An automatic fault switching circuit system

Abstract

The present invention relates to a fault automatic switching line system, which monitors the fault tripping signal of the working power supply, determines the position of the powered-off line according to the fault tripping signal; evaluates the carrying capacity of the standby line to be switched and predicts the incoming line load after switching; performs the backup power supply automatic switching operation according to the carrying capacity of the standby line and the incoming line load; after the backup power supply automatic switching operation, automatically adjusts the line access mode and load distribution strategy according to the line parameters and the whole station load situation monitored in real time. The present invention solves the problem that it is difficult to evaluate the carrying capacity of the standby line and predict the incoming line load in the prior art, resulting in insufficient accuracy and safety of the backup power supply automatic switching operation. The random forest model is trained using the basic information of the standby line and environmental factors to evaluate the carrying capacity of the standby line, improving the accuracy of the backup power supply automatic switching control. An ARIMA model is established according to the load distribution and time law characteristics of the power station to predict the incoming line load, improving the safety of the backup power supply automatic switching operation.

Description

Technical Field

[0001] The invention belongs to the technical field of transformer substations and relates to a fault automatic switching line system. Background Art

[0002] In modern society, power supply and various data communications are extremely dependent on line transmission. Whether it is household electricity, corporate production, Internet communications, financial transactions and other activities, once a line fails, it will cause serious impact. For example, power line failures may cause power outages, affect residents' lives, corporate shutdowns, and even endanger the normal operation of important places such as hospitals and transportation hubs. Communication line failures will cause network interruptions, hinder online business development, and affect people's information exchange. However, in actual operation, lines often fail due to natural factors (such as bad weather, earthquakes, etc.), human factors (such as construction damage, accidental touch, etc.) and aging. In order to quickly restore transmission when a line fails, ensure the normal progress of various activities, and reduce losses caused by failures, there is an urgent need for a fault automatic switching line system that can automatically and quickly switch transmission to a backup line after detecting a line failure, maintaining normal power supply and data communication functions.

[0003] However, when a section of the busbar in the substation loses pressure due to an incoming line fault, a busbar fault, or a fault in the upper power supply, the backup power supply automatic input device (backup automatic input) will immediately start and perform the corresponding action. The specific operation is to close the section switch connecting the two sections of the busbar. In this way, the power supply line that originally carried part of the load will instantly bear all the load of the entire station. This situation often leads to a series of adverse consequences. Because the power supply line was originally operated according to a certain load configuration, it suddenly has to bear the load of the entire station, which causes the incoming line load to increase sharply, far exceeding the range that it can carry during normal operation. This sudden increase in the incoming line load is very likely to trigger the action of the incoming line protection device, which will cause the incoming line to trip. Once the incoming line trips, the entire substation loses the power supply from the incoming line, which eventually causes a serious failure situation of the entire station losing pressure. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides a fault automatic switching line system, which aims to construct a load prediction algorithm to evaluate the carrying capacity of the standby line to be switched, and predict the incoming line load after the switching; establish an adaptive control mechanism to automatically adjust the line access mode and load distribution strategy according to the line parameters and the load situation of the entire station monitored in real time; reduce the incoming line load by adjusting the transformer tap, switching on and off the reactive compensation device, etc., so that it is kept within a safe range to prevent tripping and loss of pressure in the entire station.

[0005] The object of the present invention can be achieved by the following technical solutions:

[0006] This application provides a fault automatic switching line system, including a fault monitoring module, a load analysis module, a backup power supply automatic switching module, and an adaptive control module. The fault monitoring module, the load analysis module, the backup power supply automatic switching module, and the adaptive control module are communicatively connected. Among them:

[0007] The fault monitoring module is used to monitor the fault tripping signal of the working power supply and determine the position of the power-off line according to the fault tripping signal;

[0008] The load analysis module is used to evaluate the carrying capacity of the standby line to be switched and predict the incoming line load after switching;

[0009] The backup power supply automatic switching module is used to perform the backup power supply automatic switching action according to the carrying capacity of the standby line and the incoming line load;

[0010] The adaptive control module is used to automatically adjust the line access mode and load distribution strategy according to the line parameters and the whole station load conditions monitored in real time after the backup power supply automatic switching action.

[0011] Further, in the fault monitoring module, the monitoring of the fault tripping signal of the working power supply and the determination of the position of the power-off line according to the fault tripping signal include the following steps:

[0012] Establish a monitoring connection: correctly electrically connect and communicatively connect the fault monitoring module with the working power supply and its related protection devices;

[0013] Configure the sensor network: install current transformers, voltage transformers, and temperature sensors at the key line nodes of the working power supply for real-time collection of line data, and the line data includes current, voltage, and temperature data;

[0014] Real-time signal monitoring: receive the line data transmitted by the sensors and the protection devices in real time, and when a signal change of the line data that meets the preset fault tripping condition is found, it is determined that a fault tripping signal is monitored;

[0015] Signal verification and confirmation: when a certain fault tripping signal is initially monitored, further verify the fault tripping signal with reference to the changes in other line data at the same moment;

[0016] Collect the change data of line parameters: after monitoring the fault tripping signal, collect the change data of the relevant line data of each line;

[0017] Comparison analysis and judgment: compare and analyze the change data of the line data with the data during normal operation stored in advance, and determine the position of the power-off line in combination with the power system topology structure;

[0018] Generate a fault report: Generate a fault report based on the monitored fault trip signal and the process of determining the location of the power-off line.

[0019] Further, in the load analysis module, the evaluation of the load-bearing capacity of the standby line to be switched includes the following steps:

[0020] S1. Obtain the basic information of the standby line. The basic information includes basic parameters and rated parameters. The basic parameters include line material, cross-sectional area, and line length. The rated parameters include rated voltage, rated current, and rated power.

[0021] S2. Identify the environmental factors where the standby line is located. The environmental factors include temperature, humidity, and electromagnetic interference.

[0022] S3. Use the obtained basic information and environmental factors of the standby line as independent variables, and the load-bearing capacity of the standby line as the dependent variable to train the load-bearing capacity analysis model.

[0023] S4. Input the basic information and environmental factors of the standby line to be switched into the load-bearing capacity analysis model, and output the load-bearing capacity of the standby line to be switched.

[0024] Further, the load-bearing capacity analysis model is configured as a random forest model, including the following construction steps:

[0025] Divide the training set, validation set, and test set: Divide the preprocessed sample data into a training set, validation set, and test set. The training set is used to train the random forest model. The validation set is used to monitor and adjust the performance of the model during training, and optimize the model parameters according to the performance indicators on the validation set. The test set is used to evaluate the final performance of the model after training to verify whether the model can accurately predict new data.

