A smart reclosing circuit breaker control system

Abstract

This invention belongs to the field of electrical equipment technology. It provides an intelligent reclosing circuit breaker control system, comprising: a scenario monitoring unit, a data analysis and decision-making unit, a circuit monitoring unit, and an adaptive control unit. The control system of this invention combines real-time and planned operation data to generate dynamic fault critical characteristics and integrates analysis, achieving precise control of the circuit breaker. It can both promptly trip to avoid faults and timely close to restore power supply, reducing misoperation and power interruption, improving circuit safety and power supply continuity, and adapting to complex load scenarios.

Description

Technical Field

[0001] This invention relates to the field of electrical equipment technology, and more specifically, to an intelligent reclosing circuit breaker control system. Background Technology

[0002] In modern power systems, reclosing circuit breakers are core equipment for power distribution and circuit protection, and their performance directly affects power supply reliability, power consumption safety, and the stable operation of the power network. With the rapid development of industrial automation, smart grids, and various precision electrical equipment, power loads are becoming more diversified and dynamic. The frequency and complexity of faults such as overloads and short circuits in circuits have increased significantly, placing higher demands on the response speed, protection accuracy, and automation level of circuit breakers.

[0003] While traditional reclosing circuit breakers can achieve manual operation and automatic tripping and closing after a fault, they have significant limitations in the pre-identification and handling of overload conditions. On the one hand, most traditional devices rely on fixed current thresholds as the basis for overload judgment, making it difficult to adapt to dynamic load changes in different scenarios. For example, large motors in industrial workshops generate short-term inrush currents when starting, and multiple devices starting simultaneously in household electricity may also cause instantaneous high loads. These non-fault current fluctuations are often misjudged as overloads, leading to unnecessary tripping operations and affecting normal power supply. On the other hand, traditional systems have weak predictive capabilities for overload trends, often triggering protection mechanisms only after an overload fault has occurred, failing to take early warning or adjustment measures. This may lead to damage to circuit components due to continuous overload, or even cause more serious electrical accidents.

[0004] Therefore, how to improve the accuracy of reclosing circuit breakers in identifying overload conditions and achieve accurate prediction and adaptive handling of faults is a key issue that urgently needs to be addressed in the field of power protection equipment. Summary of the Invention

[0005] In response, the present invention provides an intelligent reclosing circuit breaker control system, electronic equipment, computer storage medium, and computer program product to solve at least one of the above-mentioned technical problems.

[0006] In a first aspect, the present invention provides an intelligent reclosing circuit breaker control system, comprising a scenario monitoring unit, a data analysis and decision-making unit, a circuit monitoring unit, and an adaptive control unit; wherein the scenario monitoring unit is used to acquire load operation data of load devices within the control scenario range of the circuit breaker, including real-time operation data and planned operation data.

[0007] The data analysis and decision-making unit is used to determine the basic fault critical characteristics based on the real-time operation data, predict the critical fluctuation characteristics based on the planned operation data and random fluctuation characteristics, and integrate the basic fault critical characteristics and critical fluctuation characteristics to obtain fault critical characteristic data for the target period.

[0008] The circuit monitoring unit is connected to the circuit and is used to collect the circuit's current, voltage, and power factor during the target time period, and extract the operating characteristic data after preprocessing.

[0009] The adaptive control unit is used to perform similarity analysis between the operating characteristic data and the fault critical characteristic data to determine whether the control conditions are met. If so, it sends a tripping control command to the circuit breaker; and if the subsequent similarity analysis results do not meet the control conditions, it sends a closing control command to the circuit breaker.

[0010] A second aspect of the present invention provides an electronic device for use in an intelligent reclosing circuit breaker control system as described in any of the preceding claims, the electronic device comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.

[0011] A third aspect of the present invention provides a computer storage medium for use in an intelligent reclosing circuit breaker control system as described in any of the preceding claims, the computer storage medium storing a computer program executable by a processor.

[0012] A fourth aspect of the present invention provides a computer program product for use in an intelligent reclosing circuit breaker control system as described in any of the preceding claims, the computer program product comprising a computer program executable by a processor.

[0013] The control system of this invention combines real-time and planned operation data to generate dynamic fault critical characteristics and integrate analysis, thereby achieving precise control of the circuit breaker. It can both promptly trip to avoid faults and timely close to restore power supply, reducing misoperation and power interruption, improving circuit safety and power supply continuity, and adapting to complex load scenarios. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1This is a schematic diagram of the structure of an intelligent reclosing circuit breaker control system disclosed in an embodiment of the present invention.

[0016] Figure 2 This is a schematic diagram of the process for determining the critical characteristics of basic faults based on real-time operating data, as disclosed in an embodiment of the present invention.

[0017] Figure 3 This is a schematic diagram of the process for predicting critical fluctuation characteristics based on planned operation data and random fluctuation characteristics, as disclosed in an embodiment of the present invention.

[0018] Figure 4 This is a schematic diagram of the process for generating a fluctuation probability distribution model disclosed in an embodiment of the present invention.

[0019] Figure 5 This is a schematic diagram of the process disclosed in this embodiment of the invention for integrating basic fault critical characteristics and critical fluctuation characteristics to obtain fault critical characteristic data for a target time period.

[0020] Figure 6 This is a schematic diagram of the process for determining whether the control conditions are met, as disclosed in an embodiment of the present invention.

[0021] Figure 7 This is a schematic diagram of the process of sending a closing control command to a circuit breaker, as disclosed in an embodiment of the present invention. Detailed Implementation

[0022] The following specific embodiments 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. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] Furthermore, the technical features involved in the different embodiments of this application described below can be combined with each other as long as they do not conflict with each other.

