An integrated measurement and control method and system for circuit breakers for power distribution IoT

Abstract

This invention discloses an integrated measurement and control method and system for circuit breakers in power distribution IoT, relating to the field of particle swarm optimization (PSO) technology. Its key technical points are: acquiring multi-source operating data of the target circuit breaker and adjacent circuit breakers topologically connected to it; aligning the multi-source operating data of the target circuit breaker and adjacent circuit breakers in time sequence, and concatenating the aligned data into a node state matrix; initializing the particle population of the PSO optimization model; updating the inertia weights of the particles based on the communication delay time; calculating the fitness function values ​​of the particles based on the node state matrix; iterating the position and velocity of the particles based on the inertia weights and fitness function values ​​until a preset stopping condition is reached, and outputting the optimal action threshold parameters. This invention improves the accuracy of circuit breaker action control and adaptability to the power distribution IoT environment, reduces the probability of maloperation and cascading tripping, and extends the service life of the equipment.

Description

Technical Field

[0001] This invention relates to the field of particle swarm optimization algorithm technology, specifically to an integrated measurement and control method and system for circuit breakers for power distribution IoT. Background Technology

[0002] In the development of distribution IoT technology, distribution IoT nodes typically integrate smart circuit breakers with data acquisition and edge computing capabilities. Faced with increasingly complex distribution network operating environments, such as the widespread integration of distributed power sources and bidirectional load fluctuations, related technologies mainly employ fixed setting parameters or rely on unified control strategies issued from the cloud to control the circuit breaker's operation. The execution logic of these technologies typically involves: collecting electrical parameters such as voltage and current from the circuit breaker and comparing them with fixed protection logic preset in the microprocessor protection device; or using general optimization algorithms, such as standard particle swarm optimization or genetic algorithms, to find a set of threshold parameters with the smallest error online, thereby controlling the circuit breaker to perform tripping or reclosing operations.

[0003] However, the execution logic of the aforementioned technologies still has some hidden and unresolved deep-seated pain points in actual power distribution IoT scenarios: First, when using general optimization algorithms to dynamically adjust action parameters, the iterative logic within the algorithm is entirely in a state of blind mathematical search, failing to take into account the physical loss limits of the circuit breaker's underlying components, such as the deterioration of contact wear due to frequent interruptions, and ignoring the inherent communication delays and spatial topological distances between power distribution IoT nodes. This disconnect between the algorithm mechanism and the underlying physical conditions makes the calculated output parameters prone to having excessively low fault tolerance in actual physical execution, leading to delayed coordination between adjacent nodes, malfunctions, or even more serious cascading tripping accidents, seriously threatening the power supply reliability of the power distribution network. Second, ignoring the communication delays and spatial topological characteristics between power distribution IoT nodes results in unreasonable particle swarm iteration parameter updates and insufficient collaborative control precision. There are inherent communication delays between power distribution IoT nodes, and these delays increase with the number of topological hops between nodes, while the optimization algorithms in the relevant technologies do not consider this characteristic. First, the particle swarm optimization (PSO) uses only fixed inertia weights and social learning factors for particle iteration, without dynamically updating the inertia weights based on communication delay time or constructing a spatial decay function to optimize the social learning factor in conjunction with network topology hop count. This results in the algorithm's output action parameters being unable to adapt to communication differences between nodes, leading to lag in coordination between adjacent circuit breakers and even cascading tripping accidents. Second, the PSO process lacks an adaptive adjustment mechanism for extreme operating conditions, making it prone to getting trapped in local optima. Furthermore, the optimal parameter output and control command generation lack safety verification. Related technologies do not consider extreme operating conditions such as sudden temperature changes in circuit breakers during the PSO process and do not set up a physical mutation mechanism. When the circuit breaker experiences a sudden temperature rise, the temperature change rate exceeds the threshold, and the algorithm tends to converge, it cannot force an update of particle positions to escape local optima. Simultaneously, after outputting the optimal action threshold parameters, the estimated breaking capacity corresponding to the parameters is not verified, and the estimated breaking capacity is not compared with the physical limit value of the circuit breaker's maximum breaking capacity before directly generating control commands, posing a risk of exceeding the physical limits of the equipment and causing equipment damage. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: An integrated measurement and control method for circuit breakers for power distribution IoT includes: Acquire multi-source operating data of the target circuit breaker and adjacent circuit breakers topologically connected to the target circuit breaker. The multi-source operating data includes at least electrical sampling data, mechanical vibration data, and contact temperature data. The multi-source operation data of the target circuit breaker and its adjacent circuit breakers are time-series aligned and then concatenated into a node state matrix. The contact wear index of the target circuit breaker is calculated based on the node state matrix, and the particle population of the particle swarm optimization model is initialized with the contact wear index as the boundary constraint condition; where each particle in the particle population represents a candidate solution of a set of action threshold parameters. During the iterative update of the particle swarm optimization model, the communication delay time between the target circuit breaker and adjacent circuit breakers is extracted, and the inertia weight of the particles is updated according to the communication delay time. The fitness function value of the particles is calculated based on the node state matrix. The fitness function value is obtained by weighting the current transient tracking error, thermal capacity redundancy, and over-trip penalty term. The position and velocity of the particles are iterated based on the inertia weight and fitness function value until the preset stopping condition is reached. The optimal action threshold parameter is output, and the control command of the target circuit breaker is generated based on the optimal action threshold parameter.

