Modularized intelligent power supply rapid switching method and system for power grid fault

Abstract

The invention provides a power grid fault-oriented modularized intelligent power supply rapid switching method and system, and relates to the field of power systems, and the method comprises the steps: collecting the parameter variation of power grid nodes, building a fault propagation velocity field based on the topological relation between the nodes, and predicting the time when a fault reaches each node; according to a node fault arrival time sequence, controlling a standby power supply module to output virtual load current to establish internal flux linkage distribution before the fault arrives; collecting a voltage drop track of the fault power supply module and a voltage rise track of the standby power supply module, and calculating a cross moment and a cross voltage value based on a flux linkage steady-state value; identifying that the fault does not reach the power supply module, calculating transmission impedance to a target load node, and establishing a virtual electrical coupling channel; and controlling the standby power supply module to output voltage at the crossing moment, and executing power supply switching when the phase difference change rate is zero and the output voltage reaches a crossing voltage value. According to the invention, fault prediction and rapid seamless switching are realized, and the power supply reliability and the switching speed are improved.

Description

Technical Field

[0001] This invention relates to power system technology, and more particularly to a modular intelligent power supply fast switching method and system for grid faults. Background Technology

[0002] Traditional power switching technologies primarily rely on a passive response mechanism after fault detection, meaning that the backup module is activated and the switching operation is performed only after a fault in the power supply module is detected. This approach suffers from significant response lag; the time from fault detection and backup module activation to successful switching is typically substantial. During this period, the load may experience voltage dips or even temporary power outages, which can severely impact sensitive loads. Furthermore, existing technologies generally suffer from long internal flux linkage establishment times during the backup power supply module's startup process. Power supply modules, especially those containing energy storage components such as inductors, require a certain amount of time to establish a steady-state flux linkage distribution during the initial startup phase. If switching is performed before the flux linkage is sufficiently established, it can lead to severe output voltage fluctuations and large inrush currents, affecting power quality and potentially impacting the load equipment. Traditional methods typically employ a fixed pre-startup time, but this approach is ill-suited to varying operating conditions and load conditions, and cannot guarantee optimal flux linkage status at the moment of switching.

[0003] Existing technologies also have significant shortcomings in terms of timing and voltage matching during power supply switching. If there is a large difference in amplitude or phase between the output voltage of the faulty power supply module and the output voltage of the backup power supply module during switching, circulating current and voltage surges will occur at the moment of switching, affecting the normal operation of the load. Traditional methods often use simple voltage threshold judgment or fixed delay control, lacking accurate tracking and prediction of voltage change trajectories on both sides, and thus failing to achieve smooth, shock-free switching. Especially when a fault causes a rapid voltage drop, it is difficult to capture the optimal switching window, easily leading to switching too early or too late. Switching too early may subject the backup module to the fault impact, while switching too late will prolong the load's voltage loss time. Summary of the Invention

[0004] This invention provides a modular intelligent power supply rapid switching method and system for power grid faults, which can solve the problems in the prior art.

[0005] A first aspect of this invention provides a modular intelligent power supply fast switching method for power grid faults, comprising:

[0006] The parameter changes of power grid nodes are collected, and a fault propagation velocity field is established based on the topological relationship between nodes based on the parameter changes. The arrival time of faults at each node is predicted based on the fault propagation velocity field to obtain the node fault arrival time sequence.

[0007] Based on the arrival sequence of node faults, the backup power supply module is controlled to output virtual load current before the fault arrives. The backup power supply module is then driven by the virtual load current to establish an internal flux distribution and obtain the steady-state value of the flux.

[0008] Collect the voltage drop trajectory of the faulty power supply module and the voltage rise trajectory of the backup power supply module, and calculate the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory based on the magnetic flux steady-state value;

[0009] Identify the power supply modules that have not been reached by the fault based on the arrival time sequence of the node fault, calculate the transmission impedance from the power supply modules that have not been reached by the fault to the target load node, and establish a virtual electrical coupling channel;

[0010] Energy is transferred through a virtual electrical coupling channel. The backup power supply module outputs voltage at the crossover moment. The phase difference rate between the backup power supply module output voltage and the actual voltage of the target load node is calculated. When the phase difference rate is zero and the output voltage reaches the crossover voltage value, power supply switching is performed.

[0011] The parameters of power grid nodes are collected, and a fault propagation velocity field is established based on the topological relationships between nodes. The arrival time of faults at each node is predicted based on the fault propagation velocity field, resulting in the following node fault arrival time sequence:

[0012] Collect the operating parameters of the power grid nodes, calculate the difference of the operating parameters in the time dimension to obtain the parameter change, calculate the rate of change of the parameter change, and identify abnormal nodes whose rate of change exceeds the normal fluctuation threshold.

[0013] Obtain the electrical distance and line impedance between the abnormal node and the other nodes in the power grid, and construct the topology matrix between nodes;

[0014] The propagation rate of fault disturbance between nodes is calculated based on the rate of change of parameter changes and the topological relationship matrix. The propagation rate is then mapped to the spatial location of the nodes to form a fault propagation velocity field.

[0015] The propagation time required for a fault to spread from an abnormal node to the remaining nodes is calculated by integrating the fault propagation velocity field along each transmission path. The predicted arrival time of the fault at each node is obtained by adding the detection time of the abnormal node to the propagation time. The nodes are then arranged in the order of the predicted times to form the node fault arrival time sequence.

[0016] Based on the node fault arrival sequence, the backup power supply module is controlled to output a virtual load current before the fault arrives. The virtual load current drives the backup power supply module to establish an internal flux linkage distribution, and the steady-state value of the flux linkage is obtained, including:

[0017] The predicted time of arrival of the fault at the node where the backup power supply module is located is determined based on the fault arrival sequence, and the time interval between the current time and the predicted time is calculated as the preparation time window.

[0018] Within the preparation time window, the power switching devices of the backup power supply module are controlled to operate according to the preset modulation strategy, so that the backup power supply module outputs virtual load current to the virtual load.

[0019] The virtual load current generates magnetic field energy in the output filter inductor and the output transformer and forms an internal magnetic flux distribution. The inductance current of the output filter inductor and the excitation current of the output transformer are collected, and the evolution curve of the internal magnetic flux distribution over time is calculated based on the inductance current and the excitation current.

[0020] When the change amplitude of the evolution curve within a continuous time period is lower than the preset convergence threshold, it is determined that the internal flux distribution has reached a stable state, and the flux values ​​corresponding to the inductor current and excitation current in the stable state are extracted as the flux steady-state values.

[0021] The inductor current of the output filter inductor and the magnetizing current of the output transformer are collected. Based on the inductor current and the magnetizing current, the evolution curve of the internal flux linkage distribution over time is calculated, including:

[0022] The inductor current of the output filter inductor and the magnetizing current of the output transformer are collected according to a fixed sampling period to obtain the inductor current sampling sequence and the magnetizing current sampling sequence.

[0023] The current change rate feature is extracted by performing time-domain differentiation on the inductor current sampling sequence. The current change rate feature is then combined with the current current sampling sequence to obtain the output filter inductor flux linkage sequence by weighted integration. The frequency domain decomposition of the excitation current sampling sequence is performed to extract the harmonic component distribution. The excitation current sampling sequence is then combined with the harmonic component distribution to obtain the output transformer flux linkage sequence by piecewise integration.

[0024] Extract the topological connection relationship and energy transfer path between the output filter inductor and the output transformer inside the backup power supply module, and construct a spatial weight matrix that reflects the magnetic flux coupling strength.

[0025] The output filter inductor flux sequence and the output transformer flux sequence are weighted and fused using a spatial weight matrix to obtain the total flux sequence, which is then mapped to a time-series coordinate system to form the evolution curve of the internal flux distribution over time.

[0026] The voltage drop trajectory of the faulty power supply module and the voltage rise trajectory of the backup power supply module are collected. Based on the steady-state value of the magnetic flux linkage, the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory are calculated, including:

[0027] The voltage values ​​of the faulty power supply module and the backup power supply module are collected, and the voltage change rate is obtained by calculating the voltage difference between adjacent sampling points.

[0028] The sampling time interval is set based on the voltage change rate. The voltage sampling sequence is obtained according to the sampling time interval. Wavelet transform is performed on the voltage sampling sequence and the voltage waveform sequence is reconstructed.

[0029] The amplitude and phase range of the voltage waveform sequence are determined based on the steady-state value of the magnetic flux linkage. Within the determined amplitude and phase ranges, the voltage waveform sequence is reconstructed to generate the voltage drop trajectory function of the fault power supply module and the voltage rise trajectory function of the backup power supply module.

[0030] Calculate the magnetic flux distribution within the overlapping interval of the voltage drop trajectory function of the fault power supply module and the voltage rise trajectory function of the backup power supply module, and extract the time point where the minimum value of the magnetic flux distribution is located as the crossover time.

[0031] Substitute the crossover moment into the voltage drop trajectory function and the voltage rise trajectory function of the fault power supply module respectively to obtain the voltage value of the drop trajectory and the voltage value of the rise trajectory. Calculate the average value of the voltage value of the drop trajectory and the voltage value of the rise trajectory as the crossover voltage value.