[0024] like Figure 1 As shown in the figure, an embodiment of the present invention discloses an intelligent reclosing circuit breaker control system 100, including a scenario monitoring unit 1001, a data analysis and decision-making unit 1002, a circuit monitoring unit 1003, and an adaptive control unit 1004; the scenario monitoring unit 1001 is used to acquire load operation data of load devices within the control scenario range of the circuit breaker, including real-time operation data and planned operation data.

[0025] The scenario monitoring unit is responsible for accurately collecting the operating data of all load devices within the circuit breaker control scenario, specifically including: (1) real-time operating data, including dynamic parameters such as the current instantaneous current, real-time power, and operating status (e.g., start-up, operation, shutdown) of the equipment, such as the current change corresponding to the real-time speed of the motor in an industrial workshop, and the real-time power consumption of the air conditioner in a home; (2) planned operating data, including pre-set information such as the preset start-up and stop time, planned running time, and load adjustment plan of the equipment, such as the schedule of production equipment in a factory (e.g., running equipment A from 8:00 to 12:00 and running equipment B from 14:00 to 18:00), and the timed start-up settings of the water heater in a home. After these data are collected through sensors, equipment interfaces, etc., they are summarized and transmitted in real time to the data analysis and decision-making unit.

[0026] The data analysis and decision-making unit 1002 is used to determine the basic fault critical characteristics based on the real-time operation data, predict the critical fluctuation characteristics based on the planned operation data and random fluctuation characteristics, and integrate the basic fault critical characteristics and critical fluctuation characteristics to obtain fault critical characteristic data for the target period.

[0027] The data analysis and decision-making unit generates critical characteristic data for judging circuit faults based on the load operation data provided by the scenario monitoring unit. Specifically, it first cleans the data (removing outliers and noise) and extracts features (such as the average operating current, maximum instantaneous current, and load duration) based on real-time operating data and the rated parameters of the load equipment (e.g., rated current and rated power). For example, for a motor with a rated current of 10A, by analyzing the fluctuation range of its current real-time operating current, the overload critical current under the current operating condition (e.g., 12A) is calculated. This means that an overload fault may occur when the real-time current continuously exceeds this value, forming the basic fault critical characteristics reflecting the current static safety boundary.

[0028] Then, the start-up and shutdown periods of the equipment and the load superposition patterns (such as the time window for multiple devices to start simultaneously) are extracted from the planned operation data. At the same time, the random fluctuation characteristics in the real-time operation data (the frequency, amplitude, and trend of current fluctuations, such as the pulse current pattern when a welding machine is working) are analyzed. Using time series prediction algorithms (such as ARIMA and LSTM), combined with the planned operation data, the load fluctuation curve within the target period is predicted. The dynamic load limit (such as the total current limit for multiple devices operating simultaneously) and the time-specific risk threshold (such as the fluctuation tolerance during peak electricity consumption periods) of multiple devices operating in coordination within this period are calculated, forming the critical fluctuation characteristics that reflect the dynamic safety boundary of the target period.

[0029] Next, a weighted allocation model (such as setting weights based on real-time requirements and equipment importance) is used to fuse basic fault critical features with critical fluctuation features. For example, for precision equipment with high real-time requirements, the weight of basic fault critical features is increased; for scenarios where multiple devices operate collaboratively, the weight of critical fluctuation features is increased, ultimately generating a set of fault critical feature data for the target time period that includes both the current static safety threshold and the dynamic early warning threshold.

[0030] The circuit monitoring unit 1003 is connected to the circuit and is used to collect the circuit's current, voltage, and power factor during the target time period, and extract the operating characteristic data after preprocessing them.

[0031] The circuit monitoring unit uses devices such as current transformers and voltage sensors connected in series in the main circuit to collect key electrical parameters such as total current, line voltage, and power factor of the circuit in real time during the target period.

[0032] The collected raw data undergoes preprocessing, including filtering (to remove high-frequency interference, such as harmonic interference in the power grid), normalization (to unify data format and magnitude for easier subsequent analysis), and smoothing (to reduce the impact of instantaneous fluctuations on the analysis results). Subsequently, characteristic quantities reflecting the circuit's operating status during the target time period are extracted from the preprocessed data. These include the slope of the real-time current curve (reflecting the current change trend; a positive slope indicates rising current), voltage stability indicators (such as whether voltage fluctuations are within allowable limits), and the real-time value and rate of change of the power factor. This forms the operating characteristic data, which is then transmitted to the adaptive control unit.

[0033] The adaptive control unit 1004 is used to perform similarity analysis between the operating characteristic data and the fault critical characteristic data to determine whether the control conditions are met. If so, it sends a tripping control command to the circuit breaker; and if the subsequent similarity analysis results do not meet the control conditions, it sends a closing control command to the circuit breaker.

[0034] The adaptive control unit is the execution module of the control system of this invention for intelligent control of the circuit breaker. Based on the comparison between the circuit operating status and critical criteria, it sends opening or closing commands to the circuit breaker. The specific process is as follows: The operating characteristic data output by the circuit monitoring unit (such as the slope of the current change curve, voltage fluctuation amplitude, power factor change rate, etc. within the target time period) is converted into an operating feature vector. Simultaneously, the fault critical characteristic data generated by the data analysis and decision-making unit (such as the static safety threshold, dynamic early warning threshold, etc. within the target time period) is converted into a critical feature vector. The similarity between the two feature vectors is calculated using algorithms such as cosine similarity and Euclidean distance.

[0035] Based on the similarity analysis results, it is determined whether the control conditions are met. If the similarity between two feature vectors reaches or exceeds the preset threshold (i.e., the operating feature vector and the critical feature vector are highly similar), it is determined that the control conditions are met (there is a risk of overload or other faults). At this time, a trip control command is sent to the circuit breaker to quickly disconnect the circuit and avoid damage to the load equipment or line due to continuous overload.