[0026] Train the random forest model: Use the divided training set to train the random forest model. During training, each decision tree grows independently based on the data in the training set, analyzes and judges the input features according to the set parameters, gradually forms its own decision path, and finally constructs a random forest model composed of multiple decision trees.

[0027] Evaluate the model performance: Use the test set to evaluate the performance of the trained random forest model. The indicators used for evaluation include mean squared error, mean absolute error, and coefficient of determination.

[0028] Further, in the load analysis module, the prediction of the incoming line load after switching includes the following steps:

[0029] Obtain the load distribution according to the switch status of all load types in the power station.

[0030] Obtain the current time regularity features, where the time regularity features include production activity features and resident life features;

[0031] Input the load distribution and time regularity features into a preset time series model to predict the sequence data of the incoming line load of the spare line within a future period of time;

[0032] Obtain the statistical information of the sequence data of the incoming line load, where the statistical information includes the maximum value, minimum value, average value, and standard deviation.

[0033] Further, the time series model is configured as an ARIMA model and includes the following steps:

[0034] Collect historical data: Collect historical data samples from relevant data sources of the power station, where the historical data samples include the incoming line load, load distribution, and time regularity features;

[0035] Test data stationarity: Use the ADF test to perform stationarity tests on the preprocessed historical data samples of the incoming line load;

[0036] Model order determination: Determine the autoregressive order and moving average order by observing the autocorrelation function graph and partial autocorrelation function graph;

[0037] Train the ARIMA model: Train the ARIMA model according to the determined autoregressive order and moving average order. During the training process, the model adjusts the autoregressive coefficient and moving average coefficient to minimize the error between the predicted value and the true value;

[0038] Evaluate model performance: Use the test set to evaluate the performance of the trained ARIMA model.

[0039] Further, in the backup power supply automatic switching module, the backup power supply automatic switching operation according to the carrying capacity of the spare line and the incoming line load includes the following steps:

[0040] Compare the load difference between the carrying capacity of the spare line and the incoming line load;

[0041] When the load difference exceeds a preset load difference threshold, suspend the backup power supply automatic switching operation;

[0042] When the load difference does not exceed the preset load difference threshold, close the corresponding sectional switch to connect the backup power supply from the faulty working power supply line to the spare line. Further, the load difference is calculated as follows:

[0043] ,

[0044] where, De Indicates the load difference; C e Indicates the carrying capacity of the standby line; J e Indicates the incoming line load.

[0045] Furthermore, when the line load exceeds the preset load threshold, the adaptive control module reduces the incoming line load by adjusting the transformer tap or switching the reactive power compensation device.

[0046] Advantages of the present invention:

[0047] (1) By monitoring the fault trip signal of the working power supply and determining the position of the de-energized line according to the fault trip signal; evaluating the carrying capacity of the standby line to be switched and predicting the incoming line load after switching; performing the backup power supply automatic switching (BPA) action according to the carrying capacity of the standby line and the incoming line load; after the BPA action, automatically adjusting the line access mode and load distribution strategy according to the real-time monitored line parameters and the whole station load situation. The present invention solves the problem that it is difficult to evaluate the bearing capacity of the standby line and predict the incoming line load in the prior art, resulting in insufficient accuracy and safety of the BPA action.

[0048] (2) Using the existing basic information of the standby line and environmental factors as independent variables and the carrying capacity of the standby line as the dependent variable to train a random forest model, realizing the prediction of the carrying capacity of the standby line based on the existing information, and improving the accuracy of the BPA control.

[0049] (3) Establishing an ARIMA model according to the load distribution and time law characteristics of the power station to predict the sequence data of the incoming line load of the standby line in a future period of time, realizing the prediction of the incoming line load according to the actual situation, and improving the safety of the BPA. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] For the convenience of those skilled in the art to understand, the present invention will be further described below with reference to the accompanying drawings.

[0051] Figure 1 It is a structural diagram of a fault automatic switching line system in the present invention.

[0052] Figure 2 It is a flowchart of constructing a random forest model in an embodiment of the present invention.

[0053] Figure 3 It is a flowchart of constructing an ARIMA model in an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0054] To further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, elaborate in detail on the specific implementation manners, structures, features and their effects of the present invention as follows.

[0055] Please refer to Figures 1 - 3 , this application provides a fault automatic switching line system, including a fault monitoring module, a load analysis module, a backup power supply automatic switching module and an adaptive control module. The fault monitoring module, the load analysis module, the backup power supply automatic switching module and the adaptive control module are communicatively connected, wherein:

[0056] The fault monitoring module is used to monitor the fault tripping signal of the working power supply and determine the position of the powered-off line according to the fault tripping signal;

[0057] In this embodiment, the fault monitoring module is a very crucial part of the fault automatic switching line system. It mainly undertakes two important tasks. One is to closely monitor the fault tripping signal of the working power supply. During the normal operation of the working power supply, through various sensors, monitoring instruments, etc. connected to the power supply and its protection device, it always pays attention to whether there is a fault tripping signal caused by the action of the protection device due to line short circuit, overload, equipment failure, etc. Once such a signal is captured, it immediately proceeds to the next step. The other is to accurately determine the position of the powered-off line based on the monitored situation. Given that there may be multiple lines drawn from the working power supply in the power system, it, by clearly understanding the system topology structure and combining information such as the parameter changes of each line at the moment of the fault, through analysis and comparison, precisely locates which line is powered off due to the fault tripping, thereby providing a key basis for subsequent system processing such as the backup power supply automatic switching action, and ensuring the rapid restoration of power supply to the corresponding load.

[0058] Further, in the fault monitoring module, the monitoring of the fault tripping signal of the working power supply and the determination of the position of the powered-off line according to the fault tripping signal include the following steps:

[0059] Establish a monitoring connection: First, correctly electrically connect and communicatively connect the fault monitoring module with the working power supply and its related protection device. Through appropriate signal lines, data lines, etc., ensure that various signals emitted by the output end of the working power supply and the protection device can be obtained, laying a foundation for subsequent monitoring work.

[0060] Configure the sensor network: Install various types of sensors, including current transformers, voltage transformers, temperature sensors, etc., at key line nodes of the working power supply, such as the incoming line end, outgoing line end, etc. These sensors are used to collect real-time parameter information such as current, voltage, and temperature on the line, so as to monitor the operating status of the working power supply from multiple angles, because abnormal changes in these parameters often accompany the occurrence of faults, and these changes may trigger fault trip signals.

[0061] Real-time signal monitoring: Start the monitoring program, and the fault monitoring module begins to receive various signal data transmitted by sensors and protection devices in real time. Continuously analyze and process these data, focusing on signal characteristics related to fault tripping, such as sudden increase in current (which may indicate a short circuit fault), sudden drop in voltage (which may be due to reasons such as line break or overload), etc. Once a signal change that meets the preset fault tripping conditions is detected, it is determined that a fault tripping signal has been monitored.