[0032] Identify power supply modules that have not yet reached the fault based on the node fault arrival time sequence, calculate the transmission impedance from the power supply module that has not yet reached the fault to the target load node, and establish a virtual electrical coupling channel, including:

[0033] Extract the predicted arrival time of each node fault from the node fault arrival time sequence, construct the node state matrix, and calculate the fault propagation time distribution from the node state matrix;

[0034] Based on the fault propagation time distribution and predicted time, power supply modules that have not reached the fault time are selected as fault-not-reached power supply modules, and the spatial location of the fault-not-reached power supply modules is determined.

[0035] The electrical parameters of the line between the power supply module that the fault has not reached and the target load node are collected. The transmission impedance is calculated based on the electrical parameters. The transmission impedance distribution is constructed using the spatial location and fault propagation time distribution. The line combination with the smallest impedance is selected as the transmission channel according to the magnitude of the transmission impedance.

[0036] A detection pulse signal is applied to the transmission channel, and the voltage and current responses generated by the detection pulse signal through the transmission channel are recorded. The amplitude ratio and phase difference are calculated, and the amplitude ratio and phase difference are used as control parameters of the compensation circuit. The compensation circuit is configured to be connected to the transmission channel to form a virtual electrical coupling channel.

[0037] Energy is transferred through a virtual electrical coupling channel. The backup power supply module outputs voltage at the crossover moment. The phase difference rate of change between the backup power supply module output voltage and the actual voltage of the target load node is calculated. When the phase difference rate of change is zero and the output voltage reaches the crossover voltage value, power supply switching is performed, including:

[0038] The voltage and current values ​​on the virtual electrical coupling channel are collected, and the voltage and current values ​​are discretely sampled according to a preset sampling frequency. The instantaneous energy transfer power sequence is calculated from the discrete sampled values.

[0039] Identify power fluctuation patterns in the power transfer power sequence, construct a predictive control sequence based on the power fluctuation patterns, map the predictive control sequence to a backup power supply module drive signal, and control the backup power supply module to output voltage at the crossover moment through the drive signal;

[0040] The output voltage of the backup power supply module and the actual voltage of the target load node are collected. The inherent components of the voltage signal are extracted using empirical mode decomposition. The phase information is separated from the inherent components to construct the phase difference distribution. The phase difference distribution is adaptively corrected by combining the power fluctuation law to obtain the phase difference change rate.

[0041] Power supply switching is performed when the phase difference change rate is zero and the output voltage reaches the cross voltage value.

[0042] A second aspect of this invention provides a modular intelligent power supply fast switching system for power grid faults, comprising:

[0043] The first unit is used to collect parameter changes of power grid nodes, establish a fault propagation velocity field based on the topological relationship between nodes based on the parameter changes, and predict the time when the fault arrives at each node based on the fault propagation velocity field to obtain the node fault arrival time sequence.

[0044] The second unit is used to control the backup power supply module to output virtual load current before the fault arrives, based on the node fault arrival time sequence. The backup power supply module is then driven by the virtual load current to establish an internal flux distribution and obtain the flux steady-state value.

[0045] The third unit is used to collect the voltage drop trajectory of the fault power supply module and the voltage rise trajectory of the backup power supply module, and calculate the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory based on the magnetic flux steady-state value.

[0046] The fourth unit is used to identify the power supply module that has not been reached due to the fault based on the arrival time sequence of the node fault, calculate the transmission impedance from the power supply module that has not been reached due to the fault to the target load node, and establish a virtual electrical coupling channel.

[0047] The fifth unit is used to transfer energy through a virtual electrical coupling channel, control the output voltage of the backup power supply module at the crossover moment, calculate the phase difference rate of change between the output voltage of the backup power supply module and the actual voltage of the target load node, and perform power supply switching when the phase difference rate of change is zero and the output voltage reaches the crossover voltage value.

[0048] A third aspect of the present invention provides an electronic device, comprising:

[0049] processor;

[0050] Memory used to store processor-executable instructions;

[0051] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0052] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0053] In this embodiment, by establishing a fault propagation velocity field to predict the arrival time of faults at each node, switching preparations can be made in advance before the fault actually affects the power supply module. Compared with the traditional method of responding after a fault occurs, this significantly improves the response speed of power supply switching, effectively avoids power interruption caused by switching delays, and improves the continuity and reliability of power grid supply. By pre-driving the backup power supply module to establish an internal flux distribution through virtual load current, and accurately calculating the crossover time and crossover voltage value of the voltage trajectory based on the steady-state value of the flux, a smooth transition in voltage amplitude and phase between the backup power supply module and the faulty power supply module is achieved. This eliminates the inrush current caused by voltage mutations and phase jumps in traditional switching methods, protects the safety of power supply equipment and load equipment, and improves the stability of power supply switching. By establishing a virtual electrical coupling channel and monitoring the phase difference change rate in real time, precise voltage synchronization between the backup power supply module and the target load node is achieved. This ensures that switching is performed when the phase difference change rate is zero and the voltage reaches the crossover value, making the switching process seamless and avoiding the adverse effects of power interruption and voltage fluctuations on the load. This significantly improves the intelligent switching performance and overall operational stability of the modular power supply system under grid fault conditions. Attached Figure Description

[0054] Figure 1 This is a flowchart illustrating the modular intelligent power supply fast switching method for power grid faults according to an embodiment of the present invention.

[0055] Figure 2 This diagram illustrates a method for obtaining the steady-state value of the flux linkage of a backup power supply module based on virtual load current, according to an embodiment of the present invention. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0058] Figure 1 This is a flowchart illustrating the modular intelligent power supply fast switching method for power grid faults according to an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes:

[0059] The parameter changes of power grid nodes are collected, and a fault propagation velocity field is established based on the topological relationship between nodes based on the parameter changes. The arrival time of faults at each node is predicted based on the fault propagation velocity field to obtain the node fault arrival time sequence.

[0060] Based on the arrival sequence of node faults, the backup power supply module is controlled to output virtual load current before the fault arrives. The backup power supply module is then driven by the virtual load current to establish an internal flux distribution and obtain the steady-state value of the flux.

[0061] Collect the voltage drop trajectory of the faulty power supply module and the voltage rise trajectory of the backup power supply module, and calculate the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory based on the magnetic flux steady-state value;

[0062] Identify the power supply modules that have not been reached by the fault based on the arrival time sequence of the node fault, calculate the transmission impedance from the power supply modules that have not been reached by the fault to the target load node, and establish a virtual electrical coupling channel;

[0063] Energy is transferred through a virtual electrical coupling channel. The backup power supply module outputs voltage at the crossover moment. The phase difference rate between the backup power supply module output voltage and the actual voltage of the target load node is calculated. When the phase difference rate is zero and the output voltage reaches the crossover voltage value, power supply switching is performed.

[0064] In one optional implementation, parameter changes at power grid nodes are collected, a fault propagation velocity field is established based on the topological relationships between nodes according to the parameter changes, and the arrival time of faults at each node is predicted based on the fault propagation velocity field, resulting in a node fault arrival time sequence including:

[0065] Collect the operating parameters of the power grid nodes, calculate the difference of the operating parameters in the time dimension to obtain the parameter change, calculate the rate of change of the parameter change, and identify abnormal nodes whose rate of change exceeds the normal fluctuation threshold.

[0066] Obtain the electrical distance and line impedance between the abnormal node and the other nodes in the power grid, and construct the topology matrix between nodes;

[0067] The propagation rate of fault disturbance between nodes is calculated based on the rate of change of parameter changes and the topological relationship matrix. The propagation rate is then mapped to the spatial location of the nodes to form a fault propagation velocity field.

[0068] The propagation time required for a fault to spread from an abnormal node to the remaining nodes is calculated by integrating the fault propagation velocity field along each transmission path. The predicted arrival time of the fault at each node is obtained by adding the detection time of the abnormal node to the propagation time. The nodes are then arranged in the order of the predicted times to form the node fault arrival time sequence.

[0069] In a specific implementation, the first step is to collect the operating parameters of each node in the power grid in real time. These operating parameters include, but are not limited to, voltage amplitude, phase angle, frequency, active power, and reactive power. High-precision synchronous phasor measurement devices are used to install data acquisition terminals at key nodes of the power grid, with a sampling frequency set to 100Hz to ensure data time synchronization and accuracy. The collected operating parameters are then subjected to time-dimension difference calculations, i.e., calculating the difference between the parameter value at the current moment and the parameter value at the previous moment, to obtain the parameter change. Taking voltage amplitude as an example, if the voltage at a node is 220kV at time t and 219.5kV at time t+10ms, then the voltage change at that node within those 10ms is -0.5kV. The rate of change of the parameter change is further calculated, i.e., the parameter change divided by the time interval. In the above example, the voltage change rate is -50kV / s. Based on historical data statistical analysis, threshold ranges for normal fluctuations of various parameters are set; for the voltage change rate, the normal fluctuation threshold is set to ±30kV / s. When the parameter change rate of a node exceeds the set threshold, the node is marked as an abnormal node and used as a potential source of failure.

