurately collected and transmitted to the cloud platform in real time, providing key data support for subsequent fault prediction, performance evaluation, and dynamic control.

[0019] Step 2: In the cloud platform, query the inspection log to obtain the inspection data of the target ring main unit at the most recent inspection time node, wherein the inspection data includes SF6 gas composition records.

[0020] Specifically, in the cloud platform, the inspection log of the ring main unit is queried, where the inspection log includes SF6 gas composition records (including gas purity, impurity content and acid value, etc.), component aging data (such as the aging degree of switches, contactors, busbars, and terminal blocks, including oxidation, wear and insulation degradation, etc.); further, the log is queried based on the unique identifier of the target ring main unit (such as the equipment number or RFID tag) to obtain the inspection data of the target ring main unit at the most recent inspection time node (the inspection node with the shortest interval from the current time node), where the inspection data includes SF6 gas composition records (gas purity, impurity content and acid value, etc.).

[0021] Step 3: Based on the inspection log and SF6 gas composition record, predict and obtain the SF6 gas composition of the target ring main unit at the current time node to obtain predicted SF6 gas data.

[0022] Specifically, based on the sample data set collected from the inspection log, supervised training is performed on the gas composition prediction plug-in constructed based on machine learning to obtain a gas composition prediction plug-in that meets the convergence constraints; on the other hand, the current interval duration is calculated based on the current time node and the most recent inspection time node; then the SF6 gas composition record and the current interval duration are input into the gas composition prediction plug-in for prediction, and the predicted SF6 gas data of the target ring network cabinet at the current time node is output.

[0023] Step 4: Based on the predicted SF6 gas data, SF6 gas concentration, cabinet air temperature, and cabinet air humidity, insulation fault probability analysis is performed in combination with the power internet to obtain a predicted insulation fault probability.

[0024] Specifically, using the predicted SF6 gas data, SF6 gas concentration, cabinet air temperature, and cabinet air humidity as constraints, the system conducts a related information search in conjunction with the power internet to obtain a dataset of similar ring main unit operation records that meet the expected conditions. It then performs an insulation fault probability analysis based on these similar ring main unit operation record datasets, outputting a predicted insulation fault probability. This method, combined with analysis of historical inspection records and real-time data, allows for more accurate predictions of the equipment's insulation status, improving the accuracy and reliability of the insulation fault probability prediction.

[0025] Step 5: If the predicted insulation failure probability exceeds the expected probability threshold, a risk warning signal is generated, and the target ring main unit is dynamically adjusted according to a predetermined plan.

[0026] Specifically, if the predicted insulation failure probability exceeds the expected probability threshold, indicating a high insulation failure risk for the device, the risk warning mechanism is triggered, generating a risk warning signal. The target ring main unit (RMU) is then adjusted according to a preset dynamic control strategy to reduce the insulation failure risk. For example, if the RMU failure probability is high, the device load and current load can be reduced to prevent overload-induced equipment failure. By combining data analysis with dynamic control, timely and effective measures can be taken when equipment failure risks arise, ensuring the stability and security of the power system.

[0027] The state evaluation and control method of a ring main unit is applied to a state evaluation and control platform of a ring main unit, which can solve the technical problems that the insulation performance evaluation method of a traditional ring main unit usually relies on a single physical quantity for judgment, cannot fully understand the insulation state of the equipment, and has low accuracy and reliability of insulation performance evaluation. Through fixed-point monitoring, the SF6 gas concentration, cabinet air temperature and cabinet air humidity of the target ring main unit are obtained and transmitted to the cloud platform; then, in the cloud platform, the inspection log is queried to obtain the inspection data of the target ring main unit at the most recent inspection time node, wherein the inspection data includes SF6 gas composition records; further based on the inspection log and SF6 gas composition records, the SF6 gas composition of the target ring main unit at the current time node is predicted to obtain predicted SF6 gas data; then, based on the predicted SF6 gas data, SF6 gas concentration, cabinet air temperature and cabinet air humidity, an insulation fault probability analysis is performed in combination with the power Internet to obtain a predicted insulation fault probability; if the predicted insulation fault probability exceeds the expected probability threshold, a risk warning signal is generated, and finally, the target ring main unit is dynamically regulated according to a predetermined plan; the above method can improve the accuracy and reliability of the insulation performance evaluation of the ring main unit, so that potential insulation fault threats can be discovered in a timely manner, and targeted means can be collected for dynamic regulation to ensure the power supply reliability and safety of the power system.

