Energy storage vsg-based passive load power supply management method and system

Abstract

The application provides a passive load power supply management method and system based on energy storage VSG, and relates to the technical field of power supply management.The method comprises the following steps: collecting electrical parameters of an energy storage VSG grid-connected point, and extracting a characteristic vector through multi-scale time-frequency analysis; performing current load disturbance mode identification, matching control parameters according to the load disturbance category, obtaining load energy demand, searching for a power supply strategy according to the load energy demand combined with energy storage state parameters, determining an energy supply timing strategy, and performing equipment power supply management control.The technical problem that in the prior art, the power supply management of the energy storage VSG to the passive load relies on passive correction of instantaneous errors, resulting in insufficient matching degree of the power supply strategy and the real demand of the load, is solved.Through multi-scale time-frequency analysis, the load disturbance characteristics are accurately identified, and forward-looking timing power distribution is performed in the energy demand direction, so that the technical effect of changing from passive error correction to active energy planning management is achieved, and the accuracy of energy storage power supply is improved.

Description

Technical Field

[0001] This invention relates to the field of power supply management technology, and specifically to a passive load power supply management method and system based on energy storage VSG. Background Technology

[0002] Virtual synchronous generators (VSGs) with energy storage, as a power electronic conversion technology that simulates the inertia and damping characteristics of synchronous generators, have been widely used in the power supply management of passive loads in microgrids. In existing technologies, the power supply control of passive loads using VSGs with energy storage typically employs a linear feedback regulation method based on the instantaneous electrical quantity deviation at the grid connection point. This involves detecting the deviation between the current grid connection point frequency and the rated frequency, and the deviation between the voltage amplitude and the rated voltage, and then calculating the power adjustment amount using a PI controller to adjust the output of the energy storage converter. This method is driven by real-time errors, resulting in control actions lagging behind the moment of disturbance occurrence. Furthermore, it can only respond to the instantaneous deviation at the current sampling moment, lacking an active identification mechanism for the internal characteristics of load power fluctuations and the ability to proactively plan for the total load energy demand and its temporal distribution within a preset time window. When various types of passive loads, such as industrial motor start-stop, welding machine switching, and reciprocating compressor periodic load changes, are connected to a microgrid, the load power exhibits complex fluctuation characteristics of multiple scales and modes. Existing linear feedback control methods cannot distinguish the inherent differences in power supply demand caused by different types of disturbances, and can only use uniform control parameters to deal with all operating conditions. This results in insufficient matching between the power supply strategy and the actual energy demand of the load, and reduces the power quality of the power supply.

[0003] In existing technologies, energy storage VSG relies on passive correction of instantaneous errors for power supply management of passive loads and lacks active identification and timing planning of load energy demand characteristics, resulting in technical problems such as insufficient matching between power supply strategy and actual load demand. Summary of the Invention

[0004] The purpose of this application is to provide a passive load power supply management method and system based on energy storage VSG, which solves the technical problems in the prior art where the power supply management of passive loads by energy storage VSG relies on passive correction of instantaneous errors and lacks active identification and timing planning of load energy demand characteristics, resulting in insufficient matching between power supply strategy and actual load demand.

[0005] In view of the above problems, this application provides a passive load power supply management method and system based on energy storage VSG.

[0006] The first aspect of this application provides a passive load power supply management method based on energy storage VSG. The method includes: real-time acquisition of electrical parameters at the grid connection point of the energy storage VSG; extraction of feature vectors of load power fluctuations through multi-scale time-frequency analysis; identification of the current load disturbance mode based on the feature vectors to determine the load disturbance category; matching of control parameters based on the load disturbance category and performing energy analysis to obtain the load energy demand; searching for power supply strategies based on the load energy demand and energy storage status parameters to determine an energy supply timing strategy; and managing and controlling the power supply of each passive load electrical device according to the energy supply timing strategy.

[0007] Optionally, electrical parameters of the VSG grid connection point are collected in real time, including three-phase voltage and three-phase current; active power and reactive power of the grid connection point are calculated based on the three-phase voltage and three-phase current; multi-scale time-frequency analysis is performed based on the active power and / or reactive power of the grid connection point to extract the energy distribution characteristics of load power fluctuations in different frequency bands and form a feature vector.

[0008] Optionally, wavelet packet multi-scale decomposition is performed on the active power signal at the grid connection point, and the signal is split into a preset number of decomposition layers to obtain multiple sub-band components covering different frequency ranges; each sub-band component is reconstructed, and the sum of squares of each reconstructed signal is calculated as the energy feature value of the corresponding frequency band; the energy feature values ​​of all sub-bands are normalized to obtain the proportion of each frequency band's energy to the total energy; the proportion values ​​are arranged and combined in order of frequency band frequency from low to high to form the feature vector.

[0009] Optionally, the feature vector is matched with the feature relationship of each preset disturbance category to obtain the feature relationship matching degree; based on the feature relationship matching degree, the preset disturbance category with the highest matching degree is determined as the current load disturbance category.

[0010] Optionally, the preset disturbance categories include at least: slow drift, step impact, periodic fluctuation, and fault transient.

[0011] Optionally, based on the current load disturbance category and the current operating status of the energy storage VSG, the required output active power and reactive power to maintain the system frequency and voltage within the allowable range within a preset time window are predicted; the predicted active power is integrated over time to obtain the active power demand; the predicted output reactive power is integrated over time to obtain the reactive power demand; and the load energy demand is obtained based on the active power demand and the reactive power demand.

[0012] Optionally, the current operating status parameters of the energy storage VSG are obtained, including the current grid connection point frequency, the current grid connection point voltage amplitude, the current output active power, and the current output reactive power. Based on the frequency deviation between the current grid connection point frequency and the rated frequency, and the voltage deviation between the current grid connection point voltage amplitude and the rated voltage, combined with the control parameters corresponding to the load disturbance category, the active power adjustment required to eliminate the frequency deviation and the reactive power adjustment required to eliminate the voltage deviation are calculated respectively. The current output active power plus the active power adjustment is used as the predicted active power within a future preset time window, and the current output reactive power plus the reactive power adjustment is used as the predicted reactive power within a future preset time window.

[0013] Optionally, based on the load energy demand, obtain the load energy demand curve of a single passive load within a future preset time window. The load energy demand curve describes the trajectory of the active and reactive power values ​​that the passive load needs to inject at different times over time. Obtain the current state parameters and supply chain characteristic parameters of the energy storage system. The supply chain characteristic parameters include the response delay time between the energy storage converter receiving the command and the actual output power, and the maximum ramp rate limit of the energy storage system's power change rate. Using the load energy demand curve as the power supply target curve, the response delay time as the time offset, and the maximum ramp rate limit as the power change constraint, perform demand-supply timing matching: shift the power supply target curve forward along the time axis by the response delay time to obtain a pre-start curve, so that the energy storage system can start up in real time. The system responds in advance to offset the impact of link delays before actual demand arrives; the pre-start curve is compared with the current maximum discharge power limit and maximum charge power limit of the energy storage system to determine whether the power demand at each moment on the pre-start curve exceeds the instantaneous supply capacity of the energy storage system; the time period exceeding the instantaneous supply capacity is marked as the power over-limit interval, and the power demand within the power over-limit interval is decomposed, wherein the excess power demand is moved forward to the available supply period before the over-limit interval for pre-supply, or moved backward to the available supply period after the over-limit interval for supplementary supply, or the power demand amplitude in the corresponding period is reduced to adapt to the supply capacity of the energy storage system; the decomposed power demand is rearranged in chronological order to establish the energy supply timing strategy, which includes the active power command value and reactive power command value at each moment.