[0036] After the circuit breaker trips, it continuously receives operational characteristic data updated by the circuit monitoring unit within the target time period, constructs a new operational characteristic vector in real time, and compares its similarity with the fault critical characteristic vector. When the similarity analysis results show that the similarity between the two characteristic vectors is lower than a preset threshold (i.e., the operational characteristic vector and the critical characteristic vector are significantly different, indicating that the fault has been eliminated), it is determined that the control conditions are not met, and a closing control command is sent to the circuit breaker to restore power supply to the circuit, ensuring that normal power is restored at a precise time after the fault is eliminated.

[0037] The control system of this invention combines real-time and planned operation data to generate dynamic fault critical characteristics and integrate analysis, thereby achieving precise control of the circuit breaker. It can both promptly trip to avoid faults and timely close to restore power supply, reducing misoperation and power interruption, improving circuit safety and power supply continuity, and adapting to complex load scenarios.

[0038] As an example, such as Figure 2 As shown, the step of determining the basic fault critical characteristics based on the real-time operating data includes: 101, extracting equipment type parameters, real-time current values, real-time power values ​​and continuous operating time from the real-time operating data, and obtaining the rated current threshold, rated power threshold and allowable overload duration corresponding to the equipment type parameters based on a preset equipment rated parameter library.

[0039] 102. Calculate the first deviation rate between the real-time current value and the rated current threshold, and the second deviation rate between the real-time power value and the rated power threshold. Combine the first deviation rate, the second deviation rate and the continuous running time to calculate the overload risk coefficient.

[0040] 103. When the overload risk coefficient reaches the preset risk threshold, the corresponding real-time current value, real-time power value and continuous running time are determined as the critical characteristics of single device basic fault.

[0041] 104. Calculate the total load critical value based on the equipment association relationship, and combine the basic fault critical characteristics of each individual equipment with the total load critical value to form the basic fault critical characteristics.

[0042] The following parameters are extracted from the real-time operation data collected by the scene monitoring unit: equipment type parameters (such as information identifying equipment attributes such as three-phase asynchronous motors and household air conditioners); real-time current value (the instantaneous current of the equipment during current operation, in A); real-time power value (the current active power of the equipment, in kW); and continuous running time (the continuous running time of the equipment from this start-up to the current moment, in min).

[0043] Meanwhile, a pre-built equipment rated parameter library is constructed, which stores the factory rated parameters of various types of equipment. Based on the extracted equipment type parameters, the corresponding benchmark thresholds are automatically matched and obtained, including: rated current threshold (the maximum current for long-term safe operation of the equipment, such as 10A); rated power threshold (the maximum power limit of the equipment design, such as 2.2kW); and allowable overload duration (the maximum duration for which the equipment can withstand short-term overload, such as 5min).

[0044] Then, calculate the first deviation rate between the real-time current value and the rated current threshold, and the second deviation rate between the real-time power value and the rated power threshold. Specifically, the first deviation rate reflects the degree of current overload, and the calculation formula is: First deviation rate = (real-time current value - rated current threshold) / rated current threshold × 100%. Wherein, if the result is negative, it indicates that there is no overload, and is taken as 0.

[0045] The second deviation rate reflects the degree of power overload, and the calculation formula is: (real-time power value - rated power threshold) / rated power threshold × 100%. Similarly, if the result is negative, it indicates that there is no overload, and the value is 0.

[0046] The overload risk coefficient is calculated based on the first deviation rate, the second deviation rate, and the continuous operating time. This quantifies the current overload risk by integrating the impact of current, power overload degree, and duration. The formula is: Overload Risk Coefficient = (First Deviation Rate × First Weight + Second Deviation Rate × Second Weight) × (Continuous Operating Time / Allowable Overload Duration). The first and second weights can be set according to the equipment characteristics. For example, for current-dominated equipment, the first weight is 0.6 and the second weight is 0.4 to emphasize the influence of current parameters.

[0047] A preset risk threshold (e.g., 0.8-0.9, adjustable according to equipment importance) is set. When the calculated overload risk coefficient reaches this preset threshold, it indicates that the single device is approaching or has reached a critical fault state. At this time, the corresponding real-time current value, real-time power value, and continuous running time are recorded as the basic critical fault characteristics. For example, if a motor has an overload risk coefficient of 0.85 when the real-time current is 12A, the real-time power is 2.5kW, and the continuous running time is 4min, then (12A, 2.5kW, 4min) is determined as the basic critical fault characteristics of this single device.

[0048] Finally, based on the device relationships (such as the superposition of loads when operating simultaneously), the total load critical value is calculated (e.g., the total current critical value is 1.1 times the sum of the current critical values ​​of each device, and the total power critical value is 1.1 times the sum of the power critical values ​​of each device; these coefficients are used to buffer coordinated fluctuations). The basic fault critical characteristics are jointly constituted by the critical parameters of all individual devices (preserving the independent safety boundary of each device) and the total load critical value (reflecting the risk of multi-device coordination).

[0049] As an example, such as Figure 3 As shown, the critical fluctuation characteristics predicted based on the planned operation data and random fluctuation characteristics include: 201, extracting the preset start and stop times, single running duration and load type correlation of each load device from the planned operation data, and constructing the device running sequence table within the target time period.

[0050] 202. Extract random fluctuation characteristics from the real-time operating data, including current fluctuation frequency, power fluctuation amplitude and fluctuation duration, and generate a fluctuation probability distribution model through statistical analysis.

[0051] 203. Based on the device runtime sequence table, the theoretical total load benchmark value at different times within the target time period is calculated using the load superposition algorithm. The theoretical total load benchmark value is dynamically corrected in combination with the fluctuation probability distribution model, and the load fluctuation range under different confidence levels is calculated.

[0052] 204. Based on the upper limit of the load fluctuation range under different confidence levels corresponding to the theoretical total load benchmark value at each time, the critical fluctuation characteristics are derived, including the time-limited total current limit, the power fluctuation safety threshold, and the load superposition warning line for multi-device collaborative operation.