[0028] Furthermore, this application also includes: The SF6 gas composition record includes oxygen content, hydrogen fluoride content, sulfur dioxide content and acid value.

[0029] Specifically, the SF6 gas composition record includes oxygen content, hydrogen fluoride content, sulfur dioxide content and acid value. Among them, an increase in oxygen content usually indicates that SF6 gas has partially decomposed under a high electric field to generate oxides (such as sulfur oxide). Excessive oxygen content may indicate gas leakage or abnormal discharge inside the equipment; excessive hydrogen fluoride content may mean that the gas is affected by discharge or high temperature, and the insulation performance of the equipment may be damaged; sulfur dioxide is one of the products of SF6 gas decomposition during arc discharge. An increase in SO2 usually indicates that the gas has undergone a large degree of thermal decomposition, which may reduce its insulation strength; acid value refers to the concentration of acidic substances in the gas, which is usually related to the moisture content and the degree of aging of the gas. A high acid value means that the gas is highly aged and may have begun to corrode electrical components and reduce insulation performance.

[0030] Furthermore, this application also includes: The current interval duration is calculated based on the most recent inspection time node and the current time node; based on the inspection log, the sample initial SF6 gas composition set, sample interval duration set and sample measured SF6 gas composition set of the same type of ring main unit within the preset area at the time of the most recent inspection are collected; the sample initial SF6 gas composition set, sample interval duration set and sample measured SF6 gas composition set are used as training data, and supervised training is performed on multiple prediction operators respectively, and a gas composition prediction plug-in is integrated to construct; the SF6 gas composition record and the current interval duration are input into the gas composition prediction plug-in, and the predicted SF6 gas data is output.

[0031] Specifically, first, the current interval duration is calculated based on the most recent inspection time node and the current time node, where the current interval duration is the time difference between the current time node and the most recent inspection time node, which is used to evaluate the duration of equipment operation since the last inspection.

[0032] Then, based on the inspection log, the sample initial SF6 gas composition, sample interval duration and sample measured SF6 gas composition of the same type of ring main unit in a preset area (e.g., within 10 square kilometers) at the time of the most recent inspection are collected, wherein the data set is obtained from the inspection record of each similar ring main unit. Since the inspection is carried out in batches by time, the equipment in the same area may have different inspection times; the sample initial SF6 gas composition refers to the SF6 gas composition data at the previous adjacent inspection time node; the sample measured SF6 gas composition refers to the SF6 gas composition data at the current inspection time node; the sample initial SF6 gas composition set, the sample interval duration set and the sample measured SF6 gas composition set are obtained.

[0033] The initial SF6 gas composition set, sample interval duration set, and measured SF6 gas composition set are further used as training data to perform supervised training on multiple prediction operators (BP neural network, random decision forest, and support vector machine), generating the first, second, and third prediction branches. Based on the principle of ensemble learning, a gas composition prediction plug-in is constructed based on the first, second, and third prediction branches. Finally, the SF6 gas composition records and the current interval duration are input into the gas composition prediction plug-in for prediction, outputting the predicted SF6 gas data for the target ring main unit at the current time point. This ensemble learning approach enables a more accurate and comprehensive prediction of the SF6 gas status of the target ring main unit, providing strong support for subsequent fault diagnosis.

[0034] Furthermore, this application also includes: Acquire multiple prediction operators, wherein the multiple prediction operators include a BP neural network, a random decision forest, and a support vector machine; use the sample initial SF6 gas composition set, the sample interval time set, and the sample measured SF6 gas composition set as training data, divide the training data into nine equal parts, select nine times with replacement, construct a first sample data set, and iterate the selection three times to obtain three sample data sets; use the three sample data sets to perform supervised training on the BP neural network, the random decision forest, and the support vector machine, respectively, until a predetermined number of training times is reached, to obtain a first prediction branch, a second prediction branch, and a third prediction branch that are in convergence; obtain a first convergence accuracy, a second convergence accuracy, and a third convergence accuracy of the first prediction branch, the second prediction branch, and the third prediction branch, and configure a credible weight ratio based on the first convergence accuracy, the second convergence accuracy, and the third convergence accuracy; based on the principle of ensemble learning, assign weights to the first prediction branch, the second prediction branch, and the third prediction branch respectively according to the credible weight ratio, and integrate and construct the gas composition prediction plug-in, wherein the output of the gas composition prediction plug-in is a weighted result of the outputs of the first prediction branch, the second prediction branch, and the third prediction branch.

