Voltage stability control method and system for power distribution network with high proportion of distributed power supply and medium

Abstract

The application discloses a high-proportion distributed power distribution network voltage stability control method and system and a medium, relates to the technical field of power grid voltage control, and comprises the following steps: constructing state information flow of a plurality of nodes of a distribution network; performing energy feedback mapping in the information time crystal structure, regarding node voltage disturbance as a crystal defect, generating a stable time lattice based on periodic reconstruction of the crystal defect; after identifying a local unstable interval, performing periodic modulation and phase-locked optimization of inverter control signals, and outputting a voltage stabilization control strategy. The application solves the technical problems that, in the prior art, in a high-proportion distributed power access scene, the voltage of a plurality of nodes of a distribution network is easily disturbed, and the accuracy of an inverter control strategy is insufficient, resulting in poor voltage stability control effect, and achieves the technical effects of realizing high-proportion distributed power distribution network voltage stability control and improving the voltage stability and control strategy accuracy of the distribution network.

Description

Technical Field

[0001] This invention relates to the field of power grid voltage control technology, specifically to a method, system, and medium for voltage stability control of high-proportion distributed power generation distribution networks. Background Technology

[0002] With the large-scale integration of distributed power sources into distribution networks, traditional centralized voltage control methods face significant challenges. On the one hand, the intermittent and fluctuating output of distributed power sources easily leads to random voltage disturbances at multiple nodes in the distribution network, making it difficult for traditional control methods to capture the dynamic changes in voltage, phase angle, and power at each node in real time. On the other hand, the increasing number of nodes and the complex topology of distribution networks mean that existing technologies lack sufficient accuracy in identifying local voltage instability intervals. The modulation of inverter control signals relies heavily on empirical parameters, lacking coordinated optimization of timing characteristics and global resonance states. This easily leads to problems such as lagging control strategies or low accuracy, which in turn causes voltage fluctuations in the distribution network to exceed the allowable range, affecting power supply reliability and the safe operation of electrical equipment.

[0003] Existing technologies suffer from technical problems such as the susceptibility of voltage fluctuations at multiple nodes in the distribution network and insufficient accuracy of inverter control strategies in scenarios with a high proportion of distributed power sources, resulting in poor voltage stability control. Summary of the Invention

[0004] This application provides a method, system, and medium for voltage stability control of distribution networks with a high proportion of distributed power sources. It addresses the technical problems in existing technologies where the voltage of multiple nodes in the distribution network is easily disturbed and the accuracy of inverter control strategies is insufficient, resulting in poor voltage stability control performance in scenarios with a high proportion of distributed power sources.

[0005] In view of the above problems, this application provides a method, system and medium for voltage stability control of high-proportion distributed power generation distribution networks.

[0006] A first aspect of this application provides a method for voltage stability control of a high-proportion distributed generation power distribution network, the method comprising:

[0007] A state information flow for multiple nodes in the distribution network is constructed, including node voltage, phase angle, power attention value, and inverter control command signal. An information time crystal structure is constructed using a time evolution operator. Energy feedback mapping is performed in the information time crystal structure, treating node voltage disturbances as crystal defects. A stable time lattice is generated based on the periodic reconstruction of the crystal defects. After identifying local unstable regions using the amplitude and frequency shift of the stable time lattice, periodic modulation and phase-locked optimization of the inverter control signal are performed to output a voltage regulation control strategy.

[0008] A second aspect of this application provides a voltage stability control system for a high-proportion distributed power generation distribution network, the system comprising:

[0009] The state information flow construction module is used to construct the state information flow of multiple nodes in the distribution network. The state information flow includes node voltage, phase angle, power attention value, and inverter control command signal. It uses a time evolution operator to construct an information time crystal structure. The stable time lattice generation module is used to perform energy feedback mapping in the information time crystal structure, treat node voltage disturbances as crystal defects, and generate a stable time lattice based on the periodic reconstruction of crystal defects. The control strategy output module is used to identify local unstable regions by utilizing the amplitude and frequency shift of the stable time lattice, and then perform periodic modulation and phase-locked optimization of the inverter control signal to output a voltage regulation control strategy.

[0010] In a third aspect of this application, a computer-readable storage medium is provided storing a computer program for executing the high-proportion distributed power generation distribution network voltage stability control method provided in this application.

[0011] One or more technical solutions provided in this application have at least the following technical effects or advantages: constructing a state information flow of multiple nodes in the distribution network; performing energy feedback mapping in the information time crystal structure, treating node voltage disturbances as crystal defects, and generating a stable time lattice based on the periodic reconstruction of crystal defects; using the amplitude and frequency shift of the stable time lattice, identifying local unstable regions, and then performing periodic modulation and phase-locked optimization of the inverter control signal to output a voltage regulation control strategy. This achieves the technical effect of realizing stable voltage control of a high-proportion distributed power distribution network, improving the voltage stability and control strategy accuracy of the distribution network. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A schematic flowchart illustrating the voltage stability control method for a high-proportion distributed power generation distribution network provided in this application embodiment.

[0014] Figure 2 A schematic diagram of the structure of a high-proportion distributed power generation distribution network voltage stability control system provided in an embodiment of this application.

[0015] Figure labeling: State information flow construction module 10, stable time lattice generation module 20, control strategy output module 30. Detailed Implementation

[0016] This application provides a method, system, and medium for voltage stability control of distribution networks with a high proportion of distributed power sources. It addresses the technical problems in existing technologies where the voltage of multiple nodes in the distribution network is easily disturbed and the accuracy of inverter control strategies is insufficient, resulting in poor voltage stability control performance in scenarios with a high proportion of distributed power sources.

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0018] Example 1, as Figure 1 As shown, this application provides a voltage stability control method for a high-proportion distributed generation power distribution network, the method comprising:

[0019] Step S100: Construct a state information flow for multiple nodes in the distribution network. The state information flow includes node voltage, phase angle, power consumption, and inverter control command signal. Construct an information time crystal structure using a time evolution operator.