[0062] Signal verification and confirmation: When a possible fault tripping signal is initially monitored, to avoid misjudgment, the fault monitoring module will further verify the signal. It may refer to the changes in other relevant parameters at the same moment, or conduct comparative analysis with historical normal operation data. For example, if only the current suddenly increases but then quickly returns to normal and no obvious abnormalities are found in other parameters, it may not be a real fault tripping signal but a temporary interference. Only after multi-faceted verification and confirmation is it finally determined that an effective fault tripping signal has been monitored.

[0063] Collect data on line parameter changes: At the moment when a fault tripping signal is monitored, the fault monitoring module will immediately collect the change data of relevant parameters of each line (all lines drawn from the working power supply) at this time. These parameters include but are not limited to current, voltage, power factor, etc. Because when a fault occurs in different lines leading to the tripping of the working power supply, the change situations of their own parameters will be different, and these change data will become an important basis for determining the position of the power-off line.

[0064] Comparative analysis and judgment: Compare and analyze the parameter change data of each line at the moment of the fault with the normal operation parameter templates and fault characteristic templates of each line pre-stored in the module. For example, if the current of a certain line drops to zero and the voltage also drops significantly at the moment of fault tripping, while the parameter changes of other lines do not conform to this characteristic, then it is very likely that this line is the power-off line. Through this detailed comparative analysis and logical judgment, combined with the understanding of the topology structure of the entire power system (the connection relationship and direction of each line), accurately determine the position of the power-off line.

[0065] Generate a fault report: Once the location of the power-off line is determined, the fault monitoring module generates a detailed fault report based on the detected fault trip signal and the entire process of determining the power-off line location. This report includes key information such as the time of the fault occurrence, the specific characteristics of the fault trip signal, the number or identification of the power-off line, etc.

[0066] Transmit fault information: Timely transmit the generated fault report to other relevant modules in the system, such as the load analysis module, the automatic standby power supply module, the adaptive control module, etc. So that these modules can quickly carry out their respective corresponding work based on the received fault information. For example, the load analysis module can evaluate the load-carrying capacity of the standby line accordingly, and the automatic standby power supply module can decide whether and how to perform the automatic standby power supply operation, etc., thereby ensuring the efficient and coordinated operation of the entire fault automatic switching line system.

[0067] The load analysis module is used to evaluate the load-carrying capacity of the standby line to be switched and predict the incoming line load after switching;

[0068] In this embodiment, the load analysis module is mainly responsible for two important tasks. One is to evaluate the load-carrying capacity of the standby line to be switched. For this purpose, it comprehensively collects the design parameters of the standby line, such as material, cross-sectional area, length, rated voltage, current, etc., and at the same time considers the operating environment factors of the line, such as temperature, electromagnetic interference, special area conditions, etc. And combined with the operating parameters such as current and voltage obtained in real time through sensors, it uses professional calculation and analysis models for dynamic evaluation to accurately determine the actual load-carrying capacity of the standby line at present.

[0069] The second is to predict the incoming line load after switching. This requires first carefully analyzing the existing load distribution of the whole station, mastering the specific usage and overall structure of different types of loads such as power, lighting, electrothermal, etc. It is also necessary to combine the specific strategies of the automatic standby power supply module and estimate according to the different line switching methods it may adopt. At the same time, fully considering the trends of the whole station load changing with time, production activities, residents' living rules, etc., using statistical analysis methods and prediction models to accurately predict the incoming line load entering the standby line after switching, so as to provide strong support for the reasonable operation and effective decision-making of the system.

[0070] Further, in the load analysis module, the evaluation of the load-carrying capacity of the standby line to be switched includes the following steps:

[0071] S1. Obtain the basic information of the standby line;

[0072] This is the primary step of the evaluation, aiming to comprehensively collect various key parameters of the standby line, which lay the foundation for accurately evaluating its load-carrying capacity subsequently.

[0073] Basic parameters: The material of the circuit (such as copper, aluminum, etc.) determines the electrical conductivity of the circuit. Different materials have different resistivity, which will affect the resistance when current passes through. The cross-sectional area is directly related to the amount of current that the circuit can allow to pass through. Generally, the larger the cross-sectional area, the relatively larger the current that can be carried. The length of the circuit cannot be ignored either. Due to the cumulative effect of resistance in a longer circuit, more electrical energy will be lost during transmission, which will in turn affect the carrying capacity.

[0074] Rated parameters: The rated voltage clarifies the normal operating voltage value that the circuit is designed to adapt to, which is the voltage standard to ensure the normal operation of the circuit and the connected equipment. The rated current limits the maximum current value that the circuit can pass through in a safe state. Exceeding this value may cause problems such as overheating and damage to the circuit. The rated power comprehensively reflects the maximum power that the circuit can transmit under the conditions of rated voltage and rated current, and is one of the important indicators to measure the carrying capacity of the circuit.

[0075] S2. Identify the environmental factors where the standby circuit is located;

[0076] In addition to the basic information of the circuit itself, the environmental factors where it is located also have an important impact on the carrying capacity, and need to be accurately identified and included in the evaluation scope.

[0077] Temperature: The level of the ambient temperature will change the resistance characteristics of the circuit material. For example, when the temperature rises, the resistance of the metal material usually increases, which means that at the same voltage, the current that the circuit can pass through will decrease, thus reducing the carrying capacity of the circuit.

[0078] Humidity: A high-humidity environment may affect the insulation performance of the circuit, making the circuit more likely to have a leakage phenomenon. This will not only cause power loss, but also may pose a safety hazard, indirectly affecting the ability of the circuit to normally carry the load.

[0079] Electromagnetic interference: If there is strong electromagnetic interference around the standby circuit, it may interfere with the transmission stability of the electrical signals in the circuit, resulting in problems such as signal distortion and reduced transmission efficiency, and thus having an adverse impact on the ability of the circuit to carry the load.

[0080] S3. Train the carrying capacity analysis model

[0081] In this step, use the obtained basic information of the standby circuit and the identified environmental factors as independent variables, and the carrying capacity of the standby circuit (including the maximum carrying current and the maximum carrying voltage) as the dependent variable to train a special carrying capacity analysis model.

[0082] Basis for model construction: Based on a large amount of experimental data, actual operation data, and theoretical analysis results, an internal logical relationship model is established among the basic information of the line, environmental factors, and carrying capacity. For example, based on the actual carrying capacity test data of lines with different materials, cross-sectional areas, and lengths under various temperature, humidity, and electromagnetic interference environments, to determine how these factors interact and ultimately affect the maximum carrying current and maximum carrying voltage of the line.