[0014] Optionally, the individual energy demand curves of each passive load within a preset future time window are obtained, and all individual energy demand curves are superimposed on the same time axis to generate a multi-load comprehensive demand distribution curve. The time overlap relationship between the demand periods of each load is detected. When multiple loads have energy demand in the same time period, the corresponding time period is marked as the demand overlap interval, and the sum of the demand amplitudes of each load in the overlap interval is calculated as the total demand amplitude of the corresponding time period. Based on the preset priority level of each load and the comparison result between the total demand amplitude and the maximum supply capacity of the energy storage system, a multi-load coordinated power supply strategy is generated: when the total demand amplitude does not exceed the maximum supply capacity of the energy storage system, the power command is superimposed according to the original demand curves of each load; when the total demand amplitude exceeds the maximum supply capacity of the energy storage system, part or all of the power demand of the low-priority loads in the overlap interval is shifted to the non-overlapping time period for peak-shifting power supply, and the power commands of each load after peak-shifting adjustment are accumulated at the same time to synthesize a unified active power total command sequence and reactive power total command sequence to obtain the energy supply timing strategy under the multi-load scenario.

[0015] A second aspect of this application provides a passive load power supply management system based on energy storage VSG. The system includes: a feature extraction module for real-time acquisition of electrical parameters at the VSG grid connection point and extraction of feature vectors for load power fluctuations through multi-scale time-frequency analysis; a disturbance category determination module for identifying the current load disturbance mode based on the feature vectors and determining the load disturbance category; an energy demand acquisition module for matching control parameters based on the load disturbance category and performing energy analysis to obtain the load energy demand; and a power supply management module for searching power supply strategies based on the load energy demand and energy storage status parameters, determining an energy supply timing strategy, and managing and controlling the power supply of each passive load electrical device according to the energy supply timing strategy.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages: By collecting electrical parameters of the VSG grid-connected point in real time, feature vectors of load power fluctuations are extracted through multi-scale time-frequency analysis. Based on these feature vectors, current load disturbance modes are identified to determine the load disturbance category. Control parameters are matched according to the load disturbance category, and energy analysis is performed to obtain the load energy demand. Based on the load energy demand and energy storage status parameters, a power supply strategy search is conducted to determine the energy supply timing strategy. Power supply management and control for each passive load electrical device are then performed according to this strategy. This achieves the technical effect of accurately identifying load disturbance characteristics through multi-scale time-frequency analysis and performing forward-looking timing power allocation based on energy demand, realizing a shift from passive error correction to proactive energy planning and management, and improving the accuracy and adaptability of energy storage power supply.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the passive load power supply management method based on energy storage VSG provided in this application.

[0020] Figure 2 The passive load energy demand curve in the passive load power supply management method based on energy storage VSG provided in this application.

[0021] Figure 3 A schematic diagram of the passive load power supply management system based on energy storage VSG provided in this application.

[0022] Figure labeling: Feature extraction module 11, disturbance category determination module 12, energy demand acquisition module 13, power supply management module 14. Detailed Implementation

[0023] This application provides a passive load power supply management method and system based on energy storage VSG, addressing the technical problems in existing technologies where energy storage VSG relies on passive correction of instantaneous errors for passive load power supply management, lacks proactive identification and timing planning of load energy demand characteristics, resulting in insufficient matching between power supply strategies and actual load demand. It achieves the technical effect of accurately identifying load disturbance characteristics through multi-scale time-frequency analysis and performing forward-looking timing power allocation based on energy demand, realizing a shift from passive error correction to proactive energy planning and management, and improving the accuracy and adaptability of energy storage power supply.

[0024] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0025] Example 1, as Figure 1 As shown, this application provides a passive load power supply management method based on energy storage VSG, the passive load power supply management method based on energy storage VSG includes: The electrical parameters of the VSG grid connection point are collected in real time, and the feature vector of load power fluctuation is extracted through multi-scale time-frequency analysis.

[0026] Furthermore, the electrical parameters of the VSG grid connection point are collected in real time, and the feature vector of load power fluctuation is extracted through multi-scale time-frequency analysis. This includes: collecting the electrical parameters of the VSG grid connection point in real time, including three-phase voltage and three-phase current; calculating the active power and reactive power of the grid connection point based on the three-phase voltage and three-phase current; and performing multi-scale time-frequency analysis based on the active power and / or reactive power of the grid connection point to extract the energy distribution characteristics of load power fluctuation in different frequency bands, thus forming a feature vector.

[0027] Specifically, by installing three-phase voltage transformers and three-phase current transformers at the electrical connection nodes between the energy storage virtual synchronous generator and the passive load power supply bus at the energy storage VSG grid connection point, real-time synchronous sampling is performed at a sampling frequency of not less than 128 times the grid base frequency (50Hz), i.e., above 6.4kHz, to obtain the electrical parameters of the energy storage VSG grid connection point, including the three-phase voltages ua(t), ub(t), uc(t) of A, B, and C, and the three-phase currents ia(t), ib(t), ic(t).

[0028] Based on the three-phase voltage and three-phase current, the real-time active power p(t) and reactive power q(t) at the grid connection point are calculated using instantaneous power theory. Wavelet packet transform is preferably used to perform multi-scale time-frequency analysis based on the active power and / or reactive power at the grid connection point, extracting the energy distribution characteristics of load power fluctuations in different frequency bands to form a feature vector that reflects the energy distribution characteristics of load fluctuations in different frequency ranges.

[0029] By performing multi-scale time-frequency analysis on the electrical parameters of the VSG grid connection point, frequency domain energy characteristics of different load disturbances are extracted to generate feature vectors, providing accurate input for subsequent disturbance identification and energy analysis. This enables the VSG to accurately adjust its control strategy according to different disturbance categories, thereby achieving adaptive intelligent power supply management.

[0030] Furthermore, multi-scale time-frequency analysis is performed to extract the energy distribution characteristics of load power fluctuations in different frequency bands, forming a feature vector. This includes: performing wavelet packet multi-scale decomposition on the active power signal at the grid connection point, splitting the signal layer by layer to a preset number of decomposition layers to obtain multiple sub-frequency band components covering different frequency ranges; reconstructing each sub-frequency band component separately, calculating the sum of squares of each reconstructed signal as the energy feature value of the corresponding frequency band; normalizing the energy feature values ​​of all sub-frequency bands to obtain the proportion of energy in each frequency band to the total energy; and arranging and combining the proportions in ascending order of frequency band frequency to form the feature vector.