[0020] Specifically, real-time data collection of multi-node operation in a high-proportion distributed power distribution network is first performed to clarify the core components of the state information flow, encompassing real-time voltage values, voltage phase angles, power parameters, and inverter output control command signals for each node. Subsequently, the collected state information flow is preprocessed to remove DC offset interference and perform normalization. A ground-state time vector field is constructed by calculating the offset trend residual. Then, multi-scale low-pass filtering and adaptive denoising algorithms are used to suppress high-frequency disturbances, and the temporal coherence features between node voltage and power parameters are extracted. These temporal coherence features are used as evolutionary drivers and input into a preset time evolution operator. The formula for the time evolution operator is as follows: ,in This is the state vector after time evolution. Characterization time The node summary status below, For time-coherent feature vectors, To drive the mapping matrix, For the length of time memory, For time-delay index, The temporal convolution kernel matrix, This is the neighborhood coupling weight coefficient matrix. Representing the neighborhood set, Characterizing the time-backtracking input components, Representing neighboring nodes In time The state vector below, For bias terms, As a nonlinear activation function, time reparameterization is performed through the internal temporal convolution kernel of the operator to generate a time vector matrix. Based on this matrix, the autocoherence spectrum of local temporal features is calculated, and the energy ratio of the principal component and the subharmonic component of the spectrum is extracted to construct a normalized information vector. Finally, using this vector as the initial lattice unit, the initial period and energy trap distribution of the information time crystal are defined to complete the construction of the information time crystal structure.

[0021] Step S200: Perform energy feedback mapping in the information time crystal structure, treat node voltage perturbations as crystal defects, and generate a stable time lattice based on the periodic reconstruction of crystal defects.

[0022] Specifically, an energy feedback mapping mechanism is initiated within the constructed information-time crystal structure: First, the energy offset of each node is calculated in real time, reflecting the deviation between the actual energy of the node and the ideal steady-state energy. Nodes with energy offsets exceeding a preset threshold are marked as abnormal nodes, triggering anomaly warnings to promptly address potential risks. Subsequently, the node voltage disturbances occurring in the distribution network operation are equated to crystal defects in the information-time crystal structure. By analyzing the time-varying pattern of these defects, the periodic characteristics of the defects are determined, such as the frequency and duration of disturbances. Based on this crystal defect period, the lattice structure of the information-time crystal is reconstructed and adjusted to correct the lattice disorder caused by voltage disturbances, generating a stable time lattice that accurately reflects the stable operation trend of the distribution network, providing a structured analysis basis for subsequent identification of unstable regions.

[0023] Step S300: After identifying the local unstable region by utilizing the amplitude and frequency shift of the stable time lattice, perform periodic modulation and phase-locked optimization of the inverter control signal, and output a voltage regulation control strategy.

[0024] Specifically, the key characteristic parameters of the stable-time lattice are first extracted: the amplitude of lattice vibration corresponds to the voltage fluctuation amplitude and frequency shift, which correspond to the fluctuation frequency change. By analyzing regions with abnormally increased amplitude and frequency shift exceeding the normal range, local unstable regions in the distribution network are accurately identified. Next, a time oscillation mapping layer synchronized with the stable-time lattice is established, mapping the modulation period of the inverter output signal to the oscillation period of the subcells of the time lattice. Spectral decomposition is performed on the oscillation trajectory of the mapping layer, and dynamic modulation units corresponding to local unstable regions are calibrated within the layer. In each dynamic modulation unit, the phase coupling strength of adjacent subcells is calculated based on the oscillation correlation, and a time phase is constructed. The system employs a dryness function and establishes an amplitude stability domain by combining the amplitude offset of the local unstable region. Based on these two factors, it performs joint iteration to achieve periodic modulation of the inverter control signal. The signal output period is adjusted to adapt to the lattice stability requirements and phase-locked optimization, ensuring that the inverter output phase is synchronized with the grid phase. Subsequently, the optimization results of each dynamic modulation unit are transformed into a local steady-state spectrum vector, which is input into the high-order harmonic channel of the steady-time lattice. By calculating the high-order characteristic resonant mode and performing global steady-state resonance scheduling, the periodic perturbation and phase correction factor of each local mapping layer are adjusted in reverse. Finally, a voltage regulation control strategy that can effectively suppress voltage fluctuations and maintain the voltage stability of the distribution network is output.

[0025] In one possible implementation, step S300 further includes:

[0026] Step S310: Establish a time oscillation mapping layer synchronized with the stable time lattice, wherein the time oscillation mapping layer maps the modulation period of the inverter output signal to the subcell oscillation period of the time lattice.

[0027] Step S320: Perform spectral decomposition on the oscillation trajectory of the time oscillation mapping layer, extract the amplitude and frequency shift characteristics of the stable time lattice, identify the local unstable region based on the extraction results, and calibrate the corresponding dynamic modulation unit in the time oscillation mapping layer.

[0028] Step S330: In each dynamic modulation unit, the phase coupling strength of adjacent subcells is calculated based on the oscillation correlation of the time oscillation mapping layer, and a time coherence function is constructed.

[0029] Step S340: Using the amplitude offset of the local unstable region, establish an amplitude stable region on the time oscillation mapping layer, and perform joint iteration based on the amplitude stable region and the time coherence function to perform periodic modulation and phase-locked optimization of the inverter control signal.

[0030] Specifically, the basic parameters of the generated stable time lattice are obtained, including the lattice principal period, the number of subcells, and the subcell reference oscillation frequency. These parameters are used as the timing synchronization reference for the mapping layer. Next, a period sampling module is deployed to collect the modulation period of the inverter output signal in real time, such as the PWM modulation period. The analog signal period is converted into a digital period parameter through AD conversion, and signal noise interference is removed. Subsequently, a period mapping algorithm model is constructed. This model has a built-in matching table between the subcell oscillation period and the inverter modulation period. Through linear interpolation or nonlinear fitting algorithms, the collected inverter modulation period value is accurately converted into the subcell oscillation period value of the stable time lattice, ensuring that the two period values ​​are in one-to-one correspondence and synchronized in real time. At the same time, a synchronization verification unit is set up to compare the deviation between the subcell oscillation period in the time oscillation mapping layer and the subcell period of the stable time lattice in real time. If the deviation exceeds a preset threshold, such as ±5%, a period calibration mechanism is triggered to adjust the conversion coefficient of the mapping algorithm until the period synchronization accuracy of the two meets the voltage control requirements of the distribution network. Finally, the construction of the time oscillation mapping layer that is synchronized with the stable time lattice and realizes the period mapping function is completed.

[0031] For the oscillation trajectory within the time oscillation mapping layer, the dynamic time-series trajectory formed by the subcell oscillation period is subjected to spectral decomposition processing. Algorithms such as Fourier transform or wavelet decomposition are used to decompose the oscillation trajectory into spectral components of different frequencies and amplitudes. Key characteristic parameters of the stable time lattice are extracted from the decomposition results, namely the amplitude characteristics reflecting the voltage fluctuation amplitude and the frequency shift characteristics reflecting the fluctuation frequency change. The extracted amplitude and frequency shift characteristics are compared with the preset normal threshold range. If the amplitude of a certain region exceeds the upper limit of normal fluctuation or the frequency shift deviates from the reference frequency, the region is determined to be a local unstable region. Subsequently, within the time oscillation mapping layer, according to the spatial and temporal range of the local unstable region, the corresponding dynamic modulation unit is calibrated. Each dynamic modulation unit corresponds to one or a group of inverter control signal channels that need to be regulated.