[0005] Furthermore, the multi-source operating data of the target circuit breaker and adjacent circuit breakers are time-aligned, specifically including: Multi-source operating data is collected through Hall sensors, Rogowski coils, piezoelectric sensors, infrared thermal imaging probes, or fiber optic grating thermometers. Among them, electrical sampling data includes three-phase voltage, current waveforms, and power factor; mechanical vibration data includes vibration frequency and amplitude; and contact temperature data includes real-time surface temperature and temperature change trend of stationary and moving contacts. After collection, the data needs to be preprocessed to remove outliers and noise data. Extract the electrical timestamp from the electrical sampling data, the mechanical timestamp from the mechanical vibration data, and the temperature timestamp from the contact temperature data; Using the electrical timestamp as the reference time axis, calculate the time offset of the mechanical timestamp and temperature timestamp relative to the reference time axis; Linear interpolation resampling is performed on mechanical vibration data and contact temperature data using time offset to obtain target aligned data synchronized with electrical timestamps; After time alignment, the multi-source runtime data is arranged in a matrix according to the time dimension, device node dimension, and data type dimension, and then stitched together into a node status matrix after standardization.

[0006] Furthermore, the contact wear index of the target circuit breaker is calculated based on the node state matrix, specifically including: Extract the historical breaking current sequence and single arc duration sequence of the target circuit breaker from the node state matrix. The historical breaking current sequence is the set of effective values ​​of fault current cleared in the historical breaking operation of the circuit breaker, and the single arc duration sequence is the set of time spans from the separation of moving and stationary contacts to the complete extinguishing of the arc in each disconnection operation. Calculate the square of the effective value of each current in the historical interruption current sequence, multiply it by the single arc duration of the corresponding interruption, and sum them up to generate the cumulative arc energy value. The contact wear index is obtained by comparing the accumulated arc energy value with the preset standard rated withstand energy value.

[0007] Furthermore, the particle population for initializing the particle swarm optimization model specifically includes: The lower limit of the maximum permissible breaking time threshold of the target circuit breaker is determined based on the contact wear index. In the preset global parameter space, parameter coordinates that are less than the lower limit of the maximum allowable break time threshold are removed to generate an effective search space; Initial position and velocity vectors are randomly generated within the effective search space to form a particle swarm.

[0008] Furthermore, the inertial weights of the particles are updated based on the communication delay time, specifically including: Monitor the communication network card message interaction status between the target circuit breaker and the adjacent circuit breaker, read the timestamp difference returned by the network layer to obtain the communication delay time, calculate the ratio of the communication delay time to the preset communication delay benchmark value, and obtain the dimensionless delay attenuation coefficient. Obtain the current iteration count and the preset maximum iteration count for the target circuit breaker; Based on the initial inertia weight and the weight adjustment amount that decreases linearly with the number of iterations, combined with the delay decay coefficient, the inertia weight of the current iteration step is obtained. Query the distribution network topology routing table stored in the memory of the target circuit breaker node, calculate the number of routing forwarding nodes between the IP addresses of the target circuit breaker and each of its adjacent circuit breakers, and obtain the network topology hop count; A spatial decay function is constructed based on the number of hops in the network topology. As the number of hops in the network topology increases, the output value of the spatial decay function decreases exponentially. The social learning factor in the particle swarm optimization model is updated using the output value of the spatial decay function. The social learning factor is used to control the speed step size of particles moving toward the global optimum.

[0009] Furthermore, the methods for calculating the fitness function value include: Calculate the thermal capacity redundancy based on the contact temperature data in the node state matrix; Compare the candidate solutions for the action threshold parameters of the target circuit breaker with the action time parameters of the previous round of adjacent circuit breakers. If the candidate solution causes the action time of the target circuit breaker to be greater than the action time of the adjacent circuit breaker, then trigger the over-trip penalty item and assign a preset penalty constant to the over-trip penalty item.

[0010] Furthermore, the process of iterating the particle's position and velocity based on inertia weights and fitness function values ​​also includes: The temperature change rate is obtained by calculating the difference between the contact temperature data at the current moment and the contact temperature data at the previous moment in the node state matrix. When the rate of temperature change is greater than the preset temperature abrupt change threshold and the population fitness variance of the particle swarm optimization model is less than the preset convergence threshold, the physical mutation mechanism is triggered. For particles in the particle population whose fitness function value is lower than the average value, a random perturbation following a Gaussian distribution is introduced to force the particle's position to be updated.

[0011] Furthermore, the optimal action threshold parameters are output, specifically including: The position coordinates of the particle corresponding to the maximum fitness function value are decoded, and the coordinate values ​​of the first dimension are extracted as the threshold of the action current amplitude, and the coordinate values ​​of the second dimension are extracted as the threshold of the reclosing delay time.

[0012] Furthermore, before generating control commands for the target circuit breaker based on the optimal action threshold parameters, the process also includes: Obtain the physical limit value of the maximum breaking capacity of the target circuit breaker under the current environment; Using electrical sampling data in the node state matrix, the estimated breaking capacity generated when the action is executed with the optimal action threshold parameters is deduced; Compare the estimated breaking capacity with the physical limit value of the maximum breaking capacity. If the estimated breaking capacity is less than or equal to the physical limit value of the maximum breaking capacity, then execute the corresponding steps of the control command to generate the target circuit breaker.

[0013] An integrated measurement and control system for circuit breakers for power distribution IoT includes: The multi-source operation data acquisition module acquires multi-source operation data of the target circuit breaker and adjacent circuit breakers topologically connected to the target circuit breaker. The multi-source operation data includes at least electrical sampling data, mechanical vibration data, and contact temperature data. The node state matrix construction module aligns the multi-source operating data of the target circuit breaker and adjacent circuit breakers in time sequence and splices them into a node state matrix. The particle swarm optimization model initialization module calculates the contact wear index of the target circuit breaker based on the node state matrix, and initializes the particle population of the particle swarm optimization model with the contact wear index as the boundary constraint condition; where each particle in the particle population represents a candidate solution of a set of action threshold parameters. The particle swarm optimization iterative update module extracts the communication delay time between the target circuit breaker and adjacent circuit breakers during the iterative update process of the particle swarm optimization model, updates the inertia weight of the particles based on the communication delay time, and calculates the fitness function value of the particles based on the node state matrix. The fitness function value is obtained by weighting the current transient tracking error, thermal capacity redundancy, and over-trip penalty term. The optimal parameter output and control command generation module iterates the position and velocity of the particles based on the inertia weight and fitness function value until the preset stopping condition is reached, outputs the optimal action threshold parameter, and generates the control command for the target circuit breaker based on the optimal action threshold parameter.