[0031] Specifically, wavelet packet transform is used to perform multi-scale decomposition of the active power signal at the grid connection point. First, the number of decomposition levels N is set, such as N=5, and a mother wavelet basis function is selected, such as the db4 or sym8 wavelet. The original active power signal p(t) is simultaneously low-pass filtered in the first level (transfer function H) and high-pass filtered (transfer function G), and then downsampled by a factor of 2 to obtain approximate coefficients A1 and detail coefficients D1. In the second level, A1 and D1 are low-pass and high-pass filtered respectively, and then downsampled by a factor of 2 to obtain four coefficients: AA2, AD2, DA2, and DD2. This process is repeated layer by layer until the Nth level, generating a total of 2... N Group wavelet packet coefficient sequence {cj k j=1,2,...,2 N}, where k represents the number of the k-th sub-band, cj k This represents the wavelet packet coefficient sequence corresponding to the j-th group and k-th sub-band after wavelet packet decomposition, where c is the wavelet packet coefficient. The length of each group of wavelet packet coefficient sequences is 1 / 2 the length of the original active power signal. N Each set of wavelet packet coefficients corresponds to a specific frequency range, the width of which is fs / 2. N+1fs is the sampling rate, and the coverage range is from 0Hz to fs / 2. Each sub-band is arranged in order of increasing center frequency to obtain multiple sub-band components covering different frequency ranges.

[0032] Then for each sub-band component cj k To perform single-branch reconstruction, only the coefficients of that set are used, and all other coefficients are set to zero. The time-domain waveform is then recovered through inverse wavelet packet transform, yielding the reconstructed signal components pk(t) corresponding to each sub-band, k=1,2,...,2 N Since the wavelet packet transform is an orthogonal transform, the original active power signal p(t) is equal to the sum of the reconstructed components, i.e. |p(t)| 2 dt=Σ |pk(t)| 2 Based on this, the energy eigenvalues ​​of each sub-band component are calculated. t is the sampling time, M is the total number of sampling points, Δt is the sampling time interval, which is equal to the reciprocal of the sampling frequency, and Ek is the energy characteristic value of the k-th sub-band, reflecting the energy magnitude of load power fluctuations at different frequency scales: low frequency band (e.g., 0~0.1Hz) corresponds to slow drift disturbances of the load, such as changes in electricity consumption between day and night; mid frequency band (0.1Hz~5Hz) corresponds to periodic fluctuation disturbances, such as periodic load changes of motors; and high frequency band (>5Hz) corresponds to step impact or transient disturbances of faults, such as starting of large motors and short circuit faults.

[0033] The energy eigenvalues ​​of all sub-bands are normalized. First, the sum of the energy eigenvalues ​​of all sub-bands is calculated. Then, the ratio of the energy eigenvalue of each sub-band to the sum of the energy is calculated to achieve normalization. These ratios are then arranged and combined in ascending order of frequency to form an eigenvector. The same steps are used for multi-scale time-frequency analysis of reactive power.

[0034] For example, a large asynchronous motor is connected to the grid connection point of an energy storage VSG. Voltage and current are collected in real time at a sampling rate of 6.4kHz, i.e., 128 sampling points per cycle, and the active power time-series signal p(t) is calculated. During a certain time period, the motor suddenly starts, simultaneously superimposed with the periodic torque pulsations of the motor during normal operation. Setting the wavelet packet decomposition level N=5, a total of 32 sub-bands are obtained, each with a bandwidth of 6.4kHz / 2. 6=100Hz. After decomposition, reconstruction, and energy calculation, the energy Ek of each frequency band is obtained. Among them, the energy of frequency band 1 (0~100Hz, corresponding to extremely low frequency slow change) is E1=200; the energy of frequency band 5 (400~500Hz, corresponding to periodic fluctuation) is E5=3500; the energy of frequency band 28 (2700~2800Hz, corresponding to the high frequency component of the impulse transient) is E28=8000, and the energy values ​​of the remaining frequency bands are all small, totaling 300. The total energy is calculated as E=200+3500+8000+300=12000. The normalized energy proportions are: r1=200 / 12000≈0.017, r5=3500 / 12000≈0.292, r28=8000 / 12000≈0.667. The final eigenvector V = [0.017, 0.000, ..., 0.292, ..., 0.667, ..., 0.000] T The feature vector has 32 dimensions. The energy distribution of the feature vector shows that the high-frequency band r28 is absolutely dominant, while the mid-frequency band r5 also has significant energy, corresponding to the periodic torque pulsation during motor operation.

[0035] By employing the complete decomposition characteristics of orthogonal wavelet packets, the energy of different types of load disturbances in the time-domain active signal is accurately separated into corresponding sub-frequency bands. Through reconstruction, energy statistics, and normalization sorting, standardized feature vectors are generated, which not only fully preserves the disturbance details of each frequency band, but also achieves the regularization and unification of feature dimensions, providing a reliable input with high discriminative power for subsequent high-precision matching and identification of different load disturbance modes.

[0036] Based on the feature vector, the current load disturbance mode is identified to determine the load disturbance category.

[0037] Furthermore, based on the feature vector, current load disturbance mode identification is performed to determine the load disturbance category, including: matching the feature vector with the feature relationships of each preset disturbance category to obtain the feature relationship matching degree; and determining the preset disturbance category with the highest matching degree as the current load disturbance category based on the feature relationship matching degree.

[0038] Furthermore, the preset disturbance categories include at least: slow drift, step impact, periodic fluctuation, and fault transient.

[0039] Specifically, for four preset load disturbances—slow drift, step impact, periodic fluctuation, and fault transient—a large number of typical load power time-series samples, generated from actual measurements or simulations, are collected and analyzed using multi-scale time-frequency analysis. The corresponding eigenvectors are extracted and denoted as standard eigenvectors. For each disturbance type, the arithmetic mean of the eigenvectors of all samples is taken to obtain the central eigenvector Vm for that disturbance type. Simultaneously, the covariance matrix Cm of the eigenvectors for that type is calculated. A one-to-one correspondence is established between the central eigenvector Vm, the covariance matrix Cm, and each disturbance type, constructing a standard feature relation library for the preset disturbance categories. Among them, slow drift refers to a slow change in load power that shows a monotonically increasing or decreasing trend over a relatively long time scale, such as preheating of industrial furnaces or the gradual switching on of office building lights during working hours. Step impact refers to a sudden change in load power that causes a large increase or decrease in a very short time, such as the direct starting of a high-power motor or the switching on of a welding machine. Periodic fluctuation refers to a load power that exhibits regular oscillation characteristics, the frequency of which is usually related to the mechanical speed of the equipment or the switching frequency of power electronic devices, such as the periodic load changes caused by the piston movement of a reciprocating compressor or the periodic pulsation driven by a frequency converter. Fault transient refers to a power transient impact with extremely high amplitude and extremely short duration caused by abnormal operating conditions such as power grid short circuits and grounding faults, the spectral characteristics of which often show wide bandwidth and high energy.