[0032] For each calibrated dynamic modulation unit, the oscillation correlation between the unit and its neighboring subcells within the time oscillation mapping layer is first analyzed. By calculating the oscillation phase difference and amplitude variation coordination degree of different subcells at the same time node, the degree of oscillation correlation between subcells is determined. Based on this oscillation correlation, the phase coupling strength calculation formula is used, and the phase coupling strength between adjacent subcells is solved based on the calculation model of cross-correlation coefficient or phase lock value, quantifying the mutual influence of subcell oscillation phases. With the phase coupling strength as the core parameter, combined with the temporal distribution characteristics of subcells, a time coherence function is constructed. This function can reflect the temporal coordination of subcell oscillations within the dynamic modulation unit in real time.

[0033] First, the amplitude offset of the identified local unstable region is obtained, which is the difference between the actual amplitude of the region and the normal amplitude threshold. Based on this offset, an amplitude stability region is defined on the time oscillation mapping layer. The boundary of the amplitude stability region is set with the goal of returning the amplitude to the normal threshold range, and the reasonable range of the subcell oscillation amplitude that needs to be maintained in this region is defined. The boundary conditions of the amplitude stability region and the constructed time coherence function are used as input parameters for joint iteration, and the iterative calculation process is started. In each iteration, the phase relationship between adjacent subcells is adjusted according to the time coherence function to achieve phase locking optimization. At the same time, the modulation period of the inverter control signal is corrected according to the amplitude stability region to make the subcell amplitude converge into the stability region. Through multiple iterations, until the amplitude deviation in the amplitude stability region meets the preset accuracy requirements and the time coherence function reaches its maximum value, that is, the subcell phase coordination is optimal, and finally the periodic modulation and phase locking optimization of the inverter control signal are completed.

[0034] In one possible implementation, step S300 further includes:

[0035] Step S350: Convert the optimization output results of each dynamic modulation unit into the local steady-state spectrum vector of the time lattice node. The local steady-state spectrum vector includes the amplitude stability domain center value component, the phase lock angle deviation component, and the oscillation frequency shift component.

[0036] Step S360: Activate the higher harmonic channel of the stable time lattice, input the local steady-state spectral vector into the higher harmonic channel, establish the energy coupling matrix between the local vector spectral vectors, and calculate the higher-order characteristic resonant mode.

[0037] Step S370: Utilize the higher-order characteristic resonant mode to perform global steady-state resonance scheduling, and perform temporal resonance synchronization between multiple local steady-state spectral vectors by minimizing the resonance energy difference.

[0038] Step S380: After obtaining the aggregation degree parameter of global resonance synchronization, reverse schedule the periodic perturbation and phase correction factor of each local time oscillation mapping layer, and output the voltage regulation control strategy.

[0039] Specifically, the process involves acquiring the output results of each dynamic modulation unit after completing the periodic modulation and phase-locking optimization of the inverter control signal. These results include the specific control parameters of each unit for amplitude stabilization, phase locking, and frequency shift adjustment. Subsequently, the optimization output results of each dynamic modulation unit are correlated and matched with the corresponding nodes of the time lattice to ensure that the control parameters of each unit are accurately mapped to their corresponding time lattice nodes. Based on this, according to the preset vector construction rules, key parameters are extracted from the optimization results and transformed into three core components of the local steady-state spectrum vector: the amplitude stability domain center value component, which corresponds to the ideal stable reference value of amplitude determined by optimization, ensuring that amplitude control has a clear target; the phase-locking angle deviation component, which quantifies the remaining deviation between the actual phase and the ideal locked phase after optimization, providing a basis for subsequent fine-tuning; and the oscillation frequency shift component, which records the offset value of the oscillation frequency relative to the reference frequency after optimization, reflecting the frequency shift control effect. Finally, a local steady-state spectrum vector exclusive to each time lattice node is formed, transforming the dispersed optimization results into a structured data form suitable for subsequent high-order operations.

[0040] Based on the current operating status of the distribution network and the timing characteristics of the steady-state time lattice, a pre-set high-order harmonic channel in the steady-state time lattice is automatically activated. This channel is specifically designed to handle energy correlation and resonance analysis between multiple nodes and can adapt to complex energy interactions in scenarios with a high proportion of distributed power sources. Next, the generated local steady-state spectrum vectors corresponding to each time lattice node are input one by one into the activated high-order harmonic channel according to the node's timing and spatial correlation. Inside the channel, an energy coupling analysis algorithm is used to calculate the energy interaction intensity between any two local steady-state spectrum vectors at different harmonic frequencies, and the intensity value is correlated with the amplitude between the vectors. The degree of coordination and phase correlation are positively correlated. An energy coupling matrix between local steady-state spectral vectors is constructed based on all interaction intensity values. The rows and columns of the matrix correspond to the local steady-state spectral vectors of different nodes, and the matrix elements represent the degree of energy coupling between the corresponding two vectors. Finally, the matrix eigenvalue decomposition tool is used to operate on the energy coupling matrix to extract the higher-order eigenvalues ​​of the matrix and their corresponding eigenvectors. These eigenvalues ​​and eigenvectors are integrated into a higher-order characteristic resonance mode. This resonance mode can accurately characterize the overall energy resonance characteristics between multiple nodes in the distribution network, providing a core analytical basis for subsequent global steady-state resonance scheduling.

[0041] Using the calculated higher-order characteristic resonance mode as the global control benchmark, this resonance mode contains ideal energy resonance state parameters among multiple nodes in the distribution network, such as resonance frequency and energy distribution ratio. First, the actual resonance energy corresponding to the local steady-state spectrum vector of each time lattice node is extracted. By comparing it node by node with the ideal resonance energy in the higher-order characteristic resonance mode, the resonance energy difference between each local steady-state spectrum vector and the ideal state is calculated. At the same time, the interaction deviation caused by the energy difference among multiple local steady-state spectrum vectors is statistically analyzed. Then, the global steady-state resonance scheduling algorithm is started. With the goal of minimizing the resonance energy difference, the key parameters of the local steady-state spectrum vector of each node are dynamically adjusted, such as fine-tuning the center value of the amplitude stability domain to optimize energy output and correcting the phase lock angle deviation to coordinate the phase relationship. Through multiple rounds of iterative adjustment, the time-domain resonance energy difference between each local steady-state spectrum vector is gradually reduced, so that the resonance state of each node gradually approaches the ideal state represented by the higher-order characteristic resonance mode. Finally, the time-domain resonance synchronization among multiple local steady-state spectrum vectors is achieved, avoiding voltage fluctuations in the distribution network caused by resonance imbalance between nodes, and laying a collaborative foundation for the subsequent output of the global voltage stabilization strategy.

