Micro-grid layered stability control method and system

Abstract

The invention discloses a micro-grid layered stability control method and system, and relates to the technical field of micro-grid control. Periodic data and instruction signals of upper and lower layer control links of a micro-grid are collected, and reverse adjustment delay amount and an idle stroke deviation value during period switching are calculated; processing to obtain a period difference index and a collaborative adaptation index containing an impact coefficient; analyzing a time sequence action relationship between the period difference index and a micro-grid adjusting link, and calculating a connection deviation coefficient of a control rhythm by counting lagging frequency and amplitude of instruction response; potential problems of a micro-grid control link can be found in advance, stable operation time is prolonged through strategy adjustment when the instability risk is low, maintenance is conducted in time before the risk rises, power failure accidents caused by layered control mismatch are reduced, and the operation stability and reliability of the micro-grid are improved especially for the micro-grid with the high new energy proportion and large fluctuation.

Description

Technical Field

[0001] This invention belongs to the field of microgrid control technology, specifically a hierarchical stability control method and system for microgrids. Background Technology

[0002] Microgrids are key systems that integrate distributed generation, energy storage and loads, and the stability of their hierarchical control architecture (including bottom-level device control, mid-level coordination and upper-level energy management) is of paramount importance. In microgrid operation, the periodic coordination between upper and lower control levels is crucial for system stability. Due to factors such as signal transmission delays and differences in equipment response, reverse regulation delays and idle distance deviations can easily occur during periodic switching. This can lead to inconsistent control rhythms and even amplify voltage frequency fluctuations, causing potential oscillation risks. Traditional control methods struggle to accurately quantify the impact of these deviations on system stability and cannot predict the duration of stability maintenance in advance. Adjustments are often made only after instability has occurred, making it difficult to guarantee the continuous and reliable operation of the microgrid. Therefore, accurately analyzing deviations during periodic switching, quantifying and assessing the system's instability risk, and then adjusting control strategies in advance are key to ensuring the stable operation of microgrids. Summary of the Invention

[0003] The purpose of this invention is to provide a hierarchical stability control method and system for microgrids to solve the problems mentioned in the background art.

[0004] A hierarchical stability control method for microgrids, comprising: Collect periodic data and command signals from the upper and lower control links of the microgrid, calculate the reverse adjustment delay and idle distance deviation during period switching, and process them to obtain the period difference index and the collaborative adaptation index including the impact coefficient. The relationship between the cycle difference index and the timing effect of the microgrid regulation link is analyzed. By statistically analyzing the lag frequency and amplitude of command response, the connection deviation coefficient of the control rhythm is calculated. Monitor the voltage and frequency fluctuations in the microgrid operation, and based on the impact coefficient in the cooperative adaptation index, quantify its amplification effect on the fluctuation amplitude to generate oscillation hazard parameters for the operating status. Based on the connection deviation coefficient of the control rhythm and the oscillation hazard parameters, the key risk parameters of hierarchical instability of the microgrid are determined; When the key risk parameters of hierarchical instability are lower than the safety threshold, the adjustment frequency and amplitude change rate of the sub-instruction sequence are calculated, and the parameter distribution characteristics are analyzed to measure the adaptation deviation of the control link. Based on the adaptation deviation of the control link and the instability threshold of similar systems in the historical operation information database, the stability maintenance period is estimated, and a predictive mechanism is triggered to adjust the control strategy or repair components.

[0005] As a further aspect of the present invention: the calculation of the reverse adjustment delay and idle distance deviation during cycle switching includes: Record the time difference between the issuance of the upper-level control command and the actual reverse adjustment action of the lower-level control link. After subtracting the fixed delay of the signal transmission itself, the actual reverse adjustment delay is obtained. Then, through time series decomposition, the trend and fluctuation of the delay are extracted. The absolute difference between the target value and the actual output value before and after the sampling cycle is switched is calculated. When the difference is continuous for several sampling cycles and there is no decreasing trend, it is determined to be a space deviation. The spectrum analysis of the difference during this period is performed to find the main deviation frequency. When processing the obtained period difference index, the trend of the delay amount is correlated with the main deviation frequency of the deviation value in the time and frequency domain, and the period difference index is obtained through correlation analysis.

[0006] As a further aspect of the present invention: the quantitative analysis of the amplification effect of the impact coefficient on the fluctuation amplitude includes: Establish a dynamic correlation map between the impact coefficient and the voltage frequency fluctuation amplitude, and classify the fluctuation amplitude samples corresponding to different impact coefficients in historical data according to operating conditions to form multiple sets of operating condition feature combinations. During real-time monitoring, the feature combination of the current working condition is matched, the correspondence between the impact coefficient and the fluctuation amplitude within the feature combination is called to obtain the initial amplification effect evaluation value, and then dynamic correction is made in combination with the transient characteristics of real-time fluctuations. The oscillation hazard parameter is the degree of deviation between the corrected amplification effect assessment value and the preset safety boundary. When the deviation enters the warning range, it is marked as a potential oscillation risk, and the corresponding operating condition characteristics are recorded.

[0007] Secondly, this application provides a microgrid hierarchical stability control system, the system comprising: The data acquisition module collects periodic data and command signals from the upper and lower control links of the microgrid, calculates the reverse adjustment delay and idle distance deviation during period switching, and processes them to obtain the period difference index and the collaborative adaptation index containing the impact coefficient. The analysis module analyzes the time-series relationship between the cycle difference index and the microgrid regulation link, and calculates the connection deviation coefficient of the control rhythm by statistically analyzing the lag frequency and amplitude of the command response. The analysis module monitors the voltage and frequency fluctuations in the microgrid operation. Based on the impact coefficient in the cooperative adaptation index, it quantitatively analyzes the amplification effect of the fluctuation amplitude and generates oscillation hazard parameters for the operating status. The module is determined based on the connection deviation coefficient of the control rhythm and the oscillation hazard parameters to identify the key risk parameters of hierarchical instability of the microgrid; The calculation module calculates the adjustment frequency and amplitude change rate of the sub-instruction sequence when the key risk parameters of hierarchical instability are lower than the safety threshold, and analyzes the parameter distribution characteristics to measure the adaptation deviation of the control link. The generation module estimates the stability maintenance period based on the adaptation deviation of the control link and the instability threshold of similar systems in the historical operation information database, and triggers a predictive mechanism for control strategy adjustment or component maintenance.