[0040] The system calls upon a pre-stored standard feature relation library to match feature vectors with feature relations for each preset perturbation category, preferably using Mahalanobis distance as the quantification of matching degree. Where Vcur is the feature vector and Dm is the Mahalanobis distance. The Mahalanobis distance uses the inverse of the covariance matrix to weight each dimension, eliminating the dimensional differences and correlation effects between the energy proportions of different frequency bands, thus improving the accuracy and reliability of the matching results. The Mahalanobis distance is converted into a matching score Sm = exp(-Dm) through a negative exponential mapping; the closer the matching score is to 1, the higher the matching degree.

[0041] After completing the matching degree calculation for the four preset disturbance categories, multiple matching degree scores are obtained. The preset disturbance category with the highest matching degree is determined as the current load disturbance category.

[0042] For example, 200 typical power time-series samples were collected for each of four types of disturbances: slow drift, step impact, periodic fluctuation, and fault transient. After extracting feature vectors through multi-scale time-frequency analysis, the arithmetic mean of each was taken to obtain four sets of central feature vectors. Only the top three frequency bands with significant energy proportions were listed, while the other dimensions were approximately zero. Among them, the central feature vector of the step impact type was V1=[r1=0.02,r5=0.20,r28=0.65]. TThe corresponding covariance matrix C1 represents the degree of dispersion of this type of sample across various frequency bands. The central eigenvector of the periodic fluctuation class is V2=[r1=0.05,r5=0.70,r28=0.10]. T The central feature vector of the slowly drifting class is V3=[r1=0.75,r5=0.08,r28=0.02]. T The central feature vector of the fault transient class is V4 = [r1 = 0.02, r5 = 0.25, r28 = 0.60]. T During online operation, the current feature vector extracted in actual measurement is matched with the feature relationship of the preset disturbance category. The Mahalanobis distance is calculated to obtain: D1=0.52, corresponding to the matching degree score S1=exp(-0.52)=0.595, D2=2.80, S2=exp(-2.80)=0.061, D3=4.10, S3=exp(-4.10)=0.017, D4=1.35, S4=exp(-1.35)=0.259. After comparison, the maximum matching degree Smax=0.595 is selected as the step impact category, and the current load disturbance category is determined to be the step impact category.

[0043] By matching the feature relationships of preset disturbance categories, rapid disturbance identification in unsupervised scenarios is achieved, avoiding the problem of poor generalization of traditional threshold determination methods, and providing accurate category basis for subsequent control parameter matching and energy demand analysis.

[0044] Control parameters are matched according to the load disturbance category, and energy analysis is performed to obtain the load energy demand.

[0045] Furthermore, control parameters are matched according to the load disturbance category, and energy analysis is performed to obtain the load energy demand, including: predicting the output active power and reactive power required to maintain the system frequency and voltage within the allowable range within a preset time window based on the current load disturbance category and the current operating state of the energy storage VSG; integrating the predicted active power over time to obtain the active power demand; integrating the predicted output reactive power over time to obtain the reactive power demand; and obtaining the load energy demand based on the active power demand and reactive power demand.

[0046] Furthermore, predicting the output active and reactive power required to maintain the system frequency and voltage within the allowable range within a future preset time window includes: acquiring the current operating status parameters of the energy storage VSG, which include the current grid connection point frequency, the current grid connection point voltage amplitude, the current output active power, and the current output reactive power; calculating the active power adjustment required to eliminate the frequency deviation and the reactive power adjustment required to eliminate the voltage deviation based on the frequency deviation between the current grid connection point frequency and the rated frequency, and the voltage deviation between the current grid connection point voltage amplitude and the rated voltage, combined with the control parameters corresponding to the load disturbance category; using the current output active power plus the active power adjustment as the predicted active power within the future preset time window, and using the current output reactive power plus the reactive power adjustment as the predicted reactive power within the future preset time window.

[0047] Specifically, the current operating status of the energy storage VSG is obtained through the frequency measurement unit, voltage transformer, and current transformer at the grid connection point. This includes the current grid connection point frequency, current grid connection point voltage amplitude, current output active power, and current output reactive power. Combined with the system reference values ​​of a 50Hz rated frequency and a 220V rated phase voltage, the real-time frequency deviation and voltage deviation are calculated. The frequency deviation is the difference between the current grid connection point frequency and the rated frequency, and the voltage deviation is the difference between the current grid connection point voltage amplitude and the rated voltage.

[0048] Based on the control parameters corresponding to the frequency deviation, voltage deviation, and load disturbance category, the active power adjustment required to eliminate the frequency deviation and the reactive power adjustment required to eliminate the voltage deviation are calculated respectively. The active power adjustment ΔP required to eliminate the frequency deviation Δf is calculated using the following formula: ΔP = -Kp Δf-J dΔf / dt, where -Kp Δf is the proportional term used to respond to the steady-state frequency deviation, where Kp is the active-frequency droop coefficient, obtained by looking up the disturbance-parameter mapping table, -J dΔf / dt is the differential term used to respond to the rate of change of frequency, i.e., inertia support. J is the virtual inertia coefficient. dΔf / dt is obtained by numerically differentiating the frequency deviation sequence, such as by real-time calculation using the three-point difference method.

[0049] Similarly, based on the reactive power-voltage control characteristics of VSG, the reactive power adjustment ΔQ required to eliminate voltage deviation ΔV is calculated using the following formula: ΔQ = -Kq ΔV-Ki ΔV dt, where -Kq ΔV is the proportional term, representing the instantaneous voltage deviation in response; Kq is the reactive power-voltage droop coefficient, obtained by looking up the disturbance-parameter mapping table; -Ki ΔV dt is the integral term, used to eliminate steady-state voltage error and ensure error-free voltage regulation, and Ki is the integral coefficient.