[0042] The aggregation degree parameter is collected in real time after multiple local steady-state spectral vectors have achieved time-domain resonance synchronization. This parameter is used to quantitatively evaluate the overall coordination degree of resonance synchronization among nodes. The closer the value is to the preset optimal threshold, the better the resonance synchronization effect among nodes, ensuring that subsequent scheduling is initiated only when the aggregation degree parameter meets the distribution network stability control requirements. Then, based on the aggregation degree parameter as feedback, reverse scheduling is performed. Combining the timing characteristics of the steady-time lattice and the initial parameters of each local time oscillation mapping layer, the periodic perturbation amount that needs to be adjusted for each mapping layer is derived in reverse. This is used to finely correct the modulation period of the inverter control signal, compensate for local period deviations and phase correction factors, and further optimize the phase coordination between subcells to reduce phase locking deviations. Finally, the derived periodic perturbation amount and phase correction factor are sent to the corresponding local time oscillation mapping layers to complete the precise scheduling of parameters at each layer. The scheduling results of all mapping layers are integrated to form a voltage regulation control strategy that can be directly applied to the distribution network inverter. This strategy includes the specific modulation period parameters and phase locking parameters of each inverter, which can achieve global and precise suppression of distribution network voltage fluctuations. Finally, it is output to the execution layer to ensure the stability of the distribution network voltage.

[0043] In one possible implementation, step S360 further includes:

[0044] Step S361: Calculate the global voltage resonance anomaly index using the higher-order characteristic resonant mode.

[0045] Step S362: Determine whether the global voltage resonance anomaly index exceeds the preset tolerance range. If it does, generate a global voltage regulation adjustment strategy.

[0046] Step S363: Adjust and compensate for global steady-state resonance scheduling according to the global voltage regulation adjustment strategy.

[0047] Specifically, the process involves obtaining calculated high-order characteristic resonant modes, which contain core parameters of the resonant characteristics between multiple nodes in the distribution network, such as high-order eigenvalues, corresponding eigenvectors, and resonant energy distribution characteristics. Subsequently, key analytical parameters are extracted from these high-order characteristic resonant modes, including the maximum deviation of the resonant frequency from the reference frequency, the dispersion of eigenvector components corresponding to each node, and the energy imbalance between different resonant modes. These parameters are then processed using a weighted summation algorithm, with weight coefficients pre-calibrated based on the voltage sensitivity of distribution network nodes, the capacity of distributed power sources, and the resonant tolerance threshold of equipment. This transforms the multi-dimensional resonant characteristic parameters into a single, quantified global voltage resonance anomaly index. This index directly reflects the degree to which the overall resonant state of the distribution network deviates from the ideal stable state; a larger value indicates a higher risk of voltage fluctuations caused by resonance anomalies, providing a quantitative basis for subsequent judgments on whether global control is necessary.

[0048] The calculated global voltage resonance anomaly index is retrieved, and a pre-set tolerance range is loaded. This tolerance range is comprehensively calibrated based on the distribution network voltage stability operation standard, the proportion of distributed power source access (which needs to be tightened in high-proportion access scenarios to improve stability), and the resonance tolerance threshold of grid equipment, clearly defining the global resonance state within a safe and controllable numerical range. Subsequently, the actual calculated global voltage resonance anomaly index is compared with the preset tolerance range: if the index value is within the tolerance range, it indicates that the current distribution network resonance state does not pose a risk to voltage stability, and no additional intervention is required; the process directly proceeds to the subsequent global steady-state resonance scheduling process. If the index value exceeds the tolerance range, it is determined that there is a global resonance anomaly risk in the distribution network, which may cause significant voltage fluctuations. At this time, combined with the anomaly root causes reflected by higher-order characteristic resonance modes, such as specific resonance frequency shifts and local node energy imbalances, a global voltage stabilization adjustment strategy is generated. This strategy clearly includes the resonance modes that need to be controlled, the direction of node energy distribution correction, and preliminary resonance suppression parameters, providing an execution basis for subsequent global steady-state resonance scheduling adjustment and compensation.

[0049] The generated global voltage regulation strategy is analyzed, and its explicit resonance correction directions are extracted, such as suppressing abnormal resonances in specific frequency bands, balancing energy distribution among multiple nodes, and focusing on the range of key control nodes, such as the distributed power source access node with the most significant resonance anomalies, and preliminary adjustment parameters, such as energy allocation correction coefficients and resonance frequency suppression thresholds. Subsequently, these analytical results are transformed into compensation control parameters for global steady-state resonance scheduling and integrated with the energy optimization algorithm and node parameter adjustment logic in the original scheduling process. For example, for abnormal resonance frequency bands that need to be suppressed first in the strategy, the weight coefficient of the energy optimization algorithm is adjusted to increase the reduction weight of resonance energy in that frequency band; for key control nodes, the comparison and correction strength between their local steady-state spectrum vector and the ideal state of higher-order characteristic resonance modes is strengthened; finally, by dynamically adjusting the iteration step size and convergence threshold of global steady-state resonance scheduling, the adjustment and compensation process is ensured to accurately match the requirements of the global voltage regulation strategy, gradually correcting the resonance anomaly state of the distribution network, and causing the resonance characteristics of each node to converge towards the ideal stable range, laying a stable global resonance foundation for subsequent reverse scheduling and voltage regulation control strategy output.

[0050] In one possible implementation, step S380 further includes:

[0051] Step S381: Execute the activation record of the dynamic modulation unit and generate the activation record database.

[0052] Step S382: Perform time vector indexing on the activation record database, and map the time vector matrix fragments corresponding to each round of voltage regulation strategy to historical lattices.

[0053] Step S383: Before re-executing the voltage regulation control strategy optimization, perform time similarity analysis using historical lattices to construct self-evolutionary feedback.

[0054] Step S384: Utilize the self-evolutionary feedback to perform a new round of voltage regulation control strategy optimization management.

[0055] Specifically, after each round of voltage regulation control strategy completes its output and applies it to the distribution network, the dynamic modulation unit activation recording process is automatically triggered. The data acquisition component extracts key information of all activated dynamic modulation units in this round of regulation, including but not limited to: the unique ID of the activated dynamic modulation unit, the activation start and end times, the location of the corresponding local instability interval in the distribution network and its initial abnormal characteristics, such as amplitude offset and frequency shift deviation, and the periodic modulation parameters of the inverter control signal executed on that unit in this round, such as modulation period adjustment and phase-locked loop optimization results, such as the phase correction factor. Subsequently, this structured information is categorized and organized according to regulation round and timestamp, and stored in a pre-set database using standardized data formats, such as JSON or tables, gradually accumulating to form an activation record database. This database can be updated in real time and supports data retrieval by modulation unit ID, regulation time, abnormal characteristic type, and other dimensions.

