A method and system for isolation protection control of intelligent ring main unit

Abstract

This application relates to the field of intelligent ring main unit technology, specifically an isolation protection control method and system for intelligent ring main units. Through multi-dimensional data anomaly verification, physical constraint temperature rise prediction, and dynamic health assessment, it significantly improves the reliability and maintenance efficiency of the ring main unit's power supply unit. The system verifies ambient temperature, cabinet internal temperature, and current data in real time to ensure input reliability; it calculates transient temperature rise coefficients and thermal inertia coefficients based on historical operating data, constructing a temperature rise prediction model that integrates physical laws to accurately predict temperature rise; the health calculation comprehensively considers actual temperature rise deviations and coefficient offsets to achieve early fault warning. When the health level falls below a threshold, the faulty unit is automatically isolated and power is switched to a healthy unit, avoiding unplanned power outages.

Description

Technical Field

[0001] This application relates to the field of intelligent ring main unit technology, specifically an isolation protection control method and system for intelligent ring main units. Background Technology

[0002] A ring main unit (RMU) is an electrical device consisting of a group of power transmission and distribution equipment (high-voltage switchgear) housed in a metal or non-metal insulated cabinet or assembled into a modular ring network power supply unit. Its core components utilize load switches and fuses, offering advantages such as simple structure, small size, low price, improved power supply parameters and performance, and enhanced power supply safety.

[0003] Ring main units can isolate / switch different power supply units. Since the operating loads of different power supply units vary, components within each unit (such as high-voltage switches, fuses, transformers, and cable joints) need to be inspected based on their actual health status to improve maintenance efficiency. However, current technology lacks effective health assessment methods, thus relying on periodic maintenance, which to some extent wastes maintenance resources. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide an isolation protection control method and system for intelligent ring main units to solve the problems in the background art.

[0005] To achieve the above objectives, this application adopts the following technical solution:

[0006] This application discloses an isolation protection and control method for an intelligent ring main unit, comprising the following steps:

[0007] Obtain the operating data of multiple power supply units in the ring main unit at a target number of historical time points, wherein the operating data includes ambient temperature, cabinet temperature and current value;

[0008] Anomaly verification is performed on the operational data to obtain the verification results;

[0009] When the verification result indicates that the operating data is normal, operating features are extracted from the operating data; the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle are calculated based on the operating data of the pre-built operating database, wherein the operating features include the load rate, current fluctuation rate and historical temperature rise of the previous historical time point at multiple historical time points, and the operating database includes operating data within a period of time at the beginning of the maintenance cycle;

[0010] The operational characteristics are input into a pre-built temperature rise prediction model based on physical constraints to obtain the predicted temperature rise at the current time point;

[0011] The actual temperature rise at the current time point is calculated, and the health of the power supply unit is calculated based on the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle; and the power supply unit is subjected to isolation protection control for maintenance purposes based on the health.

[0012] In one embodiment of this application, it further includes:

[0013] When the verification result indicates that the running data is abnormal, the running data is updated at the next time point, and the abnormality verification of the running data is performed again to obtain the verification result; until the verification result indicates that the running data is normal.

[0014] In one embodiment of this application, anomaly verification is performed on the running data to obtain a verification result, including:

[0015] Based on a pre-built data reference range, the values ​​of operating parameters at multiple time points in the operating data are checked for value anomalies to obtain the value anomaly verification results.

[0016] The cabinet temperature sequence and the ambient temperature sequence in the operation data are subjected to first-order difference to obtain a temperature first-order difference sequence; the value of each element in the temperature first-order difference sequence is compared with the preset maximum temperature change threshold to obtain the temperature jump verification result.

[0017] The first-order difference sequence of each parameter in the running data is normalized and then averaged. The average value is then compared with a preset minimum change threshold to obtain the freeze anomaly verification result.

[0018] The various parameter sequences of the operational data are time-series aligned, and a temperature rise sequence is calculated based on the ambient temperature sequence and the cabinet temperature sequence. The slope of the temperature rise sequence and the slope of the current value square sequence are extracted. Multivariate physical constraint verification is performed based on pre-configured verification rules to obtain multivariate physical constraint verification results. The verification rules include: the absolute value of the difference between the slope of the temperature rise sequence and the slope of the current value square sequence is less than a preset following threshold, and the cabinet temperature at the same time point is greater than or equal to the ambient temperature.

[0019] If the results of the abnormal value verification, the temperature jump verification, the freezing anomaly verification, and the multivariable physical constraint verification are all normal, the running data is determined to pass the anomaly verification; otherwise, it fails the anomaly verification.

[0020] In one embodiment of this application, calculating the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit for the current maintenance cycle based on the operational data of the operational database includes:

[0021] Obtain standard operating data for a period of time at the start of the current maintenance cycle, and extract standard temperature rise, standard historical temperature rise, and standard heat loss power at multiple time points from the standard operating data;

[0022] Construct a multiple linear fitting equation, wherein the mathematical expression of the multiple linear fitting equation is:

[0023]

[0024] In the formula, Indicates the temperature rise at the reference time point. Indicates the transient temperature rise coefficient. Indicates heat loss power. Indicates the thermal inertia coefficient. This indicates the historical temperature rise at the previous reference time point. Indicates the error term;

[0025] Substituting the standard temperature rise, standard historical temperature rise, and standard heat loss power at multiple time points into the multivariate linear fitting equation, and combining it with the least squares method for fitting, the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle are obtained.