[0014] This invention provides an integrated measurement and control method and system for circuit breakers for power distribution IoT, which has the following advantages: 1. This invention integrates multi-source operating data of the target circuit breaker and adjacent circuit breakers and performs time-series alignment processing to construct a node state matrix and calculate the contact wear index based on it. This index is used as the boundary constraint condition for initializing the particle swarm optimization model. This breaks the limitation of existing algorithms being disconnected from the physical loss limit of circuit breakers, enabling the algorithm optimization process to deeply adapt to the actual physical operating state of the circuit breaker, effectively avoiding the risk of control failure, improving the reliability of circuit breaker operation, reducing false operation, failure to operate and potential equipment damage, and also extending the service life of the circuit breaker. 2. To address the issues of communication delay and spatial topology characteristics between distribution IoT nodes being easily overlooked, and insufficient collaborative control accuracy, this invention extracts the communication delay time and network topology hop count between the target circuit breaker and adjacent circuit breakers, and dynamically updates the inertia weight and social learning factor of particles in the particle swarm optimization model. This allows for flexible adaptation to communication differences between nodes, effectively avoiding collaborative control lag and cascading trip accidents, and ensuring the stability of power supply in the distribution network. It is better suited to complex and ever-changing distribution IoT scenarios. 3. Simultaneously, based on multi-source operational data in the node state matrix, this invention constructs a multi-dimensional fitness function that includes current transient tracking error, thermal capacity redundancy, and over-trip penalty term. A physical mutation mechanism is added, effectively solving the problems of easily getting trapped in local optima and insufficient evaluation of candidate parameters in existing particle swarm iteration processes. This significantly improves the optimization performance of the optimization algorithm, ensuring that the output optimal action threshold parameter can balance control accuracy and operational safety, further enhancing the intelligence level of circuit breaker control. Before generating control commands, the breaking capacity is estimated through deduction and compared with the physical limit value of the circuit breaker's maximum breaking capacity, effectively solving the problem of lack of safety verification in existing control commands. This achieves closed-loop coordination between circuit breaker measurement and control, ensuring the safety of control commands and reducing equipment maintenance costs. 4. This invention provides reliable data support for subsequent parameter calculation and optimization by performing time-series alignment processing on multi-source operating data of the target circuit breaker and adjacent circuit breakers. At the same time, before generating control commands, the invention solves the problem of lack of safety verification in existing control commands by deriving and estimating the breaking capacity and comparing it with the physical limit value of the maximum breaking capacity of the circuit breaker. This achieves closed-loop coordination of circuit breaker measurement and control and ensures the safety of control commands. In summary, this solution effectively addresses many pain points in existing circuit breaker control, such as the disconnect between algorithm optimization and equipment physical state, neglect of communication and topology characteristics in collaborative control, easy getting trapped in local optima in parameter optimization, inaccurate processing of multi-source data, and lack of security verification for control commands. It achieves deep collaboration and intelligent optimization of circuit breaker measurement and control, significantly improving the reliability, safety, and intelligence level of circuit breaker operation in power distribution IoT scenarios, and providing strong support for the efficient and stable operation of power distribution IoT systems. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the steps of the present invention. Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1: Please see Figure 1 This embodiment provides an integrated measurement and control method for circuit breakers for power distribution IoT, including: S101, Multi-source Operational Data Acquisition: Acquire multi-source operational data of the target circuit breaker and adjacent circuit breakers topologically connected to the target circuit breaker. This multi-source operational data includes electrical sampling data, mechanical vibration data, and contact temperature data. Specifically, the target circuit breaker is the core circuit breaker requiring measurement and control, and adjacent circuit breakers are those directly connected to the target circuit breaker via the power distribution network topology and possessing electrical connections. This ensures that the acquired data reflects the overall operational status of the target circuit breaker and its surrounding coordinating nodes. Electrical sampling data includes three-phase voltage, current waveforms, and power factor, while mechanical vibration data includes the vibration frequency and amplitude of the operating mechanism during the circuit breaker's opening and closing processes. Contact temperature data includes real-time temperature and temperature change trends of the circuit breaker's stationary and moving contacts, providing comprehensive data support for subsequent equipment status assessment and parameter optimization.

[0018] In some embodiments, multi-source operating data of the target circuit breaker and adjacent circuit breakers topologically connected to the target circuit breaker are acquired. The multi-source operating data includes electrical sampling data, mechanical vibration data, and contact temperature data. The multi-source operating data of the target circuit breaker and adjacent circuit breakers are time-series aligned, and the aligned data is spliced ​​into a node state matrix.

[0019] Specifically, electrical sampling data containing three-phase voltage and current waveforms is collected by Hall sensors and Rogowski coils deployed inside the circuit breaker; mechanical vibration data during the opening and closing process of the circuit breaker is captured by piezoelectric sensors mounted on the circuit breaker housing; contact temperature data of the stationary and moving contacts of the circuit breaker are obtained by infrared thermal imaging probes or fiber optic grating thermometers; electrical timestamps of the electrical sampling data, mechanical timestamps of the mechanical vibration data, and temperature timestamps of the contact temperature data are extracted; the time offset of the mechanical timestamps and temperature timestamps relative to the reference time axis is calculated using the electrical timestamps as the reference time axis; the mechanical vibration data and contact temperature data are linearly interpolated and resampled using the time offsets to obtain target aligned data synchronized with the electrical timestamps; the synchronized target circuit breaker data and the data of two adjacent circuit breakers received through the IoT communication bus are matrix-arranged according to the time and spatial dimensions and spliced ​​to generate a multi-dimensional node state matrix.