[0056] For the generated activation record database, a time vector index construction process is initiated. Time vector matrix fragments generated by the time oscillation mapping layer during each round of voltage regulation strategy execution are extracted from the database. These fragments contain complete data on subcell oscillation periods, phase changes, and timing correlations for the corresponding round. A unique timestamp is added to each matrix fragment, accurate to the control initiation time, establishing a time-dimensional-centric index system to ensure rapid location of target matrix fragments based on time range and control round. Subsequently, a lattice mapping algorithm is initiated to match and calibrate each extracted time vector matrix fragment with the structural features of a stable time lattice, such as subcell arrangement rules and timing period benchmarks. The oscillation period parameters and phase coupling strength data in the matrix fragments are transformed into historical lattices conforming to the stable time lattice structure specifications. Each historical lattice completely corresponds to the timing characteristics and control correlation data of a round of voltage regulation strategy, reflecting both the resonant state changes of the distribution network under that round and tracing the corresponding dynamic modulation unit activation parameters.

[0057] When a new round of voltage regulation control strategy optimization is triggered, the process first extracts and generates time vector features corresponding to the current operating state based on real-time operating data of the current distribution network, such as voltage, phase angle, power parameters, and the current control signal status of the inverter at each node. These features are then matched with the dimensions of the time vector matrix segments constructed during the historical lattice construction. Subsequently, all stored historical lattices are retrieved from the activation record database. Using time similarity analysis algorithms such as Dynamic Time Warping (DTW), the similarity value between the current time vector features and the time vector matrix segments corresponding to each historical lattice is calculated. A higher similarity value indicates that the operating state of the distribution network corresponding to the historical lattice is closer to the current state. Based on the similarity ranking results, the top similarity lattices are selected. N is a historical lattice preset based on the distribution network node scale and historical data volume. The corresponding control parameters of these historical lattices are extracted, such as periodic modulation amount, phase correction factor, and anomaly handling logic, such as response strategies for specific resonance anomalies and data on the effectiveness of strategy implementation, such as voltage stability maintenance duration and resonance anomaly elimination efficiency. Finally, the results of the difference analysis between these extracted historical experience data and the current operating state, such as the difference in amplitude offset and frequency shift deviation between the current and historical states, are fused to construct self-evolutionary feedback information. This feedback information clarifies the range of reusable historical effective parameters for the new round of optimization, the ineffective control directions to be avoided, and adjustment suggestions for the differences in the current state.

[0058] The constructed self-evolving feedback information is fully input into the new round of voltage regulation control strategy optimization. First, the core content of the feedback information is analyzed, including the range of historically effective control parameters, such as verified effective inverter cycle modulation intervals, phase correction factor thresholds, ineffective control directions to be avoided (e.g., parameter combinations that previously exacerbated resonance anomalies), and adjustment suggestions for differences between the current and historical states (e.g., correction schemes for current amplitude offsets that are larger than historical states). Subsequently, using this feedback information as constraints and optimization basis, the optimization process is managed: on the one hand, the range of historically effective parameters is used as the initial optimization space, replacing the full-domain search without feedback, significantly reducing the parameter retrieval dimension and improving optimization efficiency; on the other hand, during the optimization iteration process, the current candidate parameters are compared with ineffective control directions in the feedback information in real time, automatically filtering parameter combinations that may cause voltage fluctuation risks, reducing trial-and-error costs. Simultaneously, based on the differences between the current and historical operating states, and combined with adjustment suggestions in the feedback information, candidate parameters are dynamically corrected. For example, when the current frequency shift deviation is larger, the frequency shift adjustment weight is appropriately increased based on historical experience, ensuring that candidate parameters both inherit historically verified effectiveness and adapt to the current actual operating state of the distribution network. Ultimately, through this optimization management that integrates self-evolutionary feedback, a new round of more precise and efficient voltage stabilization control strategies is output, further improving the adaptability and reliability of voltage stability control in the distribution network.

[0059] In one possible implementation, step S100 further includes:

[0060] Step S110: After de-DC offsetting and normalizing the state information stream, the ground state time vector field is established using the offset trend residual.

[0061] Step S120: In the ground state time vector field, multi-scale low-pass filtering and adaptive denoising algorithm are used to suppress high-frequency disturbances, and the time coherence features between node voltage and power attention are extracted.

[0062] Step S130: Input the temporal coherence features as evolution driving quantities into the temporal evolution operator, perform temporal reparameterization through the temporal convolution kernel inside the temporal evolution operator, and establish a temporal vector matrix.

[0063] Step S140: Calculate the autocoherence spectrum of local time-series features based on the time vector matrix, extract the energy ratio of the principal component and the subharmonic component of the spectrum, and construct a normalized information vector.

[0064] Step S150: Using the normalized information vector as the initial lattice unit, define the initial period and energy trap distribution of the information time crystal, and construct the information time crystal structure.

[0065] Specifically, the pre-constructed multi-node state information flow of the distribution network, including node voltage, phase angle, power parameters, and inverter control command signals, is preprocessed: a high-pass filtering algorithm is used to remove the DC offset component in each parameter, eliminating the interference of static deviation on subsequent analysis; then, the Min-Max normalization method is used to uniformly map the de-offset state information flow parameters to the [0, 1] interval, avoiding calculation deviations caused by differences in the magnitude of different parameters; subsequently, the offset trend residual of each parameter in the time dimension is calculated, that is, the difference between the actual parameter value and the fitted trend line. Based on these residual data, a ground-state time vector field is constructed. This vector field has time as the horizontal axis and the offset trend residual of each node state parameter as the vertical axis, clearly presenting the basic time-series change characteristics of the state information flow.

[0066] In the constructed ground-state time vector field, the anti-interference and feature extraction process is initiated: First, a multi-scale low-pass filtering algorithm, such as wavelet low-pass filtering, is used to gradually filter out high-frequency disturbance signals caused by distributed power source fluctuations and load changes in the vector field by setting different scale filtering windows, while retaining the low-frequency effective components that reflect the stable operation trend of the distribution network; then, an adaptive denoising algorithm, such as threshold-based adaptive wavelet denoising, is used to dynamically adjust the denoising intensity according to the signal-to-noise ratio of the local signal in the vector field, further optimizing the data quality; finally, based on the denoised ground-state time vector field, the correlation coefficients of node voltage and power at each time node are calculated, and the synergistic law of their changes over time is extracted to form a time coherence feature characterizing their correlation.