[0070] After identifying the anomalous node, it is necessary to obtain the topological relationship information between this node and the other nodes in the power grid. The physical connection structure of the power grid is extracted through the power system management platform to obtain the electrical distance and line impedance data between each node. Electrical distance is represented as the sum of the line lengths along the shortest connection path between two nodes, while line impedance is the vector sum of the impedances of each segment along the corresponding path. An n×n topological relationship matrix R is constructed with the anomalous node as the center, where n is the total number of nodes in the power grid, and the matrix element R(i,j) represents the combined relationship value between the electrical distance and line impedance from node i to node j. When there is a direct connection between two nodes, the value of R(i,j) is the actual line parameter; when there is no direct connection between two nodes, the value of R(i,j) is calculated through the path of the intermediate nodes.

[0071] Based on the parameter change rate of the anomalous nodes and the established topological relationship matrix, the propagation rate of fault disturbances between nodes is calculated. The propagation rate is directly proportional to the magnitude of parameter change and inversely proportional to the electrical distance and line impedance between nodes. For an anomalous node A, its parameter change rate is VA, and its topological relationship to node B is R(A, B). Then, the propagation rate of the fault from A to B can be expressed as a function inversely proportional to VA and R(A, B). The calculated propagation rates in each direction are mapped to the geographic spatial location information of the nodes to form a two-dimensional or three-dimensional fault propagation velocity field. This velocity field is visualized as a gradient distribution map radiating outward from the anomalous node, where the arrow direction indicates the fault propagation direction and the arrow length indicates the propagation rate magnitude.

[0072] Using the constructed fault propagation velocity field, the propagation time required for a fault to spread from the anomalous node to the remaining nodes is calculated by integral calculation along each propagation path. Specifically, for each possible path from the anomalous node to the target node, the path is divided into several small segments. The propagation time of each segment is equal to the segment length divided by the propagation rate on that segment. The propagation times of each segment are summed to obtain the propagation time of the entire path. The path with the shortest propagation time is selected from multiple possible paths as the actual fault propagation path. By adding the initial detection time of the anomalous node to the calculated propagation time, the arrival time of the fault at each node can be predicted. For example, if an anomalous node is detected with abnormal parameters at 10:15:30.000, and it is calculated that the fault propagation to node C takes 150ms, then the predicted arrival time of the fault at node C is 10:15:30.150. Based on the predicted fault arrival times, the nodes are arranged in chronological order to form the node fault arrival time sequence.

[0073] In practical applications, when a short-circuit fault occurs in a regional power grid, the first abnormal node detected is the busbar of substation A, whose voltage change rate reaches -75kV / s, exceeding the normal fluctuation threshold. The system immediately marks A as an abnormal node and obtains its topological relationship information with 15 surrounding nodes, constructing a 16×16 topological relationship matrix. Based on the voltage change rate of node A and the topological matrix, the propagation rate of the fault in each direction is calculated, forming a fault propagation velocity field centered on A. Through velocity field integration calculation, it is predicted that the fault will arrive at node B (after 45ms), node C (after 78ms), node D (after 102ms), etc., in sequence. Finally, the node fault arrival sequence is generated as ABCDEFGHIJKLMNOP, providing a timing basis for subsequent protection and control decisions.

[0074] In this embodiment, a method for establishing a fault propagation velocity field is used to accurately describe and predict the power grid fault propagation process. By fully considering the dynamic characteristics of the power grid topology and parameter changes, the propagation time from the fault source to each node can be accurately calculated, improving the accuracy of fault location and early warning. By predicting the arrival time of the fault at each node, corresponding protection measures can be activated in advance, providing a time window for rapid switching of smart power supply, effectively reducing the fault propagation range, and minimizing power outage time and economic losses.

[0075] like Figure 2 The flowchart shown illustrates the method for obtaining the steady-state value of the flux linkage of the backup power supply module based on the virtual load current in this embodiment.

[0076] In one optional implementation, based on the node fault arrival sequence, the backup power supply module is controlled to output a virtual load current before the fault arrives. The virtual load current drives the backup power supply module to establish an internal flux linkage distribution. Obtaining the steady-state value of the flux linkage includes:

[0077] The predicted time of arrival of the fault at the node where the backup power supply module is located is determined based on the fault arrival sequence, and the time interval between the current time and the predicted time is calculated as the preparation time window.

[0078] Within the preparation time window, the power switching devices of the backup power supply module are controlled to operate according to the preset modulation strategy, so that the backup power supply module outputs virtual load current to the virtual load.

[0079] The virtual load current generates magnetic field energy in the output filter inductor and the output transformer and forms an internal magnetic flux distribution. The inductance current of the output filter inductor and the excitation current of the output transformer are collected, and the evolution curve of the internal magnetic flux distribution over time is calculated based on the inductance current and the excitation current.

[0080] When the change amplitude of the evolution curve within a continuous time period is lower than the preset convergence threshold, it is determined that the internal flux distribution has reached a stable state, and the flux values ​​corresponding to the inductor current and excitation current in the stable state are extracted as the flux steady-state values.

[0081] In a power system, after identifying the arrival time sequence of a node fault, it is necessary to determine when the fault will reach the node where the backup power supply module is located. Based on the fault propagation velocity field and node fault arrival time sequence obtained by the aforementioned method, the estimated arrival time of the fault at the node where the backup power supply module is located is extracted. The current system time is compared with the predicted time, and the time difference between the two is calculated. This time difference is the available time window for the backup power supply module to perform preparatory operations. The length of the time window directly affects the adequacy of the preparatory operations. In practical applications, if the fault is expected to arrive at the node where the backup power supply module is located in 300ms, and the current time is 10:25:45.000, then the estimated arrival time of the fault is 10:25:45.300, and the preparatory time window is 300ms.

[0082] Within the defined preparation time window, the backup power supply module needs to begin pre-establishing its internal flux linkage distribution. The backup power supply module consists of a power electronic converter and corresponding control circuitry, including multiple sets of insulated-gate bipolar transistor (IGBT) power switching devices, output filter inductors, and output transformers. The control unit sends pulse-width modulation (PWM) signals to the power switching devices, causing them to switch according to a preset modulation strategy. The modulation strategy employs synchronous PWM, with a carrier frequency of 10kHz, a modulation ratio of 0.4, and a switching dead time of 1μs, enabling the backup power supply module to output a virtual load current with preset characteristics to the virtual load. The virtual load can be composed of resistive and inductive components, with resistance and inductance values ​​set to 10Ω and 5mH respectively, matching the impedance characteristics of the actual load. The amplitude of the virtual load current is set to 30% of the rated current, its frequency is consistent with the grid fundamental frequency, and its phase is synchronously adjusted according to the output current phase of the main power supply module.

[0083] When the virtual load current flows through the output filter inductor and output transformer of the backup power supply module, it generates magnetic field energy and forms a specific flux linkage distribution within these magnetic components. Flux linkage is the product of magnetic flux and the number of turns, characterizing the state of magnetic energy stored in the magnetic component. To monitor the formation process of the internal flux linkage distribution, a high-precision Hall current sensor is used to collect the inductor current of the output filter inductor and the magnetizing current of the output transformer in real time. The sampling frequency is set to 50kHz, and the accuracy is 0.1% of full scale. The collected inductor current and magnetizing current can directly reflect the flux linkage state in the magnetic components. According to the law of electromagnetic induction, the flux linkage in an inductor is equal to the product of the inductance value and the current, and the flux linkage in a transformer is equal to the product of the magnetizing inductance and the magnetizing current. Through this correspondence, the evolution curve of the internal flux linkage distribution over time is calculated. Taking a set of output filter inductors in the backup power supply module as an example, with an inductance value of 2mH, when the measured inductor current is 15A, the corresponding flux linkage value is 30mWb.

[0084] The flux linkage evolution curve reflects the dynamic changes in the magnetic energy storage state within the backup power supply module. To determine whether the internal flux linkage distribution has reached a stable state, a steady-state determination of the evolution curve is required. A 50ms observation window is set, and the maximum change in flux linkage value within the window is calculated. A preset convergence threshold is set to 2% of the rated flux linkage value. When the change in the evolution curve within three consecutive observation windows is observed to be below the preset convergence threshold, the internal flux linkage distribution is considered to have reached a stable state. At this point, the flux linkage values ​​corresponding to the inductor current and excitation current in the stable state are extracted as the steady-state flux linkage value. In actual operation, if the rated flux linkage value of the output filter inductor is 50mWb, the convergence threshold is set to 1mWb. When the change in flux linkage value within three consecutive 50ms observation windows does not exceed 1mWb, the flux linkage distribution is confirmed to have reached a steady state, and the flux linkage value at this time is recorded as the steady-state flux linkage value.