[0083] Purpose of model training: By continuously inputting different combinations of independent variable data (i.e., various combinations of basic line information and environmental factors) and corresponding dependent variable data (actual carrying capacity data), the model learns and masters the laws between these data, so as to accurately predict the carrying capacity of the corresponding standby line according to the newly input basic line information and environmental factor conditions.

[0084] S4. Input the basic information of the standby line to be switched and the environmental factors into the carrying capacity analysis model, and output the carrying capacity of the standby line to be switched.

[0085] This is the last step of the evaluation. Accurately input the specific basic information of the standby line to be switched and the data of its actual environmental factors into the trained carrying capacity analysis model.

[0086] Data input requirements: Ensure that the input data is complete, accurate, and matches the data format and type used during model training. For example, if the temperature data during model training is in degrees Celsius, then the temperature data of the standby line to be switched should also use the same unit.

[0087] Interpretation of output results: The model will quickly output the carrying capacity of the standby line to be switched according to the input data based on the laws it has learned, specifically manifested as the values of the maximum carrying current and maximum carrying voltage. These output results will directly provide a key basis for subsequent judgment on whether the standby line can meet the load requirements after switching, helping the system reasonably arrange line switching and load distribution operations to ensure the stable operation of the entire power supply system.

[0088] Furthermore, the carrying capacity analysis model is configured as a random forest model and includes the following construction steps:

[0089] I. Data collection and preprocessing stage

[0090] Collect sample data: Collect a large amount of sample data about backup lines from multiple existing backup line-related data sets and actual operation records. These data should cover basic information such as different line materials, cross-sectional areas, lengths, rated voltages, rated currents, and rated powers, as well as actual carrying capacity data under corresponding environmental factors such as temperature, humidity, and electromagnetic interference, including specific values ​​of maximum carrying current and maximum carrying voltage. Ensure that the collected data is representative and diverse, and can reflect various possible line conditions and environmental conditions, so as to provide sufficient basic data for building an accurate random forest model.

[0091] Data cleaning: Check the integrity of the collected sample data and remove data records with missing values, erroneous values, or data that are obviously inconsistent with the actual situation. For example, if the cross-sectional area value of a line in a record is negative, this is obviously inconsistent with physical reality and such data should be deleted. In the case of a small number of missing values, it can be processed according to other relevant data features or using appropriate filling methods (such as mean filling, median filling, etc.) to ensure the integrity and availability of the data.

[0092] Data standardization: Since the data value ranges of different features may vary greatly, for example, the route length may be in kilometers, while the temperature is in degrees Celsius, in order to avoid some features having a dominant influence on the model due to their large value range during the subsequent model training process, all data need to be standardized. Commonly used standardization methods include Z-score standardization, which is to subtract the mean of each feature from the value of the feature, and then divide it by the standard deviation of the feature, so that the data of all features are converted into a standard normal distribution with a mean of 0 and a standard deviation of 1, so that different features can be treated equally in model training.

[0093] 2. Feature Selection and Extraction Stage

[0094] Determine relevant features: Based on the prior knowledge of the impact on the backup line's carrying capacity and the results of data analysis, determine which features to select from the collected data as input features for the random forest model. In this case, line material, cross-sectional area, length, rated voltage, rated current, rated power, temperature, humidity, electromagnetic interference, etc. are clearly selected as relevant features because these features have been proven to have a significant impact on the backup line's carrying capacity.

[0095] Feature Encoding: For some non-numerical features, such as wire material (possible values are copper, aluminum, etc.), encoding processing is required to convert them into numerical forms so that the model can handle them. For example, the one-hot encoding method can be used. Encoding the wire material as copper as [1, 0], the wire material as aluminum as [0, 1], etc., enables different material categories to participate in model training in appropriate numerical forms.

[0096] III. Model Parameter Setting Stage

[0097] Determine the number of decision trees: The random forest model is an ensemble model composed of multiple decision trees. First, the number of decision trees included in the model needs to be determined. Too few decision trees may lead to insufficient generalization ability of the model and inability to accurately capture complex relationships in the data; too many decision trees may lead to overfitting, resulting in good performance of the model on the training data but poor performance on new data.

[0098] Generally, methods such as cross-validation can be used to determine the appropriate number of decision trees. For example, starting from a small initial value (such as 10 trees), gradually increase the number of decision trees, and at the same time observe the performance metrics (such as mean squared error, accuracy, etc.) of the model on the validation dataset. When the performance metrics no longer improve significantly, the corresponding number of decision trees is the more appropriate value.

[0099] Set other parameters: In addition to the number of decision trees, some other related parameters need to be set, such as the maximum depth of each decision tree, the minimum number of samples based on which node splitting is performed, etc. The maximum depth of each decision tree determines the growth degree of the decision tree. An overly deep decision tree may lead to overfitting. Generally, by trying different depth values (such as 3, 5, 7, etc.) and combining the performance of the model on the validation dataset, the appropriate depth can be determined. The minimum number of samples based on which node splitting is performed controls the growth speed and complexity of the decision tree. A smaller minimum number of samples may lead to an overly complex decision tree, and a larger minimum number of samples may lead to slow growth of the decision tree. Similarly, the appropriate value needs to be determined according to the actual situation and the performance on the validation dataset.

[0100] IV. Model Training Stage

[0101] Dividing the training set, validation set, and test set: The preprocessed sample data is divided into a training set, a validation set, and a test set according to a certain ratio (such as the common 7:2:1 or 8:1:1, etc.). The training set is used to train the random forest model so that the model can learn the patterns and feature relationships in the data; the validation set is used to monitor and adjust the performance of the model during the training process, and the model parameters are optimized according to the performance indicators on the validation set; the test set is used to evaluate the final performance of the model after the model training is completed to verify whether the model can accurately predict new data.

[0102] Training the random forest model: Use the divided training set to train the random forest model. During the training process, each decision tree will grow independently based on the data in the training set, analyze and judge the input features according to the set parameters (such as the basis for node splitting, maximum depth, etc.), gradually form their own decision paths, and finally construct a random forest model composed of multiple decision trees.

[0103] During the training process, closely monitor the performance indicators of the model on the validation set, such as mean squared error, accuracy, etc., and adjust the model parameters (such as the number of decision trees, maximum depth, etc.) in a timely manner according to the changes of these indicators to improve the performance and generalization ability of the model.

[0104] V. Model Evaluation and Optimization Phase

[0105] Evaluating the model performance: Use the test set to evaluate the performance of the trained random forest model. Calculate various performance indicators of the model on the test set, such as mean squared error (MSE), mean absolute error (MAE), coefficient of determination (R²), etc. The mean squared error measures the average of the squared differences between the model prediction values and the true values, and the smaller the value, the more accurate the model prediction; the mean absolute error measures the average of the absolute values between the model prediction values and the true values, and the smaller the value, the better; the coefficient of determination reflects the proportion of the model prediction results that can explain the variation of the real data, and its value ranges from 0 to 1, and the closer it is to 1, the better the prediction effect of the model.