[0035] Specifically, the purpose of constructing a gas composition prediction plug-in through multiple prediction operators is to combine the advantages of different algorithms through an integrated learning method to improve the accuracy and robustness of SF6 gas composition prediction. Among them, each prediction operator has different learning and prediction methods and can show good performance under different data modes.

[0036] First, multiple prediction operators are obtained, wherein the multiple prediction operators include BP neural network, random decision forest and support vector machine. BP neural network is a multi-layer perceptron network, which is trained by error back propagation algorithm and can learn the nonlinear relationship between input features and output results through layer-by-layer network structure; random forest is an integrated learning method, which constructs multiple decision trees and performs voting prediction. Each tree increases the diversity of the model by training a subset of data, thereby improving the stability and accuracy of the prediction; support vector machine is a classification and regression algorithm based on the principle of structural risk minimization. It performs data separation by finding the optimal hyperplane and can handle complex problems in high-dimensional space.

[0037] Next, the sample initial SF6 gas composition set, sample interval duration set and sample measured SF6 gas composition set are used as training data, and the training data are equally divided into nine parts. Nine parts are selected with replacement to construct the first sample data set; the same method is used to iterate the selection three times to obtain three sample data sets.

[0038] The three sample data sets are further used, with the initial SF6 gas composition of the sample and the sample interval duration as input, and the measured SF6 gas composition of the sample as supervision, to perform supervised training on the BP neural network, random decision forest, and support vector machine, respectively. First, the BP neural network is supervised trained using the first sample data set. First, the network weights and biases are initialized, and the predicted value is calculated through forward propagation; then, the error between the predicted value and the measured value is calculated through a loss function (such as mean square error MSE); further, the weights and biases are adjusted using a gradient descent algorithm to reduce the error; the above process is repeated until the predetermined number of training times is reached or the error reaches the convergence condition, and the first prediction branch with completed training is obtained.

[0039] Next, supervised training of the random decision forest is performed using the second sample dataset. First, a subset is randomly selected from the sample dataset to construct multiple decision trees. Next, a tree structure is constructed based on the input features (gas composition, time interval), with nodes gradually split to maximize information gain. The predictions from all decision trees are then combined through majority voting or averaging. Training is repeated until the number of trees and model performance meet the predetermined requirements, resulting in a trained second prediction branch. Separately, supervised training of the support vector machine is performed using the third sample dataset. First, an appropriate kernel function (such as a linear kernel or RBF kernel) is selected based on the data characteristics. Next, an optimal hyperplane is constructed on the sample data to minimize the error. The penalty coefficient and kernel function parameters are then adjusted based on the model error to optimize model performance. These steps are repeated until the predetermined number of training cycles or convergence conditions are reached, resulting in a trained third prediction branch.

[0040] Then, the first convergence accuracy, second convergence accuracy, and third convergence accuracy of the first prediction branch, the second prediction branch, and the third prediction branch are obtained, where the convergence accuracy is an indicator for measuring the prediction performance of each prediction branch model, such as the model input accuracy; then, the trusted weight ratio is configured according to the first convergence accuracy, the second convergence accuracy, and the third convergence accuracy, where the total weight is 1, and the higher the convergence accuracy of the branch, the greater the weight ratio, so as to improve the credibility of the final integrated prediction result. For example, the ratio of the convergence accuracy of each branch to the sum of the convergence accuracy of the three branches is set as the weight of the branch to obtain the trusted weight ratio.