[0020] This step collects different types of operational data through multiple sensors, then aligns all data based on electrical timestamps and splices them into a node state matrix, obtaining synchronous and comprehensive node operational data. This provides accurate underlying data support for subsequent parameter optimization and solves the problem of inaccurate optimization data caused by asynchronous sampling and noisy data from multiple sensors in existing technologies.

[0021] S102, Node State Matrix Construction: The multi-source operating data of the target circuit breaker and adjacent circuit breakers are time-series aligned, and the aligned data is spliced ​​into a node state matrix. Due to the differences in the acquisition equipment and acquisition frequency of different types of data, time synchronization problems are prone to occur. Therefore, all multi-source data are first time-series aligned, and then the multi-dimensional operating data of the target circuit breaker and adjacent circuit breakers are integrated through matrix splicing to form a node state matrix that can comprehensively and intuitively represent the operating status of each node, providing a standardized data carrier for subsequent steps such as contact wear calculation and particle swarm optimization.

[0022] In some embodiments, the multi-source operating data of the target circuit breaker and adjacent circuit breakers are time-aligned, and the aligned data is concatenated into a node state matrix. Specifically, this includes: first, extracting the timestamps corresponding to each type of multi-source operating data; using the electrical timestamp of the electrical sampling data as the reference time axis, because the electrical sampling data has the highest acquisition frequency and the strongest correlation with circuit breaker operation, and can most accurately reflect the real-time operating status of the circuit breaker; then calculating the time offset of the mechanical timestamp of the mechanical vibration data and the temperature timestamp of the contact temperature data relative to the reference time axis; and resampling the mechanical vibration data using linear interpolation. Time calibration is performed on the contact temperature data to ensure that the three types of data correspond one-to-one at the same time point, achieving time synchronization of multi-source data. Finally, the synchronized electrical sampling data, mechanical vibration data, and contact temperature data of the target circuit breaker are matrixed with the corresponding synchronized data of adjacent circuit breakers in the order of time dimension, equipment node dimension, and data type dimension, and spliced ​​to generate a multi-dimensional node status matrix. The rows of the matrix correspond to the time series, and the columns correspond to different types of operating data of different equipment nodes, ensuring that the matrix can completely and intuitively present the overall operating status of the target circuit breaker and surrounding collaborative nodes.

[0023] Specifically, the multi-source operating data of the target circuit breaker and adjacent circuit breakers are first preprocessed to remove outliers and noise data, avoiding the impact of invalid data on timing alignment and matrix splicing. Then, the electrical timestamp corresponding to each sampling point of the electrical sampling data is extracted to construct a uniformly distributed reference time axis. The sampling interval of the reference time axis is consistent with the acquisition interval of the electrical sampling data to ensure the accuracy of the reference time axis. Next, for the mechanical vibration data and contact temperature data, the time offset between the timestamp of each data point and the nearest time point on the reference time axis is calculated. A linear interpolation algorithm is then used to re-interpolate the mechanical vibration data and contact temperature data. Sampling is performed to generate target aligned data that is fully synchronized with the reference time axis, ensuring that the three types of data are completely matched in the time dimension. Finally, the synchronized multi-source data of the target circuit breaker and the synchronized multi-source data of adjacent circuit breakers are classified by equipment node and sequentially filled into a preset matrix template. Each equipment node corresponds to a set of column vectors in the matrix, each time point corresponds to a row vector in the matrix, and each column vector is further subdivided into three data sub-columns: electrical, mechanical, and temperature. These are then spliced ​​together to form a complete node state matrix. At the same time, the matrix is ​​standardized to convert data of different dimensions into a unified dimension, which facilitates the subsequent calculation of contact wear index and iterative calculation of particle swarm optimization model.

[0024] This step eliminates time skew in multi-source data through time alignment and integrates multi-dimensional data through matrix splicing. It solves the problems of scattered, asynchronous, and inconsistent dimensions of multi-source data in existing technologies, which lead to large errors in subsequent parameter calculations and inaccurate model iterations. The generated node state matrix can comprehensively, synchronously, and systematically present the operating status of the target circuit breaker and adjacent cooperative nodes, providing reliable underlying data support for the accurate calculation of contact wear index, reasonable initialization of particle swarm optimization model, and efficient iteration. It also provides a data foundation for subsequent multi-node collaborative control and fault prediction, further improving the coordination and accuracy of circuit breaker measurement and control.

[0025] S103, Particle Swarm Optimization Model Initialization: Calculate the contact wear index of the target circuit breaker based on the node state matrix, and initialize the particle population of the particle swarm optimization model using the contact wear index as a boundary constraint. Each particle in the particle population represents a candidate solution for a set of action threshold parameters. First, extract relevant operating data of the target circuit breaker from the node state matrix to accurately calculate the contact wear index, quantify the degree of contact wear, and then determine the physical operating limit of the circuit breaker. Then, use this wear index as a constraint to limit the effective range of action threshold parameters, avoiding candidate parameters that exceed the physical limits of the equipment during the optimization process. Finally, initialize the particle population within the effective range to ensure that the particle population meets the actual operating requirements of the circuit breaker.

[0026] In some embodiments, calculating the contact wear index of the target circuit breaker based on the node state matrix includes: extracting the historical breaking current sequence and single arc duration sequence of the target circuit breaker from the node state matrix; calculating the sum of the products of the square of the historical breaking current sequence and the corresponding single arc duration sequence to obtain the cumulative arc energy value; and comparing the cumulative arc energy value with the preset standard rated withstand energy value to obtain the contact wear index.