[0008] Compared with the prior art, the beneficial effects of the present invention are: This invention can detect potential problems in the microgrid control system in advance, extend the stable operation time by adjusting strategies when the risk of instability is low, and perform timely maintenance before the risk increases, thereby reducing power outages caused by hierarchical control mismatch. It can improve the stability and reliability of microgrids with a high proportion of renewable energy and large fluctuations. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the method framework structure of the present invention. Detailed Implementation

[0010] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0011] Regarding this, firstly: please refer to... Figure 1 This application provides a hierarchical stability control method for microgrids, comprising: The process involves collecting periodic data and command signals from the upper and lower control levels of the microgrid, calculating the reverse adjustment delay and idle distance deviation during period switching, and processing these data to obtain a period difference index and a coordination and adaptation index including an impact coefficient. It's important to note that the upper control level primarily involves macro-level commands such as energy dispatch and power allocation, while the lower control level focuses on real-time equipment adjustment and execution. Periodic data includes the control cycle duration and data sampling interval of each level, while command signals refer to the content and timing of control commands. The reverse adjustment delay during period switching is the time difference between the issuance of the command and the correction of a reverse action when the lower control level responds to an upper command and performs an action opposite to the expected adjustment direction. For example, if the upper command requests an increase in photovoltaic output, but the lower level mistakenly executes a reduction, the time from error detection to correct adjustment is the reverse adjustment delay. The idle distance deviation refers to the deviation of microgrid parameters during the window period after the control command has been issued but before the actuator has taken actual action during period switching. For instance, if the frequency increases from 50Hz to 50.2Hz during the window period, the deviation is 0.2Hz. By processing these two values, we can obtain the cycle difference index (reflecting the degree of mismatch between the upper and lower control cycles) and the coordination and adaptation index including the impact coefficient. The impact coefficient is a parameter that measures the impact intensity of the command on the microgrid during cycle switching. It can be calculated by the ratio of the command change rate to the microgrid inertia. For example, if the command power change rate is 2kW / s and the system inertia is 0.5s, then the impact coefficient is 4.

[0012] This paper analyzes the relationship between the cycle difference index and the timing effect of the microgrid regulation link. By statistically analyzing the lag frequency and amplitude of command responses, the connection deviation coefficient of the control rhythm is calculated. It should be noted that the regulation link refers to the complete control path from command issuance to equipment execution; the timing effect refers to the influence of the cycle difference on the timing of command transmission along the link. The connection deviation coefficient of the control rhythm is calculated by statistically analyzing the lag frequency (the number of times the command response is delayed per unit time) and amplitude (the length of each delay). This coefficient is the ratio of the weighted average of the lag frequencies to the theoretical response period. For example, if the weighted average of the lag frequencies is 0.3s and the theoretical response period is 1s, then the connection deviation coefficient is 0.3.

[0013] The voltage and frequency fluctuation status of the microgrid is monitored. Based on the impact coefficient in the coordination and adaptation index, its amplification effect on the fluctuation amplitude is quantitatively analyzed to generate oscillation hazard parameters for the operating status. It should be noted that when the impact coefficient is large, the impact of periodic switching may increase the originally small frequency fluctuation amplitude. For example, when the impact coefficient is 3, the fluctuation amplitude can be amplified from ±0.1Hz to ±0.3Hz. Based on this, the oscillation hazard parameters for the operating status are generated, which reflect the probability of system oscillation caused by the impact.

[0014] Based on the control rhythm connection deviation coefficient and oscillation hazard parameters, the key risk parameters for hierarchical instability of microgrids are determined; these parameters comprehensively reflect the instability risks that may be caused by the control connection problems between upper and lower layers and the impact amplification effect.

[0015] When the key risk parameters for hierarchical instability are below the safety threshold, it indicates that the current instability risk is low, but the adaptability of the control link still needs to be monitored. At this time, the adjustment frequency (the number of times a sub-instruction is executed per unit time) and amplitude change rate (the change in the adjustment amplitude of the sub-instruction per unit time) of the sub-instruction sequence are calculated. By analyzing the distribution characteristics of these parameters (such as whether they are concentrated in a certain range and how discrete they are), the adaptability deviation of the control link is measured. The higher the deviation, the worse the adaptability of the control link to the execution of instructions.

[0016] Based on the adaptation deviation of the control link and the instability threshold of similar systems in the historical operation information database (i.e., the critical value of adaptation deviation when similar systems become unstable), the stability maintenance period is estimated, that is, the time that the system can maintain stable operation under the current adaptation state, and a predictive mechanism for control strategy adjustment or component maintenance is triggered. It should be understood that if the estimated period is short, the control strategy will be adjusted, such as optimizing the command sending cycle; if the deviation is close to the instability threshold, component maintenance will be prompted to avoid further decline in adaptability due to equipment aging.

[0017] For example, in a microgrid including wind and solar power, when the cycle of the upper-level dispatch command is 10 minutes and the cycle of the lower-level inverter control is 1 minute, the cycle difference index may fluctuate due to mismatch. If a voltage frequency fluctuation of ±0.5Hz is detected due to a large impact coefficient, the oscillation risk parameter increases, while the key risk parameter for hierarchical instability remains below the safety threshold. The system will then further analyze the adjustment of the sub-commands. If the frequency of sub-command adjustments is found to be fluctuating and the adaptation deviation is large, and the estimated stability maintenance period is only 2 hours based on historical data, the system will automatically adjust the matching degree of the upper and lower-level control cycles or prompt the inverter's response module for maintenance.

[0018] The innovation of this invention lies in the fact that this process can detect potential problems in the microgrid control system in advance, extend the stable operation time through strategy adjustment when the risk of instability is low, and carry out timely maintenance before the risk increases, thereby reducing power outages caused by hierarchical control mismatch. In particular, for microgrids with a high proportion of renewable energy and large fluctuations, it can improve the stability and reliability of their operation.

[0019] Second, multi-dimensional parameter quantification (such as reverse adjustment delay, idle distance deviation, etc.) and cross-domain correlation analysis (time-frequency domain coupling, Markov chain modeling) have constructed a complete control link from deviation identification to risk warning. This helps to realize the advanced prediction of microgrid hierarchical instability risk, solves the limitations of passive response in traditional methods, and is conducive to improving the timeliness and accuracy of system stability control.

[0020] As a further implementation, the calculation of the reverse adjustment delay and idle distance deviation during cycle switching includes: Record the time difference between the issuance of the upper-level control command and the actual reverse adjustment action of the lower-level control link. After subtracting the fixed delay of the signal transmission itself, the actual reverse adjustment delay is obtained. Then, through time series decomposition, the trend and fluctuation of the delay are extracted. It should be noted that, for example, if the upper layer issues a command to increase the energy storage discharge power at 10:00:00, but the lower layer begins to execute the reverse action of reducing the power at 10:00:02, the time difference between the two is 2 seconds. Since the signal transmission itself has a fixed delay of 0.1 seconds, subtracting this portion, the actual reverse adjustment delay is 1.9 seconds. Then, through time series decomposition, this delay is broken down into a trend part that changes with the cycle switching pattern (e.g., the delay increases by 0.2 seconds with each cycle switching) and a random fluctuation part (e.g., fluctuations of ±0.1 seconds caused by instantaneous interference).