[0041] Ensemble learning improves prediction accuracy and stability by combining the outputs of multiple prediction models. The contribution of each model is weighted according to its reliability (i.e., convergence accuracy), and the results are combined to achieve better prediction results. Next, the first, second, and third prediction branches are weighted according to their respective credibility weights. A gas composition prediction plug-in is constructed based on the weighted combination of the first, second, and third prediction branches. The output of the gas composition prediction plug-in is the weighted sum of the outputs of the first, second, and third prediction branches. In this way, the ensemble learning method can fully leverage the strengths of each prediction model, reduce the overfitting or bias that may be caused by a single model, and effectively improve the accuracy and reliability of SF6 gas composition predictions.

[0042] Furthermore, this application also includes: Read the current load ratio of the target ring main unit; use the target ring main unit as the equipment retrieval constraint, the current load ratio as the operation retrieval constraint, and the predicted SF6 gas data, SF6 gas concentration, cabinet air temperature, and cabinet air humidity as conditional retrieval constraints; based on a predetermined similarity comparison threshold, perform information retrieval in combination with the power Internet to obtain similar ring main unit operation record data that meets a preset number, and construct a similar ring main unit operation record data set; based on the similar ring main unit operation record data set, count the number of insulation faults within a predetermined time range, calculate the insulation fault ratio, and set the insulation fault ratio as the predicted insulation failure probability.

[0043] Specifically, first, the current load ratio of the target ring main unit is obtained through the real-time monitoring system. The load ratio is usually provided by the power monitoring system, which reflects the current load situation of the ring main unit. The load ratio is the ratio between the current actual load carried by the ring main unit and its rated load.

[0044] Next, the target ring main unit is used as the device retrieval constraint, that is, the specific identity of the target ring main unit (ring main unit device number, device type, etc.) is determined according to the device ID or device name, ensuring that only data records related to the device are queried; the current load ratio is used as the operation retrieval constraint, that is, according to the current load ratio of the target ring main unit, the retrieval conditions are set, and only the records of ring main units operating under similar load conditions are retrieved, such as setting a load ratio range (such as 80%-90% load equipment) to ensure that the operating status of the selected equipment is similar to that of the target equipment; the predicted SF6 gas data, SF6 gas concentration, cabinet air temperature and cabinet air humidity are used as conditional retrieval constraints, that is, device records operating under similar environments to the target ring main unit are screened. A predetermined similarity comparison threshold is set, that is, the device records that are most similar to the target ring main unit under the above-mentioned constraints are screened out based on this standard; then, based on the predetermined similarity comparison threshold, information retrieval is performed in combination with the power Internet, and the similarity between devices can be evaluated based on calculation methods such as Euclidean distance, cosine similarity, and weighted average method, and sample data that meets the predetermined similarity comparison threshold is selected to obtain similar ring main unit operation record data greater than or equal to a preset number (such as 1000), and to construct a similar ring main unit operation record data set.

[0045] Then, based on the similar RMU operation record dataset, the number of insulation faults within a predetermined timeframe (e.g., within 6 hours of operation) is counted to obtain the total number of insulation faults. The ratio of the total number of insulation faults to the total number of data in the similar RMU operation record dataset is used as the predicted insulation fault probability for the target RMU. The predicted insulation fault probability is used as the probability of a future insulation fault in the target RMU for risk assessment and early warning. By statistically analyzing the operation records of similar RMUs, the insulation fault probability of the target RMU can be effectively predicted, and this probability can be used as a basis for implementing risk warnings and dynamic control measures. This method can improve the accuracy and timeliness of RMU equipment management, ensuring the reliability and safety of the power system.

[0046] Furthermore, this application also includes: If the predicted insulation failure probability is greater than the expected probability threshold, a risk warning signal is generated, and the probability deviation between the predicted insulation failure probability and the expected probability threshold is calculated to determine the failure probability deviation; the load adjustment ratio is determined based on the failure probability deviation analysis, and the target ring network cabinet is dynamically adjusted based on the load adjustment ratio.

[0047] Specifically, the predicted insulation failure probability is compared with a set expected probability threshold. This threshold is typically a critical value based on historical data and equipment safety standards. If the predicted failure probability exceeds this threshold, it indicates a high risk to the equipment and requires appropriate action. If the predicted insulation failure probability exceeds the expected probability threshold, a risk warning signal is immediately generated.