[0027] Specifically, the completed node state matrix is ​​analyzed; the historical interruption current sequence representing the current waveform and the single arc duration sequence representing the arc duration are separated; the historical interruption current sequence refers to the set of effective values ​​of fault current cleared by the circuit breaker in historical interruption operations; the single arc duration sequence refers to the set of time spans from the separation of moving and stationary contacts to the complete extinguishing of the arc in each disconnection operation; the effective values ​​of each current in the historical interruption current sequence are squared and multiplied by the single arc duration of the corresponding interruption; all multiplications are summed to output the cumulative arc energy value representing the degree of contact erosion; the standard rated withstand energy value given in the factory nameplate is extracted; the cumulative arc energy value is divided by the standard rated withstand energy value to obtain a ratio between 0 and 1 as the contact wear index. For example, when the calculated ratio is 0.85, it indicates that 85% of the contact life has been consumed.

[0028] This step extracts historical breaking current and arc duration, calculates the cumulative arc energy value, and compares it with the standard withstand energy value to obtain a wear index that accurately reflects the degree of contact wear. This makes the ambiguous contact aging process quantifiable and solves the problems in existing technologies that make it impossible to accurately determine the contact wear state and difficult to grasp the boundary of the circuit breaker's arc extinguishing capability.

[0029] In some embodiments, initializing the particle population of the particle swarm optimization model with the contact wear index as the boundary constraint includes: determining the lower limit of the maximum allowable breaking time threshold of the target circuit breaker based on the contact wear index; removing parameter coordinates that are less than the lower limit of the maximum allowable breaking time threshold in a preset global parameter space to generate an effective search space; and randomly generating initial position vectors and velocity vectors within the effective search space to form the particle population.

[0030] Specifically, the contact wear index calculated in the previous step is read; a preset attenuation mapping table is called to find the lower limit of the maximum allowable breaking time threshold corresponding to the current contact wear index; for example, when the contact wear index reaches 0.85, the table is consulted to find that its mechanical action mechanism requires at least 40 milliseconds to completely disconnect the current due to aging, that is, the lower limit of the maximum allowable breaking time threshold is set to 40 milliseconds; in the global parameter space containing 0 to 100 milliseconds, a matrix filtering operator is used to forcibly remove all parameter coordinates with a time dimension less than 40 milliseconds; in the remaining effective search space, an initial position vector and velocity vector are generated by a pseudo-random number generator to complete the instantiation of the particle swarm.

[0031] This step determines the lower limit of the breaking time threshold by using the contact wear index, eliminates invalid parameter space, and generates a particle population within the effective range. This results in an initial particle set that conforms to the current physical state of the circuit breaker, allowing the algorithm to avoid dangerous parameters from the beginning. This solves the problem in the prior art where the particle population initialization is blind and easily exceeds the physical limits of the equipment, leading to the circuit breaker refusing to operate or failing to break.

[0032] S104, Particle Swarm Iterative Update: During the iterative update of the particle swarm optimization model, the communication delay time between the target circuit breaker and adjacent circuit breakers is extracted, and the inertial weights of the particles are updated according to the communication delay time. The fitness function value of the particles is calculated based on the node state matrix. The fitness function value is obtained by weighting the current transient tracking error, thermal capacity redundancy, and over-trip penalty term. During the iteration process, the communication status between distribution IoT nodes is adapted in real time, and the particle inertial weights are dynamically adjusted to ensure that the algorithm adapts to the differences in network communication. At the same time, through the multi-dimensional fitness function, the merits of each set of action threshold candidate parameters are comprehensively evaluated, taking into account the accuracy of the parameters, equipment safety, and node coordination, providing a scientific basis for the iteration of particle position and velocity.

[0033] In some embodiments, updating the inertial weight of a particle based on the communication delay time includes: obtaining the current iteration number of the target circuit breaker and a preset maximum iteration number; calculating the ratio of the communication delay time to a preset communication delay reference value to obtain a delay attenuation coefficient; subtracting the initial inertial weight from the weight adjustment amount that decreases linearly with the iteration number, and multiplying the result of the subtraction with the delay attenuation coefficient to obtain the inertial weight of the current iteration step.

[0034] Specifically, the process involves monitoring the TCP / IP packet interaction status of the underlying communication network card; reading the timestamp difference returned by the network layer to obtain the communication latency; extracting the current iteration number of the internal controller of the particle swarm optimization model; dividing the communication latency by a pre-configured baseline value, such as 10 milliseconds, to output a dimensionless latency decay coefficient; for example, when the actual communication latency is 20 milliseconds, the latency decay coefficient is calculated to be 2; calculating the weight adjustment amount that decreases with the number of iterations and deducting this adjustment amount from the initial inertial weights; and multiplying the intermediate weight matrix after deduction by the latency decay coefficient to output the final inertial weights for the current iteration step.

[0035] This step obtains iterative parameters adapted to the communication state by acquiring the communication delay time and the number of iterations, calculating the delay attenuation coefficient and the adjusted inertia weight, enabling the algorithm to adapt to changes in network delay. This solves the problems in existing technologies that ignore network delay, leading to outdated data misleading optimization and asynchronous node collaboration.

[0036] In some embodiments, the method for calculating the fitness function value includes: calculating the thermal capacity redundancy based on the contact temperature data in the node state matrix; comparing the candidate solution of the action threshold parameter of the target circuit breaker with the previous round action time parameter of the adjacent circuit breaker; if the candidate solution causes the action time of the target circuit breaker to be greater than the action time of the adjacent circuit breaker, then triggering the over-trip penalty term and assigning a preset penalty constant to the over-trip penalty term.

[0037] Specifically, the contact temperature data at the current moment in the node state matrix is ​​analyzed; thermal capacity redundancy refers to the physical margin that the circuit breaker can continue to absorb heat in the current state without exceeding its insulation and mechanical damage critical temperature; the thermal capacity redundancy is obtained by subtracting the current contact temperature data from the preset maximum withstand temperature threshold; candidate solutions in the action time dimension of the current particle are extracted; the action time parameters of the previous round broadcast by the adjacent circuit breakers through the communication bus are read; Boolean logic comparison is performed, and when it is determined that the estimated action time of the target circuit breaker at this level is longer than the action time of the adjacent circuit breaker at the next lower level, it is determined that a timing reversal has occurred; the over-trip penalty term variable is activated and assigned a very large constant value, such as 9999, so that the overall fitness of the candidate solution deteriorates sharply and it is eliminated.