[0053] The following key information is extracted from the planned operation data collected by the scene monitoring unit: preset start and stop times (e.g., device A is scheduled to start at 8:00 and stop at 12:00, and device B is scheduled to start at 9:30 and stop at 17:00); single run duration (calculated from the start and stop times, e.g., device A runs for 4 hours at a time); load type correlation (reflecting the operational dependencies or conflicts between devices, e.g., "device C needs to stop when device D starts", "device E and device F can run simultaneously").

[0054] Based on the above information, the operating status of all devices within the target time period (e.g., the next 24 hours) is arranged in chronological order to form a visual device operation sequence table. For example, during the period from 9:00 to 9:30, only device A is running; during the period from 9:30 to 12:00, device A and device B are running simultaneously, clearly showing the device operation combinations at each time.

[0055] Furthermore, key features reflecting random load changes are extracted from real-time operational data: current fluctuation frequency (the number of times the current exceeds the stable range per unit time, such as 5 fluctuations per hour for a certain device); power fluctuation amplitude (the maximum percentage deviation of power from the average value, such as ±8% for a certain device); and fluctuation duration (the duration of each fluctuation, such as a fluctuation lasting 10 seconds). Statistical analysis (such as calculating the mean, variance, and probability density) is performed on these extracted key features to generate a fluctuation probability distribution model. For example, analysis of historical data reveals that the power fluctuation amplitude of device A follows a normal distribution, with 90% of the fluctuation amplitude concentrated within ±5% and 95% concentrated within ±6%.

[0056] Furthermore, based on the equipment operation sequence table, a load superposition algorithm is used to calculate the theoretical total load baseline value at different times within the target period. The calculation formula is: Theoretical total load baseline value = Σ(Single equipment rated power × Operating status coefficient). Here, the operating status coefficient is 1 during equipment operation (included in the total load) and 0 during non-operational periods (not included in the total load). For example, if equipment A has a rated power of 2kW and equipment B has a rated power of 3kW, and they operate simultaneously from 9:30 to 12:00, the theoretical total load baseline value = 2 × 1 + 3 × 1 = 5kW.

[0057] Furthermore, by combining the fluctuation probability distribution model with the theoretical total load baseline value as a reference, the load fluctuation range under different confidence levels is calculated. For example, based on the power fluctuation amplitude (±6%) corresponding to a 95% confidence level in the fluctuation probability distribution model, the fluctuation range corresponding to a 5kW baseline value is 5×(1±6%)=4.7kW-5.3kW; at a 90% confidence level, the fluctuation amplitude is ±5%, corresponding to a range of 4.75kW-5.25kW. Different confidence levels correspond to different range widths; the higher the confidence level, the wider the range, covering more extreme fluctuation scenarios.

[0058] Finally, for each moment within the target period, the upper limit of the load fluctuation range under different confidence levels of the theoretical total load benchmark is selected as the critical fluctuation characteristic for that moment. For example, the theoretical total load benchmark is 5kW for the period 9:30-12:00, the upper limit of the fluctuation range at 95% confidence level is 5.3kW, and the upper limit of the fluctuation range at 90% confidence level is 5.25kW. These upper limits correspond to the critical standards for different risk levels.

[0059] Critical fluctuation characteristics specifically include: time-limited total current limit (calculated from total power, such as 24A corresponding to 5.3kW); power fluctuation safety threshold (the relative deviation between the upper limit and the theoretical benchmark value, such as +6% at 95% confidence level); and load superposition warning line for multi-device coordinated operation (i.e., directly using the upper limit of power as the coordinated control threshold of the total load of multiple devices, such as the warning threshold corresponding to 5.3kW).

[0060] In this embodiment, the critical fluctuation feature is directly based on the upper limit of the load fluctuation range. This retains the reference significance of the theoretical total load benchmark value (clarifying the relative source of fluctuation) while focusing on the upper limit threshold related to overload risk. This makes the logic of the dynamic critical standard clearer and takes into account both the determinism of planned operation and the randomness of actual operation, providing a precise dynamic boundary for predicting overload risk.

[0061] As an example, such as Figure 4 As shown, random fluctuation features are extracted from the real-time operating data, including current fluctuation frequency, power fluctuation amplitude and fluctuation duration, and a fluctuation probability distribution model is generated through statistical analysis, including: 301, current time series data and power time series data within a preset time window are extracted from the real-time operating data, and feature values ​​are extracted from the dynamic fluctuation segment data, including current fluctuation frequency, power fluctuation amplitude and fluctuation duration.

[0062] 302. Collect feature values ​​of at least a preset number of preset time windows to form a sample set. Use kernel density estimation to fit the probability density of the current fluctuation frequency, power fluctuation amplitude, and fluctuation duration in the sample set to generate their respective marginal distribution models.

[0063] 303. A joint probability distribution model of the above three eigenvalues ​​is constructed using the Copula function, and this model is used as the fluctuation probability distribution model.

[0064] First, current time-series data (a sequence of current values ​​recorded in chronological order) and power time-series data (a sequence of power values ​​recorded in chronological order) within a preset time window (e.g., 5 minutes) are extracted from the real-time operational data. Steady-state segments, such as periods of stable operation with fluctuations less than 2% within 30 consecutive seconds, are filtered out, retaining only dynamic fluctuation segments—times in which current or power shows significant fluctuations—to ensure that subsequent analysis focuses on the actual fluctuation state. It is understood that the real-time operational data collected by the aforementioned scenario monitoring unit is not operational data at a single moment, but rather operational data over a period of time.