[0085] In a real-world case study of a power distribution network, fault propagation prediction determined that a fault would reach the node containing the backup power supply module in 250ms. The control system immediately initiated a preparatory procedure, controlling the power switching devices of the backup power supply module to perform pulse width modulation (PWM) with a carrier frequency of 10kHz and a modulation ratio of 0.4, outputting a peak current of 12A to the virtual load. The virtual load current generated a magnetic field and established a flux linkage distribution in the output filter inductor and transformer. Monitoring devices recorded that the output filter inductor current stabilized at 8.5A, and the output transformer excitation current stabilized at 1.2A, corresponding to flux linkage values ​​of 17mWb and 6mWb, respectively. The flux linkage evolution curve showed that the flux linkage distribution reached a steady state 150ms after the virtual load current injection, with flux linkage changes of less than 0.8mWb in three consecutive observation windows. The system recorded the steady-state flux linkage value to prepare for a smooth subsequent switchover.

[0086] In this embodiment, by pre-establishing the internal flux distribution of the backup power supply module, the problem of flux establishment lag in traditional power switching methods is effectively solved, significantly improving the power switching speed and stability under grid fault conditions. This method utilizes the time window provided by fault prediction to complete the pre-storage of magnetic energy in the backup module before the actual arrival of the fault, avoiding current surges and voltage fluctuations during switching and protecting sensitive load equipment. Simultaneously, through virtual load technology, the backup module can be pre-magnetized without affecting normal power supply, improving system resource utilization efficiency.

[0087] In one optional implementation, the inductor current of the output filter inductor and the magnetizing current of the output transformer are collected, and the evolution curve of the internal flux linkage distribution over time is calculated based on the inductor current and the magnetizing current, including:

[0088] The inductor current of the output filter inductor and the magnetizing current of the output transformer are collected according to a fixed sampling period to obtain the inductor current sampling sequence and the magnetizing current sampling sequence.

[0089] The current change rate feature is extracted by performing time-domain differentiation on the inductor current sampling sequence. The current change rate feature is then combined with the current current sampling sequence to obtain the output filter inductor flux linkage sequence by weighted integration. The frequency domain decomposition of the excitation current sampling sequence is performed to extract the harmonic component distribution. The excitation current sampling sequence is then combined with the harmonic component distribution to obtain the output transformer flux linkage sequence by piecewise integration.

[0090] Extract the topological connection relationship and energy transfer path between the output filter inductor and the output transformer inside the backup power supply module, and construct a spatial weight matrix that reflects the magnetic flux coupling strength.

[0091] The output filter inductor flux sequence and the output transformer flux sequence are weighted and fused using a spatial weight matrix to obtain the total flux sequence, which is then mapped to a time-series coordinate system to form the evolution curve of the internal flux distribution over time.

[0092] During the pre-activation process of the backup power supply module, it is necessary to accurately calculate the evolution of the internal flux linkage distribution, which requires the acquisition and processing of relevant current data. A high-precision Hall current sensor is used to acquire the inductor current of the output filter inductor and the excitation current of the output transformer. A fixed sampling period of 20μs is set, corresponding to a sampling frequency of 50kHz, to ensure that the time resolution of the acquired signal meets the requirements for rapid switching. A 16-bit analog-to-digital converter is selected as the acquisition device, with a full-scale current set to ±50A to ensure that the accuracy of the acquired data reaches above 0.1%. During data acquisition, bandpass filtering is used to eliminate high-frequency interference and DC bias; the passband range of the filter is set to 20Hz to 20kHz. The acquired inductor current sampling sequence and excitation current sampling sequence each contain 5000 consecutive data points, covering the complete flux linkage establishment process.

[0093] The time-domain differential operation is performed on the acquired inductor current sampling sequence to extract the current rate of change feature. The time-domain differential uses the central difference method, calculating the difference between the current values ​​of two adjacent sampling points divided by the sampling period to obtain the current rate of change at each sampling moment. To eliminate the noise amplification effect introduced by the differential operation, a low-pass filter is applied to smooth the calculation results, with the filter cutoff frequency set to 5kHz. The extracted current rate of change feature reflects the instantaneous characteristics of the dynamic change process of the inductor current, including information such as the current rise rate, fall rate, and current inflection points. The inductor current sampling sequence is combined with the current rate of change feature, and a weighted integration method is used to calculate the output filter inductor flux linkage sequence. The weighting factor is dynamically adjusted according to the magnitude of the current rate of change, assigning higher weights to regions with larger rates of change to ensure higher calculation accuracy in rapidly changing current regions. The integration uses the trapezoidal rule, multiplying the weighted average of the current values ​​of adjacent sampling points by the sampling period, and considering the nonlinear characteristics of the inductance value changing with the current, to obtain the flux linkage value of the output filter inductor at each moment, forming the filter inductor flux linkage sequence.

[0094] Frequency domain decomposition is performed on the sampling sequence of the output transformer excitation current to extract the harmonic component distribution. A Fast Fourier Transform (FFT) is used to convert the time-domain current signal to the frequency domain, and the amplitude and phase of each frequency component are calculated. The fundamental frequency and major harmonic components are extracted from the spectrum, typically considering the fundamental, 3rd, 5th, and 7th harmonics, which have a significant impact on the transformer excitation process. Based on the characteristics of the harmonic component distribution, the linear and saturation regions in the excitation current are identified and divided into multiple characteristic segments. Different integration methods are used for each segment: a simple trapezoidal integral is used in the linear region, and a piecewise nonlinear corrected integral is used in the saturation region, fully considering the influence of core saturation on the flux linkage. A nonlinear correction factor for the transformer excitation curve is introduced during the integration process; this correction factor is determined based on the pre-measured transformer magnetization characteristic curve. The flux linkage value of the output transformer at each moment is calculated through piecewise integration, forming a transformer flux linkage sequence.

[0095] The topological connections and energy transfer paths of the output filter inductor and output transformer within the backup power supply module are extracted. The circuit topology is analyzed to identify the direct electrical connections and indirect magnetic field coupling between the filter inductor and transformer. Electrical connections are described by wire connection methods and node sharing, while magnetic field coupling is determined by the distance between components and their relative orientation. Based on the topology analysis results, a spatial weight matrix reflecting the strength of magnetic flux coupling is constructed. This matrix has an n×n dimension, where n is the total number of magnetic components, and the matrix elements represent the magnetic flux coupling coefficients between different magnetic components. The coupling coefficient is determined based on the physical distance between components, their relative orientation, and the magnetic shielding. Component pairs that are physically close, have consistent orientations, and are unshielded have higher coupling coefficients. In practical applications, the coupling coefficient of two closely placed, unshielded inductors can reach 0.8 to 0.9; while with magnetic shielding or when the distance is greater, the coupling coefficient can be as low as below 0.1.

[0096] The flux linkage sequences of the output filter inductors and the output transformers are weighted and fused using a spatial weight matrix. At each moment, the flux linkage value of each magnetic element is processed by matrix operations to obtain a comprehensive flux linkage value considering coupling effects. The weighted fusion process considers electromagnetic interference and energy transfer between elements, more accurately reflecting the internal magnetic energy distribution state of the backup power supply module. The result of the fusion operation forms a total flux linkage sequence. This sequence is mapped to a time-series coordinate system, with the horizontal axis representing time and the vertical axis representing the total flux linkage value, yielding an evolution curve of the internal flux linkage distribution over time. This curve visually illustrates the establishment process of the flux linkage from its initial state to a stable state, including the rising phase, oscillation phase, and stable phase. In a specific implementation, a backup power supply module contains four output filter inductors and two output transformers. The acquisition period is set to 20 μs, accumulating 5000 data points covering a 100 ms duration. The flux linkage evolution curve calculated by the above method shows that the flux linkage value starts from 0, reaches a peak of 21 mWb at 35 ms, and after a small oscillation, it tends to stabilize at 60 ms with a steady-state value of 19.5 mWb and an oscillation amplitude of less than 0.5 mWb.

[0097] In this embodiment, by acquiring and processing inductor current and excitation current with high precision, combined with topology analysis, the temporal evolution of flux distribution within the backup power supply module is accurately calculated, providing a precise basis for subsequent flux pre-establishment and rapid switching. This overcomes the limitations of traditional flux calculations that neglect nonlinear factors and coupling effects. Through techniques such as time-domain differential feature extraction, frequency-domain harmonic analysis, nonlinear integral correction, and spatial coupling matrix fusion, the nonlinear characteristics and mutual coupling effects of magnetic components are comprehensively considered, significantly improving the accuracy and practicality of flux calculation.

[0098] In one optional implementation, the voltage drop trajectory of the faulty power supply module and the voltage rise trajectory of the backup power supply module are collected, and the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory are calculated based on the magnetic flux steady-state value, including:

[0099] The voltage values ​​of the faulty power supply module and the backup power supply module are collected, and the voltage change rate is obtained by calculating the voltage difference between adjacent sampling points.

[0100] The sampling time interval is set based on the voltage change rate. The voltage sampling sequence is obtained according to the sampling time interval. Wavelet transform is performed on the voltage sampling sequence and the voltage waveform sequence is reconstructed.