[0026] In one embodiment of this application, when the verification result indicates that the running data is normal, extracting running features from the running data includes:

[0027] Based on the aforementioned operational data, the most recent... Load rate at historical time points Current fluctuation rate Historical warming at the previous historical point in time , wherein the load rate The current fluctuation rate and the historical temperature rise at the previous point in time The mathematical expression is:

[0028]

[0029]

[0030]

[0031] In the formula, Indicates a historical time point index. For the number of historical time points, Rated current, For time points The current value, The maximum current value for the running data. The minimum current value for the operating data. For time points The temperature inside the cabinet, For time points The ambient temperature value.

[0032] In one embodiment of this application, the method for constructing the temperature rise prediction model based on physical constraints includes:

[0033] S1, obtain operating data samples of multiple power supply units of the ring main unit in a healthy state, wherein the operating data samples include the values ​​of various operating parameters at multiple historical time points;

[0034] S2, the running data sample is filtered to obtain filtered data; and the abnormal time distribution in the filtered data is removed to obtain preprocessed data;

[0035] S3, the preprocessed data is segmented to obtain multiple segmented samples, wherein the last historical time point of each segmented sample is the sample reference time point.

[0036] S4, extract operating feature samples from the segmented samples, wherein the operating feature samples include load rate sequence samples, current fluctuation rate samples and historical temperature rise samples;

[0037] S5, using the temperature rise sample corresponding to the sample baseline time point as the label, and combining it with the running feature sample to construct training samples, thereby obtaining the training dataset;

[0038] S6. Select training samples for the current training batch from the training dataset and input the training samples for the current training batch into the LSTM model to obtain the predicted temperature rise;

[0039] S7, calculate the loss between the predicted temperature rise and the label based on the pre-built physical constraint-based loss function, and adjust the parameters of the LSTM model based on the loss direction propagation and the gradient descent method.

[0040] S8. Repeat steps S6-S7 until training is complete, and obtain a temperature rise prediction model based on physical constraints.

[0041] In one embodiment of this application, the mathematical expression of the loss function based on physical constraints is:

[0042]

[0043]

[0044]

[0045] In the formula, For the total loss, For MSE loss, For physical loss, As a weighting factor, The number of samples in a single training batch. For sample index, Indicates the first The predicted temperature rise corresponding to each sample Indicates the first The label of each sample This is the design value for the transient temperature rise coefficient. This is the design value for the thermal inertia coefficient. For the first The heat loss power corresponding to each sample For the first The historical temperature rise of each sample.

[0046] In one embodiment of this application, the actual temperature rise at the current time point is calculated, and the health of the power supply unit is calculated based on the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle, including:

[0047] Based on the actual temperature rise at the current time point Predicted temperature rise at the current time point Calculate the first health level The mathematical expression is:

[0048]

[0049] In the formula, Indicates the maximum temperature rise;

[0050] Transient temperature rise coefficient based on the current maintenance cycle Thermal inertia coefficient during the current maintenance cycle Calculate the second health level Among them, the second health level The mathematical expression is:

[0051]

[0052] In the formula, This is the design value for the transient temperature rise coefficient. The design value is based on the reference thermal inertia coefficient;

[0053] For the first health level and the second health level The health status of the power supply unit is obtained by weighting the values. Among them, the health of the power supply unit The mathematical expression is:

[0054]

[0055] In the formula, As the first weight, It is the second weight.

[0056] In one embodiment of this application, the power supply unit is subjected to isolation and protection control for maintenance purposes based on the health status, including:

[0057] The health status is compared with a preset health status threshold. When the health status is less than the preset health status threshold, the incoming load switch of the current power supply unit is disconnected, and the interconnection switch of the adjacent power supply unit is closed to switch to the healthy adjacent power supply unit for power supply.

[0058] This application also provides an isolation and protection control system for an intelligent ring main unit, characterized in that it includes:

[0059] The acquisition module is used to acquire the operating data of multiple power supply units in the ring main unit at a target number of historical time points, wherein the operating data includes ambient temperature, cabinet temperature and current value;

[0060] The verification module is used to perform anomaly verification on the running data and obtain the verification result.

[0061] The feature extraction module is used to extract operating features from the operating data when the verification result indicates that the operating data is normal; and to calculate the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle based on the operating data of the pre-built operating database. The operating features include the load rate, current fluctuation rate and historical temperature rise of the previous historical time point at multiple historical time points. The operating database includes operating data within a certain period of time at the beginning of the maintenance cycle.

[0062] The temperature rise prediction module is used to input the operating characteristics into a pre-built temperature rise prediction model based on physical constraints to obtain the predicted temperature rise at the current time point;

[0063] The protection control module is used to calculate the actual temperature rise at the current time point, and calculate the health status of the power supply unit based on the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle; and to perform isolation protection control on the power supply unit for maintenance purposes based on the health status.