[0106] Optimizing the model: If the performance indicators of the model on the test set are not ideal, the model needs to be optimized. Possible optimization measures include: increasing the amount of training data, readjusting the model parameters (such as the number of decision trees, maximum depth, etc.), trying different feature combinations or extracting new features, etc. Through continuous evaluation and optimization, the random forest model can achieve better performance and accurately predict its carrying capacity based on the input basic information of the standby line and environmental factors, providing a reliable basis for the line switching decision in the automatic fault switching line system.

[0107] Furthermore, in the load analysis module, the predicting the incoming line load after switching includes the following steps:

[0108] Obtain the load distribution based on the switching status of all load types in the power station;

[0109] Load type analysis: The area served by the power station usually covers various types of loads, such as industrial loads (various production equipment), commercial loads (electrical equipment in shopping malls, office buildings, etc.), and residential loads (household appliances, etc.). The electricity consumption characteristics and patterns of each load type are different. Industrial loads may have a relatively large and concentrated electricity consumption during the production period during the day on weekdays, commercial loads have a relatively stable electricity demand during business hours, and residential loads show regular patterns such as morning and evening peaks following the daily life schedule of residents.

[0110] Switching status monitoring: By monitoring the switching status (on or off) related to each load type, it is possible to understand in real time which loads are consuming electricity and which are idle. For example, in a factory, by monitoring the switching status of equipment in each production workshop, the number and type of currently operating equipment can be determined, and thus the electricity consumption scale of industrial loads can be inferred; for a residential community, monitoring the opening and closing status of switches in the electricity meter boxes of each household can generally grasp the electricity consumption of residential loads.

[0111] Load distribution determination: By comprehensively analyzing the monitoring results of the switching status of all load types, the current load distribution of the power station can be accurately obtained. That is, it is clear how much electricity is consumed by different regions and different types of loads at the current moment, and the proportion they account for in the total load, thus providing a basic data based on the current actual electricity consumption situation for subsequent prediction of the incoming line load.

[0112] Obtain the current time regularity characteristics, where the time regularity characteristics include production activity characteristics and residential life characteristics;

[0113] Production activity characteristics: For loads involved in production and operation activities such as industry and commerce, it is necessary to deeply understand the time regularity of their production activities. For example, manufacturing factories may implement a shift system of day shift, middle shift, and night shift, with different production intensities and equipment opening situations in different shifts. The day shift is usually the busiest production period with all equipment fully open, and the electricity consumption will reach a peak; while some service enterprises may only operate during specific periods on weekdays, and their electricity demand also shows corresponding time regularities. By analyzing the time arrangements, shift rotations, seasonal production adjustments, etc. of these production activities, the influence law of production activities on the load of the power station can be grasped.

[0114] Resident living characteristics: The electricity consumption pattern of residents also has an important impact on the load of the power station. Generally speaking, there is a small peak in residents' electricity consumption after getting up in the morning for activities such as preparing breakfast and washing up; from evening to night is the peak electricity consumption period, when people get off work and turn on various electrical appliances such as lighting, TV, air conditioner, and water heater; the electricity consumption on weekends may also be different from that on weekdays. On weekends, due to longer staying at home, electrical appliances may be used more frequently, and the load may change compared to weekdays. Analyzing these characteristics such as the daily routine of residents, the differences between weekends and weekdays helps to comprehensively understand the impact of residents' living on the load of the power station.

[0115] Integration of overall time pattern characteristics: Integrate various time pattern characteristics such as production activity characteristics and resident living characteristics to form a complete description of the current time pattern. This description can reflect the electricity consumption change pattern of different types of loads in the service area of the entire power station based on time factors during the current time period, providing important input information related to the time dimension for subsequent prediction using the time series model.

[0116] Input the load distribution and time pattern characteristics into a preset time series model to predict the sequence data of the incoming load of the standby line for a period of time in the future;

[0117] Principle of time series model: The time series model is a statistical model that predicts future data based on historical data and the chronological order. It assumes that there is a certain regularity and correlation in the data over time, and infers the future development trend by analyzing past data patterns. In this scenario, using the obtained load distribution (based on the current actual electricity consumption) and time pattern characteristics (reflecting the electricity consumption change pattern of different loads based on time) as inputs, the time series model can uncover the hidden time series relationships behind these data, and then predict the incoming load of the standby line for a period of time in the future.

[0118] Prediction process: Accurately input the integrated load distribution and time pattern characteristic data into a preset time series model. The model will analyze and process the input data according to its internally set algorithms and parameters. For example, it may identify the change trend of the incoming load in the future for a period of time under similar past load distributions and time patterns, and make predictions based on this. The model will output the sequence data of the incoming load of the standby line for a period of time in the future (such as the next hour, day, or week, etc., and the specific time span can be set according to actual needs). This sequence data presents the expected incoming load of the standby line at different future times.

[0119] Obtain the statistical information of the incoming load sequence data, where the statistical information includes the maximum value, minimum value, average value, and standard deviation.

[0120] Maximum and minimum values: Obtaining the maximum and minimum values in the sequence data of the incoming line load can intuitively understand the extreme situations that the incoming line load of the standby line may encounter in a future period. The maximum value represents the highest load that may occur during the prediction period, which is crucial for determining whether the standby line can withstand such a high load; the minimum value reflects the lowest load that may occur, helps to understand the operation of the standby line during light load periods, and can also be used to analyze the fluctuation range of the load.

[0121] Average value: Calculating the average value of the sequence data of the incoming line load can obtain the average level of the incoming line load of the standby line in a future period. The average value is an indicator that comprehensively reflects the overall situation of the future load. It can give us a more balanced understanding of the future load-bearing pressure of the standby line, helps to consider the average load situation during system planning and decision-making, so as to reasonably arrange operations such as line switching.

[0122] Standard deviation: The standard deviation measures the degree of dispersion of the sequence data of the incoming line load relative to the average value. A larger standard deviation indicates that the load fluctuates greatly in a future period, and there may be some sudden high or low load situations, which need to be paid attention to during system design and operation. For example, additional measures may be needed to cope with large load fluctuations to ensure the stable operation of the standby line. By obtaining these statistical information, we can understand the situation of the incoming line load of the future standby line more comprehensively and deeply, and provide more sufficient decision-making basis for the effective operation of the automatic fault-switching line system.