[0101] The amplitude and phase range of the voltage waveform sequence are determined based on the steady-state value of the magnetic flux linkage. Within the determined amplitude and phase ranges, the voltage waveform sequence is reconstructed to generate the voltage drop trajectory function of the fault power supply module and the voltage rise trajectory function of the backup power supply module.

[0102] Calculate the magnetic flux distribution within the overlapping interval of the voltage drop trajectory function of the fault power supply module and the voltage rise trajectory function of the backup power supply module, and extract the time point where the minimum value of the magnetic flux distribution is located as the crossover time.

[0103] Substitute the crossover moment into the voltage drop trajectory function and the voltage rise trajectory function of the fault power supply module respectively to obtain the voltage value of the drop trajectory and the voltage value of the rise trajectory. Calculate the average value of the voltage value of the drop trajectory and the voltage value of the rise trajectory as the crossover voltage value.

[0104] In the event of a power grid failure, it is crucial to accurately determine the switching timing between the faulty power supply module and the backup power supply module. This necessitates the acquisition and processing of voltage trajectory information from both modules. A high-precision voltage acquisition device is used to simultaneously monitor the output voltage of both the faulty and backup power supply modules. The voltage acquisition device comprises a differential amplifier, a sample-and-hold circuit, and an analog-to-digital converter (ADC). The differential amplifier gain is set to 0.1, suitable for power supply systems with a rated voltage of 380V. The sample-and-hold circuit utilizes a rail-to-rail input / output operational amplifier, ensuring a hold time error of less than 0.01%. The ADC employs 16-bit precision, with an initial sampling rate of 10kHz. During the initial sampling phase, the acquired voltage values ​​are processed to calculate the difference between adjacent sampling points. A sliding window method is used in the calculation, with a window length of 10 sampling points and a sliding step of 1 sampling point. The maximum difference in voltage values ​​within each window is calculated and divided by the corresponding time interval to obtain the voltage change rate for that window. The window length is chosen to balance computational accuracy and real-time performance; a window that is too short is prone to noise interference, while a window that is too long reduces the responsiveness to rapid changes.

[0105] Based on the calculated voltage change rate, the sampling time interval is dynamically adjusted. An adaptive sampling strategy is adopted: when the voltage change rate is higher than a preset threshold, the sampling time interval is shortened to capture rapidly changing details; when the voltage change rate is lower than the preset threshold, the sampling time interval is appropriately extended to reduce the amount of data. The voltage change rate threshold is set to 5% of the rated voltage per millisecond, corresponding to approximately 19V / ms for a 380V system. When the detected change rate exceeds this threshold, the sampling interval is adjusted to 50μs, and the sampling rate is increased to 20kHz; when the change rate is lower than 20% of this threshold, the sampling interval is adjusted to 200μs, and the sampling rate is reduced to 5kHz. Voltage sampling sequences are obtained according to the adjusted sampling time intervals, forming two sets of sampling sequences for the faulty power supply module and the backup power supply module respectively. Median filtering is used to remove outliers, and the filtering window size is set to 5 sampling points. Wavelet transform processing is performed on the filtered voltage sampling sequences, using the db4 wavelet basis function, with a decomposition level of 4, to extract low-frequency approximation coefficients and high-frequency detail coefficients. Noise reduction of high-frequency detail coefficients is achieved through a thresholding method, using a soft thresholding approach where the threshold value is adaptively determined based on the standard deviation of the detail coefficients. The processed wavelet coefficients are then used for signal reconstruction, resulting in a smooth voltage waveform sequence free of noise and abrupt interference.

[0106] Based on the obtained steady-state flux linkage value, the amplitude and phase ranges of the voltage waveform sequence are determined. The steady-state flux linkage value reflects the state of magnetic energy stored in the power system and corresponds to the amplitude and phase of the voltage waveform. Ideally, the flux linkage is proportional to the integral of the voltage, and the phase leads the voltage by 90 degrees. Based on this relationship and combined with system parameters, the theoretical voltage amplitude range matching the steady-state flux linkage value is calculated. For a steady-state flux linkage value of 20mWb, the corresponding voltage amplitude range is 90% to 110% of the rated voltage, i.e., 342V to 418V. The phase range depends on the phase angle of the steady-state flux linkage value and is typically adjusted within ±30 degrees. Within the determined amplitude and phase ranges, the voltage waveform sequence is reconstructed in a targeted manner. A piecewise polynomial fitting method is used to smoothly interpolate the voltage values ​​between sampling points. The physical characteristics of voltage changes are considered during the fitting process to ensure that the fitted curve conforms to the basic laws of the transient process of the power system. The continuous functions obtained through fitting are used as the voltage drop trajectory function of the fault power supply module and the voltage rise trajectory function of the standby power supply module, respectively.

[0107] The overlapping interval in time between the voltage drop trajectory function of the fault power supply module and the voltage rise trajectory function of the backup power supply module is calculated. The overlapping interval is determined by judging the intersection of the domains of the two trajectory functions; in power switching scenarios, this interval is typically tens to hundreds of milliseconds. Within the determined overlapping interval, sampling is performed with a high time resolution and a time step of 10 μs. The voltage values ​​corresponding to the two trajectory functions at each sampling point are calculated. Based on the voltage values ​​and time information, the instantaneous magnetic flux distribution within the overlapping interval is calculated. The magnetic flux calculation considers the voltage-time integral relationship and uses the trapezoidal integration method, with the integration step size consistent with the sampling time step. The magnetic flux distribution curve reflects the change in the system's magnetic energy state during the switching process. The trough of the curve corresponds to the moment of minimum magnetic energy fluctuation, which is the optimal switching point. By searching for the minimum point of the magnetic flux distribution curve, the corresponding time coordinate is extracted as the crossover moment. If the minimum value of the magnetic flux distribution occurs 35 ms after the start of the overlapping interval, this moment is determined as the crossover moment of the power switching.

[0108] The determined crossover time is substituted into the voltage drop trajectory function of the faulty power supply module and the voltage rise trajectory function of the backup power supply module, respectively, to calculate the drop trajectory voltage value and the rise trajectory voltage value. Ideally, these two voltage values ​​should be equal, indicating seamless switching; however, in actual systems, due to factors such as measurement errors and model approximations, the two values ​​usually have slight differences. The arithmetic mean of the drop trajectory voltage value and the rise trajectory voltage value is calculated as the final crossover voltage value. This value represents the system voltage level at the moment of switching and is an important indicator for evaluating the smoothness of switching. In a certain power grid fault scenario, the crossover time is determined to be 87ms after the fault occurs. At this time, the drop trajectory voltage value of the faulty power supply module is 342V, and the rise trajectory voltage value of the backup power supply module is 338V. The calculated crossover voltage value is 340V, which is approximately 89.5% of the rated voltage, meeting the normal operating requirements of the load equipment.

[0109] In this embodiment, by collecting and analyzing the voltage trajectories of the faulty power supply module and the backup power supply module, and combining this with the steady-state value of the system flux linkage, the optimal time and voltage value for power switching were accurately calculated, providing key technical support for achieving smooth and disturbance-free power switching. Taking into full account the electromagnetic energy transfer laws in the power system, high-precision extraction and analysis of the voltage trajectory were achieved through techniques such as adaptive sampling, wavelet transform noise reduction, and parameter range constraints. The switching time determined by the principle of minimum flux distribution can effectively reduce electromagnetic transient impacts during the switching process and lower the risk of system oscillation.

[0110] In one optional implementation, identifying the fault-free power supply module based on the node fault arrival time sequence, calculating the transmission impedance from the fault-free power supply module to the target load node, and establishing a virtual electrical coupling channel includes:

[0111] Extract the predicted arrival time of each node fault from the node fault arrival time sequence, construct the node state matrix, and calculate the fault propagation time distribution from the node state matrix;

[0112] Based on the fault propagation time distribution and predicted time, power supply modules that have not reached the fault time are selected as fault-not-reached power supply modules, and the spatial location of the fault-not-reached power supply modules is determined.

[0113] The electrical parameters of the line between the power supply module that the fault has not reached and the target load node are collected. The transmission impedance is calculated based on the electrical parameters. The transmission impedance distribution is constructed using the spatial location and fault propagation time distribution. The line combination with the smallest impedance is selected as the transmission channel according to the magnitude of the transmission impedance.

[0114] A detection pulse signal is applied to the transmission channel, and the voltage and current responses generated by the detection pulse signal through the transmission channel are recorded. The amplitude ratio and phase difference are calculated, and the amplitude ratio and phase difference are used as control parameters of the compensation circuit. The compensation circuit is configured to be connected to the transmission channel to form a virtual electrical coupling channel.

[0115] When a power grid fault occurs, the propagation of the fault signal in the network exhibits temporal characteristics. By extracting the predicted arrival time of the fault at each node from the fault arrival time sequence, power supply modules that have not yet arrived due to the fault can be identified. Based on fault detection devices pre-deployed at key nodes, the arrival time data of the fault signal at each monitoring point is acquired. The fault detection devices employ high-speed sampling voltage and current sensors with a sampling frequency of 10MHz to ensure accurate capture of the details of the fault waveform. Feature extraction is performed on the acquired voltage and current waveforms to identify the arrival time point of the fault wavefront, with an accuracy of up to 100ns.