[0064] The beneficial effects of this application are as follows: The isolation protection control method and system for intelligent ring main units disclosed in this application significantly improve the reliability and maintenance efficiency of the ring main unit power supply unit through multi-dimensional data anomaly verification, physical constraint temperature rise prediction, and dynamic health assessment. The system verifies ambient temperature, cabinet temperature, and current data in real time to ensure input reliability; it calculates transient temperature rise coefficients and thermal inertia coefficients based on historical operating data, constructs a temperature rise prediction model that integrates physical laws, and accurately predicts temperature rise; the health calculation comprehensively considers actual temperature rise deviations and coefficient offsets to achieve early fault warnings. When the health level is below a threshold, the faulty unit is automatically isolated and power is switched to a healthy unit to avoid unplanned power outages. This application reduces reliance on manual inspections, optimizes preventative maintenance cycles, effectively reduces operation and maintenance costs, and the physical constraint model improves prediction accuracy, avoids the overfitting risk of purely data-driven systems, and ensures the safe and stable operation of the power grid. Attached Figure Description

[0065] The present application will be further described below with reference to the accompanying drawings and embodiments:

[0066] Figure 1 This is a structural framework diagram of an intelligent ring main unit system shown in one embodiment of this application;

[0067] Figure 2 This is a network topology diagram of an intelligent ring network cabinet in one embodiment of this application;

[0068] Figure 3 This is a flowchart illustrating an isolation protection control method for an intelligent ring main unit in one embodiment of this application;

[0069] Figure 4 This is a schematic diagram of the health calculation process in one embodiment of this application;

[0070] Figure 5 This is a structural diagram of an isolation protection control system for an intelligent ring main unit, as shown in one embodiment of this application. Detailed Implementation

[0071] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0072] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the layers related to this application and are not drawn according to the actual number, shape and size of the layers in the actual implementation. In the actual implementation, the form, number and proportion of each layer can be arbitrarily changed, and the layer layout may also be more complex.

[0073] Numerous details are explored in the following description to provide a more thorough explanation of embodiments of this application; however, it will be apparent to those skilled in the art that embodiments of this application may be practiced without these specific details.

[0074] Figure 1 This is a structural framework diagram of an intelligent ring main unit system shown in one embodiment of this application, as follows: Figure 1 As shown, the intelligent ring network cabinet includes multiple power supply units (T1-T4), each power supply unit includes a transformer, and the power supply unit is connected to the power supply line through the incoming load switch. The power supply unit is also equipped with an intelligent data acquisition terminal; the load output terminals of adjacent power supply units are connected through a tie switch.

[0075] Figure 2 Here is a network topology diagram of the intelligent ring network cabinet in one embodiment of this application, such as... Figure 2 As shown, the incoming load switch and tie switch are both connected to the edge controller. The edge controller and the intelligent acquisition terminal are connected to the cloud server through the network to realize remote monitoring and control.

[0076] Figure 3 This is a flowchart illustrating an isolation protection control method for an intelligent ring main unit in one embodiment of this application, as shown below. Figure 3 As shown, the isolation protection control method for an intelligent ring main unit in this application mainly includes the following steps:

[0077] S310, acquire the operating data of multiple power supply units in the ring main unit at a target number of historical time points, wherein the operating data includes ambient temperature, cabinet temperature and current value, and in some cases, voltage value also needs to be collected;

[0078] Specifically, the intelligent data acquisition terminal periodically collects the operating data of the power supply unit inside the ring network cabinet. The operating data includes sampling points from multiple time points prior to the current time point. Each sampling point includes the ambient temperature, cabinet temperature, and current value. The table below shows an example of the operating data.

[0079] Table 1. Example of runtime data

[0080]

[0081] S320, Perform anomaly verification on the running data and obtain the verification result;

[0082] In this application, anomaly verification is not simply "denoising," but a multi-dimensional reliability assessment that integrates measurement boundaries, temporal characteristics, and physical laws. Only data that passes rigorous verification can be used to construct reliable temperature rise prediction models and health indicators. The verification process in this application includes:

[0083] S321, based on a pre-built data reference range, perform anomaly checks on the values ​​of operating parameters at multiple time points in the operating data to obtain anomaly check results;

[0084] In this application, a physically reasonable range is set for each parameter to verify whether the data is reasonable. The table below is the data reference range table for this application, and the reference range for each parameter is shown in the table below:

[0085] Table 2. Example of Data Reference Range

[0086]

[0087] If data that meets the anomaly criteria exists, it is marked as a range anomaly.

[0088] S322, Perform a first-order difference on the cabinet temperature sequence and the ambient temperature sequence in the operating data to obtain a first-order temperature difference sequence; compare each element value in the first-order temperature difference sequence with a preset maximum temperature change threshold to obtain a temperature jump verification result.

[0089] The temperature sequence inside the cabinet and the ambient temperature sequence are first-order differencing to obtain the rate of temperature change (i.e., the first-order difference temperature sequence), which is then compared with a preset maximum temperature change threshold. Since temperature cannot change abruptly (it usually undergoes a change process), any abrupt changes are marked as abrupt anomalies.

[0090] S323, after normalizing the first-order difference sequence of each parameter in the running data, take the average value, and compare the average value with the preset minimum change threshold to obtain the freeze anomaly verification result;

[0091] Different parameters (temperature, current, voltage) have different dimensions, making direct comparison meaningless. Therefore, normalization (such as min-max normalization) is used to make the first-order difference sequences of each parameter comparable on the same scale.

[0092] Then, the average value is calculated. If the first-order difference remains unchanged for a long period, it indicates that the sensor reading has been frozen for an extended period. In this case, the data is marked as a frozen anomaly.