[0065] The following characteristic values ​​are extracted from the dynamic fluctuation segments of current time series data and power time series data: Current fluctuation frequency: the number of times the current value exceeds the steady-state mean ±3% per unit time (e.g., per minute) (e.g., if there are 12 fluctuations in a 5-minute window, the frequency is 2.4 times / minute), reflecting the intensity of the fluctuation.

[0066] Power fluctuation amplitude: Calculate the percentage of the difference between the peak power and the steady-state mean in each fluctuation relative to the steady-state mean (e.g., if the power increases from 2kW to 2.3kW in a certain fluctuation, the amplitude is 15%), and take the 90th percentile value of all fluctuation amplitudes (i.e., 90% of the fluctuation amplitudes are less than this value) as the characteristic value to avoid interference from extreme outliers.

[0067] Fluctuation duration: Record the duration of each fluctuation from exceeding the threshold to returning to the threshold (e.g., a fluctuation lasts for 15 seconds), and take the average of all durations as the feature value to reflect the continuity of the fluctuation.

[0068] Next, feature values ​​for at least a preset number (e.g., 10) of preset time windows are collected. Each window corresponds to a set of current fluctuation frequency, power fluctuation amplitude, and fluctuation duration, forming a sample set containing sufficient statistics. Kernel density estimation is used to fit the probability density of the three features in the sample set: for the current fluctuation frequency sample data, a marginal distribution model is generated, for example, showing a distribution pattern where "most windows have fluctuation frequencies concentrated between 1-3 times / minute, and a few windows reach more than 5 times / minute." Similarly, marginal distribution models for power fluctuation amplitude and fluctuation duration are generated, each model individually describing the probability distribution characteristics of a certain feature, for example, power fluctuation amplitude is mostly concentrated within ±5%, with a few reaching ±8%.

[0069] The Copula function is used to connect the marginal distribution models of the three features mentioned above to construct a joint probability distribution model. The role of the Copula function is to capture the correlation between the three features, such as the co-correlation between high current fluctuation frequency and high power fluctuation amplitude, and the co-correlation between long-term fluctuation and high-frequency fluctuation scenarios, rather than simply superimposing a single distribution.

[0070] The construction process is described as follows: 1. For the three features of time-limited total current limit (X), power fluctuation safety threshold (Y), and multi-device collaborative load superposition early warning line (Z), edge distribution models are constructed respectively.

[0071] Collect historical data (such as feature values ​​at various times over the past 3 months), remove outliers, and standardize. Use the Kolmogorov-Smirnov test to determine the marginal distribution type of each feature, including: the total current limit value (X) follows a normal distribution. The power fluctuation safety threshold (Y) follows a log-normal distribution. The load superimposed warning line (Z) follows a Gamma distribution. .

[0072] The parameters of each marginal distribution are solved using the maximum likelihood estimation method, for example... .

[0073] 2. Select an appropriate Copula function based on the correlation between the three features (measured by the Pearson or Kendal coefficient): For low correlation scenarios (correlation coefficient < 0.3): use the Gaussian Copula, whose joint distribution function is: ;in, It is a multivariate normal distribution function. The correlation coefficient matrix, , is the cumulative distribution function of the marginal distribution.

[0074] For high-correlation scenarios (correlation coefficient ≥ 0.3): t-Copula is selected to preserve the correlation structure of Gaussian Copula, while capturing tail correlation through the degree of freedom parameter.

[0075] 3. Estimating Copula Function Parameters: Based on the maximum likelihood estimation method, the Copula parameters (such as the correlation coefficient matrix of the Gaussian Copula) are solved with the objective of maximizing the joint log-likelihood function. ): ;in, For sample size, For Copula parameters.

[0076] The optimal Copula model is selected using either AIC (Akaike Information Criterion) or BIC (Bayesian Information Criterion). The smaller the value, the better the model fit.

[0077] 4. Constructing a joint probability distribution model: Combining the marginal distribution with the Copula function, we obtain the joint probability distribution of the three features: ;in, The marginal cumulative distribution functions for the three features are respectively. For the selected Copula function, These are the parameters of the Copula function.

[0078] The final joint probability distribution model is the complete fluctuation probability distribution model, which can reflect the distribution law of individual features and describe the mutual influence between features. For example, the model can directly output the probability of occurrence of the composite event "current fluctuation frequency > 3 times / minute, power fluctuation amplitude > 5%, fluctuation duration > 10 seconds".

[0079] As an example, such as Figure 5 As shown, the basic fault critical features and critical fluctuation features are integrated to obtain fault critical feature data for the target time period, including: 401, the first weight coefficient of the basic fault critical features is set based on the real-time requirements of the equipment, and the second weight coefficient of the critical fluctuation features is set based on the complexity of multi-equipment collaborative operation.

[0080] 402. The basic fault critical characteristics and critical fluctuation characteristics are normalized, and the comprehensive fault critical index at each time point within the target time period is calculated using a weighted fusion algorithm. The fault critical characteristic data is generated based on the comprehensive fault critical index.

[0081] First, set the weighting coefficients for the basic fault critical characteristics and critical fluctuation characteristics, as follows: Dynamically set the first weighting coefficient of the basic fault critical characteristics according to the real-time requirements of the equipment: For equipment with high real-time response requirements (such as medical monitors and industrial robot controllers), instantaneous overload may directly cause safety accidents or damage to precision components, so the first weighting coefficient is set to a higher value (such as 0.6-0.7) to prioritize the real-time safety boundary of a single device. For equipment with low real-time requirements (such as ventilation fans and general lighting systems), the impact of short-term overload is relatively small, so the first weighting coefficient is set to a lower value (such as 0.3-0.4).