[0048] Then, the probability deviation between the predicted insulation failure probability and the expected probability threshold is calculated to determine the failure probability deviation. The failure probability deviation is the difference between the predicted insulation failure probability and the expected probability threshold, which is used to quantify the severity of the risk. The load adjustment ratio is then determined based on the failure probability deviation analysis, wherein there is a correlation between the load adjustment ratio and the failure probability deviation. The greater the deviation, the greater the load ratio that needs to be reduced. The adjustment ratio can be calculated using a linear function or other rules. Finally, the target ring network cabinet is dynamically regulated based on the load adjustment ratio, that is, the working state of the ring network cabinet is adjusted according to the calculated load adjustment ratio, such as by allocating part of the load to other ring network cabinets to reduce the working pressure of the target ring network cabinet and reduce the risk of failure. In this way, insulation failures of the ring network cabinet can be effectively prevented and the safety and stability of the power system can be guaranteed.

[0049] Furthermore, this application also includes: Based on the inspection data of the target ring network cabinet at the most recent inspection time node, a component aging feature set is obtained; the equipment aging degree is analyzed according to the component aging feature set to determine the comprehensive aging coefficient; a probability compensation coefficient is calculated according to the regional ambient temperature, the current load ratio and the comprehensive aging coefficient, and the initial probability threshold is compensated to obtain the expected probability threshold, wherein the initial probability threshold is 1%, and the probability compensation coefficient is negatively correlated with the regional ambient temperature, the current load ratio and the comprehensive aging coefficient.

[0050] Specifically, based on the inspection data of the target ring main unit at the most recent inspection time, a set of component aging characteristics is obtained, such as the aging characteristics of circuit breakers and transformers. This includes component aging data (such as contact wear, insulation material aging, number of switch operations, etc.), performance degradation data (such as electrical impedance, conductivity reduction, circuit breaker operation delay, etc.), and other physical characteristics (such as wear of mechanical components and cracks in insulation boards). The aging degree of the equipment is then analyzed based on this set of component aging characteristics. Based on the collected aging characteristics, a multi-index comprehensive analysis method is used to determine the aging degree of the target ring main unit equipment. For example, a weighted scoring method is used to assign a weighted score to each aging characteristic to obtain an overall aging score for the equipment, indicating the aging degree of the equipment.

[0051] The probability compensation coefficient is then calculated based on the regional ambient temperature, the current load ratio and the comprehensive aging coefficient, which can be calculated through an empirical formula or a regression model. The probability compensation coefficient is negatively correlated with the regional ambient temperature, the current load ratio and the comprehensive aging coefficient, that is, a higher temperature will accelerate equipment aging, thereby increasing the probability of failure, and a higher load will make the equipment more prone to failure. As the degree of equipment aging increases, the probability of failure will also increase; that is, when the temperature, load ratio or aging degree increases, the compensation coefficient will decrease, thereby reducing the failure probability threshold, and the compensation coefficient is in the range of 0 to 1.

[0052] An initial probability threshold of 1% is obtained. This initial probability threshold is then compensated based on a probability compensation coefficient. This coefficient is multiplied by the initial probability threshold, and the product of the two is used as the expected probability threshold. This solution dynamically adjusts the expected failure probability threshold for a device by comprehensively considering factors such as device aging, ambient temperature, and load ratio, thereby achieving more accurate and reliable failure risk warnings.

[0053] In summary, the ring main unit status assessment and control method provided in this application has the following technical effects: Through fixed-point monitoring, the SF6 gas concentration, cabinet air temperature and cabinet air humidity of the target ring main unit are obtained and transmitted to the cloud platform; then, in the cloud platform, the inspection log is queried to obtain the inspection data of the target ring main unit at the most recent inspection time node, wherein the inspection data includes SF6 gas composition records; further based on the inspection log and SF6 gas composition records, the SF6 gas composition of the target ring main unit at the current time node is predicted to obtain predicted SF6 gas data; then, based on the predicted SF6 gas data, SF6 gas concentration, cabinet air temperature and cabinet air humidity, an insulation fault probability analysis is performed in combination with the power Internet to obtain a predicted insulation fault probability; if the predicted insulation fault probability exceeds the expected probability threshold, a risk warning signal is generated, and finally, the target ring main unit is dynamically regulated according to a predetermined plan; the above method can improve the accuracy and reliability of the insulation performance evaluation of the ring main unit, so that potential insulation fault threats can be discovered in a timely manner, and targeted means can be collected for dynamic regulation to ensure the power supply reliability and safety of the power system.