[0093] S324, perform time-series alignment of the various parameter sequences of the operating data, calculate the temperature rise sequence based on the ambient temperature sequence and the cabinet temperature sequence, extract the slope of the temperature rise sequence and the slope of the current value square sequence; and perform multivariate physical constraint verification based on pre-configured verification rules to obtain multivariate physical constraint verification results, wherein the verification rules include: the absolute value of the difference between the slope of the temperature rise sequence and the slope of the current value square sequence is less than a preset following threshold, and the cabinet temperature at the same time point is greater than or equal to the ambient temperature;

[0094] Specifically, the reasonableness of the data is verified using the electrical-thermal coupling relationship:

[0095] (1) Verification of the causal relationship between temperature rise and current

[0096] According to Joule's law, temperature rise should be proportional to I²R; if the current increases significantly (e.g., >50% of the rated current) but the cabinet temperature does not rise (or even falls), it is abnormal. Conversely, if the current is close to 0 but the cabinet temperature continues to rise, it may indicate an external heat source or sensor malfunction.

[0097] (2) The temperature inside the cabinet is greater than or equal to the ambient temperature (when there is no active cooling).

[0098] If the temperature inside the cabinet is lower than the ambient temperature and there is no refrigeration device → physically impossible, mark as abnormal.

[0099] S325, if the abnormal value verification result, the temperature jump verification result, the freezing abnormal verification result, and the multivariable physical constraint verification result all indicate that the data is normal, the running data is determined to pass the abnormal verification; otherwise, the abnormal verification is not passed.

[0100] S326, when the verification result indicates that the running data is abnormal, the running data is updated at the next time point, and the abnormality verification of the running data is performed again to obtain the verification result; until the verification result indicates that the running data is normal.

[0101] If any anomalies are found, the running data will be updated using a rolling update method until the running data passes the verification. Only then will the normal running data be used to execute the following process.

[0102] S330, when the verification result indicates that the operating data is normal, extract operating features from the operating data; calculate the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle based on the operating data of the pre-built operating database, wherein the operating features include the load rate, current fluctuation rate and historical temperature rise of the previous historical time point at multiple historical time points, and the operating database includes operating data within a period of time at the beginning of the maintenance cycle;

[0103] In this application, the maintenance cycle of each power supply unit is independent of each other and is determined by the health status. A new maintenance cycle is started after one maintenance is performed.

[0104] When the operating data is normal, this application uses a physical constraint-based LSTM model to predict the theoretical temperature rise and compares it with the actual temperature rise to calculate the health status.

[0105] The data required for temperature rise prediction using a physically constrained LSTM model includes: Current fluctuation rate Historical temperature rise at the previous historical point in time .

[0106] The methods for extracting the above features include:

[0107] (1) Transient heating coefficient and thermal inertia coefficient

[0108] (1-1) Obtain standard operating data for a period of time at the beginning of the current maintenance cycle, and extract standard temperature rise, standard historical temperature rise and standard heat loss power at multiple time points from the standard operating data;

[0109] The health status in this application is constructed from two dimensions: first, the reversible lifespan decay due to the accumulation of dust, consumables, etc. within each maintenance cycle; and second, the irreversible lifespan decay due to the oxidation of the material surface (joints, insulators, etc.). At the end of each maintenance cycle, it is assumed that the health status returns to the optimal state of the current cycle, and standard temperature rise, standard historical temperature rise, and standard heat loss power are collected at multiple time points to fit the transient temperature rise coefficient and thermal inertia coefficient that produce changes, so as to determine the health status loss value of the irreversible part.

[0110] (1-2) Construct a multiple linear fitting equation, wherein the mathematical expression of the multiple linear fitting equation is:

[0111]

[0112] In the formula, Indicates the temperature rise at the reference time point. Indicates the transient temperature rise coefficient. Indicates heat loss power. Indicates the thermal inertia coefficient. This indicates the historical temperature rise at the previous reference time point. Indicates the error term;

[0113] The heat loss power in this application The formula for calculation is: , This is the current value. This is the equivalent resistance of the power supply unit.

[0114] The derivation of the above formula is as follows:

[0115] The dynamic process of temperature rise in electrical equipment is described by the heat balance equation:

[0116] Based on the fundamental principles of heat transfer, the thermal dynamics of the power supply unit can be described by a lumped-parameter thermal model. Let the temperature rise of the power supply unit relative to the ambient temperature be... Its heat capacity is The equivalent heat dissipation coefficient is The internal heat loss power is According to the law of conservation of energy, the heat absorbed by the heat capacity per unit time is equal to the heat generated minus the heat lost, that is:

[0117]

[0118] Mode It is a first-order linear differential equation, with a heat dissipation coefficient. This reflects the comprehensive ability of the power supply unit to dissipate heat to the environment through conduction, convection, and other means, including its heat capacity. This characterizes its ability to store heat.

[0119] To facilitate discretized sampling applications, a backward difference approximation of the derivative is used. Let the sampling time interval be... Then we have:

[0120]

[0121] The formula Substitute into formula ,get:

[0122]

[0123] The results were:

[0124]

[0125] Therefore, we get:

[0126]

[0127] make , By adding an error term, we can obtain the mathematical expression for the multiple linear fitting equation as follows:

[0128]

[0129] The transient temperature rise coefficient integrates the effects of heat capacity, heat dissipation coefficient, and sampling interval, and essentially reflects the combined effect of heat conduction and heat dissipation. The thermal inertia coefficient determines the degree to which the temperature rise at the previous moment affects the current moment. Its value is between 0 and 1 and is related to the heat capacity and the sampling interval. The larger the value, the stronger the thermal inertia, and the slower the temperature changes.