[0067] The extracted temporal coherence features are used as the core evolutionary driving force and input into the preset temporal evolution operator. Temporal convolution kernel matrix inside the operator Temporal coherence features are reparameterized by integrating historical time series information with current feature data through convolutional kernel weights corresponding to different time delay indices l, transforming one-dimensional temporal coherence features into a multi-dimensional time series matrix. Finally, the state vectors at each time point output by the time evolution operator are combined column by column in chronological order to establish a time vector matrix containing time series correlation information of multiple nodes in the distribution network.

[0068] Local time-series feature analysis is conducted based on the time vector matrix. The time vector matrix is ​​decomposed in the frequency domain using short-time Fourier transform, and the autocoherence spectrum of the time-series features for each local time segment is calculated. This spectrum visually displays the distribution and correlation of different frequency components in the time-series dimension. The principal spectral components, corresponding to the dominant operating frequency of the distribution network, and the subharmonic components, corresponding to the harmonic characteristics caused by distributed generation, are separated from the autocoherence spectrum. The energy values ​​of the two types of components are calculated, and their energy proportions are determined. Based on the energy proportion results, the feature parameters in the time vector matrix are normalized, redundant features with excessively low energy proportions are removed, core features are retained, and a normalized information vector with unified dimensions and prominent features is constructed.

[0069] Normalized information vectors are used as the initial lattice units of the information time crystal. The initial period of the information time crystal is defined in conjunction with the operating cycle of the distribution network, such as the output cycle of distributed power sources and the load change cycle, to ensure that the crystal period matches the actual time-series characteristics of the distribution network. At the same time, the energy trap distribution of the information time crystal is set according to the stability threshold of the state parameters of each node. The energy trap depth is positively correlated with the stability sensitivity of the parameters. For example, the higher the stability sensitivity of voltage parameters, the greater the corresponding trap depth, so that the crystal structure can prioritize the stabilization of key parameters. Based on the initial lattice units, the initial period, and the energy trap distribution, a complete information time crystal structure is constructed. This structure can dynamically reflect the temporal evolution law of the state information flow of the distribution network.

[0070] In one possible implementation, step S130 further includes:

[0071] The time evolution operator is as follows:

[0072] .

[0073] in, This is the state vector after time evolution. Characterization time The node summary status below, For time-coherent feature vectors, To drive the mapping matrix, For the length of time memory, For time-delay index, The temporal convolution kernel matrix, This is the neighborhood coupling weight coefficient matrix. Representing the neighborhood set, Characterizing the time-backtracking input components, Representing neighboring nodes In time The state vector below, For bias terms, It is a non-linear activation function.

[0074] The calculated state vectors are combined column by column to form a time vector matrix.

[0075] Specifically, the mathematical expression of the time evolution operator, namely Furthermore, the physical meaning and function of each parameter in the formula were clearly defined. It represents the node state vector output after time evolution, which comprehensively reflects the overall operating status of multiple nodes in the distribution network after evolution; Characterization at time step The summation of the state vectors of all nodes is the basic input for evolutionary computation; The extracted temporal coherence feature vector serves as the core variable driving the evolution of the operator, carrying the temporal correlation information between node voltage and power attention. It is the driving mapping matrix, used to map temporally coherent eigenvectors. Transform it into a dimension that fits the state vector to achieve effective coupling between features and state; Defined as the time memory length, it determines the range of historical time-series data that the operator can trace back to; l is the time delay index, used to traverse historical data at different backtracking times. This is a time-delayed index used to iterate through historical data at different backtracking points; It is a temporal convolution kernel matrix, with different indices. The corresponding matrix elements represent the weights of the historical state vectors under different time delays, and the integration of historical time series information is achieved through convolution operations; This is the neighborhood coupling weight coefficient matrix, used to quantify the degree of state influence between adjacent nodes; Representing the neighborhood set, Calculate the state difference between the neighboring nodes and the current node, and then... After weighting, the data is incorporated into the evolutionary process, reflecting the spatial correlation characteristics between nodes; It is a time rewind The state vector at each time step provides historical time-series input for the operator; Representing neighboring nodes In time The state vector is the key data for calculating the state differences between nodes; This is a bias term used to fine-tune the operator output and improve the accuracy of evolution calculations; Nonlinear activation functions, such as Sigmoid or ReLU functions, are used to introduce nonlinear transformations, enabling the operator to fit the complex nonlinear evolution of the distribution network state information flow.

[0076] After clarifying the operator formula and parameter definitions, the method for constructing the time vector matrix is ​​specified: calculate all state vectors output by the time evolution operator at different time steps. The data are combined column-wise in chronological order to form a matrix whose dimensions match the number of time steps and the dimensions of node state parameters—that is, a time vector matrix. This matrix fully preserves the evolution trajectory of the distribution network state information flow in the time dimension and the correlation characteristics between nodes, providing a structured data foundation for subsequent calculation of local time-series feature autocoherence maps and construction of standardized information vectors.

[0077] In one possible implementation, step S200 further includes:

[0078] Step S210: When performing energy feedback mapping in the information-time crystal structure, calculate the energy offset of each node.

[0079] Step S220: Mark nodes whose energy offset exceeds a preset threshold as abnormal nodes and perform abnormal early warning management.

[0080] Specifically, when performing energy feedback mapping on the constructed information-time crystal structure, the energy potential distribution data of each node in the crystal structure is first extracted to obtain the ideal stable energy level corresponding to each node. This level is determined based on the node's rated voltage, rated power parameters, and the overall stable operation standard of the distribution network, and is dynamically calibrated in conjunction with the capacity of the distributed power sources connected to the node. Subsequently, by collecting the current operating parameters of each node, including node voltage, phase angle, active power, and reactive power, these parameters are substituted into a preset real-time energy calculation model for the node, such as an energy calculation model based on power time-domain integration or voltage-impedance relationship, to obtain the actual energy state value of each node. Finally, the difference between the actual energy state value of each node and the corresponding ideal stable energy level is calculated. The result is the energy offset of the node. This offset can intuitively reflect the direction and degree of the node's energy deviation from the ideal stable state. A positive value indicates that the energy is too high, and a negative value indicates that the energy is too low, providing a core quantitative basis for subsequent abnormal node identification.