[0021] The absolute difference between the target value and the actual output value before and after the sampling cycle is switched is calculated. When the difference is continuous for several sampling cycles and there is no decreasing trend, it is determined to be a space deviation. The spectrum analysis of the difference during this period is performed to find the main deviation frequency. It should be noted that, for example, if the target value of the instruction before the cycle switch is 80kW of photovoltaic power output, the actual output value after the switch will fluctuate between 75kW and 78kW, and the absolute difference between the two is 2kW to 5kW. When the difference remains around 3kW for three consecutive sampling periods (each period is 10 seconds) without showing a decreasing trend, it can be identified as a space travel deviation. Spectral analysis of the difference within these 30 seconds reveals that the deviation is mainly concentrated at a frequency of 0.5Hz, indicating that the space travel deviation is sufficiently affected by periodic interference at this frequency.

[0022] When processing the period difference index, the trend of delay (e.g., an increase of 0.2 seconds per switch) is correlated with the main deviation frequency (0.5Hz) of the deviation value in the time-frequency domain. The period difference index is obtained through correlation analysis. Specifically, the delay trend data and deviation frequency data are transformed into the same time-frequency coordinate system, and the correlation coefficient between the two at the same time and frequency points is calculated. The higher the correlation coefficient, the stronger the correlation between the two. The final comprehensive value obtained through this correlation analysis is the period difference index.

[0023] It should be noted that, for example, by calculating the correlation coefficient between the two within the same time interval, when the coefficient is greater than 0.8, it indicates that there is a strong correlation between the delay trend and the deviation frequency. The resulting cycle difference index can more accurately reflect the degree of mismatch between the upper and lower control cycles. If the correlation is high, the index value will increase, indicating that the cycle mismatch problem is prominent.

[0024] For example, in a microgrid including wind power, the upper-level dispatch cycle is 15 minutes, and the lower-level wind turbine control cycle is 2 minutes. During a cycle switch, the upper-level command requires the wind turbines to increase speed, but the lower-level command first decelerates for 1.8 seconds (after deducting a 0.1-second transmission delay, the actual reverse adjustment delay is 1.7 seconds), and the delay increases by 0.15 seconds with each switch. Simultaneously, the target wind speed is 12 m / s, but the actual output fluctuates between 10.5 and 11 m / s, showing no decrease over four consecutive sampling cycles. Spectrum analysis shows the main deviation frequency is 0.3 Hz. Correlating the delay trend with the 0.3 Hz deviation frequency, if the correlation coefficient reaches 0.85, the cycle difference index increases accordingly, clearly demonstrating the regulation disorder caused by the mismatch between the upper and lower-level cycles.

[0025] This calculation method can accurately capture anomalies during cycle switching, providing reliable data for subsequent analysis of control rhythm connection deviations, etc., making the coordination between upper and lower level controls more coordinated during the microgrid regulation process, reducing power fluctuations caused by cycle mismatch, and improving the stability of microgrid operation.

[0026] As a further implementation method, the quantitative analysis of the amplification effect of the impact coefficient on the fluctuation amplitude includes: A dynamic correlation map between the impact coefficient and the voltage frequency fluctuation amplitude is established. Samples of fluctuation amplitudes corresponding to different impact coefficients in historical data are classified according to operating conditions, forming multiple sets of operating condition characteristic combinations. It should be noted that the dynamic correlation map refers to a visual or data-driven model that reflects the correspondence between the two under different operating conditions. This is achieved by first collecting historical operating data of the microgrid, extracting voltage frequency fluctuation amplitudes (e.g., ±0.1Hz, ±0.3Hz, ±0.5Hz) corresponding to different impact coefficients (e.g., 1.2, 1.8, 2.5), and then classifying them according to operating conditions. It should be understood that operating conditions are mainly divided based on the proportion of renewable energy output (e.g., wind power accounting for 20% or 35%) and load type (e.g., mainly industrial load or mainly residential load). Samples of the impact coefficient and fluctuation amplitude under the same operating condition are grouped together, forming multiple sets of operating condition characteristic combinations, such as combinations like "30% wind power + industrial load" or "25% wind power + residential load".

[0027] During real-time monitoring, the system matches the characteristic combination of the current operating condition. For example, if the sensor collects data showing that wind power output accounts for 32% and industrial load accounts for 65%, the system can match the closest characteristic combination of "wind power 30% + industrial load". The system then calls the correspondence between the impact coefficient and the fluctuation amplitude within this characteristic combination to obtain the initial amplification effect assessment value, and then dynamically corrects it by combining it with the transient characteristics of the real-time fluctuation. It should be noted that, for example, in this characteristic combination, an impact coefficient of 1.5 corresponds to a fluctuation amplitude of ±0.2Hz, and an impact coefficient of 2.0 corresponds to ±0.4Hz. If the current real-time impact coefficient is 1.8, the initial amplification effect assessment value can be obtained by interpolation, specifically (1.8-1.5) / (2.0-1.5)×(0.4-0.2)+0.2=±0.32Hz. Then, dynamic correction is performed by combining the transient characteristics of real-time fluctuations. The transient characteristics include the rise rate of the fluctuation (e.g., 0.1Hz / second, 0.2Hz / second) and the duration (e.g., 2 seconds, 5 seconds). For example, when the rise rate of the fluctuation is fast (0.2Hz / second), a correction of 10% needs to be added to the initial evaluation value, and finally the corrected amplification effect evaluation value is ±0.35Hz.

[0028] The oscillation hazard parameter is the degree of deviation between the corrected amplification effect assessment value and the preset safety boundary. When the deviation enters the warning range, it is marked as a potential oscillation risk, and the corresponding operating characteristics are recorded. It should be noted that the preset safety boundary is set according to microgrid operating standards, such as ±0.3Hz. The deviation is the difference between the assessment value and the safety boundary. In this example, the corrected assessment value is ±0.35Hz, and the deviation is ±0.05Hz. When the deviation enters the warning range (e.g., ±0.03Hz to ±0.1Hz), the system will mark this state as a potential oscillation risk and record the corresponding operating characteristics (wind power 32% + industrial load 65%) to provide a basis for subsequent adjustments.