[0130] When the lifespan of the ring main unit power supply unit undergoes irreversible degradation, The value will increase slowly. The resistance decreases slowly. This is because as the ring main unit ages, the contact surfaces inside, such as contacts, busbar connections, and cable joints, undergo oxidation, corrosion, or loosening, leading to increased contact resistance. Heat loss power. As the contact resistance increases, from the perspective of equivalent thermal resistance, the increase in contact resistance is equivalent to an increase in local thermal resistance, resulting in a higher temperature rise for the same power loss. Increased. Some components inside the ring main unit that participate in heat storage (such as insulation materials and fillers) may age, crack, or decompose under long-term thermal stress, leading to a decrease in their effective heat capacity. Decrease, due to Therefore, it can be inferred that It then descends slowly.

[0131] (1-3) Substitute the standard temperature rise, standard historical temperature rise and standard heat loss power at multiple time points into the multivariate linear fitting equation, and combine it with the least squares method to fit the equation to obtain the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle.

[0132] In this application, the collected samples are substituted into the aforementioned multiple linear fitting equation to obtain the actual transient temperature rise coefficient of the power supply unit after each maintenance. and thermal inertia coefficient The constructed transient temperature rise coefficient and thermal inertia coefficient The sequence can be used to reflect the lifespan degradation of ring main units.

[0133] (2) Operational characteristics

[0134] (2-1) Calculate the most recent data based on the aforementioned operational data Load rate at historical time points Current fluctuation rate Historical warming at the previous historical point in time , wherein the load rate The current fluctuation rate and the historical temperature rise at the previous point in time The mathematical expression is:

[0135]

[0136]

[0137]

[0138] In the formula, Indicates a historical time point index. For the number of historical time points, Rated current, For time points The current value, The maximum current value for the running data. The minimum current value for the operating data. For time points The temperature inside the cabinet, For time points The ambient temperature value.

[0139] S340, The input features are input into a pre-built temperature rise prediction model based on physical constraints to obtain the predicted temperature rise at the current time point;

[0140] In this application, a temperature rise prediction model based on physical constraints is pre-constructed, and the specific process includes:

[0141] S1, obtain operating data samples of multiple power supply units of the ring main unit in a healthy state, wherein the operating data samples include the values ​​of various operating parameters at multiple historical time points;

[0142] Specifically, for equipment that has been in operation for ≤1 year, operational data samples are automatically collected via smart terminals. The composition of the operational data samples is consistent with the operational data described above, and will not be repeated here.

[0143] S2, the running data sample is filtered to obtain filtered data; and the abnormal time distribution in the filtered data is removed to obtain preprocessed data;

[0144] This application applies a sliding median filter to the sequences of multiple parameters to eliminate transient interference (such as current spikes caused by switching operations). Furthermore, data from outlier time points are removed based on a 3-standard-deviation principle. It is worth noting that if data for one parameter at time point A is removed, then data for all other parameters at time point A must also be removed to ensure temporal alignment.

[0145] S3, the preprocessed data is segmented to obtain multiple segmented samples, wherein the last historical time point of each segmented sample is the sample reference time point.

[0146] Each sample contains N = 3 consecutive time points. , This is the baseline time point.

[0147] S4, extract operating feature samples from the segmented samples, wherein the operating feature samples include load rate sequence samples, current fluctuation rate samples and historical temperature rise samples;

[0148] The method for extracting feature samples is as described above and will not be repeated here.

[0149] S5, using the temperature rise sample corresponding to the sample baseline time point as the label, and combining it with the running feature sample to construct training samples, thus obtaining the training dataset;

[0150] S6. Select training samples for the current training batch from the training dataset and input the training samples for the current training batch into the LSTM model to obtain the predicted temperature rise;

[0151] Temperature rise is a continuous process over time, therefore this application chooses to use an LSTM model that can extract features of temporal change as the training model.

[0152] S7. Calculate the loss between the predicted temperature rise and the label based on a pre-built physical constraint-based loss function, and perform backpropagation based on the loss, then adjust the parameters of the LSTM model using gradient descent.

[0153] The mathematical expression for the loss function in this application is:

[0154]

[0155]

[0156]

[0157] In the formula, For the total loss, For MSE loss, For physical constraint loss, As a weighting factor, The number of samples in a single training batch. For sample index, Indicates the first The predicted temperature rise corresponding to each sample Indicates the first The label of each sample This is the design value for the transient temperature rise coefficient. This is the design value for the thermal inertia coefficient. For the first Each sample corresponds to a heat loss power. For the first The historical temperature rise of each sample.

[0158] In the above loss function, the MSE loss term is used. Ensure the temperature rise predicted by the model Compared with the measured label Consistent. And incorporating physical constraints. This ensures that the model's output satisfies the heat balance equation derived earlier. .

[0159] The model trained using the aforementioned loss function constraints achieves a balance between prediction accuracy and safety. It is the optimal model based on the actual needs of the ring main unit scenario.

[0160] S8. Repeat steps S6-S7 until training is complete, and obtain a temperature rise prediction model based on physical constraints.

[0161] Finally, when the number of training iterations reaches the target value, or when the loss value is less than the threshold and no longer changes significantly, the training is complete, and a temperature rise prediction model based on physical constraints is obtained.

[0162] In the model described above, during the real-time prediction phase, the load rate at three time points is combined with other features to construct the input vector, which can be represented as: By inputting the input vector into the model, the predicted temperature rise at the current time point can be obtained. .

[0163] S350, calculate the actual temperature rise at the current time point, and calculate the health status of the power supply unit based on the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle; and perform isolation protection control on the power supply unit for maintenance purposes based on the health status.