[0116] A node state matrix is ​​constructed based on the fault arrival time of each node. Each row of this matrix represents a monitoring node, and each column represents a time point. A value of 1 indicates that the fault has reached the node at that time point, while a value of 0 indicates that the fault has not reached the node. In smart distribution networks, 12 to 24 monitoring nodes are typically set up, forming a node state matrix of 12×100 to 24×100, covering a 10ms time window after the fault occurs. Based on this matrix, a spatiotemporal analysis method is used to calculate the fault propagation time distribution. The fault signal propagation time difference between adjacent nodes is calculated, and combined with the physical distance between nodes, the fault propagation speed is obtained. The fault propagation speed is approximately 80% to 90% of the speed of electromagnetic waves in conductors, approximately 2.4 × 10⁻⁶. 8 Up to 2.7×10 8 Meters per second. Based on the spatial location of each node and the measured propagation time, the fault source is located using the triangulation principle, and the propagation time of the fault to each unaffected node is predicted.

[0117] Based on the fault propagation time distribution and predicted time, power supply modules that have not yet reached their fault arrival time are selected as fault-not-arrived power supply modules. A safety time threshold of 1ms is set, and power supply modules whose predicted fault arrival time exceeds the current time plus the safety threshold are marked as fault-not-arrived power supply modules. The setting of the safety time threshold takes into account prediction error and operation response time, ensuring sufficient time for switching operations before the actual arrival of the fault. The spatial location of the fault-not-arrived power supply modules is determined, and their three-dimensional coordinate values ​​are recorded. The origin of the coordinate system is set at the location of the main substation of the distribution network, and the coordinate values ​​are obtained by measuring the actual physical distance of the module from the origin, accurate to the centimeter level. By combining the spatial location information with the fault propagation path analysis, the expected time for the fault wavefront to reach each module can be calculated, forming a complete fault spatiotemporal distribution map.

[0118] In a certain power distribution network scenario, a short-circuit fault is detected in area A. The fault signal is 2.5 × 10 8 The fault propagates outward at a speed of meters per second. According to calculations, the fault is predicted to reach power supply module 3 in area B after 2.4ms and power supply module 5 in area C after 3.6ms. These two modules are determined to be power supply modules where the fault has not yet reached, located at coordinates (120, 350, 0) and (180, 420, 0) respectively.

[0119] Electrical parameters of the lines between the power supply module and the target load node before the fault reaches are collected, including line resistance, inductance, capacitance, and mutual inductance values. The data acquisition method uses a network analyzer for frequency sweep testing, with a test frequency range from 50Hz to 10kHz, to obtain impedance characteristics at different frequencies. For long-distance lines, a segmented measurement method is used, dividing the long line into multiple shorter segments for measurement and then calculating the combined values. Transmission impedance is calculated based on the collected electrical parameters. The transmission impedance considers the frequency characteristics of the line, using frequency domain analysis to calculate the impedance value at the power frequency and its harmonic frequencies. The impedance calculation uses a π-type equivalent circuit model of the line, comprehensively considering the effects of series impedance and parallel admittance. A three-dimensional impedance space is constructed using spatial location and fault propagation time distribution, with the horizontal and vertical axes representing geographical locations and the vertical axis representing impedance magnitude. Lines with the lowest impedance are selected as transmission channels based on their transmission impedance magnitude. Impedance ranking considers both impedance amplitude and phase angle, prioritizing lines with small impedance amplitudes and phase angles close to zero. When multiple paths have similar impedances, the path with the shorter physical distance and later fault arrival time is prioritized. In the example above, the transmission impedance from power supply module 3 to the target load is calculated to be 0.8 + j0.6 ohms, and the transmission impedance from power supply module 5 to the target load is 1.2 + j0.9 ohms. Therefore, the line between power supply module 3 and the target load is selected as the transmission channel.

[0120] A detection pulse signal is applied to the selected transmission channel to detect the dynamic response characteristics of the line. The pulse signal is in square wave form, with an amplitude of 10% of the rated voltage, a pulse width of 100μs, and a repetition frequency of 1kHz. A high-precision oscilloscope is used to simultaneously record the voltage and current waveforms of the pulse signal at both ends of the transmission channel, with a sampling rate of 100MHz and a recording duration of 10ms, covering a complete 10 pulse cycles. Waveform analysis is used to calculate the amplitude ratio and phase difference between the input and output signals. The amplitude ratio is obtained by measuring the ratio of the output voltage to the input voltage, and the phase difference is calculated by measuring the time difference between the zero-crossing points of the output and input voltage waveforms.

[0121] In transmission lines, typical amplitude attenuation ranges from 5% to 20%, and phase lag ranges from 10° to 30°. The calculated amplitude ratio and phase difference are used as control parameters for the compensation circuit. The compensation circuit is composed of an adjustable gain amplifier and a phase adjustment circuit, with a gain adjustment range of 0.5 to 2 times and a phase adjustment range of -45° to +45°. The gain value is set based on the reciprocal of the measured amplitude ratio, and the phase adjustment angle is set based on the negative value of the phase difference, achieving amplitude and phase compensation of the signal. The compensation circuit is connected to the input of the transmission channel, forming a virtual electrical coupling channel with active compensation function. The signal transmitted through this channel, after compensation, can restore the original signal amplitude and phase characteristics at the load end, reducing transmission loss and distortion. In practical applications, the amplitude ratio measured on the transmission channel from power supply module 3 to the target load is 0.92, and the phase difference is -15°. Based on this, the gain of the compensation circuit is set to 1.087, and the phase compensation is set to +15°, achieving accurate signal compensation.

[0122] In this embodiment, by analyzing the arrival time sequence of node faults, the safe power supply module during the propagation of power grid faults is accurately identified. Combined with impedance characteristic assessment and signal compensation techniques, a virtual electrical coupling channel with low loss and low distortion characteristics is established, providing a reliable path for power supply switching in the event of a power grid fault. This technology overcomes the limitations of traditional fixed physical line connections, realizing intelligent path selection and optimization based on the dynamic characteristics of fault propagation. Through an active signal compensation mechanism, signal attenuation and phase distortion problems in long-distance transmission are effectively overcome, ensuring the stability of power quality during switching. The establishment of the virtual electrical coupling channel provides flexible and varied path selection for power transmission in fault situations, significantly improving the adaptability and recovery speed of the power supply system in the face of complex faults, and providing technical assurance for the continuous and stable power supply to critical loads.

[0123] In one optional implementation, energy is transferred through a virtual electrical coupling channel, the backup power supply module outputs voltage at the crossover moment, the phase difference rate of change between the backup power supply module output voltage and the actual voltage of the target load node is calculated, and power supply switching is performed when the phase difference rate of change is zero and the output voltage reaches the crossover voltage value, including:

[0124] The voltage and current values ​​on the virtual electrical coupling channel are collected, and the voltage and current values ​​are discretely sampled according to a preset sampling frequency. The instantaneous energy transfer power sequence is calculated from the discrete sampled values.

[0125] Identify power fluctuation patterns in the power transfer power sequence, construct a predictive control sequence based on the power fluctuation patterns, map the predictive control sequence to a backup power supply module drive signal, and control the backup power supply module to output voltage at the crossover moment through the drive signal;

[0126] The output voltage of the backup power supply module and the actual voltage of the target load node are collected. The inherent components of the voltage signal are extracted using empirical mode decomposition. The phase information is separated from the inherent components to construct the phase difference distribution. The phase difference distribution is adaptively corrected by combining the power fluctuation law to obtain the phase difference change rate.

[0127] Power supply switching is performed when the phase difference change rate is zero and the output voltage reaches the cross voltage value.

[0128] After the virtual electrical coupling channel is established, energy needs to be transferred through this channel to control the backup power supply module to switch at the optimal time. Therefore, voltage and current values ​​on the virtual electrical coupling channel are collected. High-precision voltage transformers and current transformers are used to collect the voltage and current signals on the channel, respectively. The voltage transformer ratio is 380:1 with an accuracy class of 0.2; the current transformer ratio is 100:1 with an accuracy class of 0.2S. The collected analog signals are processed by a signal conditioning circuit, including voltage tracking, filtering, and amplification. The signal conditioning circuit uses a fourth-order Butterworth low-pass filter with a cutoff frequency of 5kHz and an attenuation characteristic of -24dB / octave, ensuring effective suppression of high-frequency interference while retaining the effective components of the signal. The conditioned signal is sent to a high-speed data acquisition card for analog-to-digital conversion, and the voltage and current values ​​are discretely sampled according to a preset sampling frequency. The sampling frequency is set to 50kHz to meet the Nyquist sampling theorem requirements, accurately capturing the power frequency and its harmonic components. Each cycle collects 1000 data points, achieving a time resolution of 20μs, meeting the accuracy requirements of rapid switching control. The instantaneous energy transfer power sequence is calculated from the discrete sampled values ​​using the product of voltage and current values. Considering the characteristics of a three-phase system, the power of each phase is calculated separately and then summed to obtain the total power. Linear interpolation is used to compensate for potential data gaps during sampling, ensuring the continuity of the power sequence. The accuracy of power calculation affects the accuracy of energy transfer assessment; high-precision multipliers and accumulators are used to control the calculation error within 0.5%.