[0038] This step calculates the thermal capacity redundancy by measuring the contact temperature and then compares the action time trigger penalty to obtain a fitness evaluation result that takes into account both equipment safety and coordination. This eliminates unqualified candidate parameters and solves the problem of the single fitness evaluation in the existing technology, which is prone to equipment overheating and over-tripping, leading to large-scale power outages.

[0039] In some embodiments, after updating the inertial weights of the particles according to the communication delay time, the method further includes: extracting the network topology hop count between the target circuit breaker and each adjacent circuit breaker providing the node state matrix; constructing a spatial decay function based on the network topology hop count, wherein the output value of the spatial decay function decreases exponentially as the network topology hop count increases; and updating the social learning factor in the particle swarm optimization model using the output value of the spatial decay function, wherein the social learning factor is used to control the velocity step size of the particles approaching the global optimal position.

[0040] Specifically, the distribution network topology routing table stored in the node's memory is queried; the number of routing forwarding nodes between the IP address of the target circuit breaker and the IP address of the adjacent circuit breaker providing data is calculated to obtain the network topology hop count; the network topology hop count is used as the independent variable with the base being a negative exponent of the natural constant e to calculate the spatial decay function value; for example, when the hop count is 1, the decay value is 0.36, and when the hop count is 3, the decay value drops to 0.04; this spatial decay function value is directly multiplied into the social learning factor term in the velocity update equation of the particle swarm optimization model.

[0041] This step obtains the learning parameters for the appropriate node distance by acquiring the network topology hop count, constructing a spatial decay function, and updating the social learning factor with the decay value. This allows the algorithm to prioritize referencing data from nearby and highly correlated nodes, solving the problems of blindly referencing all node data, being susceptible to interference from external faults, and wasting computing power in existing technologies.

[0042] In some embodiments, the process of iterating the position and velocity of particles based on inertia weights and fitness function values ​​further includes: calculating the difference between the contact temperature data at the current moment and the contact temperature data at the previous moment in the node state matrix to obtain the temperature change rate; triggering a physical mutation mechanism when it is determined that the temperature change rate is greater than a preset temperature abrupt change threshold and the population fitness variance of the particle swarm optimization model is less than a preset convergence threshold; and introducing a random perturbation quantity following a Gaussian distribution to particles in the particle population whose fitness function value is lower than the average value to force the particle position to be updated.

[0043] Specifically, within each algorithm loop, the contact temperature data of the current loop and the previous loop are read in real time; a subtraction operation is performed to calculate the transient change slope of the contact temperature, i.e., the temperature change rate; the variance of the fitness values ​​of all particles in the current particle population is calculated; it is determined whether the variance is less than the convergence threshold that indicates the algorithm is trapped in a local optimum; when the variance is minimal and the temperature change rate exceeds a mutation threshold such as 15 degrees Celsius per second, the physical mutation interruption function is activated; the set of particles whose current fitness ranking is in the bottom 50% is selected; the Gaussian noise generation function is called to superimpose a random perturbation with a density of 0.1 onto the coordinate matrix of these particles, forcibly breaking them out of the local extremum trap.

[0044] This step calculates the rate of temperature change to determine whether the algorithm has converged, triggers a physical mutation mechanism to add perturbation to the particles, and obtains the particle position that can escape the local optimum. This allows the algorithm to continue to find the optimum normally when the temperature changes drastically. When faced with hidden overloads such as high-resistance grounding causing slow temperature rise, it can instantly jump out of the premature convergence state, ensuring the agile response capability of the protection system.

[0045] S105, Optimal Parameter Output and Control Command Generation: Iterate the particle position and velocity based on inertia weight and fitness function value until a preset stopping condition is reached, output the optimal action threshold parameter, and generate control commands for the target circuit breaker based on the optimal action threshold parameter; continuously iterate the particle position and velocity until the preset stopping condition is met, such as reaching the maximum number of iterations or particle swarm convergence, and select the action threshold parameter with the best fitness; then, combined with the real-time operating status of the target circuit breaker, convert the optimal parameter into a directly executable control command to achieve precise control of the target circuit breaker.

[0046] In some embodiments, the candidate solutions for the action threshold parameters represented by the particle include the action current amplitude threshold and the reclosing delay time threshold; outputting the optimal action threshold parameters includes: decoding the position coordinates of the particle corresponding to the maximum fitness function value, extracting the coordinate value of the first dimension as the action current amplitude threshold, and extracting the coordinate value of the second dimension as the reclosing delay time threshold.

[0047] Specifically, when the algorithm meets the maximum number of iterations or the error convergence condition, it locks the global optimal particle with the highest fitness function value in the current population; reads the multi-dimensional coordinate matrix vector of the optimal particle; separates the floating-point value corresponding to the first index from the vector and converts it into the action current amplitude threshold in amperes; separates the value corresponding to the second index from the vector and converts it into the reclosing delay time threshold in milliseconds; and packages the decoded two-dimensional parameters into a preset message structure for output.

[0048] This step locks the optimal particle, decodes its coordinates to obtain the current and time thresholds, and obtains specific parameters that can be directly used for circuit breaker control. This transforms the abstract particle coordinates into actual executable control parameters, solving the problem in the prior art that the particle coordinates cannot be directly adapted to the circuit breaker execution interface and the search range is redundant. It accurately corresponds to the physical execution interface of the circuit breaker with reclosing function and reduces the redundant search range of the parameter solution space.