[0029] For a specific example, in a microgrid operating under the condition of "40%+ solar PV and commercial load," historical data shows that an impact factor of 1.6 corresponds to a fluctuation amplitude of ±0.22Hz, and an impact factor of 2.1 corresponds to ±0.42Hz. During real-time monitoring, the current operating condition is "38%+ solar PV and commercial load," matching the above combination. The real-time impact factor is 1.9, and the initial assessment value calculated through interpolation is ±0.34Hz. Due to the real-time fluctuation rise rate of 0.18Hz / second, the corrected assessment value is ±0.37Hz. With a preset safety boundary of ±0.3Hz and a deviation of ±0.07Hz, the system enters the warning zone, immediately marking the potential risk and recording the current solar PV and commercial load ratio data. This quantitative analysis method can accurately capture the amplification pattern of the impact coefficient on the fluctuation amplitude, avoiding misjudgments or omissions caused by relying solely on experience. In scenarios where the output of new energy sources fluctuates significantly, such as when photovoltaic output drops sharply due to cloud cover, the amplification effect of the impact coefficient can be quickly assessed, potential oscillation risks can be identified in advance, and equipment protection actions or power outages caused by excessive fluctuation amplitude can be reduced, further improving the stability of microgrid operation.

[0030] In some of the embodiments described above in this application, a step is proposed to determine the key risk parameters of hierarchical instability of a microgrid based on the connection deviation coefficient of the control rhythm and the oscillation hazard parameter, so as to accurately identify the instability risk in the hierarchical control of the microgrid.

[0031] However, in this process, there are still problems in how to comprehensively consider the dynamic relationship between the connection deviation coefficient and the oscillation hazard parameter, and how to generate key risk parameters that can reflect both the current risk level and the trend of risk changes. In particular, when the fluctuations of the two overlap or are misaligned, simple parameter superposition may not be able to accurately characterize the risk features, which can easily lead to misjudgment or omission of instability risk.

[0032] To address this, this application further proposes obtaining time-series variation curves of the control rhythm connection deviation coefficient and oscillation hazard parameters, marking the overlapping and misaligned fluctuation intervals of the two curves on the time axis; for the overlapping fluctuation interval, analyzing the co-directional variation law of the fluctuation amplitudes of the two, and calculating the synchronization ratio of amplitude changes; for the misaligned fluctuation interval, recording the corresponding state of the oscillation hazard parameter when the connection deviation coefficient reaches its peak, and the corresponding state of the connection deviation coefficient when the oscillation hazard parameter reaches its peak, obtaining the alternating variation law of the two; combining the synchronization ratio and the alternating variation law, generating a layered instability key risk parameter that can simultaneously reflect the current risk level and risk change trend.

[0033] The technical solution of this application improves the accuracy and comprehensiveness of key risk parameters for stratified instability by distinguishing between the overlapping and misaligned fluctuation ranges of the connection deviation coefficient and the oscillation hazard parameter, and deeply analyzing the variation patterns of the two in different ranges. It quantifies the degree of coordinated fluctuation in the overlapping range by using the synchronization ratio, and captures the risk transmission characteristics of the misaligned range by using the alternating variation pattern, thus achieving a dynamic characterization of risk and providing a more reliable basis for subsequent adjustments to stability control strategies.

[0034] Specifically, by marking intervals on the time-series change curves, the system can clearly identify the interaction patterns between the connection deviation coefficient and the oscillation hazard parameter. In the overlapping fluctuation interval, the co-directional change of the two often means the superposition and amplification of risks; the higher the synchronization ratio, the greater the degree of risk synergy enhancement. In the misaligned fluctuation interval, the alternating peaks of the two reflect the transmission and transformation of risks in different control dimensions. For example, the peak of the connection deviation may trigger the aggravation of subsequent oscillation hazards, or the peak of the oscillation hazard may have a counteracting effect on the stability of the control rhythm.

[0035] This method, through detailed analysis of the dynamic correlations between parameters, helps to avoid the limitations of single parameters or simple superposition, and improves the system's ability to perceive complex risk situations. Especially when the operating state of the microgrid changes rapidly, this method can promptly capture the evolution trend of risks, providing strong support for predicting instability risks and taking early intervention measures, thereby improving the stability and reliability of the microgrid's hierarchical control.

[0036] In the technical solution of this application, the marking of overlapping and misaligned fluctuation intervals can be achieved in various ways. For example, a sliding window method can be used, setting a time window threshold. When both curves show fluctuations exceeding the threshold within a certain window, it is determined to be an overlapping fluctuation interval; otherwise, it is a misaligned fluctuation interval. Another method is to use the Dynamic Time Warping (DTW) algorithm, which automatically identifies overlapping intervals with consistent fluctuation trends and misaligned intervals with differing trends by calculating the similarity between the two curves.

[0037] When calculating the synchronization ratio, Pearson correlation coefficient or Spearman rank correlation coefficient can be used to quantify the linear correlation between the fluctuation amplitudes of the two components within the overlapping interval. The higher the correlation coefficient, the greater the synchronization ratio. For extracting the alternating change pattern, peak detection algorithms, such as the gradient-based peak identification method, can be used to mark the peak points of the connection deviation coefficient and the oscillation hazard parameter. Then, time difference analysis can be used to determine the alternation sequence and interval characteristics of the two components.

[0038] The accuracy of interval marking directly affects the analytical precision of synchronization ratio and alternation pattern, while the quantitative results of synchronization ratio and alternation pattern jointly determine the generation of key risk parameters for stratified instability. This multi-dimensional analysis approach helps to more comprehensively integrate risk information on connection deviations and oscillation hazards, making the generated key risk parameters more closely reflect the actual operating state of the microgrid.

[0039] This technical solution distinguishes between overlapping and misaligned fluctuation ranges, combining the risk synergy intensity reflected by the synchronization ratio with the risk transmission path revealed by the alternating change pattern. This constructs a layered instability key risk parameter that can both quantify the current risk level and predict the evolution trend. This helps to deeply capture the nonlinear evolutionary essence of risk in microgrid layered control, from synergistic amplification to alternating transmission, realizing a shift from "passively identifying risks" to "actively predicting risk chains." Its innovation lies in transforming the temporal correlation between parameters into quantifiable risk characteristics, revealing the risk transmission mechanism hidden in complex fluctuations, thus providing a precise optimization basis for control strategies based on the risk source.

[0040] In some embodiments of this application, how to effectively handle the complex characteristics of overlapping and misaligned fluctuation ranges during the generation of stratified instability risk parameters, and how to associate these characteristics with the risk triggering sequence, so as to accurately predict the transformation node of risk from potential to manifestation, especially when the fluctuation range presents a complex superposition or misalignment state, simple parameter extraction may not fully reflect the dynamic changes of risk, leading to deviations in the prediction of transformation nodes.

[0041] In response, this application further proposes to introduce wave propagation speed difference analysis for overlapping wave ranges, calculate the connection deviation coefficient and the leading edge arrival time difference of the oscillation hazard parameter; for misaligned wave ranges, mark the interval period of the peak occurrence of the two, and establish the matching relationship between the interval period and historical instability cases; combine the time difference and interval period characteristics to generate key risk parameters containing the risk triggering sequence, which can predict the transformation node of risk from potential to apparent.