[0129] The calculated power transfer sequence is processed to identify power fluctuation patterns. Power sequence analysis employs a joint time-frequency analysis method, including Fast Fourier Transform (FFT) and wavelet analysis. The FFT window length is selected as 10 power frequency cycles (200 ms), and the Hamming window type is chosen to effectively suppress spectral leakage. Wavelet analysis uses the db4 wavelet basis with a 5-level decomposition layer, enabling the extraction of power fluctuation characteristics at different time scales. Spectral analysis identifies periodic components in the power sequence, typically including the power frequency fundamental wave, second harmonic components, and low-frequency fluctuation components. In practical systems, power frequency fluctuations are mainly caused by load fluctuations, with characteristic frequencies usually below 10 Hz; second harmonic components are mainly caused by unbalanced and nonlinear loads, with a frequency of 100 Hz; low-frequency fluctuations are related to changes in the grid's operating state, with a frequency range of 0.1 Hz to 1 Hz. Based on the identified power fluctuation patterns, a predictive control sequence is constructed. The predictive control uses an autoregressive moving average model, with the autoregressive term at order 4 and the moving average term at order 2, predicting the power change trend within the next 100 ms based on historical power data. The model parameters are estimated online using the recursive least squares method to adapt to dynamic changes in the system.

[0130] The predictive control sequence is mapped to a backup power supply module drive signal. The mapping process considers the dynamic response characteristics of the backup module, including response delay and ramp-up rate limitations. The drive signal uses pulse width modulation (PWM) with a carrier frequency of 20kHz and a modulation depth range of 0% to 95%, ensuring the output voltage accurately tracks the target value. The drive signal controls the backup power supply module's output voltage at the crossover moment, achieving a control accuracy of ±1% of the rated voltage. In a practical application scenario, power sequence analysis identified a power fluctuation main period of 20ms and a fluctuation amplitude of ±5% of the rated power. Based on this, the constructed predictive control sequence accurately controls the backup module's output voltage, enabling it to reach the target voltage value of 340V at the predetermined crossover moment of 85ms.

[0131] The output voltage of the backup power supply module and the actual voltage of the target load node are collected, and their phase difference is compared. The acquisition device simultaneously samples two voltage signals at a sampling rate of 100kHz to ensure a phase resolution better than 0.2 degrees. The intrinsic components of the voltage signal are extracted using Empirical Mode Decomposition (EMD). EMD is an adaptive signal processing method that can decompose complex signals into a finite number of intrinsic mode functions (EMFs). The decomposition process is set to 10 rounds of filtering, with a residual threshold of 0.1% of the signal's root mean square value, typically yielding 5 to 8 EMFs. Phase information is separated from the intrinsic components using a Hilbert transform-based extraction method to calculate the instantaneous phase of each EMF. The accuracy of phase extraction directly affects the accuracy of the switching timing; by optimizing the endpoint processing method of the Hilbert transform, the phase extraction error is controlled within ±0.5 degrees. A phase difference distribution is constructed, and the phase difference between the corresponding EMFs of the backup power supply module's output voltage and the actual voltage of the load node is calculated. A weighted average is then performed according to energy weights to obtain the comprehensive phase difference. The phase difference distribution is adaptively corrected by combining the power fluctuation law. The correction method takes into account the influence of power change on phase, and the correction coefficient is proportional to the power change rate.

[0132] The phase difference change rate was calculated from the time series of the phase difference by dividing the difference in phase difference between adjacent sampling points by the sampling interval. To eliminate the influence of noise, a moving average was applied to the calculation results, with an average window length of 10 sampling points. This processing yielded an accurate phase difference change rate curve, reflecting the process of phase synchronization between the backup power supply module and the load node voltage. In actual testing, the initial phase difference was 25 degrees; as the control process progressed, the phase difference gradually decreased, and the change rate gradually approached zero from the initial -1.2 degrees / ms.

[0133] Power supply switching is executed when the phase difference change rate is detected to be zero and the output voltage reaches the crossover voltage value. The condition for zero phase difference change rate is set at an absolute value of the change rate for five consecutive sampling points all being less than 0.05 degrees / ms, avoiding misjudgments due to instantaneous fluctuations. The condition for the output voltage reaching the crossover voltage value is set at a voltage deviation not exceeding ±2%, i.e., for a 340V crossover voltage, the allowable range is 333.2V to 346.8V. Switching operation is triggered when both conditions are met simultaneously. The switching process uses a bidirectional controllable semiconductor switch with a switching action time of less than 100μs. To prevent instantaneous overvoltages generated during the switching process, a metal oxide surge arrester is installed in parallel, with a protection voltage of 1.5 times the rated voltage.

[0134] Before the switchover, a final safety check is performed, including verifying the matching degree between the backup module's output frequency and the load frequency, with a frequency difference threshold set at ±0.1Hz. After the safety check passes, the controller issues a switchover command, first closing the backup power supply path and then disconnecting the faulty power supply path; the entire switchover process is completed within 1ms. To monitor the switchover quality, voltage, current, and power waveforms are recorded during the switchover process, with the sampling rate increased to 500kHz and the recording duration 50ms before and after the switchover, for a total of 100ms. In a power grid fault scenario, the system detected that the phase difference change rate decreased to 0.03 degrees / ms and remained stable, while the backup module's output voltage reached 338.5V, meeting the switchover conditions and successfully achieving a seamless switchover. During the switchover process, the load voltage fluctuation did not exceed 3% of the rated value.

[0135] In this embodiment, a smooth power supply switchover under grid fault conditions is achieved by transferring energy through a virtual electrical coupling channel and precisely controlling the switching time. This method overcomes the impact problems caused by phase asynchrony and voltage mismatch in traditional switching technologies. Through real-time power analysis and phase tracking technology, it ensures that the switching operation is performed at the optimal time. The method of extracting the inherent components of the voltage signal using empirical mode decomposition effectively solves the phase extraction problem under conditions containing harmonics and noise interference, improving the accuracy of phase synchronization judgment. The predictive control strategy based on power fluctuation law realizes precise control of the output characteristics of the backup power supply module, ensuring a smooth voltage transition at the switching point. The design of using zero phase difference change rate as the switching trigger condition ensures energy balance during the switching process, effectively avoiding power surges and voltage fluctuations at the moment of switching, providing high-quality uninterrupted power supply to sensitive loads, and significantly improving the power supply reliability and continuity of the power system in the face of faults.

[0136] A second aspect of the present invention provides a modular intelligent power supply fast switching system for power grid faults, the system comprising:

[0137] The first unit is used to collect parameter changes of power grid nodes, establish a fault propagation velocity field based on the topological relationship between nodes based on the parameter changes, and predict the time when the fault arrives at each node based on the fault propagation velocity field to obtain the node fault arrival time sequence.

[0138] The second unit is used to control the backup power supply module to output virtual load current before the fault arrives, based on the node fault arrival time sequence. The backup power supply module is then driven by the virtual load current to establish an internal flux distribution and obtain the flux steady-state value.

[0139] The third unit is used to collect the voltage drop trajectory of the fault power supply module and the voltage rise trajectory of the backup power supply module, and calculate the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory based on the magnetic flux steady-state value.

[0140] The fourth unit is used to identify the power supply module that has not been reached due to the fault based on the arrival time sequence of the node fault, calculate the transmission impedance from the power supply module that has not been reached due to the fault to the target load node, and establish a virtual electrical coupling channel.

[0141] The fifth unit is used to transfer energy through a virtual electrical coupling channel, control the output voltage of the backup power supply module at the crossover moment, calculate the phase difference rate of change between the output voltage of the backup power supply module and the actual voltage of the target load node, and perform power supply switching when the phase difference rate of change is zero and the output voltage reaches the crossover voltage value.

[0142] A third aspect of the present invention provides an electronic device, comprising:

[0143] processor;

[0144] Memory used to store processor-executable instructions;

[0145] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0146] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0147] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A modular intelligent power supply fast switching method for grid faults, characterized in that, include: The parameter changes of power grid nodes are collected, and a fault propagation velocity field is established based on the topological relationship between nodes based on the parameter changes. The arrival time of faults at each node is predicted based on the fault propagation velocity field to obtain the node fault arrival time sequence. Based on the arrival sequence of node faults, the backup power supply module is controlled to output virtual load current before the fault arrives. The backup power supply module is then driven by the virtual load current to establish an internal flux distribution and obtain the steady-state value of the flux. Collect the voltage drop trajectory of the faulty power supply module and the voltage rise trajectory of the backup power supply module, and calculate the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory based on the magnetic flux steady-state value; Identify the power supply modules that have not been reached by the fault based on the arrival time sequence of the node fault, calculate the transmission impedance from the power supply modules that have not been reached by the fault to the target load node, and establish a virtual electrical coupling channel; Energy is transferred through a virtual electrical coupling channel. The backup power supply module outputs voltage at the crossover moment. The phase difference rate between the backup power supply module output voltage and the actual voltage of the target load node is calculated. When the phase difference rate is zero and the output voltage reaches the crossover voltage value, power supply switching is performed.