[0049] In some embodiments, before generating control instructions for the target circuit breaker based on the optimal action threshold parameters, the method further includes: obtaining the physical limit value of the maximum breaking capacity of the target circuit breaker in the current environment; using electrical sampling data in the node state matrix to deduce the estimated breaking capacity generated when the action is performed using the optimal action threshold parameters; comparing the estimated breaking capacity with the physical limit value of the maximum breaking capacity, and if the estimated breaking capacity is less than or equal to the physical limit value of the maximum breaking capacity, then performing the step of generating control instructions for the target circuit breaker.

[0050] Specifically, the industrial control system configured on the circuit breaker body reads the factory-calibrated maximum breaking capacity physical limit value; extracts the short-circuit current feedforward waveform from the node state matrix and substitutes it into the calculation to obtain the estimated fault current peak value after the short circuit occurs; multiplies the estimated fault current peak value by the system operating voltage to calculate the estimated breaking capacity when performing a tripping action under this threshold parameter; performs digital comparison using conditional branch statements; when the estimated breaking capacity does not exceed the physical limit value of the equipment, opens the control port, compiles the parameters into low-level breaking and reclosing control commands and sends them to the operating mechanism.

[0051] This step obtains safe and feasible control parameters by acquiring the maximum breaking capacity of the equipment, extrapolating and comparing the estimated breaking capacity, and ensuring that the control commands do not exceed the physical limits of the equipment. This solves the problem in existing technologies where directly issuing optimal parameters can easily exceed the equipment's carrying capacity and cause equipment damage. It provides a final hard isolation defense based on transient physical extreme values, further improving the method's fault tolerance and equipment safety under extreme fault conditions.

[0052] Example 2: Please see Figure 2 Based on Example 1, this embodiment also provides an integrated measurement and control system for circuit breakers for power distribution IoT, including: The multi-source operation data acquisition module acquires multi-source operation data of the target circuit breaker and adjacent circuit breakers topologically connected to the target circuit breaker. The multi-source operation data includes electrical sampling data, mechanical vibration data, and contact temperature data. The node state matrix construction module performs time-series alignment of the multi-source operating data of the target circuit breaker and adjacent circuit breakers, and then concatenates the aligned data into a node state matrix. The particle swarm optimization model initialization module calculates the contact wear index of the target circuit breaker based on the node state matrix, and initializes the particle population of the particle swarm optimization model with the contact wear index as the boundary constraint condition. Each particle in the particle population represents a candidate solution of a set of action threshold parameters. The particle swarm optimization iterative update module extracts the communication delay time between the target circuit breaker and adjacent circuit breakers during the iterative update process of the particle swarm optimization model, updates the inertia weight of the particles based on the communication delay time, and calculates the fitness function value of the particles based on the node state matrix. The fitness function value is obtained by weighting the current transient tracking error, thermal capacity redundancy, and over-trip penalty term. The optimal parameter output and control command generation module iterates the position and velocity of the particles based on the inertia weight and fitness function value until the preset stopping condition is reached, outputs the optimal action threshold parameter, and generates control commands for the target circuit breaker based on the optimal action threshold parameter.

[0053] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. 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, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0054] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0055] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An integrated measurement and control method for circuit breakers for power distribution IoT, characterized in that, The method includes: Acquire multi-source operating data of the target circuit breaker and adjacent circuit breakers topologically connected to the target circuit breaker. The multi-source operating data includes at least electrical sampling data, mechanical vibration data, and contact temperature data. The multi-source operation data of the target circuit breaker and its adjacent circuit breakers are time-series aligned and then concatenated into a node state matrix. The contact wear index of the target circuit breaker is calculated based on the node state matrix, and the particle population of the particle swarm optimization model is initialized with the contact wear index as the boundary constraint condition; where each particle in the particle population represents a candidate solution of a set of action threshold parameters. During the iterative update of the particle swarm optimization model, the communication delay time between the target circuit breaker and adjacent circuit breakers is extracted, and the inertia weight of the particles is updated according to the communication delay time. The fitness function value of the particles is calculated based on the node state matrix. The fitness function value is obtained by weighting the current transient tracking error, thermal capacity redundancy, and over-trip penalty term. The position and velocity of the particles are iterated based on the inertia weight and fitness function value until the preset stopping condition is reached. The optimal action threshold parameter is output, and the control command of the target circuit breaker is generated based on the optimal action threshold parameter.

2. The integrated measurement and control method for circuit breakers for power distribution IoT as described in claim 1, characterized in that, The timing alignment of the target circuit breaker's multi-source operating data with that of adjacent circuit breakers includes: Multi-source operating data is collected through Hall sensors, Rogowski coils, piezoelectric sensors, infrared thermal imaging probes, or fiber optic grating thermometers. Among them, electrical sampling data includes three-phase voltage, current waveforms, and power factor; mechanical vibration data includes vibration frequency and amplitude; and contact temperature data includes real-time surface temperature and temperature change trend of stationary and moving contacts. After collection, the data needs to be preprocessed to remove outliers and noise data. Extract the electrical timestamp from the electrical sampling data, the mechanical timestamp from the mechanical vibration data, and the temperature timestamp from the contact temperature data; Using the electrical timestamp as the reference time axis, calculate the time offset of the mechanical timestamp and temperature timestamp relative to the reference time axis; Linear interpolation resampling is performed on mechanical vibration data and contact temperature data using time offset to obtain target aligned data synchronized with electrical timestamps; After time alignment, the multi-source runtime data is arranged in a matrix according to the time dimension, device node dimension, and data type dimension, and then stitched together into a node status matrix after standardization.