[0054] Example 2: Based on the same invention concept as the state evaluation and control method of a ring main unit in the above embodiment, this application also provides a state evaluation and control platform for a ring main unit. Please refer to the attached Figure 2 ,include: The data monitoring module 11 is used for fixed-point monitoring to obtain the SF6 gas concentration, air temperature and air humidity in the target ring network cabinet, and transmit them to the cloud platform; the inspection data acquisition module 12 is used to query the inspection log in the cloud platform to obtain the inspection data of the target ring network cabinet at the most recent inspection time node, wherein the inspection data includes the SF6 gas composition record; the gas composition prediction module 13 is used to predict and obtain the SF6 gas composition of the target ring network cabinet at the current time node based on the inspection log and SF6 gas composition record, and obtain the predicted SF6 gas data; the insulation fault probability analysis module 14 is used to perform insulation fault probability analysis based on the predicted SF6 gas data, SF6 gas concentration, air temperature and air humidity in the cabinet, combined with the power Internet, to obtain the predicted insulation fault probability; the dynamic control module 15 is used to generate a risk warning signal if the predicted insulation fault probability exceeds the expected probability threshold, and dynamically control the target ring network cabinet according to a predetermined plan.

[0055] Furthermore, the ring main unit status assessment and control platform is also used for: the SF6 gas composition record includes oxygen content, hydrogen fluoride content, sulfur dioxide content and acid value.

[0056] Furthermore, the status assessment and control platform of the ring main unit is also used to: calculate the current interval duration based on the most recent inspection time node and the current time node; based on the inspection log, collect the sample initial SF6 gas composition set, sample interval duration set and sample measured SF6 gas composition set of the same type of ring main unit within the preset area at the time of the most recent inspection; use the sample initial SF6 gas composition set, sample interval duration set and sample measured SF6 gas composition set as training data, perform supervised training on multiple prediction operators respectively, and integrate and construct a gas composition prediction plug-in; input the SF6 gas composition record and the current interval duration into the gas composition prediction plug-in, and output the predicted SF6 gas data.

[0057] Furthermore, the state evaluation and control platform of the ring main unit is also used to: obtain multiple prediction operators, wherein the multiple prediction operators include BP neural network, random decision forest and support vector machine; use the sample initial SF6 gas composition set, sample interval time set and sample measured SF6 gas composition set as training data, and divide the training data into nine equal parts, select nine times with replacement to construct a first sample data set, and iterate three times to obtain three sample data sets; use the three sample data sets to supervise the BP neural network, random decision forest and support vector machine respectively until the predetermined number of training times is reached, and obtain the sample data sets in the state of the ring main unit. The converged first prediction branch, second prediction branch and third prediction branch; obtaining the first convergence accuracy, second convergence accuracy and third convergence accuracy of the first prediction branch, the second prediction branch and the third prediction branch, and configuring the trusted weight ratio according to the first convergence accuracy, the second convergence accuracy and the third convergence accuracy; based on the principle of ensemble learning, weighting the first prediction branch, the second prediction branch and the third prediction branch respectively according to the trusted weight ratio, and integrating to construct the gas composition prediction plug-in, wherein the output of the gas composition prediction plug-in is the weighted result of the output of the first prediction branch, the second prediction branch and the third prediction branch.

[0058] Furthermore, the status assessment and control platform of the ring main unit is also used to: read the current load ratio of the target ring main unit; use the target ring main unit as the equipment retrieval constraint, the current load ratio as the operation retrieval constraint, and the predicted SF6 gas data, SF6 gas concentration, cabinet air temperature and cabinet air humidity as conditional retrieval constraints; based on a predetermined similarity comparison threshold, information retrieval is performed in combination with the power Internet to obtain similar ring main unit operation record data that meets a preset number, and construct a similar ring main unit operation record data set; based on the similar ring main unit operation record data set, the number of insulation faults within a predetermined time range is counted, the insulation fault ratio is calculated, and the insulation fault ratio is set as the predicted insulation failure probability.