[0082] This can be understood as the real-time requirement of equipment not being the requirement of a single device, but rather the equivalent requirement of the real-time requirements of multiple devices of various types. For example, each device in the control scenario is first classified according to its real-time requirement (e.g., medical equipment and industrial controllers are "high", lighting and ordinary motors are "low"), and assigned a quantified value (e.g., high = 3, medium = 2, low = 1). A weighted sum is then performed based on the device's load percentage, with the formula: Equivalent Real-Time Requirement = Σ(Single Device Real-Time Quantified Value × Device Load Percentage). The equivalent value is then divided into intervals (e.g., 1-1.5 → 0.3-0.4, 2-2.5 → 0.6-0.7).

[0083] A second weighting coefficient is set based on the complexity of multi-device collaborative operation to determine the critical fluctuation characteristics. Specifically: When three or more devices in the system are running simultaneously (such as a factory production line operating at full load or a commercial building experiencing peak electricity consumption), the fluctuation risk caused by load superposition increases significantly, and the second weighting coefficient is set to a higher value (such as 0.6-0.7) to strengthen the focus on dynamic load superposition; when only one or two devices are running independently (such as only the refrigerator and television working in a household), the load relationship is simple, and the second weighting coefficient is set to a lower value (such as 0.3-0.4). In addition, a weight constraint needs to be configured, that is, the sum of the first weighting coefficient and the second weighting coefficient should be 1 to ensure the normalization of weight allocation and avoid double calculation or weight imbalance.

[0084] Next, because the physical dimensions and numerical ranges of the basic fault critical characteristics (such as a single device's critical current of 12A and critical power of 2.5kW) and critical fluctuation characteristics (such as a total current limit of 24A and a total power threshold of 5.3kW) differ significantly (for example, the critical value for a single device and the total threshold for multiple devices may differ by 1-2 orders of magnitude), direct fusion would lead to one type of characteristic dominating the calculation results. Therefore, this embodiment first uses normalization to convert both types of characteristics into dimensionless exponents (with values ​​ranging from 0 to 1), eliminating the influence of dimensions. An example of the normalization process is illustrated below:

[0085] The basic fault critical characteristic index = single device critical value / rated parameter of the device. For example, if a motor has a critical current of 12A and a rated current of 10A, then the index = 12 / 10 = 1.2, which is linearly scaled to the 0-1 range. The critical fluctuation characteristic index = upper limit of the fluctuation range / theoretical total load baseline value. For example, if the upper limit of total power fluctuation is 5.3kW and the baseline value is 5kW, then the index = 5.3 / 5 = 1.06, which is similarly mapped to the 0-1 range.

[0086] Furthermore, based on the normalized feature index and weighting coefficient, the comprehensive fault criticality index for each moment within the target time period is calculated. The calculation formula is: Comprehensive Fault Criticality Index = Basic Fault Critical Feature Index × First Weighting Coefficient + Critical Fluctuation Feature Index × Second Weighting Coefficient. For example: In a scenario where a precision device (first weighting coefficient is 0.6) operates collaboratively with four other devices (second weighting coefficient is 0.4), if the basic fault critical feature is 0.8 (close to the critical state of a single device) and the critical fluctuation feature is 0.9 (close to the upper limit of total load fluctuation), then the comprehensive index = 0.8 × 0.6 + 0.9 × 0.4 = 0.84.

[0087] Finally, taking the comprehensive fault criticality index as the core and combining it with the electrical parameter benchmark of the target period, a multi-dimensional fault criterion is generated: Time-segmented safety threshold: Multiply the comprehensive fault criticality index by the theoretical total current / power benchmark value at the corresponding time (e.g., comprehensive fault criticality index 0.84 × theoretical total current 25A = 21A) to obtain the actual safety threshold for that period.

[0088] Dynamic early warning curve: The comprehensive fault critical index at each moment is connected in time sequence to form an early warning line that is dynamically adjusted with load changes (e.g., if the index rises during peak electricity consumption periods, the early warning line will rise synchronously).

[0089] Collaborative criterion matrix: Establishes a multi-dimensional threshold combination of current, power, and duration (e.g., current exceeds the threshold, power exceeds the threshold, and duration > 3 seconds). When the comprehensive fault criticality index exceeds 0.8, a more stringent criterion combination is automatically activated.

[0090] As an example, such as Figure 6As shown, the similarity analysis of the operating feature data and the fault critical feature data to determine whether the control conditions are met includes: 501, mapping the operating feature data to a real-time feature vector, mapping the fault critical feature data to a critical feature vector, and using a similarity algorithm to calculate the similarity between the real-time feature vector and the critical feature vector.

[0091] 501. By comparing the similarity with a preset similarity range, the overload risk level is determined, and when the overload risk level is higher than the level threshold, the control condition is determined to be met.

[0092] The operational characteristic data (real-time monitored dynamic parameters such as current, power, fluctuation frequency, and duration) are mapped into real-time feature vectors (e.g., [18A, 4.2kW, 3 times / minute, 8 seconds]) according to preset dimensions (such as current value, power value, number of fluctuations within 5 minutes, and duration of a single fluctuation). Each dimension corresponds to the real-time value of a specific operational parameter.

[0093] Fault critical feature data (time-segmented safety thresholds, dynamic early warning lines, collaborative criterion thresholds, etc.) are mapped to critical feature vectors (e.g., [21A, 5.3kW, 5 times / minute, 10 seconds]) along the same dimensions, with each dimension corresponding to the safety critical value of the parameter. Cosine similarity and Euclidean distance similarity algorithms are used to calculate the matching degree between the two vectors.

[0094] The cosine similarity algorithm is used to calculate the degree of matching, or similarity, between two real-time feature vectors and the critical feature vector. For example, the formula for cosine similarity is: Similarity = (Real-time vector · Critical vector) / (||Real-time vector|| × ||Critical vector||). Here, the numerator is the dot product of the two vectors (reflecting directional consistency), and the denominator is the product of the magnitudes of the two vectors (eliminating the influence of magnitude). The closer the result is to 1, the more consistent the direction of the real-time operating state with the critical state (higher risk); the closer it is to 0, the greater the difference between the two (lower risk).