[0042] The technical solution of this application improves the comprehensiveness and accuracy of key risk parameters through targeted analysis of different types of fluctuation ranges. By introducing fluctuation propagation speed difference analysis and establishing a matching relationship between the interval period and historical cases, it achieves in-depth mining of the dynamic characteristics of risk. By combining time difference and interval period to generate parameters containing the risk trigger sequence, it solves the problem of inaccurate prediction of risk transformation nodes and provides a reliable basis for timely prevention and control measures.

[0043] Specifically, for overlapping fluctuation ranges, the system analyzes the differences in the propagation speed of the fluctuations to calculate the difference in the arrival time of the leading edge of the connection deviation coefficient and the oscillation hazard parameter. This time difference reflects the interaction rhythm of the two fluctuations during the superposition process and is an important basis for judging the intensity of risk superposition. For misaligned fluctuation ranges, the system marks the interval period of the peak occurrence and matches it with historical instability cases. Similar fluctuation patterns can be found from historical data, thereby inferring the current risk development trend.

[0044] Subsequently, by fusing the aforementioned time difference and interval characteristics, the generated key risk parameters contain the timing information of risk triggering. These parameters can clearly show the time points in which a risk gradually develops from a potential state to an apparent state. For example, in overlapping fluctuation ranges, when the time difference between the arrival of the leading edge of the connection deviation coefficient and the oscillation hazard parameter narrows to a certain threshold, it may indicate that the risk is about to manifest; while in misaligned fluctuation ranges, when the peak interval period matches the key period in historical instability cases, it may also trigger risk transformation.

[0045] This method improves the system's ability to perceive and predict stratified instability risks through detailed analysis and temporal correlation of fluctuation range characteristics. Especially when facing complex and volatile fluctuation states, it can promptly capture key signals of risk transformation, providing strong support for prevention and control decisions and helping to reduce losses caused by delayed risk prediction.

[0046] In the technical solution of this application, the wave propagation speed difference analysis can be implemented in various ways. For example, signal processing algorithms such as cross-correlation analysis or Hilbert transform can be used to calculate the propagation speed of the wave signal, thereby obtaining the arrival time difference of the leading edge. Marking the peak interval period can be achieved using peak detection algorithms, such as threshold-based extreme value detection or wavelet transform peak extraction methods. When establishing the matching relationship between the interval period and historical instability cases, pattern recognition algorithms such as k-nearest neighbor classification or support vector machines can be used to compare the current interval period with the periodic features in historical cases.

[0047] The technical solution of this application, in addressing the problem of generating stratified instability risk parameters, first achieves accurate capture of different fluctuation characteristics by separately processing overlapping and misaligned fluctuation intervals. This step effectively processes complex fluctuation data from multiple monitoring points, extracting key features such as time differences and interval periods through targeted algorithms, laying the foundation for subsequent risk parameter generation.

[0048] Based on these characteristics, the system integrates time differences and interval periods to generate key risk parameters that include the timing of risk triggers. For example, in the monitoring of a certain project, the time difference between the connection deviation coefficient of the overlapping fluctuation range and the arrival time of the leading edge of the oscillation hazard parameter is 5 seconds, while the peak interval period of the misaligned fluctuation range is 10 seconds. By matching with historical instability cases, it was found that when the time difference is less than 8 seconds and the interval period is in the range of 8-12 seconds, the probability of risk transformation is relatively high. Therefore, the generated key risk parameters will mark this timing feature as the early warning node for the transformation from potential risk to manifestation.

[0049] Suppose that in a layered structure monitoring system, overlapping and misaligned fluctuation intervals are detected. For the overlapping fluctuation interval, cross-correlation analysis calculates the arrival time of the leading edge of the connection deviation coefficient to be 20 seconds, and the arrival time of the leading edge of the oscillation hazard parameter to be 23 seconds, with a time difference of 3 seconds. For the misaligned fluctuation interval, a threshold-based extreme value detection algorithm marks the peak occurrence times at 15 seconds and 25 seconds, with an interval of 10 seconds.

[0050] The system matched the 3-second time difference and 10-second interval with historical instability cases and found that in historical cases, when similar fluctuation characteristics occurred, if the time difference was less than 5 seconds and the interval was between 8 and 12 seconds, the risk would manifest on average after 30 seconds. Therefore, the generated key risk parameters will include time-series information that "risk transformation may occur approximately 30 seconds after the current moment," thus buying valuable time for the implementation of prevention and control measures.

[0051] This method enables the accurate generation of key risk parameters for layered instability in complex volatile environments, effectively predicts risk transformation points, improves the initiative and effectiveness of risk prevention and control, and ensures the stable operation of related structures or systems.

[0052] In some embodiments of this application, a scheme is proposed to generate key risk parameters containing risk triggering timing by combining time difference and interval periodic characteristics, in order to predict the transformation node of risk from potential to apparent.

[0053] However, in the process of predicting risk transition points, the dynamic changes in time difference and interval periods often exhibit complex nonlinear characteristics. Simply analyzing these surface features makes it difficult to accurately capture the critical state of risk transition. Furthermore, when real-time loads experience sudden changes, the predicted transition point may deviate from the actual situation. Especially when the microgrid's operating status fluctuates drastically, this deviation may affect the timeliness and effectiveness of risk prevention and control measures.

[0054] In response, this application further proposes to perform chaotic characteristic analysis on time difference and interval period, and extract the stable trajectory pattern presented by the two in dynamic changes; when the trajectory pattern enters the preset critical region, the prediction result of the conversion node is corrected by combining the probability of occurrence of real-time load change signal.

[0055] The technical solution of this application, by introducing chaotic characteristic analysis, delves into the nonlinear laws governing the dynamic changes in time difference and interval periods, improving the accuracy of stable trajectory pattern recognition and providing a more reliable basis for capturing the critical state of risk transformation. By combining the probability of real-time load mutation signals to correct the prediction results, the interference caused by load mutations is effectively offset, further improving the accuracy of transformation node prediction.

[0056] Specifically, chaotic characteristic analysis can reveal the inherent patterns hidden in the seemingly disordered changes of time differences and intervals. By calculating chaotic characteristic quantities such as the maximum Lyapunov exponent and correlation dimension, stable trajectory patterns in their dynamic changes can be identified. When they enter a pre-defined critical region, it means that the risk is about to transform from a potential state to a manifest state.

[0057] Meanwhile, real-time load mutations are a significant factor affecting the stable operation of microgrids, potentially accelerating or delaying the risk transformation process. By statistically analyzing the probability of real-time load mutation signals, a high probability indicates a substantial impact of load changes on risk transformation, necessitating early correction of the initially predicted transformation points. Conversely, a low probability allows for appropriate delays in correction, ensuring the predictions better reflect actual operating conditions.