[0164] Finally, the health of the power supply unit is calculated using the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle, all calculated above. The health consists of two parts: (1) the reversible maintenance health predicted by the model; and (2) the irreversible lifetime health. Figure 4 This is a schematic diagram of the health calculation process in one embodiment of this application, such as... Figure 4 As shown, the specific calculation process is as follows:

[0165] S351, based on the actual temperature rise at the current time point Predicted temperature rise at the current time point Calculate the first health level The mathematical expression is:

[0166]

[0167] In the formula, Indicates the maximum temperature rise;

[0168] The first health level reflects the current maintenance health of the equipment. These types of problems can usually be resolved through cleaning, tightening, and maintenance, and are reversible. Its core idea is to measure the deviation between the real-time thermal state and the expected state.

[0169] The residual is the actual temperature rise compared to expectations; a large residual indicates a higher health level. The corresponding value becomes smaller. This is usually caused by poor contact, contact oxidation, dust clogging the air duct, or transient overload. These problems can usually be restored after maintenance (such as tightening and cleaning) and are therefore considered reversible.

[0170] S352, transient temperature rise coefficient based on the current maintenance cycle. Thermal inertia coefficient during the current maintenance cycle Calculate the second health level Among them, the second health level The mathematical expression is:

[0171]

[0172] In the formula, The reference transient temperature rise coefficient for the ring main unit power supply unit in a brand-new state. The reference thermal inertia coefficient for the ring main unit power supply unit in a brand-new state;

[0173] Second Health This reflects the inherent aging degree of the equipment. These problems are caused by the accumulation of material aging and physical wear, and are irreversible. The core idea is to measure the permanent changes in the inherent thermal properties.

[0174] and These are the standard parameters of the equipment in its brand-new condition, representing its factory settings or gold standard. and These are the actual parameters derived from the current period data, representing the current intrinsic physical characteristics of the device.

[0175] Its logic of change is as follows:

[0176] Increased thermal resistance leads to the drying of thermal grease, wear of contact surfaces, and corrosion of the heat sink. This means a permanent decrease in the device's heat generation / dissipation efficiency.

[0177] Reduced heat capacity leads to decomposition of insulating materials and carbonization of internal fillers. This means that the equipment's "thermal buffering" capacity is permanently weakened.

[0178] It can reflect the differences between the transient temperature rise coefficient, thermal inertia coefficient, and reference values, and is normalized to the [0,1] interval using an exponential function. The exponential function has the characteristic of being "more sensitive to larger deviations." When the parameter deviations accumulate to a certain extent, the health status will decline rapidly, which is very consistent with the physical law that the probability of failure increases sharply at the end of the life.

[0179] S353, regarding the first health level and the second health level The health status of the power supply unit is obtained by weighting the values. Among them, the health of the power supply unit The mathematical expression is:

[0180]

[0181] In the formula, As the first weight, It is the second weight.

[0182] Finally, through weighted aggregation, a single health index that comprehensively reflects the equipment status is obtained. This index is used to indicate whether the current power supply unit requires maintenance.

[0183] Finally, based on the health status, the power supply unit is subjected to isolation and protection control for maintenance purposes, including:

[0184] The health status is compared with a preset health status threshold. When the health status is less than the preset health status threshold, the incoming load switch of the current power supply unit is disconnected, and the interconnection switch of the adjacent power supply unit is closed to switch to the healthy adjacent power supply unit for power supply.

[0185] Combination Figure 1 The ring main unit structure, when the cloud server analyzes and determines that the current power supply unit's health is low, remote commands are used to disconnect the incoming load switch of the current power supply unit and close the interconnection switch of the adjacent power supply unit that is in a healthy state, allowing the adjacent unit to supply power. Additionally, the current power supply unit is marked as unhealthy, and maintenance work orders are distributed through the management system, realizing a maintenance process guided by health requirements.

[0186] This application discloses an isolation and protection control method for an intelligent ring main unit. Through multi-dimensional data anomaly verification, physical constraint temperature rise prediction, and dynamic health assessment, it significantly improves the reliability and maintenance efficiency of the ring main unit's power supply unit. The system verifies ambient temperature, cabinet temperature, and current data in real time to ensure input reliability. Based on historical operating data, it calculates the transient temperature rise coefficient and thermal inertia coefficient, constructing a temperature rise prediction model that integrates physical laws to accurately predict temperature rise. The health calculation comprehensively considers actual temperature rise deviation and coefficient offset to achieve early fault warning. When the health level falls below a threshold, the faulty unit is automatically isolated and power is switched to a healthy unit, avoiding unplanned power outages. This application reduces reliance on manual inspections and optimizes preventative maintenance cycles, effectively reducing operation and maintenance costs. Simultaneously, the physical constraint model improves prediction accuracy, avoids the overfitting risk of purely data-driven systems, and ensures the safe and stable operation of the power grid.

[0187] like Figure 5 As shown, this application also provides an isolation protection control system for an intelligent ring main unit, comprising:

[0188] The acquisition module is used to acquire the operating data of multiple power supply units in the ring main unit at a target number of historical time points, wherein the operating data includes ambient temperature, cabinet temperature and current value;

[0189] The verification module is used to perform anomaly verification on the running data and obtain the verification result.

[0190] The feature extraction module is used to extract operating features from the operating data when the verification result indicates that the operating data is normal; and to calculate the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle based on the operating data of the pre-built operating database. The operating features include the load rate, current fluctuation rate and historical temperature rise of the previous historical time point at multiple historical time points. The operating database includes operating data within a certain period of time at the beginning of the maintenance cycle.