[0081] The system retrieves the calculated energy offsets of each node and loads a preset energy offset tolerance threshold. This threshold is set comprehensively based on the functional attributes of different nodes in the distribution network, such as load nodes, distributed power supply access nodes, equipment tolerance capabilities (e.g., transformers, inverters), energy fluctuation tolerance range, and voltage stability control requirements. It also supports configuring more stringent differentiated thresholds for important nodes, such as backbone network nodes and sensitive load nodes. Subsequently, the energy offset of each node is compared one by one with the corresponding threshold. If the absolute value of the energy offset of a node exceeds the preset threshold, the node is marked as an abnormal node, and a data set including the abnormal node number, location, and current energy offset is generated. The system generates structured anomaly records, including displacement values, threshold standards, and timestamps of anomaly occurrences. Finally, it initiates anomaly warning and reporting management processes: on the one hand, anomaly records are stored in the system alarm database in real time for easy tracing and analysis; on the other hand, anomaly information is pushed to distribution network maintenance personnel through pre-defined warning channels, such as pop-up notifications on the operation and maintenance monitoring platform, SMS notifications, and audible and visual alarms. This includes preliminary speculation on the cause of the anomaly, such as sudden changes in distributed power output or a sudden increase in node load, along with suggested handling directions. This ensures that maintenance personnel can promptly grasp the anomaly and take intervention measures to prevent the spread of unstable states from abnormal nodes from affecting the overall voltage stability of the distribution network.

[0082] Example 2, based on the same inventive concept as the high-proportion distributed power generation distribution network voltage stability control method in the aforementioned examples, such as... Figure 2 As shown, this application provides a voltage stability control system for a high-proportion distributed power generation distribution network. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0083] The state information flow construction module 10 is used to construct the state information flow of multiple nodes in the distribution network. The state information flow includes node voltage, phase angle, power attention value and inverter control command signal, and uses time evolution operator to construct information time crystal structure.

[0084] The stable time lattice generation module 20 is used to perform energy feedback mapping in the information time crystal structure, treat node voltage perturbations as crystal defects, and generate a stable time lattice based on the periodic reconstruction of crystal defects.

[0085] The control strategy output module 30 is used to identify local unstable regions by utilizing the amplitude and frequency shift of the stable time lattice, and then perform periodic modulation and phase-locked optimization of the inverter control signal to output a voltage regulation control strategy.

[0086] Furthermore, the system is also used to implement the following functions:

[0087] A time oscillation mapping layer synchronized with the stable time lattice is established, which maps the modulation period of the inverter output signal to the oscillation period of the subcell of the time lattice. Spectral decomposition is performed on the oscillation trajectory of the time oscillation mapping layer to extract the amplitude and frequency shift characteristics of the stable time lattice. After identifying local unstable regions based on the extraction results, the corresponding dynamic modulation units are calibrated within the time oscillation mapping layer. In each dynamic modulation unit, the phase coupling strength of adjacent subcells is calculated based on the oscillation correlation of the time oscillation mapping layer, and a time coherence function is constructed. Using the amplitude offset of the local unstable regions, an amplitude stability region is established on the time oscillation mapping layer. Based on the amplitude stability region and the time coherence function, joint iteration is performed to execute periodic modulation and phase-locked optimization of the inverter control signal.

[0088] Furthermore, the system is also used to implement the following functions:

[0089] The optimization output of each dynamic modulation unit is transformed into a local steady-state spectrum vector of the time lattice node. The local steady-state spectrum vector includes the amplitude stability domain center value component, the phase lock angle deviation component, and the oscillation frequency shift component. The higher-order harmonic channel of the stable time lattice is activated, and the local steady-state spectrum vector is input into the higher-order harmonic channel to establish an energy coupling matrix between the local spectrum vectors and calculate the higher-order characteristic resonance mode. The higher-order characteristic resonance mode is used to perform global steady-state resonance scheduling, and the temporal resonance synchronization between multiple local steady-state spectrum vectors is performed by minimizing the resonance energy difference. After obtaining the aggregation degree parameter of global resonance synchronization, the periodic perturbation and phase correction factor of each local time oscillation mapping layer are reverse-scheduled to output the voltage regulation control strategy.

[0090] Furthermore, the system is also used to implement the following functions:

[0091] The global voltage resonance anomaly index is calculated using the higher-order characteristic resonant mode; it is determined whether the global voltage resonance anomaly index exceeds a preset tolerance range. If it does, a global voltage regulation adjustment strategy is generated; and global steady-state resonance scheduling adjustment and compensation are performed according to the global voltage regulation adjustment strategy.

[0092] Furthermore, the system is also used to implement the following functions:

[0093] The activation record of the dynamic modulation unit is executed to generate an activation record database; the activation record database is indexed by time vector, and the time vector matrix fragments corresponding to each round of voltage regulation strategy are mapped to historical lattices; before the voltage regulation control strategy optimization is executed again, time similarity analysis is performed using historical lattices to construct self-evolutionary feedback; the self-evolutionary feedback is used to manage the optimization of the new round of voltage regulation control strategy.

[0094] Furthermore, the system is also used to implement the following functions:

[0095] After the state information stream is de-DC offset and normalized, a ground-state time vector field is established using the offset trend residual. In the ground-state time vector field, multi-scale low-pass filtering and adaptive denoising algorithms are used to suppress high-frequency disturbances, and the temporal coherence features between node voltage and power attention quantities are extracted. The temporal coherence features are used as evolution driving quantities input to the time evolution operator, and time reparameterization is performed through the temporal convolution kernel inside the time evolution operator to establish a time vector matrix. Based on the time vector matrix, the autocoherence spectrum of local temporal features is calculated, and the energy ratio of the principal component and the subharmonic component of the spectrum is extracted to construct a normalized information vector. Using the normalized information vector as the initial lattice unit, the initial period and energy trap distribution of the information time crystal are defined to construct the information time crystal structure.

[0096] Furthermore, the system is also used to implement the following functions:

[0097] The time evolution operator is as follows:

[0098] ;in, This is the state vector after time evolution. Characterization time The node summary status below, For time-coherent feature vectors, To drive the mapping matrix, For the length of time memory, For time-delay index, The temporal convolution kernel matrix, This is the neighborhood coupling weight coefficient matrix. Representing the neighborhood set, Characterizing the time-backtracking input components, Representing neighboring nodes In time The state vector below, For bias terms, The nonlinear activation function is used to combine the calculated state vectors column-wise to form a time vector matrix.

[0099] Furthermore, the system is also used to implement the following functions:

[0100] When performing energy feedback mapping in the information time crystal structure, the energy offset of each node is calculated; nodes whose energy offset exceeds a preset threshold are marked as abnormal nodes, and abnormal early warning and management are implemented.

[0101] Example 3: Based on the same inventive concept as the high-proportion distributed power generation distribution network voltage stability control method in the foregoing examples, this example provides a computer-readable storage medium for storing software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the high-proportion distributed power generation distribution network voltage stability control method in this application. The processor executes various functional applications and data processing of the computer device by running the software programs, instructions, and modules stored in the memory, thereby realizing the aforementioned high-proportion distributed power generation distribution network voltage stability control method.