[0058] For example, during the operation of a microgrid, chaotic characteristic analysis revealed that the trajectory pattern of time difference and interval period gradually approached the critical region, with the initial prediction of a transition node in 10 minutes. At this point, the system detected a 70% probability of a real-time load mutation signal, indicating that the load mutation might accelerate the risk transition. Therefore, the transition node was revised to 8 minutes later, providing more time for staff to take preventative measures.

[0059] In the technical solution of this application, chaotic characteristic analysis can employ phase space reconstruction technology to transform a one-dimensional time series with time difference and interval period into a trajectory in a high-dimensional phase space, and then extract stable trajectory patterns by calculating relevant chaotic feature quantities. The statistical probability of real-time load mutation signals can be calculated based on a sliding time window, specifically by calculating the ratio of the number of occurrences of load mutation signals within the window to the total number of signals.

[0060] The accuracy of chaotic characteristic analysis directly affects the quality of stable trajectory pattern extraction, and thus the judgment of critical regions; while the calculation accuracy of the probability of real-time load mutation signals determines the rationality of transformation node correction. The effective combination of the two forms a closed-loop prediction and correction method, which helps to improve the reliability of risk transformation node prediction.

[0061] This technical solution analyzes chaotic characteristics to uncover the deep dynamic patterns of time difference and interval period, and combines real-time load mutation signal probability to correct the transformation node. This not only improves the prediction accuracy but also overcomes the limitations of traditional linear analysis in dealing with nonlinear risk evolution. By applying chaos theory to the prediction of microgrid risk transformation nodes, it breaks the dependence on surface feature analysis of parameters. At the same time, it introduces real-time load mutation probability as a correction factor to construct a dynamic adaptive prediction model. This achieves a leap from static feature analysis to dynamic pattern mining, and from single parameter dependence to multi-factor collaborative correction. It provides a new technical path for the proactive prevention and control of microgrid hierarchical instability risks, effectively improving the safety and stability of microgrid operation.

[0062] In some of the embodiments described above in this application, a scheme is proposed to correct the prediction results of the conversion node by combining the occurrence probability of real-time load change signals, which improves the accuracy of prediction to a certain extent.

[0063] However, in the process of correcting the prediction results of the conversion node, the correction is based solely on the probability of the occurrence of the load mutation signal, failing to fully consider the impact of the load mutation signal intensity on the trajectory pattern. Furthermore, the fixed boundary of the preset critical region of the trajectory pattern makes it difficult to adapt to the dynamic changes in the trajectory pattern under different intensities of load mutation, which may lead to deviations in the corrected prediction values ​​of the conversion node. In particular, when the intensity of the load mutation signal varies significantly, the fixed critical region boundary makes the judgment of whether the trajectory pattern has entered a critical state inaccurate, thus affecting the correction effect.

[0064] In response, this application further proposes to analyze the correlation between the intensity of load mutation signal and trajectory mode offset; to set the dynamic boundary of the critical region of trajectory mode according to the offset level, and to adjust the prediction value of conversion node accordingly.

[0065] The technical solution of this application establishes an intrinsic connection between the load mutation signal strength and the trajectory pattern offset by analyzing the correlation between the two, providing a basis for setting the dynamic boundary of the critical region. Setting the dynamic boundary in stages according to the offset allows the critical region to be flexibly adjusted based on the actual offset of the trajectory pattern, thereby more accurately determining whether the trajectory pattern has entered a critical state. Based on this, the predicted value of the conversion node is also more accurate.

[0066] Specifically, different load surge signal intensities have varying impacts on the trajectory patterns formed by time differences and intervals, resulting in different offsets in the trajectory patterns. Through extensive experiments and data analysis, a functional relationship or mapping table between the load surge signal intensity and the trajectory pattern offset can be established, clarifying the range of offset variation under different intensities. For example, when the load surge signal intensity is S1, the trajectory pattern offset is between O1 and O2; when the intensity is S2, the offset is between O3 and O4, and so on.

[0067] The trajectory pattern offset is graded according to its magnitude, such as slight offset, moderate offset, and severe offset. For each offset level, a corresponding dynamic boundary for the critical region is set. When the offset is slight, the critical region boundary is appropriately expanded to avoid false positives; when the offset is severe, the critical region boundary is contracted inward to more promptly capture risk conversion signals. Then, based on the adjusted dynamic boundary and whether the trajectory pattern enters that boundary, the predicted value of the conversion node is adjusted accordingly.

[0068] For example, in the operation of a microgrid, the initial predicted transition node was 15 minutes later. Analysis revealed that the trajectory pattern offset corresponding to the current load surge signal strength was moderate. According to the settings, the dynamic boundary of the critical region had shrunk somewhat compared to the default boundary. Since the trajectory pattern was already close to the shrunk dynamic boundary, the predicted transition node was adjusted to 12 minutes later, which better reflects the actual risk evolution.

[0069] In the technical solution of this application, regression analysis methods, such as linear regression and nonlinear regression, can be used to analyze the correlation between the intensity of load mutation signals and trajectory pattern offsets. A correlation model between the two can be obtained by fitting historical data. Clustering algorithms, such as K-means clustering, can be used to classify offset data into different levels based on offset levels. The dynamic boundary of the critical region can be set by combining expert experience and historical case data to establish a correspondence between offset levels and boundary adjustment amounts.

[0070] The accuracy of the correlation between the intensity of load mutation signals and trajectory pattern offset determines the rationality of offset grading, which in turn directly affects the setting accuracy of the dynamic boundary of the critical region, ultimately influencing the adjustment effect of the predicted value of the conversion node. Each link is closely interconnected, collectively improving the accuracy of the predicted conversion node.

[0071] This technical solution analyzes the correlation between the intensity of load mutation signals and trajectory pattern offset, and adjusts the predicted value of the transition node by setting dynamic boundaries of the critical region according to the offset level, thereby further improving the accuracy and adaptability of the prediction. Furthermore, it breaks through the limitation of correcting solely based on the probability of load mutation signal occurrence, deeply exploring the intrinsic relationship between signal strength and trajectory pattern offset, and innovatively proposing a dynamic boundary setting mechanism. This allows the critical region to flexibly change according to the trajectory pattern offset, achieving a leap from static boundary judgment to dynamic boundary adjustment. This provides a new approach for more accurately predicting risk transition nodes and effectively enhances the pertinence and timeliness of microgrid risk prevention and control.

[0072] In some embodiments of this application, a scheme is proposed to set the dynamic boundary of the critical region of the trajectory mode according to the offset level, and adjust the prediction value of the conversion node accordingly, which further improves the accuracy and adaptability of the prediction.