[0095] Based on system security requirements, multiple similarity intervals are preset. Taking the three-level interval as an example: High-risk interval: similarity 0.8-1.0 (real-time status is close to or reaches the critical threshold).

[0096] Warning range: similarity 0.5-0.8 (real-time status fluctuates but has not reached the critical point).

[0097] Safe range: similarity 0-0.5 (real-time state is far from the critical threshold).

[0098] When the similarity falls into the high-risk range (e.g., 0.9), the overload risk level is determined to be "high." This is higher than the preset threshold (e.g., 0.8), and the control conditions are met, triggering a protection mechanism (e.g., issuing an overload warning, automatically adjusting the load, or tripping). When the similarity falls into the warning range (e.g., 0.6), the risk level is "medium." It does not reach the threshold and the control conditions are not met, but load optimization suggestions will be generated (e.g., reducing the power of non-critical equipment). When the similarity falls into the safe range (e.g., 0.3), the risk level is "low," and the control conditions are not met, maintaining the current operating state.

[0099] As an example, such as Figure 7 As shown, when the subsequent similarity analysis results do not meet the control conditions, a closing control command is sent to the circuit breaker again, including: 601, after the opening control command is sent and executed, a continuous monitoring cycle is started, and the circuit's operating characteristic data is re-acquired in each monitoring cycle and mapped into a new real-time feature vector; wherein, the cycle length of the continuous monitoring cycle is determined comprehensively based on the reason for opening, equipment type and load characteristics.

[0100] 602. Calculate the similarity between the new real-time feature vector and the fault critical feature data to determine whether the control conditions are no longer met. If so, generate and send a closing control command to the circuit breaker.

[0101] After the tripping control command is executed, the system immediately starts a continuous monitoring mechanism, repeatedly collecting real-time operating data of the circuit at a set cycle length (e.g., 5-30 seconds). The cycle length of the continuous monitoring is determined by considering the following factors: Tripping cause: If the tripping is caused by a momentary slight overload (e.g., similarity 0.8-0.9), the fault may subside quickly, so the cycle length is set to a shorter value (e.g., 5-10 seconds); if it is caused by a severe overload (e.g., similarity ≥0.95) or continuous fluctuations, the cycle length is extended (e.g., 20-30 seconds) to avoid premature closing and secondary faults.

[0102] Equipment type and load characteristics: For equipment with frequent start-stop and rapid load fluctuations (such as stamping machines and high-frequency motors), the cycle duration should be short (3-5 seconds) to adapt to dynamic changes; for equipment with stable loads and slow response (such as heating furnaces and central air conditioning), the cycle duration should be longer (15-20 seconds) to reduce invalid monitoring.

[0103] The cycle duration is determined by combining the above two factors. For example: (1) Set a quantitative coefficient K1 for the reason of the circuit breaker tripping. The mapping rule is: K1=1.0 when there is a slight overload (similarity 0.8-0.9); K1=1.5 when there is a moderate overload (similarity 0.9-0.95); K1=2.0 (basic cycle × 2.0) when there is a severe overload or continuous fluctuation (similarity ≥ 0.95).

[0104] (2) Set the quantization coefficient K2 for equipment type and load characteristics. The mapping rules are: K2=1.0 when the start and stop are frequent and the fluctuation is fast (such as stamping machine tool, high frequency motor); K2=1.5 when the fluctuation characteristics are moderate (such as ventilation system, conveyor belt); K2=2.0 when the load is stable and the response is slow (such as heating furnace, central air conditioning).

[0105] The final cycle length is obtained by weighted multiplication of the above two factors: Cycle length = Equipment base cycle (K1) * K2. The equipment base cycle can be, for example, 5 seconds, 10 seconds, 15 seconds, etc.

[0106] Then, within each monitoring cycle, operational characteristic data such as current, power, fluctuation frequency, and duration are re-collected and mapped to new real-time feature vectors according to the same dimensions as before the trip (e.g., current value, power value, number of fluctuations within 5 minutes, duration of a single fluctuation). Also, based on the fault critical characteristic data corresponding to the previously determined target time period, a critical feature vector (e.g., [21A, 5.3kW, 5 times / minute, 10 seconds]) is generated, and the similarity algorithm used before the trip is applied to calculate the similarity between the new real-time feature vector and the critical feature vector. By comparing the new similarity with a preset range, it is determined whether the control conditions are no longer met: if the similarity falls into a safe range (e.g., 0.3) and remains stable within this range for 2-3 consecutive cycles (excluding instantaneous fluctuation interference), then the control conditions are determined to no longer be met, meaning there is currently no overload risk. At this time, the adaptive control unit automatically generates a closing control command, including a closing delay time (e.g., 2 seconds to reduce instantaneous current surge), and sends it to the circuit breaker to perform the closing operation and restore power supply to the circuit. If the similarity remains in the warning or high-risk range (e.g., 0.6), the circuit breaker will remain in the tripped state, and the above process will be repeated in the next monitoring cycle.

[0107] This invention also provides an electronic device for use in an intelligent reclosing circuit breaker control system as described in any of the preceding claims. The electronic device includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.

[0108] This invention also provides a computer storage medium for use in an intelligent reclosing circuit breaker control system as described in any of the preceding claims, the computer storage medium storing a computer program executable by a processor.

[0109] This invention also provides a computer program product for use in an intelligent reclosing circuit breaker control system as described in any of the preceding claims, the computer program product comprising a computer program executable by a processor.