[0050] The disturbance-parameter mapping table is pre-established through offline simulation or engineering tuning. Different types of load disturbances have different dynamic response requirements for the VSG. For slow drift disturbances, the VSG has power tracking capability over a longer time scale, focusing on steady-state error-free regulation. For step-impact disturbances, the VSG needs to provide instantaneous inertia support to suppress frequency abrupt changes, focusing on enhancing the virtual inertia J. For periodic fluctuation disturbances, the VSG has damping absorption capability at specific frequency points, focusing on matching the damping coefficient D and setting the notch frequency of the filter. For fault transient disturbances, the VSG needs to quickly enter current-limiting mode and provide voltage support, focusing on dynamically amplifying the reactive power compensation coefficient. For example, a detailed electromagnetic transient simulation model including energy storage VSG, typical load, and power grid is built in an offline simulation platform such as Matlab or PSCAD. Multiple parameter combinations are set for each type of disturbance, and frequency sweep simulation is performed. The dynamic response indicators of the system frequency and voltage under each parameter set are recorded, including maximum frequency deviation, maximum voltage deviation, recovery time, overshoot, and number of oscillations. The parameter combination that minimizes the comprehensive performance index (i.e., the weighted sum of squares of all deviations) is searched in the feasible parameter space and used as the initial recommended parameters for that disturbance category. Then, these initial parameters are deployed to the actual energy storage VSG control system. Hardware-in-the-loop testing or field load testing is conducted under the condition of connecting a real load. The parameters are fine-tuned based on the measured frequency / voltage response waveforms. If the field test finds that the overshoot is too large under a certain parameter set, Kp is appropriately decreased or D is increased. This process is repeated multiple times until the dynamic response meets the requirements of standards such as the energy storage converter technical specifications. The optimal parameter combinations for each disturbance category, after simulation optimization and engineering verification, are then solidified into a disturbance-parameter mapping table. Each disturbance category in the disturbance-parameter mapping table corresponds to a set of VSG control parameters {J,D,Kp,Kq,Tf}, where J is the virtual inertia in kg·m. 2 , where D is the damping coefficient (N·m·s / rad), Kp is the active power-frequency droop coefficient (W / Hz, used for primary frequency regulation), Kq is the reactive power-voltage droop coefficient (Var / V, used for primary voltage regulation), and Tf is the low-pass filter time constant.

[0051] After obtaining the power adjustment amounts ΔP and ΔQ at the current moment, they are superimposed on the current output active power Pout and reactive power Qout, respectively, to obtain the predicted active power Pref(t) and the predicted reactive power Qref(t) within the preset time window Tw. The preset time window Tw is preferably 5-10 cycles. Pref(t) = Pout + ΔP, Qref(t) = Qout + ΔQ.

[0052] Then, the predicted active power within the predicted future time window is numerically integrated over time to obtain the active power demand Ep. Similarly, by integrating the predicted output reactive power over time, the reactive power demand Eq is obtained. The functional and non-functional energy demands are integrated to form the load energy demand, which serves as the target basis for energy storage dispatch.

[0053] By transforming the identified qualitative disturbance categories into quantitative energy demand indicators that can be directly used for energy storage scheduling, power prediction is completed based on the adaptive control parameters for different disturbances. Then, the precise conversion from instantaneous power to cumulative energy is achieved through time-domain integration. This not only avoids the problem of poor adaptability of traditional fixed-parameter prediction in multi-disturbance scenarios, but also provides accurate quantitative input for the power supply planning of subsequent energy storage VSG, ensuring the effectiveness and accuracy of power supply management.

[0054] Based on the load energy demand and energy storage status parameters, a power supply strategy search is performed to determine the energy supply timing strategy, and the power supply management and control of each passive load electrical equipment is carried out in accordance with the energy supply timing strategy.

[0055] Furthermore, based on the load energy demand and energy storage status parameters, a power supply strategy search is performed to determine the energy supply timing strategy, including: obtaining the load energy demand curve of a single passive load within a future preset time window based on the load energy demand, the load energy demand curve describing the trajectory of the active and reactive power values ​​that the passive load needs to inject at different times as a function of time; obtaining the current status parameters and supply chain characteristic parameters of the energy storage system, the supply chain characteristic parameters including the response delay time between the energy storage converter receiving the command and the actual output power, and the maximum ramp rate limit of the energy storage system power change rate; using the load energy demand curve as the power supply target curve, the response delay time as the time offset, and the maximum ramp rate limit as the power change constraint, performing demand-supply timing matching: shifting the power supply target curve forward along the time axis by the response delay time. A pre-start curve is obtained by delaying the time, allowing the energy storage system to respond in advance before the actual demand arrives to offset the impact of link delay. The pre-start curve is compared with the current maximum discharge power limit and maximum charge power limit of the energy storage system to determine whether the power demand at each moment on the pre-start curve exceeds the instantaneous supply capacity of the energy storage system. The time period exceeding the instantaneous supply capacity is marked as the power over-limit interval, and the power demand within the power over-limit interval is decomposed. Specifically, the excess power demand is either moved forward to the available supply period before the over-limit interval for pre-supply, or moved backward to the available supply period after the over-limit interval for supplementary supply, or the power demand amplitude in the corresponding period is reduced to adapt to the supply capacity of the energy storage system. The decomposed power demand is rearranged in chronological order to establish the energy supply timing strategy, which includes the active power command value and reactive power command value at each moment.

[0056] Specifically, for a single passive load, based on the active and reactive power demands within the load's energy requirements, these demands are ordered chronologically to obtain a load energy demand curve for the individual passive load within a preset future time window. This load energy demand curve describes the trajectory of the active and reactive power values ​​that the passive load needs to inject at different times over time. For example, for a single air conditioning passive load, its active and reactive power demand time-series data are extracted, arranged chronologically, and a preset time window of 100ms is used. Ten time points are divided with a sampling interval of 10ms each, yielding the active and reactive power demands at each time point: t=0ms: active power 0.42kW, reactive power 0.18kvar; t=10ms: active power 0.43kW, reactive power 0.18kvar; t=20ms: active power 0.41kW, reactive power 0.19kvar; t=30ms: active power 0.44kW, reactive power 0.17kvar; t=40ms: active power 0.44kW, reactive power 0.17kvar; t=40ms: active power 0.42kW, reactive power 0.18kvar; t=40ms: active power 0.42kW, reactive power 0.18kvar; t=20ms: active power 0.41kW, reactive power 0.19kvar; t=30ms: active power 0.44kW, reactive power 0.17kvar; t=40ms: active power 0.44kW, reactive power 0.17kvar; t=40ms: active power 0.42kW, reactive power 0.18kvar. The active power is 0.43kW and reactive power is 0.18kvar, t=50ms; the active power is 0.42kW and reactive power is 0.19kvar, t=60ms; the active power is 0.41kW and reactive power is 0.18kvar, t=70ms; the active power is 0.43kW and reactive power is 0.17kvar, t=80ms; the active power is 0.44kW and reactive power is 0.18kvar, t=90ms; the active power is 0.42kW and reactive power is 0.19kvar. Based on the above ordered active and reactive power demand data, plot the load energy demand curve for this passive load within a 100ms time window, as follows: Figure 2 As shown, the horizontal axis represents time (ms), the left Y-axis represents P / kW (corresponding to the solid line of active power), and the right Y-axis represents Q / kvar (corresponding to the dashed line of reactive power). The energy demand curve of the air conditioning load fully reflects the continuous change of the active and reactive power required by the passive load of the air conditioning at each sampling moment in the preset time window over time.