[0073] However, after adjusting the predicted values ​​for transformation nodes, the impact of load mutations is not limited to trajectory pattern shifts. Their propagation in the control link can trigger a series of chain reactions. Existing solutions do not track this propagation path and the critical nodes of irreversible shifts, resulting in a lack of dynamic calibration for the boundary adjustment magnitude in critical regions. This means that as the impact of load mutations continues to propagate, boundary adjustments may fail to adapt to new risk states in a timely manner, thus affecting the reliability of subsequent transformation node predictions. In particular, when the impact of load mutations propagates to critical nodes in the control link and causes irreversible shifts, if the boundary adjustment magnitude is not calibrated accordingly, it may lead to a lag in the assessment of risk evolution.

[0074] In response, this application further proposes to track the transmission path of the impact of load mutations in the control link, mark the key transmission nodes with irreversible offsets, and dynamically calibrate the boundary adjustment range of the critical region by combining the response characteristics of the key transmission nodes.

[0075] The technical solution of this application, by tracing the transmission path of the impact of load mutations and marking key transmission nodes, can accurately locate the core points of influence of load mutations, providing a clear target for the dynamic calibration of boundary adjustment amplitude. Calibration combined with the response characteristics of key transmission nodes ensures that the boundary adjustment of the critical region closely follows the dynamic changes of the impact of load mutations, further improving the timeliness and accuracy of conversion node prediction.

[0076] Specifically, after a load surge occurs, its impact is transmitted sequentially along each link in the control link, such as from the load monitoring module to the signal processing module, and then to the risk analysis module. By constructing a topology model of the control link, the transmission path of the load surge's impact can be tracked in real time, while the offset status of each node can be monitored. When the offset of a node exceeds a preset irreversible threshold (i.e., it cannot be restored to the normal range by its own adjustment after offset), it is marked as a critical transmission node of irreversible offset.

[0077] Different critical transmission nodes possess different response characteristics, such as response speed and anti-interference capability. For example, one critical transmission node may have a fast response speed, rapidly generating offset upon the arrival of a sudden load change; while another node may have a slower response speed, but once offset occurs, it is difficult to reverse. Based on these response characteristics, a correlation model is established between the boundary adjustment amplitude and the response parameters of the critical transmission nodes. When the response characteristics of the critical transmission nodes change, the boundary adjustment amplitude of the critical region is dynamically calibrated in a timely manner.

[0078] For example, in a microgrid, after adjusting the predicted value of the conversion node to 12 minutes, the system began tracking the propagation path of the impact of load mutations. It discovered an irreversible offset at the signal processing module node, with the node's response characteristics showing a rapid increase in offset rate. Based on this characteristic, the system determined that the impact of load mutations would accelerate its propagation to subsequent stages. Therefore, the boundary adjustment range of the critical region was increased by 20%, further shrinking the boundary. This allowed the predicted value of the conversion node to be recalibrated to 10 minutes later, making it more closely reflect the actual risk development trend.

[0079] In the technical solution of this application, the transmission path of the impact of load mutation can be tracked using graph theory algorithms. Each node in the control link is considered a vertices of a graph, and the connections between nodes are considered edges. The transmission path is tracked through breadth-first search or depth-first search. Key transmission nodes with irreversible offsets can be identified by setting an offset threshold and combining it with the node's historical recovery data. The boundary adjustment range of the dynamic calibration critical region can be achieved using an adaptive control algorithm, continuously optimizing the adjustment range based on the real-time response data of the key transmission nodes.

[0080] The accuracy of tracking the transmission path affected by sudden load changes determines the accuracy of marking critical transmission nodes, while the depth of analysis of the response characteristics of critical transmission nodes directly affects the effectiveness of dynamic calibration of boundary adjustment amplitudes. These interconnected components form a closed-loop adjustment mechanism capable of adapting to changes in load fluctuations in real time.

[0081] This technical solution further optimizes the prediction of transition nodes by tracking the transmission path of load mutation effects, marking key nodes, and dynamically calibrating the boundary adjustment range based on their response characteristics. It also breaks through the limitation of only adjusting the boundary for trajectory mode offset, extending it to the entire transmission process of load mutation effects in the control link. Furthermore, it proposes a dynamic calibration method based on the response characteristics of key transmission nodes, enabling the boundary adjustment of the critical region to change from passively adapting to trajectory offset to actively tracking the transmission of effects. This achieves a leap from static adjustment to dynamic tracking calibration, providing a more advanced technical means for the prediction of risk transition nodes under complex control links, and effectively enhancing the microgrid's rapid response capability to load mutation risks.

[0082] Secondly, this invention also proposes a hierarchical stability control system for microgrids, the system comprising: The data acquisition module collects periodic data and command signals from the upper and lower control links of the microgrid, calculates the reverse adjustment delay and idle distance deviation during period switching, and processes them to obtain the period difference index and the collaborative adaptation index containing the impact coefficient. The analysis module analyzes the time-series relationship between the cycle difference index and the microgrid regulation link, and calculates the connection deviation coefficient of the control rhythm by statistically analyzing the lag frequency and amplitude of the command response. The analysis module monitors the voltage and frequency fluctuations in the microgrid operation. Based on the impact coefficient in the cooperative adaptation index, it quantitatively analyzes the amplification effect of the fluctuation amplitude and generates oscillation hazard parameters for the operating status. The module is determined based on the connection deviation coefficient of the control rhythm and the oscillation hazard parameters to identify the key risk parameters of hierarchical instability of the microgrid; The calculation module calculates the adjustment frequency and amplitude change rate of the sub-instruction sequence when the key risk parameters of hierarchical instability are lower than the safety threshold, and analyzes the parameter distribution characteristics to measure the adaptation deviation of the control link. The generation module estimates the stability maintenance period based on the adaptation deviation of the control link and the instability threshold of similar systems in the historical operation information database, and triggers a predictive mechanism for control strategy adjustment or component maintenance.

[0083] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A hierarchical stability control method for microgrids, characterized in that, include: Collect periodic data and command signals from the upper and lower control links of the microgrid, calculate the reverse adjustment delay and idle distance deviation during period switching, and process them to obtain the period difference index and the collaborative adaptation index including the impact coefficient. The relationship between the cycle difference index and the timing effect of the microgrid regulation link is analyzed. By statistically analyzing the lag frequency and amplitude of command response, the connection deviation coefficient of the control rhythm is calculated. Monitor the voltage and frequency fluctuations in the microgrid operation, and based on the impact coefficient in the cooperative adaptation index, quantify its amplification effect on the fluctuation amplitude to generate oscillation hazard parameters for the operating status. Based on the connection deviation coefficient of the control rhythm and the oscillation hazard parameters, the key risk parameters of hierarchical instability of the microgrid are determined; When the key risk parameters of hierarchical instability are lower than the safety threshold, the adjustment frequency and amplitude change rate of the sub-instruction sequence are calculated, and the parameter distribution characteristics are analyzed to measure the adaptation deviation of the control link. Based on the adaptation deviation of the control link and the instability threshold of similar systems in the historical operation information database, the stability maintenance period is estimated, and a predictive mechanism is triggered to adjust the control strategy or repair components.