[0191] The temperature rise prediction module is used to input the input features into a pre-built temperature rise prediction model based on physical constraints to obtain the predicted temperature rise at the current time point;

[0192] The protection control module is used to calculate the actual temperature rise at the current time point, and calculate the health status of the power supply unit based on the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle; and to perform isolation protection control on the power supply unit for maintenance purposes based on the health status.

[0193] This application discloses an isolation and protection control system for an intelligent ring main unit. Through multi-dimensional data anomaly verification, physical constraint temperature rise prediction, and dynamic health assessment, it significantly improves the reliability and maintenance efficiency of the ring main unit's power supply unit. The system verifies ambient temperature, cabinet temperature, and current data in real time to ensure input reliability. Based on historical operating data, it calculates the transient temperature rise coefficient and thermal inertia coefficient, constructing a temperature rise prediction model that integrates physical laws to accurately predict temperature rise. The health calculation comprehensively considers actual temperature rise deviation and coefficient offset to achieve early fault warning. When the health level falls below a threshold, it automatically isolates the faulty unit and switches power supply to a healthy unit, avoiding unplanned power outages. This application reduces reliance on manual inspections and optimizes preventative maintenance cycles, effectively reducing operation and maintenance costs. Simultaneously, the physical constraint model improves prediction accuracy, avoids the overfitting risk of purely data-driven systems, and ensures the safe and stable operation of the power grid.

[0194] This embodiment also provides an electronic terminal, including: a processor and a memory;

[0195] The memory is used to store computer programs, and the processor is used to execute the computer programs stored in the memory so that the terminal performs any of the methods in this embodiment.

[0196] As will be understood by those skilled in the art, the computer-readable storage medium described in this embodiment allows for the implementation of all or part of the steps in the above method embodiments by computer program-related hardware. The aforementioned computer program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0197] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver, and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication between them. The memory is used to store computer programs, the communication interface is used to perform communication, and the processor and the transceiver are used to run the computer programs, so that the electronic terminal performs the steps of the above method.

[0198] In this embodiment, the memory may include random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage device.

[0199] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0200] In the above embodiments, although the present application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art based on the foregoing description. The embodiments of the present application are intended to cover all such substitutions, modifications, and variations falling within the broad scope of the appended claims.

[0201] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for isolating, protecting, and controlling an intelligent ring main unit, characterized in that, Including the following steps: Obtain the operating data of multiple power supply units in the ring main unit at a target number of historical time points, wherein the operating data includes ambient temperature, cabinet temperature and current value; Anomaly verification is performed on the operational data to obtain the verification results; When the verification result indicates that the operating data is normal, operating features are extracted from the operating data; the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle are calculated based on the operating data of the pre-built operating database, wherein the operating features include the load rate, current fluctuation rate and historical temperature rise of the previous historical time point at multiple historical time points, and the operating database includes operating data within a period of time at the beginning of the maintenance cycle; The operational characteristics are input into a pre-built temperature rise prediction model based on physical constraints to obtain the predicted temperature rise at the current time point; The actual temperature rise at the current time point is calculated, and the health of the power supply unit is calculated based on the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle; and the power supply unit is subjected to isolation protection control for maintenance purposes based on the health.

2. The isolation protection control method for an intelligent ring main unit according to claim 1, characterized in that, Also includes: When the verification result indicates that the running data is abnormal, the running data is updated at the next time point, and the abnormality verification of the running data is performed again to obtain the verification result. Until the verification result indicates that the running data is normal.

3. The isolation protection control method for an intelligent ring main unit according to claim 1, characterized in that, Anomaly verification is performed on the aforementioned operational data to obtain the verification results, including: Based on a pre-built data reference range, the values ​​of operating parameters at multiple time points in the operating data are checked for value anomalies to obtain the value anomaly verification results. The cabinet temperature sequence and the ambient temperature sequence in the operation data are subjected to first-order difference to obtain a temperature first-order difference sequence; the value of each element in the temperature first-order difference sequence is compared with the preset maximum temperature change threshold to obtain the temperature jump verification result. The first-order difference sequence of each parameter in the running data is normalized and then averaged. The average value is then compared with a preset minimum change threshold to obtain the freeze anomaly verification result. The various parameter sequences of the operational data are time-series aligned, and a temperature rise sequence is calculated based on the ambient temperature sequence and the cabinet temperature sequence. The slope of the temperature rise sequence and the slope of the current value square sequence are extracted. Multivariate physical constraint verification is performed based on pre-configured verification rules to obtain multivariate physical constraint verification results. The verification rules include: the absolute value of the difference between the slope of the temperature rise sequence and the slope of the current value square sequence is less than a preset following threshold, and the cabinet temperature at the same time point is greater than or equal to the ambient temperature. If the results of the abnormal value verification, the temperature jump verification, the freezing anomaly verification, and the multivariable physical constraint verification are all normal, the running data is determined to pass the anomaly verification; otherwise, it fails the anomaly verification.