[0057] The current state parameters of the energy storage system are obtained through the Battery Management System (BMS), including the current State of Charge (SOC), the current State of Health (SOH), the current maximum discharge power limit, and the current maximum recharge power limit. Simultaneously, supply chain characteristic parameters are obtained, including: the response delay time between the energy storage converter (PCS) receiving a command and the actual output power, and the maximum ramp rate limit of the energy storage system's power change rate. The response delay time is composed of the PCS's communication link delay, PWM modulation delay, IGBT switching delay, etc., and is, for example, 2~10ms, obtained through factory testing. The maximum ramp rate limit of the energy storage system's power change rate is the maximum change in energy storage output power per unit time, limited by the PCS's DC-side capacitor charging and discharging rate and battery current change rate, typically 10%~30% of the rated power, and is provided by the PCS manufacturer.

[0058] Using the load energy demand curve as the target power supply curve, the response delay time as the time offset, and the maximum ramp rate limit as the power change constraint, demand-supply timing matching is performed: First, the target power supply curve is shifted forward along the time axis by the response delay time to obtain the pre-start curve. By advancing the command issuance time by the response delay time, the energy storage system completes power output precisely at the same moment as the actual load demand arrives, thereby offsetting the impact of link delay and achieving demand-supply synchronization. Then, the pre-start curve is compared with the current maximum discharge power limit and maximum charge power limit of the energy storage system. For each moment, it is determined whether the power demand on the pre-start curve exceeds the instantaneous supply capacity of the energy storage system. If the power demand on the pre-start curve is greater than the instantaneous supply capacity of the energy storage system, it is determined that there is a power over-limit at that moment, and the time period exceeding the instantaneous supply capacity is marked as the power over-limit interval.

[0059] Then, the power demand within the power over-limit range is decomposed and processed. Peak shaving and valley filling are implemented over time to adapt the power demand to the immediate supply capacity of the energy storage system. When there is available supply before the over-limit range—that is, the duration during which the pre-start-up curve is not exceeded and the energy storage system has sufficient remaining capacity to accommodate the excess—the excess power demand is shifted forward to the time before the over-limit range for pre-supply. This means releasing some electricity before the peak load demand arrives, smoothing out the peak and filling the valley period. When there is sufficient supply capacity after the over-limit range, the excess is shifted backward to after the over-limit range, delaying the supply of some electricity after the peak, thus achieving peak shaving and valley filling. When there is no available remaining capacity before or after the over-limit range, the power demand within the over-limit range is limited according to the maximum supply capacity of the energy storage system, reducing the peak amplitude to adapt to the immediate supply capacity of the energy storage system.

[0060] After decomposing all out-of-limit zones, the adjusted power demand sequence is rearranged in chronological order to generate a complete energy supply timing strategy. This strategy includes active and reactive power command values ​​to be sent to the energy storage converter at each time point. The energy supply timing strategy is then transmitted to the execution unit of the energy storage converter via a communication interface, such as Ethernet or fiber optic communication. The converter adjusts its output power according to the command values ​​at the corresponding times, achieving precise power supply management and control for passive load electrical equipment.

[0061] By offsetting hardware link delays through pre-start timing offsets and by detecting and decomposing over-limit operations, the power supply needs of the load are met to the maximum extent while ensuring that the energy storage system does not operate beyond its power limit. Furthermore, the optimal allocation of limited energy storage capacity is achieved through multi-load peak-shifting scheduling, thereby improving the accuracy, adaptability, and reliability of power supply.

[0062] Furthermore, the method also includes: acquiring the individual energy demand curves of each passive load within a preset future time window, superimposing all individual energy demand curves on the same time axis to generate a multi-load comprehensive demand distribution curve; detecting the time overlap relationship between the demand periods of each load, and when multiple loads have energy demand in the same time period, marking the corresponding time period as the demand overlap interval, and calculating the sum of the demand amplitudes of each load in the overlap interval as the total demand amplitude of the corresponding time period; generating a multi-load coordinated power supply strategy based on the preset priority level of each load and the comparison result between the total demand amplitude and the maximum supply capacity of the energy storage system: when the total demand amplitude does not exceed the maximum supply capacity of the energy storage system, outputting power commands according to the original demand curves of each load; when the total demand amplitude exceeds the maximum supply capacity of the energy storage system, shifting part or all of the power demand of low-priority loads in the overlap interval to non-overlapping time periods for peak-shifting power supply, and accumulating the power commands of each load after peak-shifting adjustment at the same time to synthesize a unified active power total command sequence and reactive power total command sequence to obtain an energy supply timing strategy under multi-load scenarios.

[0063] Specifically, when multiple passive loads exist, the individual energy demand curves of each passive load within a preset future time window are obtained. Then, all individual energy demand curves are superimposed point-by-point on the same time axis to generate a multi-load comprehensive demand distribution curve. The time overlap relationship of the load demands in the multi-load comprehensive demand distribution curve is detected. When multiple loads simultaneously have energy demands within the same time period, the corresponding time period is marked as the demand overlap interval, and the sum of the demand amplitudes of each load within the overlap interval is calculated and used as the total demand amplitude for the corresponding time period.

[0064] Based on the comparison between the preset priority level of each load and the total demand amplitude and the maximum supply capacity of the energy storage system, a multi-load coordinated power supply strategy is generated: When the total demand amplitude does not exceed the maximum supply capacity of the energy storage system, the output power command is superimposed according to the original demand curve of each load; when the total demand amplitude exceeds the maximum supply capacity of the energy storage system, part or all of the power demand of low-priority loads in the overlapping interval is shifted to the non-overlapping period for peak-shaving power supply, ensuring that the power supply of high-priority loads is not affected. After peak-shaving adjustment, the power commands of each load are accumulated at the same time and synthesized into a unified active power total command sequence and reactive power total command sequence to obtain the energy supply timing strategy under multi-load scenarios. Then, the energy supply timing strategy is sent to the execution unit of the energy storage converter through a communication interface, such as Ethernet or fiber optic communication. The converter adjusts its output power at the corresponding time according to the command value, realizing precise power supply management and coordinated control of multiple passive load electrical devices. The preset priority levels of loads are determined based on core safety dimensions, production / operation criticality dimensions, and general auxiliary dimensions. Loads directly related to system safety and personal safety are classified as Level 1, such as relay protection devices, emergency fire-fighting equipment, and critical monitoring sensors in passive microgrids. Loads that directly determine the continuity of core production processes are classified as Level 2, such as core process motors in industrial scenarios and life support equipment in medical scenarios. Loads that only provide auxiliary experience and non-core functions are classified as Level 3, such as general lighting and non-essential office auxiliary equipment. When power exceeds limits, the energy demand of high-level loads is strictly guaranteed according to the priority order to achieve optimal allocation of energy storage capacity in passive scenarios.

[0065] By accurately identifying overlapping periods of multiple load demands, instructions are directly superimposed when the total power does not exceed the limit. When the limit is exceeded, low load demands are shifted according to priority to avoid energy storage overload and ensure power supply to high priority loads, thereby improving equipment safety, accuracy and adaptability of energy storage power supply.