[0059] Furthermore, the status assessment and control platform of the ring network cabinet is also used to: generate a risk warning signal if the predicted insulation failure probability is greater than the expected probability threshold, and calculate the probability deviation between the predicted insulation failure probability and the expected probability threshold to determine the failure probability deviation; determine the load adjustment ratio based on the failure probability deviation analysis, and dynamically control the target ring network cabinet based on the load adjustment ratio.

[0060] Furthermore, the status assessment and control platform of the ring network cabinet is also used to: obtain a component aging feature set based on the inspection data of the target ring network cabinet at the most recent inspection time node; analyze the degree of equipment aging according to the component aging feature set to determine the comprehensive aging coefficient; calculate a probability compensation coefficient according to the regional ambient temperature, the current load ratio and the comprehensive aging coefficient, compensate the initial probability threshold, and obtain the expected probability threshold, wherein the initial probability threshold is 1%, and the probability compensation coefficient is negatively correlated with the regional ambient temperature, the current load ratio, and the comprehensive aging coefficient.

[0061] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The state evaluation and control method and specific examples of a ring network cabinet in the aforementioned embodiment one are also applicable to the state evaluation and control platform of a ring network cabinet in this embodiment. Through the aforementioned detailed description of the state evaluation and control method of a ring network cabinet, those skilled in the art can clearly understand the state evaluation and control platform of a ring network cabinet in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the platform disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0062] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0063] Obviously, those skilled in the art may make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A method for status evaluation and control of a ring main unit, characterized in that: Methods include: Fixed-point monitoring obtains the SF6 gas concentration, air temperature and air humidity of the target ring network cabinet, and transmits them to the cloud platform; In the cloud platform, query the inspection log to obtain the inspection data of the target ring main unit at the most recent inspection time node, wherein the inspection data includes the SF6 gas composition record; Based on the inspection log and the SF6 gas composition record, the SF6 gas composition of the target ring main unit at the current time node is predicted and obtained to obtain the predicted SF6 gas data; According to the predicted SF6 gas data, SF6 gas concentration, air temperature in the cabinet and air humidity in the cabinet, insulation fault probability analysis is performed in combination with the power internet to obtain a predicted insulation fault probability; If the predicted insulation failure probability exceeds the expected probability threshold, a risk warning signal is generated, and the target ring main unit is dynamically adjusted according to a predetermined plan.

2. A method for status evaluation and control of a ring main unit according to claim 1, characterized in that: The SF6 gas composition record includes oxygen content, hydrogen fluoride content, sulfur dioxide content and acid value.

3. A method for status evaluation and control of a ring main unit according to claim 2, characterized in that: Based on the inspection log and the SF6 gas composition record, the SF6 gas composition of the target ring main unit at the current time node is predicted and obtained to obtain the predicted SF6 gas data, including: The current interval duration is calculated based on the most recent inspection time node and the current time node; Based on the inspection log, collect the sample initial SF6 gas composition set, sample interval duration set and sample measured SF6 gas composition set of the same type of ring main unit in the preset area during the most recent inspection; The sample initial SF6 gas composition set, the sample interval time set and the sample measured SF6 gas composition set are used as training data, and supervised training is performed on multiple prediction operators respectively, and a gas composition prediction plug-in is integrated to be constructed; The SF6 gas composition record and the current interval duration are input into the gas composition prediction plug-in, and the predicted SF6 gas data is output.