[0123] Furthermore, the time series model, configured as an ARIMA (Autoregressive Integrated Moving Average) model, includes the following steps:

[0124] I. Data collection and preparation stage

[0125] Collect historical data:

[0126] Collect historical data on the incoming line load and related load distribution, time pattern characteristics, etc. from relevant data sources such as the monitoring system of the power station and the electricity meter records. These data should cover a long enough time period to reflect the load changes in various situations such as different seasons, weekdays and weekends, and different production activity cycles. For example, collect hourly or daily incoming line load data for the past year or even several years, as well as the load distribution situation (the electricity consumption ratio of various types of loads, etc.) and time pattern characteristics (production activity arrangements, residents' living schedules, etc.) at the corresponding moments.

[0127] Data cleaning and preprocessing:

[0128] Perform an integrity check on the collected historical data, and eliminate data records with missing values, error values, or those that are clearly not in line with the actual situation. For a small number of missing values, appropriate filling methods can be used, such as mean filling, median filling, or interpolation filling based on the data before and after, to ensure the integrity of the data.

[0129] Standardize the data to make the value ranges of different features comparable. Common standardization methods such as Z-score standardization, that is, subtract the mean of each feature from the value of the feature, and then divide by the standard deviation of the feature, to transform the data into a standard normal distribution form with a mean of 0 and a standard deviation of 1, so that each data can be treated equally in subsequent model training and avoid affecting the model performance due to excessive differences in data magnitudes.

[0130] II. Data Stationarity Test Phase

[0131] Test the data stationarity:

[0132] Use statistical test methods, such as the ADF (Augmented Dickey - Fuller) test, to perform a stationarity test on the preprocessed historical data of the incoming line load. Stationarity refers to the property that the statistical characteristics (such as mean, variance, etc.) of the data do not change over time. If the data is not stationary, directly applying the ARIMA model may lead to inaccurate prediction results.

[0133] The ADF test determines whether the data is stationary by calculating the test statistic and comparing it with the corresponding critical value. If the test statistic is less than the critical value, the null hypothesis is rejected, that is, the data is considered stationary; conversely, if the test statistic is greater than the critical value, the null hypothesis is accepted, indicating that the data is non-stationary.

[0134] Data differencing processing:

[0135] If it is found through the ADF test that the incoming line load data is non-stationary, the data needs to be differenced to make it reach a stationary state. Differencing is to subtract the data of the previous moment from the current data, and through multiple differencing operations (first-order differencing, second-order differencing, etc.), until the data passes the stationarity test.

[0136] III. Model Order Determination Phase

[0137] Determine the autoregressive order (p):

[0138] Preliminarily determine the autoregressive order by observing the autocorrelation function (ACF) graph and the partial autocorrelation function (PACF) graph. The autocorrelation function describes the correlation between a time series and its own lagged values, and the partial autocorrelation function describes the correlation between a time series and its own lagged values after controlling the influence of intermediate lagged values.

[0139] Generally speaking, the autocorrelation function gradually decays to zero after a certain number of lags, while the partial autocorrelation function suddenly truncates or rapidly decays to zero after a certain order. Based on these characteristics, the possible range of values for the autoregressive order can be initially determined. For example, if the ACF plot shows a slow decay trend and the PACF plot suddenly truncates after 3 lags, then it can be initially considered to be set around 3.

[0140] Determine the moving average order (q):

[0141] Similarly, the ACF plot and PACF plot are used to determine the moving average order. The moving average model fits the time series data by taking the moving average of the error terms, and the method for determining its order is similar to that of the autoregressive order.

[0142] Observe the decay pattern of the data in the ACF plot and PACF plot, and combine the understanding of the characteristics of the incoming line load data to determine the possible range of values for the moving average order. For example, if the ACF plot shows a rapid decay trend after 2 lags and the PACF plot also has an obvious change after 2 lags, then it can be considered to be set around 2.

[0143] Comprehensively determine the model order:

[0144] After initially determining the possible range of values for the autoregressive order and the moving average order, different combinations are tried, and the ARIMA models under different combinations are evaluated using the validation dataset.

[0145] IV. Model training stage

[0146] Divide the training set and the test set:

[0147] Divide the historical data of the incoming line load that has been preprocessed and whose stationarity and model order have been determined into a training set and a test set according to a certain ratio (such as the common 7:3 or 8:2, etc.). The training set is used to train the ARIMA model so that the model can learn the regularities and characteristic relationships in the data; the test set is used to evaluate the final performance of the model after the model training is completed to verify whether the model can accurately predict new data.

[0148] Train the ARIMA model:

[0149] Use the divided training set to train the ARIMA model. According to the determined autoregressive order, moving average order, and differencing order (if differencing is performed), input the data in the training set into the ARIMA model and train it according to the algorithm and parameter settings of the model. During the training process, the model will adjust parameters such as autoregressive coefficients and moving average coefficients to minimize the error between the predicted value and the true value, thereby learning the internal regularities and changing trends of the incoming line load data.

[0150] V. Model Evaluation and Optimization Phase

[0151] Evaluate the model performance:

[0152] Use the test set to evaluate the performance of the trained ARIMA model. Calculate various performance metrics of the model on the test set, such as mean squared error (MSE), mean absolute error (MAE), coefficient of determination (R²), etc.

[0153] Optimize the model:

[0154] If the performance metrics of the model on the test set are not satisfactory, the model needs to be optimized. Possible optimization measures include: increasing the amount of training data, readjusting the model order (such as re - determining the autoregressive order and moving average order), trying different differencing methods (if the data requires differencing), or further feature extraction and processing of the data, etc.

[0155] Through continuous evaluation and optimization, the ARIMA model can achieve better performance, accurately predict the incoming load of the standby line within a certain period in the future according to the input load distribution and time - law characteristics, and provide a reliable basis for the effective operation of the automatic fault - switching line system.

[0156] The backup power supply automatic - switching module is used to perform the backup power supply automatic - switching action according to the bearing capacity of the standby line and the incoming load.

[0157] In this embodiment, the backup power supply automatic - switching module plays a key role in the automatic fault - switching line system. It mainly works based on the bearing capacity of the standby line provided by the load analysis module (covering indicators such as maximum bearing current and maximum bearing voltage, which define the electrical load limit that the line can withstand) and the predicted incoming load.

[0158] First, the backup power supply automatic - switching module fully grasps the bearing capacity of the standby line and determines the load range it can accept. At the same time, it receives and understands the situation of the incoming load, which is a dynamic value affected by various factors such as the load distribution of the whole station and the electricity consumption law.

[0159] Then, compare and analyze the incoming load with the bearing capacity of the standby line. If the incoming load is within the acceptable range, from the perspective of load bearing, the backup power supply automatic - switching action is feasible; otherwise, the switching strategy needs to be reconsidered, such as adjusting the load distribution, etc., to ensure compliance with the bearing capacity requirements.