[0110] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0111] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An intelligent reclosing circuit breaker control system, characterized in that, It includes a scenario monitoring unit, a data analysis and decision-making unit, a circuit monitoring unit, and an adaptive control unit. The scenario monitoring unit is used to acquire load operation data of load devices within the control scenario range of the circuit breaker, including real-time operation data and planned operation data. The data analysis and decision-making unit is used to determine basic fault critical characteristics based on the real-time operation data, predict critical fluctuation characteristics based on the planned operation data and random fluctuation characteristics, and integrate the basic fault critical characteristics and critical fluctuation characteristics to obtain fault critical characteristic data for the target time period. The circuit monitoring unit is connected to the circuit and is used to collect the current, voltage and power factor of the circuit during the target time period, and extract the operating characteristic data after preprocessing. The adaptive control unit is used to perform similarity analysis between the operating characteristic data and the fault critical characteristic data to determine whether the control conditions are met. If so, it sends a tripping control command to the circuit breaker. And if the subsequent similar analysis results do not meet the control conditions, a closing control command will be sent to the circuit breaker; Determining critical characteristics of basic faults based on the real-time operating data includes: extracting equipment type parameters, real-time current values, real-time power values, and continuous operating time from the real-time operating data; matching and obtaining the rated current threshold, rated power threshold, and allowable overload duration corresponding to the equipment type parameters based on a preset equipment rated parameter library; calculating a first deviation rate between the real-time current value and the rated current threshold, and a second deviation rate between the real-time power value and the rated power threshold; calculating an overload risk coefficient by combining the first deviation rate, the second deviation rate, and the continuous operating time; when the overload risk coefficient reaches a preset risk threshold, determining the corresponding real-time current value, real-time power value, and continuous operating time as critical characteristics of basic faults for a single device; calculating the total load critical value based on the device correlation relationship; and combining the critical characteristics of basic faults for each single device with the total load critical value to constitute the critical characteristics of basic faults.

2. The intelligent reclosing circuit breaker control system according to claim 1, characterized in that: The critical fluctuation characteristics are predicted based on the planned operation data and random fluctuation features, including: extracting the preset start and stop times, single operation duration, and load type correlation of each load device from the planned operation data to construct a device operation sequence table for the target time period; extracting random fluctuation features from the real-time operation data, including current fluctuation frequency, power fluctuation amplitude, and fluctuation duration, and generating a fluctuation probability distribution model through statistical analysis; calculating the theoretical total load benchmark value at different times within the target time period using a load superposition algorithm based on the device operation sequence table, dynamically correcting the theoretical total load benchmark value in combination with the fluctuation probability distribution model, and calculating the load fluctuation range under different confidence levels; and deriving the critical fluctuation characteristics based on the upper limit of the load fluctuation range under different confidence levels corresponding to the theoretical total load benchmark value at each time, including the time-specific total current limit value, power fluctuation safety threshold, and load superposition warning line for multi-device collaborative operation.

3. The intelligent reclosing circuit breaker control system according to claim 2, characterized in that: Random fluctuation features, including current fluctuation frequency, power fluctuation amplitude, and fluctuation duration, are extracted from the real-time operating data. A fluctuation probability distribution model is generated through statistical analysis, including: extracting current time-series data and power time-series data within a preset time window from the real-time operating data; extracting feature values, including current fluctuation frequency, power fluctuation amplitude, and fluctuation duration, from the dynamic fluctuation segment data; collecting feature values ​​from at least a preset number of preset time windows to form a sample set; using kernel density estimation to fit the probability density of current fluctuation frequency, power fluctuation amplitude, and fluctuation duration in the sample set to generate their respective marginal distribution models; and constructing a joint probability distribution model of the above three feature values ​​using a Copula function, which is then used as the fluctuation probability distribution model.

4. The intelligent reclosing circuit breaker control system according to claim 3, characterized in that: The basic fault critical features and critical fluctuation features are integrated to obtain fault critical feature data for the target time period, including: setting a first weight coefficient for the basic fault critical features based on the real-time requirements of the equipment, and setting a second weight coefficient for the critical fluctuation features based on the complexity of multi-equipment collaborative operation; normalizing the basic fault critical features and critical fluctuation features, and using a weighted fusion algorithm to calculate the comprehensive fault critical index at each time point within the target time period, and generating the fault critical feature data based on the comprehensive fault critical index.

5. The intelligent reclosing circuit breaker control system according to claim 1, characterized in that: The process of performing similarity analysis between the operational feature data and the fault critical feature data to determine whether the control conditions are met includes: mapping the operational feature data to a real-time feature vector, mapping the fault critical feature data to a critical feature vector, calculating the similarity between the real-time feature vector and the critical feature vector using a similarity algorithm, determining the overload risk level by comparing the similarity with a preset similarity interval, and determining that the control conditions are met when the overload risk level is higher than the level threshold.

6. The intelligent reclosing circuit breaker control system according to claim 5, characterized in that: If the subsequent similarity analysis results do not meet the control conditions, a closing control command is sent to the circuit breaker again, including: after the opening control command is sent and executed, starting a continuous monitoring cycle, re-collecting the circuit's operating characteristic data in each monitoring cycle, and mapping it into a new real-time feature vector; wherein, the cycle length of the continuous monitoring cycle is determined comprehensively based on the reason for opening, equipment type and load characteristics; calculating the similarity between the new real-time feature vector and the fault critical characteristic data to determine whether the control conditions are no longer met, and if so, generating and sending a closing control command to the circuit breaker.

7. An electronic device, applied to the intelligent reclosing circuit breaker control system as described in any one of claims 1-6, characterized in that: The electronic device includes: at least one processor, a memory, and a computer program stored in the memory and capable of running on the at least one processor.

8. A computer storage medium, applied to the intelligent reclosing circuit breaker control system as described in any one of claims 1-6, characterized in that: The computer's storage medium stores computer programs that can be executed by a processor.

9. A computer program product, applied to the intelligent reclosing circuit breaker control system as described in any one of claims 1-6, characterized in that: This computer program product contains a computer program that can be executed by a processor.

Images