2. The method according to claim 1, characterized in that, The parameters of power grid nodes are collected, and a fault propagation velocity field is established based on the topological relationships between nodes. The arrival time of faults at each node is predicted based on the fault propagation velocity field, resulting in the following node fault arrival time sequence: Collect the operating parameters of the power grid nodes, calculate the difference of the operating parameters in the time dimension to obtain the parameter change, calculate the rate of change of the parameter change, and identify abnormal nodes whose rate of change exceeds the normal fluctuation threshold. Obtain the electrical distance and line impedance between the abnormal node and the other nodes in the power grid, and construct the topology matrix between nodes; The propagation rate of fault disturbance between nodes is calculated based on the rate of change of parameter changes and the topological relationship matrix. The propagation rate is then mapped to the spatial location of the nodes to form a fault propagation velocity field. The propagation time required for a fault to spread from an abnormal node to the remaining nodes is calculated by integrating the fault propagation velocity field along each transmission path. The predicted arrival time of the fault at each node is obtained by adding the detection time of the abnormal node to the propagation time. The nodes are then arranged in the order of the predicted times to form the node fault arrival time sequence.

3. The method according to claim 1, characterized in that, Based on the node fault arrival sequence, the backup power supply module is controlled to output a virtual load current before the fault arrives. The virtual load current drives the backup power supply module to establish an internal flux linkage distribution, and the steady-state value of the flux linkage is obtained, including: The predicted time of arrival of the fault at the node where the backup power supply module is located is determined based on the fault arrival sequence, and the time interval between the current time and the predicted time is calculated as the preparation time window. Within the preparation time window, the power switching devices of the backup power supply module are controlled to operate according to the preset modulation strategy, so that the backup power supply module outputs virtual load current to the virtual load. The virtual load current generates magnetic field energy in the output filter inductor and the output transformer and forms an internal magnetic flux distribution. The inductance current of the output filter inductor and the excitation current of the output transformer are collected, and the evolution curve of the internal magnetic flux distribution over time is calculated based on the inductance current and the excitation current. When the change amplitude of the evolution curve within a continuous time period is lower than the preset convergence threshold, it is determined that the internal flux distribution has reached a stable state, and the flux values ​​corresponding to the inductor current and excitation current in the stable state are extracted as the flux steady-state values.

4. The method according to claim 3, characterized in that, The inductor current of the output filter inductor and the magnetizing current of the output transformer are collected. Based on the inductor current and the magnetizing current, the evolution curve of the internal flux linkage distribution over time is calculated, including: The inductor current of the output filter inductor and the magnetizing current of the output transformer are collected according to a fixed sampling period to obtain the inductor current sampling sequence and the magnetizing current sampling sequence. The current change rate feature is extracted by performing time-domain differentiation on the inductor current sampling sequence. The current change rate feature is then combined with the current current sampling sequence to obtain the output filter inductor flux linkage sequence by weighted integration. The frequency domain decomposition of the excitation current sampling sequence is performed to extract the harmonic component distribution. The excitation current sampling sequence is then combined with the harmonic component distribution to obtain the output transformer flux linkage sequence by piecewise integration. Extract the topological connection relationship and energy transfer path between the output filter inductor and the output transformer inside the backup power supply module, and construct a spatial weight matrix that reflects the magnetic flux coupling strength. The output filter inductor flux sequence and the output transformer flux sequence are weighted and fused using a spatial weight matrix to obtain the total flux sequence, which is then mapped to a time-series coordinate system to form the evolution curve of the internal flux distribution over time.

5. The method according to claim 1, characterized in that, The voltage drop trajectory of the faulty power supply module and the voltage rise trajectory of the backup power supply module are collected. Based on the steady-state value of the magnetic flux linkage, the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory are calculated, including: The voltage values ​​of the faulty power supply module and the backup power supply module are collected, and the voltage change rate is obtained by calculating the voltage difference between adjacent sampling points. The sampling time interval is set based on the voltage change rate. The voltage sampling sequence is obtained according to the sampling time interval. Wavelet transform is performed on the voltage sampling sequence and the voltage waveform sequence is reconstructed. The amplitude and phase range of the voltage waveform sequence are determined based on the steady-state value of the magnetic flux linkage. Within the determined amplitude and phase ranges, the voltage waveform sequence is reconstructed to generate the voltage drop trajectory function of the fault power supply module and the voltage rise trajectory function of the backup power supply module. Calculate the magnetic flux distribution within the overlapping interval of the voltage drop trajectory function of the fault power supply module and the voltage rise trajectory function of the backup power supply module, and extract the time point where the minimum value of the magnetic flux distribution is located as the crossover time. Substitute the crossover moment into the voltage drop trajectory function and the voltage rise trajectory function of the fault power supply module respectively to obtain the voltage value of the drop trajectory and the voltage value of the rise trajectory. Calculate the average value of the voltage value of the drop trajectory and the voltage value of the rise trajectory as the crossover voltage value.

6. The method according to claim 1, characterized in that, Identify power supply modules that have not yet reached the fault based on the node fault arrival time sequence, calculate the transmission impedance from the power supply module that has not yet reached the fault to the target load node, and establish a virtual electrical coupling channel, including: Extract the predicted arrival time of each node fault from the node fault arrival time sequence, construct the node state matrix, and calculate the fault propagation time distribution from the node state matrix; Based on the fault propagation time distribution and predicted time, power supply modules that have not reached the fault time are selected as fault-not-reached power supply modules, and the spatial location of the fault-not-reached power supply modules is determined. The electrical parameters of the line between the power supply module that the fault has not reached and the target load node are collected. The transmission impedance is calculated based on the electrical parameters. The transmission impedance distribution is constructed using the spatial location and fault propagation time distribution. The line combination with the smallest impedance is selected as the transmission channel according to the magnitude of the transmission impedance. A detection pulse signal is applied to the transmission channel, and the voltage and current responses generated by the detection pulse signal through the transmission channel are recorded. The amplitude ratio and phase difference are calculated, and the amplitude ratio and phase difference are used as control parameters of the compensation circuit. The compensation circuit is configured to be connected to the transmission channel to form a virtual electrical coupling channel.

7. The method according to claim 1, characterized in that, Energy is transferred through a virtual electrical coupling channel. The backup power supply module outputs voltage at the crossover moment. The phase difference rate of change between the backup power supply module output voltage and the actual voltage of the target load node is calculated. When the phase difference rate of change is zero and the output voltage reaches the crossover voltage value, power supply switching is performed, including: The voltage and current values ​​on the virtual electrical coupling channel are collected, and the voltage and current values ​​are discretely sampled according to a preset sampling frequency. The instantaneous energy transfer power sequence is calculated from the discrete sampled values. Identify power fluctuation patterns in the power transfer power sequence, construct a predictive control sequence based on the power fluctuation patterns, map the predictive control sequence to a backup power supply module drive signal, and control the backup power supply module to output voltage at the crossover moment through the drive signal; The output voltage of the backup power supply module and the actual voltage of the target load node are collected. The inherent components of the voltage signal are extracted using empirical mode decomposition. The phase information is separated from the inherent components to construct the phase difference distribution. The phase difference distribution is adaptively corrected by combining the power fluctuation law to obtain the phase difference change rate. Power supply switching is performed when the phase difference change rate is zero and the output voltage reaches the cross voltage value.

8. A modular intelligent power supply fast switching system for grid faults, used to implement the method of any one of claims 1-7, characterized in that, include: The first unit is used to collect parameter changes of power grid nodes, establish a fault propagation velocity field based on the topological relationship between nodes based on the parameter changes, and predict the time when the fault arrives at each node based on the fault propagation velocity field to obtain the node fault arrival time sequence. The second unit is used to control the backup power supply module to output virtual load current before the fault arrives, based on the node fault arrival time sequence. The backup power supply module is then driven by the virtual load current to establish an internal flux distribution and obtain the flux steady-state value. The third unit is used to collect the voltage drop trajectory of the fault power supply module and the voltage rise trajectory of the backup power supply module, and calculate the crossover time and crossover voltage value of the voltage drop trajectory and voltage rise trajectory based on the magnetic flux steady-state value. The fourth unit is used to identify the power supply module that has not been reached due to the fault based on the arrival time sequence of the node fault, calculate the transmission impedance from the power supply module that has not been reached due to the fault to the target load node, and establish a virtual electrical coupling channel. The fifth unit is used to transfer energy through a virtual electrical coupling channel, control the output voltage of the backup power supply module at the crossover moment, calculate the phase difference rate of change between the output voltage of the backup power supply module and the actual voltage of the target load node, and perform power supply switching when the phase difference rate of change is zero and the output voltage reaches the crossover voltage value.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images