2. The microgrid hierarchical stability control method according to claim 1, characterized in that, The calculation of the reverse adjustment delay and idle deviation during cycle switching includes: Record the time difference between the issuance of the upper-level control command and the actual reverse adjustment action of the lower-level control link. After subtracting the fixed delay of the signal transmission itself, the actual reverse adjustment delay is obtained. Then, through time series decomposition, the trend and fluctuation of the delay are extracted. The absolute difference between the target value and the actual output value before and after the sampling cycle is switched is calculated. When the difference is continuous for several sampling cycles and there is no decreasing trend, it is determined to be a space deviation. The spectrum analysis of the difference during this period is performed to find the main deviation frequency. When processing the obtained period difference index, the trend of the delay amount is correlated with the main deviation frequency of the deviation value in the time and frequency domain, and the period difference index is obtained through correlation analysis.

3. The microgrid hierarchical stability control method according to claim 1, characterized in that, Quantitative analysis of the amplification effect of the impact coefficient on the amplitude of fluctuations includes: Establish a dynamic correlation map between the impact coefficient and the voltage frequency fluctuation amplitude, and classify the fluctuation amplitude samples corresponding to different impact coefficients in historical data according to operating conditions to form multiple sets of operating condition feature combinations. During real-time monitoring, the feature combination of the current working condition is matched, the correspondence between the impact coefficient and the fluctuation amplitude within the feature combination is called to obtain the initial amplification effect evaluation value, and then dynamic correction is made in combination with the transient characteristics of real-time fluctuations. The oscillation hazard parameter is the degree of deviation between the corrected amplification effect assessment value and the preset safety boundary. When the deviation enters the warning range, it is marked as a potential oscillation risk, and the corresponding operating condition characteristics are recorded.

4. The microgrid hierarchical stability control method according to claim 1, characterized in that, Analyzing parameter distribution characteristics to calculate the adaptation deviation of the control link includes: The adjustment frequency of the sub-instruction sequence is statistically analyzed on a minute-by-minute basis. A Markov chain model of the frequency distribution is constructed, and the stability of the adjustment rhythm is evaluated through the state transition probability. The amplitude change rate is calculated by the ratio of the amplitude difference between adjacent sub-instructions to the time interval. After taking the absolute value, quantile analysis is performed to determine the probability of extreme change rates. The fit deviation is a comprehensive index obtained by correlating the degree of dispersion of the steady-state distribution of the Markov chain model with the probability of extreme change rates. The magnitude of this index directly reflects the degree of fit deviation.

5. The microgrid hierarchical stability control method according to claim 1, characterized in that, Determining key risk parameters for hierarchical instability in microgrids includes: Obtain the time-series variation curves of the control rhythm connection deviation coefficient and the oscillation risk parameter, and mark the overlapping fluctuation range and the misalignment fluctuation range of the two curves on the time axis; For overlapping fluctuation intervals, analyze the co-directional change pattern of the fluctuation amplitudes of the two, and calculate the synchronization ratio of amplitude changes; For the misalignment fluctuation range, record the corresponding state of the oscillation hazard parameter when the connection deviation coefficient reaches its peak, and the corresponding state of the connection deviation coefficient when the oscillation hazard parameter reaches its peak, and obtain the alternating change pattern of the two. By combining the synchronization ratio and the alternation pattern, a stratified instability key risk parameter is generated that can simultaneously reflect the current risk level and the trend of risk change.

6. The microgrid hierarchical stability control method according to claim 5, characterized in that: When generating key risk parameters for stratified instability, the following are also included: For overlapping fluctuation intervals, fluctuation propagation speed difference analysis is introduced to calculate the connection deviation coefficient and the time difference of arrival of the leading edge of the oscillation hazard parameter; For the misaligned fluctuation range, mark the interval period between the occurrence of the peaks of the two, and establish a matching relationship between the interval period and historical instability cases; By combining time difference and interval period characteristics, key risk parameters containing the risk triggering sequence are generated. These parameters can predict the transformation node from potential to apparent risk.

7. The microgrid hierarchical stability control method according to claim 6, characterized in that, When predicting conversion nodes: Chaotic characteristics analysis was performed on time difference and interval period to extract stable trajectory patterns exhibited by both in dynamic changes; When the trajectory pattern enters the preset critical region, the prediction result of the conversion node is corrected by combining the probability of the occurrence of real-time load change signals.

8. The microgrid hierarchical stability control method according to claim 7, characterized in that, When correcting the prediction results of the conversion node: Analyze the correlation between the intensity of load abrupt change signal and trajectory pattern offset; The dynamic boundary of the critical region of the trajectory mode is set according to the offset level, and the prediction value of the conversion node is adjusted accordingly.

9. A microgrid hierarchical stability control method according to claim 8, characterized in that, After adjusting the predicted value of the conversion node: Track the transmission path of load mutation effects in the control link and mark critical transmission nodes with irreversible offsets; By combining the response characteristics of key transmission nodes, the boundary adjustment range of the critical region is dynamically calibrated.

10. A microgrid hierarchical stability control system, applicable to the microgrid hierarchical stability control method according to any one of claims 1 to 9, characterized in that, The system includes: The data acquisition module collects periodic data and command signals from the upper and lower control links of the microgrid, calculates the reverse adjustment delay and idle distance deviation during period switching, and processes them to obtain the period difference index and the collaborative adaptation index containing the impact coefficient. The analysis module analyzes the time-series relationship between the cycle difference index and the microgrid regulation link, and calculates the connection deviation coefficient of the control rhythm by statistically analyzing the lag frequency and amplitude of the command response. The analysis module monitors the voltage and frequency fluctuations in the microgrid operation. Based on the impact coefficient in the cooperative adaptation index, it quantitatively analyzes the amplification effect of the fluctuation amplitude and generates oscillation hazard parameters for the operating status. The module is determined based on the connection deviation coefficient of the control rhythm and the oscillation hazard parameters to identify the key risk parameters of hierarchical instability of the microgrid; The calculation module calculates the adjustment frequency and amplitude change rate of the sub-instruction sequence when the key risk parameters of hierarchical instability are lower than the safety threshold, and analyzes the parameter distribution characteristics to measure the adaptation deviation of the control link. The generation module estimates the stability maintenance period based on the adaptation deviation of the control link and the instability threshold of similar systems in the historical operation information database, and triggers a predictive mechanism for control strategy adjustment or component maintenance.

Images