4. A method for status evaluation and control of a ring main unit according to claim 3, characterized in that: The sample initial SF6 gas composition set, the sample interval time set and the sample measured SF6 gas composition set are used as training data, and supervised training is performed on multiple prediction operators respectively, and a gas composition prediction plug-in is integrated and constructed, including: Acquire multiple prediction operators, wherein the multiple prediction operators include BP neural network, random decision forest and support vector machine; The sample initial SF6 gas composition set, the sample interval time set and the sample measured SF6 gas composition set are used as training data, and the training data is equally divided into nine parts, selected nine times with replacement to construct a first sample data set, and iteratively selected three times to obtain three sample data sets; Using the three sample data sets, supervised training is performed on the BP neural network, the random decision forest and the support vector machine respectively until a predetermined number of training times is reached to obtain a first prediction branch, a second prediction branch and a third prediction branch that are in convergence; Obtaining a first convergence accuracy, a second convergence accuracy, and a third convergence accuracy of the first prediction branch, the second prediction branch, and the third prediction branch, and configuring a trusted weight ratio according to the first convergence accuracy, the second convergence accuracy, and the third convergence accuracy; Based on the principle of ensemble learning, the first prediction branch, the second prediction branch and the third prediction branch are weighted respectively according to the proportion of the trusted weights, and the gas composition prediction plug-in is integrated to construct the gas composition prediction plug-in, wherein the output of the gas composition prediction plug-in is the weighted result of the outputs of the first prediction branch, the second prediction branch and the third prediction branch.

5. The method for status evaluation and control of a ring main unit according to claim 1, characterized in that: According to the predicted SF6 gas data, SF6 gas concentration, air temperature in the cabinet and air humidity in the cabinet, insulation fault probability analysis is performed in combination with the power internet to obtain predicted insulation fault probability, including: Read the current load ratio of the target ring main unit; Taking the target ring main unit as the equipment retrieval constraint, the current load ratio as the operation retrieval constraint, the predicted SF6 gas data, SF6 gas concentration, the air temperature in the cabinet and the air humidity in the cabinet as the conditional retrieval constraints, based on the predetermined similarity comparison threshold, information retrieval is performed in combination with the power Internet to obtain similar ring main unit operation record data that meets the preset number, and construct a similar ring main unit operation record data set; Based on the similar ring main unit operation record data set, the number of insulation faults within a predetermined time range is counted, the insulation fault ratio is calculated, and the insulation fault ratio is set as the predicted insulation fault probability.

6. A method for status evaluation and control of a ring main unit according to claim 5, characterized in that: If the predicted insulation fault probability exceeds the expected probability threshold, a risk warning signal is generated, and the target ring main unit is dynamically regulated according to a predetermined plan, including: If the predicted insulation failure probability is greater than the expected probability threshold, a risk warning signal is generated, and the probability deviation between the predicted insulation failure probability and the expected probability threshold is calculated to determine the failure probability deviation; The load adjustment ratio is determined according to the fault probability deviation analysis, and the target ring main unit is dynamically regulated based on the load adjustment ratio.

7. A method for status evaluation and control of a ring main unit according to claim 6, characterized in that: Get the expected probability threshold, including: Based on the inspection data of the target ring main unit at the most recent inspection time node, obtain the component aging feature set; Perform equipment aging degree analysis based on the component aging feature set to determine a comprehensive aging coefficient; The probability compensation coefficient is calculated according to the regional ambient temperature, the current load ratio and the comprehensive aging coefficient, and the initial probability threshold is compensated to obtain the expected probability threshold, wherein the initial probability threshold is 1%, wherein the probability compensation coefficient is negatively correlated with the regional ambient temperature, the current load ratio and the comprehensive aging coefficient.

8. A ring main unit status evaluation and control platform, characterized in that: The steps for implementing the method for status evaluation and control of a ring main unit according to any one of claims 1 to 7 include: The data monitoring module is used to monitor the SF6 gas concentration, air temperature and air humidity of the target ring network cabinet at a fixed point and transmit them to the cloud platform; The inspection data acquisition module is used to query the inspection log in the cloud platform to obtain the inspection data of the target ring network cabinet at the most recent inspection time node, wherein the inspection data includes SF6 gas composition records; A gas composition prediction module is used to predict and obtain the SF6 gas composition of the target ring main unit at the current time node based on the inspection log and the SF6 gas composition record, and obtain the predicted SF6 gas data; An insulation fault probability analysis module is used to perform insulation fault probability analysis based on the predicted SF6 gas data, SF6 gas concentration, air temperature in the cabinet, and air humidity in the cabinet in combination with the power Internet to obtain a predicted insulation fault probability; The dynamic control module is used to generate a risk warning signal if the predicted insulation failure probability exceeds the expected probability threshold, and dynamically control the target ring network cabinet according to a predetermined plan.

Images