4. The isolation protection control method for an intelligent ring main unit according to claim 1, characterized in that, Based on the operational data in the operational database, the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit for the current maintenance cycle are calculated, including: Obtain standard operating data for a period of time at the start of the current maintenance cycle, and extract standard temperature rise, standard historical temperature rise, and standard heat loss power at multiple time points from the standard operating data; Construct a multiple linear fitting equation, wherein the mathematical expression of the multiple linear fitting equation is: In the formula, Indicates the temperature rise at the reference time point. Indicates the transient temperature rise coefficient. Indicates heat loss power. Indicates the thermal inertia coefficient. This indicates the historical temperature rise at the previous reference time point. Indicates the error term; Substituting the standard temperature rise, standard historical temperature rise, and standard heat loss power at multiple time points into the multivariate linear fitting equation, and combining it with the least squares method for fitting, the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle are obtained.

5. The isolation protection control method for an intelligent ring main unit according to claim 1, characterized in that, When the verification result indicates that the running data is normal, running features are extracted from the running data, including: Based on the aforementioned operational data, the most recent... Load rate at historical time points Current fluctuation rate Historical warming at the previous historical point in time , wherein the load rate The current fluctuation rate and the historical temperature rise of the previous cycle The mathematical expression is: In the formula, Indicates a historical time point index. For the number of historical time points, Rated current, For time points The current value, The maximum current value for the running data. The minimum current value for the operating data. For time points The temperature inside the cabinet, For time points The ambient temperature value.

6. The isolation protection control method for an intelligent ring main unit according to claim 1, characterized in that, The method for constructing the temperature rise prediction model based on physical constraints includes: S1, obtain operating data samples of multiple power supply units of the ring main unit in a healthy state, wherein the operating data samples include the values ​​of various operating parameters at multiple historical time points; S2, the running data sample is filtered to obtain filtered data; and the data corresponding to abnormal times in the filtered data is removed to obtain preprocessed data; S3, the preprocessed data is segmented to obtain multiple segmented samples, wherein the last historical time point of each segmented sample is the sample reference time point. S4, extract operating feature samples from the segmented samples, wherein the operating feature samples include load rate sequence samples, current fluctuation rate samples and historical temperature rise samples; S5, using the temperature rise sample corresponding to the sample baseline time point as the label, and combining it with the running feature sample to construct training samples, thereby obtaining the training dataset; S6. Select training samples for the current training batch from the training dataset and input the training samples for the current training batch into the LSTM model to obtain the predicted temperature rise; S7, calculate the loss between the predicted temperature rise and the label based on the pre-built physical constraint-based loss function, and adjust the parameters of the LSTM model based on the loss direction propagation and the gradient descent method. S8. Repeat steps S6-S7 until training is complete, and obtain a temperature rise prediction model based on physical constraints.

7. The isolation protection control method for an intelligent ring main unit according to claim 6, characterized in that, The mathematical expression for the loss function based on physical constraints is as follows: In the formula, For the total loss, For MSE loss, For physical loss, As a weighting factor, The number of samples in a single training batch. For sample index, Indicates the first The predicted temperature rise corresponding to each sample Indicates the first The label of each sample This is the design value for the transient temperature rise coefficient. This is the design value for the thermal inertia coefficient. For the first Each sample corresponds to a heat loss power. For the first The historical temperature rise of each sample.

8. The isolation protection control method for an intelligent ring main unit according to claim 1, characterized in that, Calculate the actual temperature rise at the current time point, and calculate the health status of the power supply unit based on the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle, including: Based on the actual temperature rise at the current time point Predicted temperature rise at the current time point Calculate the first health level The mathematical expression is: In the formula, Indicates the maximum temperature rise; Transient temperature rise coefficient based on the current maintenance cycle Thermal inertia coefficient during the current maintenance cycle Calculate the second health level Among them, the second health level The mathematical expression is: In the formula, This is the design value for the transient temperature rise coefficient. The design value is based on the reference thermal inertia coefficient; For the first health level and the second health level The health status of the power supply unit is obtained by weighting the values. Among them, the health of the power supply unit The mathematical expression is: In the formula, As the first weight, It is the second weight.

9. The isolation protection control method for an intelligent ring main unit according to claim 1, characterized in that, Based on the health status, the power supply unit is subjected to isolation and protection control for maintenance purposes, including: The health status is compared with a preset health status threshold. When the health status is less than the preset health status threshold, the incoming load switch of the current power supply unit is disconnected, and the interconnection switch of the adjacent power supply unit is closed to switch to the healthy adjacent power supply unit for power supply.

10. An isolation protection control system for an intelligent ring main unit, characterized in that, include: The acquisition module is used to acquire the operating data of multiple power supply units in the ring main unit at a target number of historical time points, wherein the operating data includes ambient temperature, cabinet temperature and current value; The verification module is used to perform anomaly verification on the running data and obtain the verification result. The feature extraction module is used to extract operating features from the operating data when the verification result indicates that the operating data is normal; and to calculate the transient temperature rise coefficient and thermal inertia coefficient of the power supply unit in the current maintenance cycle based on the operating data of the pre-built operating database. The operating features include the load rate, current fluctuation rate and historical temperature rise of the previous historical time point at multiple historical time points. The operating database includes operating data within a certain period of time at the beginning of the maintenance cycle. The temperature rise prediction module is used to input the operating characteristics into a pre-built temperature rise prediction model based on physical constraints to obtain the predicted temperature rise at the current time point; The protection control module is used to calculate the actual temperature rise at the current time point, and calculate the health status of the power supply unit based on the actual temperature rise at the current time point, the predicted temperature rise at the current time point, the transient temperature rise coefficient of the current maintenance cycle, and the thermal inertia coefficient of the current maintenance cycle; and to perform isolation protection control on the power supply unit for maintenance purposes based on the health status.

Images