3. The integrated measurement and control method for circuit breakers for power distribution IoT as described in claim 1, characterized in that, The contact wear index of the target circuit breaker is calculated based on the node state matrix, specifically including: Extract the historical breaking current sequence and single arc duration sequence of the target circuit breaker from the node state matrix. The historical breaking current sequence is the set of effective values ​​of fault current cleared in the historical breaking operation of the circuit breaker, and the single arc duration sequence is the set of time spans from the separation of moving and stationary contacts to the complete extinguishing of the arc in each disconnection operation. Calculate the square of the effective value of each current in the historical interruption current sequence, multiply it by the single arc duration of the corresponding interruption, and sum them up to generate the cumulative arc energy value. The contact wear index is obtained by comparing the accumulated arc energy value with the preset standard rated withstand energy value.

4. The integrated measurement and control method for circuit breakers for power distribution IoT as described in claim 1, characterized in that, Initialize the particle population for the particle swarm optimization model, specifically including: The lower limit of the maximum permissible breaking time threshold of the target circuit breaker is determined based on the contact wear index. In the preset global parameter space, parameter coordinates that are less than the lower limit of the maximum allowable break time threshold are removed to generate an effective search space; Initial position and velocity vectors are randomly generated within the effective search space to form a particle swarm.

5. The integrated measurement and control method for circuit breakers for power distribution IoT as described in claim 1, characterized in that, The inertial weights of particles are updated based on the communication delay time, specifically including: Monitor the communication network card message interaction status between the target circuit breaker and the adjacent circuit breaker, read the timestamp difference returned by the network layer to obtain the communication delay time, calculate the ratio of the communication delay time to the preset communication delay benchmark value, and obtain the dimensionless delay attenuation coefficient. Obtain the current iteration count and the preset maximum iteration count for the target circuit breaker; Based on the initial inertia weight and the weight adjustment amount that decreases linearly with the number of iterations, combined with the delay decay coefficient, the inertia weight of the current iteration step is obtained. Query the distribution network topology routing table stored in the memory of the target circuit breaker node, calculate the number of routing forwarding nodes between the IP addresses of the target circuit breaker and each of its adjacent circuit breakers, and obtain the network topology hop count; A spatial decay function is constructed based on the number of hops in the network topology. As the number of hops in the network topology increases, the output value of the spatial decay function decreases exponentially. The social learning factor in the particle swarm optimization model is updated using the output value of the spatial decay function. The social learning factor is used to control the speed step size of particles moving toward the global optimum.

6. The integrated measurement and control method for circuit breakers for power distribution IoT as described in claim 1, characterized in that, The methods for calculating the fitness function value include: Calculate the thermal capacity redundancy based on the contact temperature data in the node state matrix; Compare the candidate solutions for the action threshold parameters of the target circuit breaker with the action time parameters of the previous round of adjacent circuit breakers. If the candidate solution causes the action time of the target circuit breaker to be greater than the action time of the adjacent circuit breaker, then trigger the over-trip penalty item and assign a preset penalty constant to the over-trip penalty item.

7. The integrated measurement and control method for circuit breakers oriented towards power distribution IoT as described in claim 1, characterized in that, The process of iterating the position and velocity of particles based on inertia weights and fitness function values ​​also includes: The temperature change rate is obtained by calculating the difference between the contact temperature data at the current moment and the contact temperature data at the previous moment in the node state matrix. When the rate of temperature change is greater than the preset temperature abrupt change threshold and the population fitness variance of the particle swarm optimization model is less than the preset convergence threshold, the physical mutation mechanism is triggered. For particles in the particle population whose fitness function value is lower than the average value, a random perturbation following a Gaussian distribution is introduced to force the particle's position to be updated.

8. The integrated measurement and control method for circuit breakers for power distribution IoT as described in claim 1, characterized in that, Output the optimal action threshold parameters, specifically including: The position coordinates of the particle corresponding to the maximum fitness function value are decoded, and the coordinate values ​​of the first dimension are extracted as the threshold of the action current amplitude, and the coordinate values ​​of the second dimension are extracted as the threshold of the reclosing delay time.

9. The integrated measurement and control method for circuit breakers for power distribution IoT as described in claim 1, characterized in that, Before generating control commands for the target circuit breaker based on the optimal action threshold parameters, the following steps are also included: Obtain the physical limit value of the maximum breaking capacity of the target circuit breaker under the current environment; Using electrical sampling data in the node state matrix, the estimated breaking capacity generated when the action is executed with the optimal action threshold parameters is deduced; Compare the estimated breaking capacity with the physical limit value of the maximum breaking capacity. If the estimated breaking capacity is less than or equal to the physical limit value of the maximum breaking capacity, then execute the corresponding steps of the control command to generate the target circuit breaker.

10. An integrated measurement and control system for circuit breakers for power distribution IoT, characterized in that, An integrated measurement and control method for circuit breakers for power distribution IoT as described in any one of claims 1-9 includes: The multi-source operation data acquisition module acquires multi-source operation data of the target circuit breaker and adjacent circuit breakers topologically connected to the target circuit breaker. The multi-source operation data includes at least electrical sampling data, mechanical vibration data, and contact temperature data. The node state matrix construction module aligns the multi-source operating data of the target circuit breaker and adjacent circuit breakers in time sequence and splices them into a node state matrix. The particle swarm optimization model initialization module calculates the contact wear index of the target circuit breaker based on the node state matrix, and initializes the particle population of the particle swarm optimization model with the contact wear index as the boundary constraint condition; where each particle in the particle population represents a candidate solution of a set of action threshold parameters. The particle swarm optimization iterative update module extracts the communication delay time between the target circuit breaker and adjacent circuit breakers during the iterative update process of the particle swarm optimization model, updates the inertia weight of the particles based on the communication delay time, and calculates the fitness function value of the particles based on the node state matrix. The fitness function value is obtained by weighting the current transient tracking error, thermal capacity redundancy, and over-trip penalty term. The optimal parameter output and control command generation module iterates the position and velocity of the particles based on the inertia weight and fitness function value until the preset stopping condition is reached, outputs the optimal action threshold parameter, and generates the control command for the target circuit breaker based on the optimal action threshold parameter.

Images