[0160] Finally, when the conditions for the automatic switching-in of the standby power supply are met, that is, the incoming line load is appropriate, the line is free of faults, the connection is normal, etc., the module will trigger an action to perform operations such as closing the sectionalizing switch and disconnecting the connection to the faulty working power supply, so as to achieve a fast and accurate switching of the load from the faulty working power supply line to the standby line, ensuring the continuity and stability of the power supply and reducing the impact of power outages.

[0161] Further, in the automatic switching-in of the standby power supply module, the performing of the automatic switching-in of the standby power supply according to the carrying capacity of the standby line and the incoming line load includes the following steps:

[0162] Compare the load difference between the carrying capacity of the standby line and the average value of the incoming line load;

[0163] In the operation process of the automatic switching-in of the standby power supply module, the key step to be carried out first is to compare the load difference between the carrying capacity of the standby line and the incoming line load. Here, it is specifically mentioned that the average value of the incoming line load is taken because the incoming line load may be a value that changes dynamically over time, and taking its average value can comprehensively reflect the general situation of the load to be switched to the standby line to a certain extent.

[0164] By comparing the carrying capacity index of the standby line (such as the maximum carrying current, the maximum power that can be carried corresponding to the maximum carrying voltage, etc.) with the average value of the incoming line load, the specific load difference value between the two is calculated. This difference value can intuitively present the remaining capacity of the standby line when undertaking the incoming line load or whether there is an overload risk, etc. For example, if the maximum power that the standby line can carry is 1000 kW and the average value of the incoming line load is 800 kW, then the load difference is 200 kW, which indicates that the standby line has a certain carrying capacity margin in theory.

[0165] When the load difference exceeds the preset load difference threshold, suspend the automatic switching-in of the standby power supply action;

[0166] After obtaining the load difference value, it will be compared with the preset load difference threshold. This preset load difference threshold is a key index set based on factors such as the safe operation requirements of the system, the performance characteristics of the line, and past experience.

[0167] When the load difference exceeds the threshold, it means that the incoming load is either too high relative to the carrying capacity of the backup line, which may cause the backup line to be overloaded, or too low, which may cause some power quality problems such as power factor (for example, the power factor may decrease when lightly loaded, affecting the efficiency of the power grid). In order to avoid these potential risks from causing adverse effects on the system, such as damage to the backup line and unstable power supply, the backup automatic transfer module will decisively suspend the backup automatic transfer action at this time. After the action is suspended, the system may further analyze the reasons, such as re-evaluating the prediction accuracy of the incoming load, checking the actual status of the backup line, etc., so as to take appropriate measures to adjust this unbalanced situation in the future.

[0168] When the load difference does not exceed a preset load difference threshold, the corresponding section switch is closed to connect the backup power supply from the failed working power supply line to the backup line.

[0169] On the contrary, when it is found through comparison that the load difference does not exceed the preset load difference threshold, it means that from the perspective of load bearing, the current backup line has sufficient capacity to undertake the incoming load that is about to be switched, and this acceptance is within the reasonable range set by the system and will not cause the various risks mentioned above.

[0170] In this case, the standby automatic transfer module will close the corresponding section switches according to the established procedures. The operation of these section switches is the key link to accurately and quickly connect the backup power supply from the faulty working power supply line to the backup line. By closing these switches, the load originally powered by the faulty working power supply can be successfully transferred to the backup line, and the backup power supply continues to provide power to it, thereby ensuring the continuity of power supply and minimizing the power outage time and impact on electrical equipment caused by the working power failure. Furthermore, the load difference is calculated as follows:

[0171] ,

[0172] In the formula, D e Indicates load difference; C e Indicates the carrying capacity of the backup line; J e Indicates the incoming line load.

[0173] The adaptive control module is used to automatically adjust the line access mode and load distribution strategy according to the line parameters and the load situation of the entire station monitored in real time after the standby automatic switching action.

[0174] In this embodiment, the adaptive control module operates based on the information monitored in real time after the backup power supply automatic switching operation is completed. On the one hand, it closely monitors line parameters such as voltage, current, resistance, reactance, power factor, etc., so as to understand the actual operating status of each line. For example, voltage can reflect the power supply level, and current reflects the load degree, etc. On the other hand, it comprehensively grasps the load situation of the entire substation, covering the total load, distribution, nature, and change trend, etc., to clarify the power consumption scale, whether there is overload and imbalance, the characteristics of different types of loads, and the dynamic change law of the load.

[0175] Based on these real-time monitoring situations, the module will automatically adjust the line access mode and load distribution strategy. In terms of the access mode, if there are abnormalities such as large voltage fluctuations in a certain standby line, its access point will be changed or the connection sequence will be adjusted, etc., to ensure stable power supply. In terms of the load distribution strategy, when the total load of the entire substation increases or there is local overload, part of the load will be transferred to achieve balanced distribution; if the power supply quality of the sensitive load line is not good, the load distribution of other lines will be adjusted to give priority to ensuring its power supply, so as to ensure the stable operation of the power system, improve the power utilization efficiency, and meet the power supply requirements of different loads.

[0176] Furthermore, when the line load exceeds the preset load threshold, the adaptive control module reduces the incoming line load by adjusting the transformer tap or switching the reactive power compensation device.

[0177] In this embodiment, during the operation of the adaptive control module, when it is monitored that the line load exceeds the preset load threshold, this indicates that the load situation of the current line has reached a critical state that requires intervention and regulation. The preset load threshold is a key index set based on factors such as the designed carrying capacity of the line, past operation experience, and the requirement to ensure the stable operation of the power system. Once the line load breaks through this threshold, a series of problems such as line overload, excessive voltage fluctuation, and increased power loss may be caused, which will in turn affect the normal power supply of the power system and the safe operation of electrical equipment.

[0178] (1) Adjusting the transformer tap to reduce the incoming line load:

[0179] Principle of transformer tap: The transformer tap refers to the adjustable connection point set on the transformer winding. By changing the position of the tap, the transformation ratio of the transformer can be changed, thereby affecting the magnitude of the output voltage and current. When the line load is too high, the adaptive control module can utilize this characteristic for adjustment.

[0180] Adjustment process and effect: Specifically, when it is found that the line load exceeds the threshold, the module will issue an instruction to adjust the transformer tap. For example, if the transformer tap is adjusted in the step-down direction, the output voltage of the transformer will decrease. With the load resistance relatively fixed, the decrease in voltage will cause the current in the line to decrease, thereby reducing the incoming line load. Such an adjustment method can relieve the overload pressure on the line to a certain extent, bring the line load back to a more reasonable range, and ensure the normal operation of the line.