[0066] Example 2 is based on the same inventive concept as the passive load power supply management method based on energy storage VSG in the previous examples, such as... Figure 3 As shown, this application provides a passive load power supply management system based on energy storage VSG, wherein the passive load power supply management system based on energy storage VSG includes: The feature extraction module 11 is used to collect electrical parameters of the VSG grid connection point in real time and extract the feature vector of load power fluctuation through multi-scale time-frequency analysis; the disturbance category determination module 12 is used to identify the current load disturbance mode based on the feature vector and determine the load disturbance category; the energy demand acquisition module 13 is used to match control parameters based on the load disturbance category and perform energy analysis to obtain the load energy demand; the power supply management module 14 is used to search for power supply strategies based on the load energy demand and the energy storage status parameters, determine the energy power supply timing strategy, and perform power supply management and control of each passive load electrical device according to the energy power supply timing strategy.

[0067] Furthermore, the feature extraction module 11 is also used to: collect electrical parameters of the VSG grid connection point in real time, including three-phase voltage and three-phase current; calculate the active power and reactive power of the grid connection point based on the three-phase voltage and three-phase current; and perform multi-scale time-frequency analysis based on the active power and / or reactive power of the grid connection point to extract the energy distribution characteristics of load power fluctuations in different frequency bands and form a feature vector.

[0068] Furthermore, the feature extraction module 11 is also used to: perform wavelet packet multi-scale decomposition on the active power signal at the grid connection point, split the signal layer by layer to a preset number of decomposition layers, and obtain multiple sub-band components covering different frequency ranges; reconstruct each sub-band component separately, calculate the sum of squares of each reconstructed signal as the energy feature value of the corresponding frequency band; normalize the energy feature values ​​of all sub-bands to obtain the proportion of each frequency band energy to the total energy; and arrange and combine each proportion value in order of frequency band frequency from low to high to form the feature vector.

[0069] Furthermore, the disturbance category determination module 12 is also used to: match the feature vector with the feature relationship of each preset disturbance category to obtain the feature relationship matching degree; and determine the preset disturbance category with the highest matching degree as the current load disturbance category based on the feature relationship matching degree.

[0070] Furthermore, the disturbance category determination module 12 is also used to: include at least the following preset disturbance categories: slow drift, step impact, periodic fluctuation, and fault transient.

[0071] Furthermore, the energy demand acquisition module 13 is also used to: predict the output active power and reactive power required to maintain the system frequency and voltage within the allowable range within a preset time window based on the current load disturbance category and the current operating status of the energy storage VSG; integrate the predicted active power over time to obtain the active power demand; integrate the predicted output reactive power over time to obtain the reactive power demand; and obtain the load energy demand based on the active power demand and the reactive power demand.

[0072] Furthermore, the energy demand acquisition module 13 is also used to: acquire the current operating status parameters of the energy storage VSG, the current operating status parameters including the current grid connection point frequency, the current grid connection point voltage amplitude, the current output active power, and the current output reactive power; calculate the active power adjustment required to eliminate the frequency deviation and the reactive power adjustment required to eliminate the voltage deviation based on the frequency deviation between the current grid connection point frequency and the rated frequency, and the voltage deviation between the current grid connection point voltage amplitude and the rated voltage, combined with the control parameters corresponding to the load disturbance category; use the current output active power plus the active power adjustment as the predicted active power within a future preset time window, and use the current output reactive power plus the reactive power adjustment as the predicted reactive power within a future preset time window.

[0073] Furthermore, the power supply management module 14 is also used to: obtain the load energy demand curve of a single passive load within a future preset time window based on the load energy demand, wherein the load energy demand curve describes the trajectory of the active power value and reactive power value that the passive load needs to inject at different times as a function of time; obtain the current state parameters and supply chain characteristic parameters of the energy storage system, wherein the supply chain characteristic parameters include the response delay time between the energy storage converter receiving the command and the actual output of electrical energy, and the maximum ramp rate limit of the power change rate of the energy storage system; and perform demand-supply timing matching using the load energy demand curve as the power supply target curve, the response delay time as the time offset, and the maximum ramp rate limit as the power change constraint: shift the power supply target curve forward along the time axis by the response delay time to obtain the pre-startup curve. This allows the energy storage system to respond in advance before actual demand arrives to offset the impact of link delays. The pre-start curve is compared with the current maximum discharge power limit and maximum charge power limit of the energy storage system to determine whether the power demand at each moment on the pre-start curve exceeds the instantaneous supply capacity of the energy storage system. The time periods exceeding the instantaneous supply capacity are marked as power over-limit intervals, and the power demand within these intervals is decomposed. Specifically, the excess power demand is either moved forward to a supplyable period before the over-limit interval for pre-supply, or moved backward to a supplyable period after the over-limit interval for supplementary supply, or the power demand amplitude within the corresponding period is reduced to match the energy storage system's supply capacity. The decomposed power demand is then rearranged in chronological order to establish the energy supply timing strategy, which includes active power command values ​​and reactive power command values ​​at each moment.

[0074] Furthermore, the power supply management module 14 is also used to: acquire the individual energy demand curves of each passive load within a future preset time window, superimpose all individual energy demand curves on the same time axis to generate a multi-load comprehensive demand distribution curve; detect the time overlap relationship between the demand periods of each load, and when multiple loads have energy demand in the same time period, mark the corresponding time period as the demand overlap interval, and calculate the sum of the demand amplitudes of each load in the overlap interval as the total demand amplitude of the corresponding time period; generate a multi-load coordinated power supply strategy based on the preset priority level of each load and the comparison result between the total demand amplitude and the maximum supply capacity of the energy storage system: when the total demand amplitude does not exceed the maximum supply capacity of the energy storage system, output power commands according to the original demand curves of each load; when the total demand amplitude exceeds the maximum supply capacity of the energy storage system, shift part or all of the power demand of low-priority loads in the overlap interval to non-overlapping time periods for peak-shifting power supply, and accumulate the power commands of each load after peak-shifting adjustment at the same time to synthesize a unified active power total command sequence and reactive power total command sequence to obtain the energy supply timing strategy under the multi-load scenario.

[0075] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The passive load power supply management method and specific examples based on energy storage VSG in the aforementioned embodiment 1 are also applicable to the passive load power supply management system based on energy storage VSG in this embodiment. Through the foregoing detailed description of the passive load power supply management method based on energy storage VSG, those skilled in the art can clearly understand the passive load power supply management system based on energy storage VSG in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0076] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0077] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A passive load power supply management method based on energy storage VSG, characterized in that, include: Real-time acquisition of electrical parameters at the VSG grid connection point of energy storage, and extraction of feature vectors of load power fluctuations through multi-scale time-frequency analysis; Based on the feature vector, the current load disturbance mode is identified to determine the load disturbance category; Control parameters are matched according to the load disturbance category, and energy analysis is performed to obtain the load energy demand; Based on the load energy demand and energy storage status parameters, a power supply strategy search is performed to determine the energy supply timing strategy, and the power supply management and control of each passive load electrical equipment is carried out in accordance with the energy supply timing strategy.