[0102] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0103] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application. This specification and drawings are merely illustrative examples of this application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope of this application. Thus, if these modifications and variations of this application fall within the scope of this application and its equivalents, this application intends to include these modifications and variations.

Claims

1. A voltage stability control method for a high-proportion distributed power generation distribution network, characterized in that, The method includes: A state information flow of multiple nodes in the distribution network is constructed, which includes node voltage, phase angle, power attention value and inverter control command signal. An information time crystal structure is constructed using time evolution operator. Energy feedback mapping is performed in the information-time crystal structure, node voltage perturbations are treated as crystal defects, and a stable time lattice is generated based on the periodic reconstruction of crystal defects. By utilizing the amplitude and frequency shift of the stable time lattice, after identifying the local unstable region, periodic modulation and phase-locked optimization of the inverter control signal are performed to output a voltage regulation control strategy. Constructing information-time crystal structures using time evolution operators, including: After the state information stream is de-DC offset and normalized, the ground state time vector field is established using the offset trend residual. In the ground state time vector field, multi-scale low-pass filtering and adaptive denoising algorithm are used to suppress high-frequency disturbances and extract the time coherence features between node voltage and power attention. The temporal coherence features are used as evolutionary drivers and input into the temporal evolution operator. Temporal reparameterization is performed through the temporal convolution kernel inside the temporal evolution operator to establish a temporal vector matrix. Based on the time vector matrix, calculate the autocoherence spectrum of local time-series characteristics, extract the energy ratio of the principal component and the subharmonic component of the spectrum, and construct a normalized information vector; Using the normalized information vector as the initial lattice unit, the initial period and energy trap distribution of the information time crystal are defined, and the information time crystal structure is constructed.

2. The high-proportion distributed power generation distribution network voltage stability control method as described in claim 1, characterized in that, Performing periodic modulation and phase-locked optimization of the inverter control signals includes: A time oscillation mapping layer synchronized with the stable time lattice is established, which maps the modulation period of the inverter output signal to the subcell oscillation period of the time lattice. The oscillation trajectory of the time oscillation mapping layer is spectrally decomposed to extract the amplitude and frequency shift characteristics of the stable time lattice. After identifying the local unstable regions based on the extraction results, the corresponding dynamic modulation units are calibrated within the time oscillation mapping layer. In each dynamic modulation unit, the phase coupling strength of adjacent subcells is calculated based on the oscillation correlation of the time oscillation mapping layer, and a time coherence function is constructed. By utilizing the amplitude offset of the local unstable region, an amplitude stability region is established on the time oscillation mapping layer. Based on the amplitude stability region and the time coherence function, a joint iteration is performed to execute periodic modulation and phase-locked optimization of the inverter control signal.

3. The voltage stability control method for a high-proportion distributed power generation distribution network as described in claim 2, characterized in that, Output voltage regulation control strategies include: The optimization output results of each dynamic modulation unit are converted into local steady-state spectrum vectors of time lattice nodes. The local steady-state spectrum vectors include the amplitude stability domain center value component, the phase lock angle deviation component, and the oscillation frequency shift component. Activate the higher harmonic channels of the stable time lattice, input the local steady-state spectrum vector into the higher harmonic channels, establish the energy coupling matrix between the local steady-state spectrum vectors, and calculate the higher-order characteristic resonance modes; Global steady-state resonance scheduling is performed using the higher-order characteristic resonant modes, and temporal resonance synchronization between multiple local steady-state spectral vectors is achieved by minimizing the resonant energy difference. After obtaining the aggregation degree parameter of global resonance synchronization, the periodic perturbation and phase correction factor of each local time oscillation mapping layer are reverse-scheduled to output the voltage regulation control strategy.

4. The high-proportion distributed power generation distribution network voltage stability control method as described in claim 3, characterized in that, After calculating the higher-order characteristic resonance modes, the following is included: The global voltage resonance anomaly index is calculated using the aforementioned higher-order characteristic resonant modes; Determine whether the global voltage resonance anomaly index exceeds the preset tolerance range; if it does, generate a global voltage regulation adjustment strategy. Adjustment and compensation are performed for global steady-state resonance scheduling based on the global voltage regulation adjustment strategy.

5. The voltage stability control method for a high-proportion distributed power generation distribution network as described in claim 2, characterized in that, After outputting the voltage regulation control strategy, it includes: Execute the activation record of the dynamic modulation unit and generate the activation record database; The activation record database is indexed by time vector, and the time vector matrix fragments corresponding to each round of voltage regulation strategy are mapped to historical lattices. Before re-executing the voltage regulation control strategy for optimization, time similarity analysis is performed using historical lattices to construct self-evolutionary feedback. The self-evolutionary feedback is used to conduct a new round of optimization management of voltage regulation control strategy.

6. The voltage stability control method for a high-proportion distributed power generation distribution network as described in claim 1, characterized in that, Temporal reparameterization is performed using the temporal convolution kernel within the temporal evolution operator to establish a temporal vector matrix, including: The time evolution operator is as follows: in, This is the state vector after time evolution. Characterization time The node summary status below, For time-coherent feature vectors, To drive the mapping matrix, For the length of time memory, For time-delay index, The temporal convolution kernel matrix, This is the neighborhood coupling weight coefficient matrix. Representing the neighborhood set, Characterizing the time-backtracking input components, Representing neighboring nodes In time The state vector below, For bias terms, It is a non-linear activation function; The calculated state vectors are combined column by column to form a time vector matrix.

7. The voltage stability control method for a high-proportion distributed power generation distribution network as described in claim 1, characterized in that, Performing energy feedback mapping in the information-time crystal structure further includes: When performing energy feedback mapping in the information-time crystal structure, the energy offset of each node is calculated; Nodes whose energy offset exceeds a preset threshold are marked as abnormal nodes, and abnormal early warning and management are implemented.

8. A voltage stability control system for a high-proportion distributed power generation distribution network, characterized in that, The system is used to implement the high-proportion distributed power generation distribution network voltage stability control method according to any one of claims 1-7, the system comprising: The state information flow construction module is used to construct the state information flow of multiple nodes in the distribution network. The state information flow includes node voltage, phase angle, power attention value and inverter control command signal. The information time crystal structure is constructed using time evolution operator. A stable time lattice generation module is used to perform energy feedback mapping in the information time crystal structure, treat node voltage perturbations as crystal defects, and generate a stable time lattice based on the periodic reconstruction of crystal defects. The control strategy output module is used to identify local unstable regions by utilizing the amplitude and frequency shift of the stable time lattice, and then perform periodic modulation and phase-locked optimization of the inverter control signal to output a voltage regulation control strategy.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the high-proportion distributed power distribution network voltage stability control method as described in any one of claims 1-7.

Images