[0181] (2) Switching the reactive power compensation device to reduce the incoming line load:

[0182] Function of the reactive power compensation device: The reactive power compensation device is mainly used to compensate the reactive power in the power grid. In the power system, the existence of reactive power will cause the current to increase, thereby increasing the line loss and reducing the power factor. Switching the reactive power compensation device can change the reactive power distribution in the power grid, thus affecting the load condition of the line.

[0183] Switching operation and influence: When the line load exceeds the load threshold, the adaptive control module will decide whether to switch the reactive power compensation device and how to switch it according to the current reactive power condition of the power grid. If the current reactive power of the power grid is insufficient, the module may switch in the reactive power compensation device to increase the reactive power supply in the power grid, so that the power factor in the power grid is improved. With the voltage and active power relatively fixed, the increase in the power factor will cause the current to decrease, thereby reducing the incoming line load. On the contrary, if there is an excess of reactive power in the power grid, the module may switch out some of the reactive power compensation devices to optimize the reactive power distribution and also achieve the purpose of reducing the incoming line load.

[0184] The above is only a preferred embodiment of the present invention and does not impose any form of limitation on the present invention. Although the present invention has been disclosed as above with a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content to form an equivalent embodiment with equivalent changes within the scope of the technical solution of the present invention. However, as long as it does not depart from the content of the technical solution of the present invention, any brief modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A fault automatic switching line system, characterized in that: It includes a fault monitoring module, a load analysis module, a standby automatic transfer module and an adaptive control module, wherein the fault monitoring module, the load analysis module, the standby automatic transfer module and the adaptive control module are communicatively connected, wherein: The fault monitoring module is used to monitor the fault tripping signal of the working power supply and determine the position of the power-off line according to the fault tripping signal; The load analysis module is used to evaluate the carrying capacity of the standby line to be switched and predict the incoming line load after switching, and includes the following steps: Obtain load distribution based on the switching conditions of all load types in the power station; Acquire current time regularity characteristics, wherein the time regularity characteristics include production activity characteristics and resident life characteristics; Input the load distribution and time regularity characteristics into a preset time series model to predict the incoming load sequence data of the standby line in a future period of time; Obtaining statistical information of the incoming line load sequence data, wherein the statistical information includes a maximum value, a minimum value, an average value, and a standard deviation; The standby automatic switching module is used to perform standby automatic switching actions according to the standby line carrying capacity and incoming line load; The adaptive control module is used to automatically adjust the line access mode and load distribution strategy according to the line parameters and the load situation of the whole station monitored in real time after the standby automatic switching action; In the load analysis module, the evaluation of the carrying capacity of the standby line to be switched includes the following steps: S1. Obtain basic information of the standby line, wherein the basic information includes basic parameters and rated parameters; the basic parameters include line material, cross-sectional area and line length; the rated parameters include rated voltage, rated current and rated power; S2. Identify environmental factors of the backup line, where the environmental factors include temperature, humidity and electromagnetic interference; S3, using the acquired basic information of the backup line and environmental factors as independent variables and the carrying capacity of the backup line as the dependent variable, training the carrying capacity analysis model; S4, inputting the basic information of the standby line to be switched and the environmental factors thereof into the carrying capacity analysis model, and outputting the carrying capacity of the standby line to be switched; The carrying capacity analysis model is configured as a random forest model and includes the following construction steps: Divide the preprocessed sample data into training set, validation set and test set. The training set is used to train the random forest model; the validation set is used to monitor and adjust the performance of the model during the training process, and optimize the parameters of the model according to the performance indicators on the validation set; the test set is used to evaluate the final performance of the model after the model training is completed, so as to verify whether the model can accurately predict new data; Training the random forest model: Use the divided training set to train the random forest model. During the training process, each decision tree grows independently based on the data in the training set, analyzes and judges the input features according to the set parameters, gradually forms their own decision paths, and finally constructs a random forest model composed of multiple decision trees; Evaluate model performance: Use the test set to evaluate the performance of the trained random forest model. The evaluation indicators include mean square error, mean absolute error, and determination coefficient.

2. A fault automatic switching line system according to claim 1, characterized in that: In the fault monitoring module, the fault tripping signal of the monitoring working power supply is monitored, and the position of the power-off line is determined according to the fault tripping signal, including the following steps: Establish monitoring connection: Make correct electrical and communication connections between the fault monitoring module and the working power supply and its related protection devices; Configure a sensor network: install current transformers, voltage transformers and temperature sensors at key line nodes of the working power supply to collect line data in real time, including current, voltage and temperature data; Real-time signal monitoring: Receive line data from sensors and protection devices in real time. When a signal change in line data that meets the preset fault tripping conditions is found, it is determined that a fault tripping signal has been detected. Signal verification and confirmation: After a fault trip signal is initially monitored, the fault trip signal is further verified by referring to the changes in other line data at the same time; Collect line parameter change data: after detecting the fault trip signal, collect the change data of the relevant line data of each line; Comparative analysis and judgment: Compare and analyze the line data changes with the pre-stored normal operation data, and determine the location of the power outage line in combination with the power system topology; Generate fault report: Generate fault report based on the monitored fault trip signal and the process of determining the location of the power-off line.

3. A fault automatic switching line system according to claim 1, characterized in that: The time series model is configured as an ARIMA model, comprising the following steps: Collect historical data: collect historical data samples from relevant data sources of the power station, the historical data samples including incoming line load, load distribution and time regularity characteristics; Test data stability: Use ADF test to test the stability of the pre-processed historical data samples of incoming line load; Model order determination: Determine the autoregressive order and moving average order by observing the autocorrelation function graph and partial autocorrelation function graph; Training ARIMA model: The ARIMA model is trained according to the determined autoregressive order and moving average order. During the training process, the model adjusts the autoregressive coefficient and the moving average coefficient to minimize the error between the predicted value and the true value. Evaluate model performance: Use the test set to evaluate the performance of the trained ARIMA model.

4. A fault automatic switching line system according to claim 1, characterized in that: In the standby automatic switching module, the standby automatic switching action is performed according to the standby line carrying capacity and the incoming line load, including the following steps: Compare the load difference between the backup line carrying capacity and the incoming line load; When the load difference exceeds a preset load difference threshold, the standby automatic switching action is suspended; When the load difference does not exceed a preset load difference threshold, the corresponding section switch is closed to connect the backup power supply from the failed working power supply line to the backup line.

5. A fault automatic switching line system according to claim 4, characterized in that: The load difference is calculated as follows: ， In the formula, D e Indicates load difference; C e Indicates the carrying capacity of the backup line; J e Indicates the incoming line load.

6. A fault automatic switching line system according to claim 1, characterized in that: The adaptive control module reduces the incoming line load by adjusting the transformer tap or switching on and off the reactive compensation device when the line load exceeds a preset load threshold.

Images