2. The passive load power supply management method based on energy storage VSG according to claim 1, characterized in that, Real-time acquisition of electrical parameters at the VSG grid connection point for energy storage, and extraction of feature vectors for load power fluctuations through multi-scale time-frequency analysis, including: Real-time acquisition of electrical parameters at the VSG grid connection point, including three-phase voltage and three-phase current; Calculate the active and reactive power at the grid connection point based on the three-phase voltage and three-phase current. Based on the active power and / or reactive power at the grid connection point, multi-scale time-frequency analysis is performed to extract the energy distribution characteristics of load power fluctuations in different frequency bands, forming a feature vector.

3. The passive load power supply management method based on energy storage VSG according to claim 2, characterized in that, Multi-scale time-frequency analysis is performed to extract the energy distribution characteristics of load power fluctuations in different frequency bands, forming a feature vector, including: Wavelet packet multi-scale decomposition is performed on the active power signal at the grid connection point, and the signal is split into a preset number of decomposition layers to obtain multiple sub-band components covering different frequency ranges. Each sub-band component is reconstructed separately, and the sum of squares of each reconstructed signal is calculated as the energy characteristic value of the corresponding frequency band. The energy characteristic values ​​of all sub-bands are normalized to obtain the proportion of energy in each band to the total energy. The feature vector is formed by arranging and combining the proportional values ​​in order of increasing frequency band.

4. The passive load power supply management method based on energy storage VSG according to claim 1, characterized in that, Based on the feature vector, the current load disturbance mode is identified to determine the load disturbance category, including: The feature vector is matched with the feature relationships of each preset disturbance category to obtain the feature relationship matching degree. Based on the matching degree of the feature relationship, the preset disturbance category with the highest matching degree is determined as the current load disturbance category.

5. The passive load power supply management method based on energy storage VSG according to claim 4, characterized in that, The preset disturbance categories include at least: slow drift, step impact, periodic fluctuation, and fault transient.

6. The passive load power supply management method based on energy storage VSG according to claim 1, characterized in that, Control parameters are matched according to the load disturbance category, and energy analysis is performed to obtain the load energy demand, including: Based on the current load disturbance category and the current operating status of the energy storage VSG, predict the output active power and reactive power required to maintain the system frequency and voltage within the allowable range within the preset time window in the future; The predicted active power is integrated over time to obtain the active power demand. The predicted output reactive power is integrated over time to obtain the reactive power demand. The load energy requirement is obtained based on the functional energy requirement and the non-functional energy requirement.

7. The passive load power supply management method based on energy storage VSG according to claim 6, characterized in that, Predict the output active and reactive power required to maintain system frequency and voltage within allowable ranges within a preset time window, including: Obtain the current operating status parameters of the energy storage VSG, including the current grid connection point frequency, the current grid connection point voltage amplitude, the current output active power, and the current output reactive power; Based on the frequency deviation between the current grid connection point frequency and the rated frequency, and the voltage deviation between the current grid connection point voltage amplitude and the rated voltage, and in conjunction with the control parameters corresponding to the load disturbance category, the active power adjustment required to eliminate the frequency deviation and the reactive power adjustment required to eliminate the voltage deviation are calculated respectively. The current output active power plus the active power adjustment amount is used as the predicted active power within the future preset time window, and the current output reactive power plus the reactive power adjustment amount is used as the predicted reactive power within the future preset time window.

8. The passive load power supply management method based on energy storage VSG according to claim 1, characterized in that, Based on the load energy demand and energy storage status parameters, a power supply strategy search is performed to determine the energy supply timing strategy, including: Based on the load energy demand, obtain the load energy demand curve of a single passive load within a future preset time window. The load energy demand curve describes the trajectory of the active power value and reactive power value that the passive load needs to inject at different times as a function of time. Acquire the current state parameters and supply chain characteristic parameters of the energy storage system. The supply chain characteristic parameters include the response delay time between the energy storage converter receiving the command and the actual output of electrical energy, and the maximum ramp rate limit of the power change rate of the energy storage system. Using the load energy demand curve as the target power supply curve, the response delay time as the time offset, and the maximum ramp rate limit as the power variation constraint, demand-supply timing matching is performed: The power supply target curve is shifted forward along the time axis by the response delay time to obtain the pre-start curve, so that the energy storage system can respond in advance before the actual demand arrives to offset the impact of link delay. The pre-start-up curve is compared with the current maximum discharge power limit and maximum charge power limit of the energy storage system to determine whether the power demand at each moment on the pre-start-up curve exceeds the instantaneous supply capacity of the energy storage system. The time period exceeding the immediate supply capacity is marked as the power over-limit interval, and the power demand within the power over-limit interval is decomposed. Specifically, the excess power demand is moved forward to the available supply period before the over-limit interval for pre-supply, or moved backward to the available supply period after the over-limit interval for supplementary supply, or the power demand amplitude in the corresponding period is reduced to adapt to the supply capacity of the energy storage system. The decomposed power demand is rearranged in chronological order to establish the energy supply timing strategy, which includes active power command values ​​and reactive power command values ​​at each time point.

9. The passive load power supply management method based on energy storage VSG according to claim 8, characterized in that, Also includes: The individual energy demand curves of each passive load within a preset future time window are obtained separately, and all individual energy demand curves are superimposed on the same time axis to generate a multi-load comprehensive demand distribution curve. The system detects the time overlap between different load demand periods. When multiple loads have energy demand in the same period, the corresponding period is marked as the demand overlap interval, and the sum of the demand amplitudes of each load in the overlap interval is calculated as the total demand amplitude for the corresponding period. Based on the preset priority level of each load and the comparison between the total demand magnitude and the maximum supply capacity of the energy storage system, a multi-load coordinated power supply strategy is generated: When the total demand amplitude does not exceed the maximum supply capacity of the energy storage system, the output power command is superimposed on the original demand curve of each load. When the total demand exceeds the maximum supply capacity of the energy storage system, some or all of the power demand of low-priority loads in the overlapping period will be shifted to non-overlapping periods for peak-shifting power supply. The power commands of each load after peak-shifting adjustment will be accumulated at the same time and synthesized into a unified active power total command sequence and reactive power total command sequence to obtain the energy supply timing strategy under multi-load scenarios.

10. A passive load power supply management system based on energy storage VSG, characterized in that, The steps for implementing the passive load power supply management method based on energy storage VSG according to any one of claims 1 to 9, wherein the passive load power supply management system based on energy storage VSG comprises: The feature extraction module is used to collect electrical parameters of the VSG grid connection point in real time and extract the feature vector of load power fluctuation through multi-scale time-frequency analysis; The disturbance category determination module is used to identify the current load disturbance mode based on the feature vector and determine the load disturbance category; The energy demand acquisition module is used to match control parameters according to the load disturbance category and perform energy analysis to obtain the load energy demand. The power supply management module is used to search for power supply strategies based on the load energy demand and energy storage status parameters, determine the energy supply timing strategy, and manage and control the power supply of each passive load electrical device according to the energy supply timing strategy.

Images