An off-grid wind-solar direct-current microgrid power imbalance hierarchical self-recovery protection control method

Abstract

The application relates to the technical field of direct-current microgrid protection control, and discloses a kind of off-grid wind and light direct-current microgrid power imbalance hierarchical self-healing protection control methods, the method includes collecting off-grid wind and light direct-current microgrid total power generation power, total load power and direct-current bus voltage time sequence sampling data, imbalance deviation characteristics are obtained based on sliding evaluation period sampling data, power imbalance grade is determined;According to the power, bus voltage offset direction determined by statistics imbalance cause, combined with imbalance grade and cause from the preset action resource pool to select candidate protection action;When power imbalance grade is the first imbalance grade, skip the test verification, determine the fault branch and isolate through group intelligence algorithm optimization;When power imbalance grade is the second imbalance grade, output small amplitude test control quantity, collect electrical feedback data to complete test verification, and execute the candidate protection action completely through the back;Return to the sampling step after execution is completed.The application realizes hierarchical self-healing protection, and improves the reliability of microgrid operation.

Description

Technical Field

[0001] This invention relates to the field of off-grid microgrid control technology, specifically to a hierarchical self-healing protection and control method for power imbalance in off-grid wind-solar DC microgrids. Background Technology

[0002] Off-grid wind-solar DC microgrids are widely used in scenarios without a large power grid, such as islands, remote rural areas, and field operation sites. Wind and solar power generation is subject to strong random fluctuations, and frequent load switching can easily cause system power imbalance, resulting in a sharp deviation in DC bus voltage. Power imbalance protection and control is a key technology for off-grid DC microgrids.

[0003] Patent CN115833741A discloses a power coordination control scheme for an islanded DC microgrid. This scheme can mitigate some wind and solar power fluctuations, but it has significant drawbacks: the control strategy only relies on instantaneous sampled data for power compensation, lacks a time-series sliding evaluation window, and fails to classify power imbalance into first and second imbalance levels, making it unable to distinguish between normal small fluctuations. Using a uniform compensation strategy for different imbalance conditions leads to frequent charging and discharging of energy storage due to minor disturbances, and serious faults cannot be quickly isolated, making it difficult to balance power supply continuity and equipment safety.

[0004] Patent CN116092427A proposes a DC microgrid voltage hierarchical protection system that achieves multi-level over-limit protection based on bus voltage. This scheme introduces multi-threshold voltage criteria, but it is a rigid open-loop protection architecture that relies solely on real-time voltage to execute alarms or trips. It does not combine generation and load power to comprehensively determine the cause of imbalance and lacks a trial-and-verify closed-loop self-healing adjustment mechanism. Instantaneous sampling noise and short-term power surges can easily cause protection malfunctions, leading to direct tripping for mild to moderate imbalances and unnecessary power outages.

[0005] The aforementioned comparative documents share a common shortcoming: the control schemes are all based on instantaneous sampling data for decision-making, and only use the sampling information at the current moment to carry out protection judgments; they fail to fully explore the temporal variation patterns of power generation, load, and bus voltage within the cycle, do not distinguish between the two types of imbalance causes of over-generation and under-supply, have poor targeted handling, limited anti-interference capabilities, and are prone to false protection and false tripping due to short-term disturbances. Summary of the Invention

[0006] Existing off-grid wind-solar DC microgrids have several shortcomings in power imbalance protection and control: protection decisions rely solely on instantaneous sampling signals, failing to utilize the temporal variations of total generating power, total load power, and DC bus voltage; they lack a sliding evaluation window timing statistical filtering mechanism, making them susceptible to sampling noise and short-term power surges, resulting in frequent false protection and tripping; they fail to classify power imbalance levels or differentiate between excess generation and insufficient power supply imbalances, employing a uniform handling strategy for different disturbance conditions, leading to insufficient targeted protection actions; in mild to moderate power imbalance conditions, they lack a trial-and-error flexible adjustment process, directly outputting complete protection control commands, which can easily cause severe DC bus voltage oscillations; severe power imbalances lack rapid isolation branches, resulting in delayed fault handling and a high risk of system instability; and overall, they lack a closed-loop monitoring mechanism, failing to continuously track the microgrid's operating status after protection actions are completed.

[0007] Therefore, the purpose of this invention is to overcome the shortcomings of the prior art and provide a hierarchical self-healing protection control method for power imbalance in off-grid wind-solar DC microgrids. This method achieves sliding evaluation window timing statistics, hierarchical classification and judgment of power imbalance, trial verification and adjustment under light and moderate operating conditions, rapid branch isolation for severe faults, and closed-loop continuous monitoring, while taking into account both the power supply continuity of off-grid microgrids and the safety of equipment operation.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A hierarchical self-healing protection and control method for power imbalance in off-grid wind-solar DC microgrids includes the following steps:

[0010] Collect time-series sampling data of the electrical quantities of the off-grid wind-solar DC microgrid, as well as the branch operation data of each microgrid branch; perform sliding statistical processing on the time-series sampling data of the electrical quantities of the microgrid using a pre-configured sliding evaluation window, extract imbalance deviation features, and determine the power imbalance level and cause of the off-grid wind-solar DC microgrid based on the imbalance deviation features;

[0011] When the power imbalance level is the first imbalance level, a set of removable branches and the operating characteristics of each removable branch are generated based on the branch operation data; a swarm intelligence optimization model is constructed with the goal of minimizing the microgrid operation cost; the ant colony algorithm is used to iteratively optimize the swarm intelligence optimization model to obtain the fault branch isolation scheme, determine the fault branches to be removed, and perform the fault branch isolation operation.

[0012] When the power imbalance level is the second imbalance level, a set of candidate protection actions and their execution priorities are matched and filtered from a preset action resource pool. According to the order of the execution priorities, trial control quantities are output to the candidate protection actions in sequence and trial verification is performed: electrical feedback data is collected within a preset verification time window to determine whether the trial verification of the candidate protection action passes; if it passes, the current candidate protection action is executed; if it fails, the candidate protection action with the next execution priority is selected and the trial verification is performed again. If all candidate protection actions in the candidate protection action set fail the trial verification, the level is upgraded to the first imbalance level and the fault branch isolation operation is performed.

[0013] Prioritizing the collection of microgrid operating electrical quantities at each sampling moment according to a preset sampling period, time-series sampling data of microgrid operating electrical quantities is continuously collected. Simultaneously, branch operating data of each microgrid branch is collected to ensure time alignment between the branch operating data and the time-series sampling data of microgrid operating electrical quantities. A pre-configured sliding evaluation window is used to perform sliding statistical processing on the time-series sampling data of microgrid operating electrical quantities. For each sampling moment within the sliding evaluation window, the difference between total power generation and total load power is calculated to obtain the instantaneous power difference at that sampling moment. Statistical calculations are performed on the instantaneous power differences at each sampling moment within the window to obtain the window statistical power difference. Time-series data of DC bus voltage within the window is extracted, and the average value of the DC bus voltage within the window is calculated. The difference between the average value of the DC bus voltage and the rated reference voltage of the DC bus is used as the bus voltage offset. The imbalance deviation characteristics are obtained by combining the window statistical power difference and the bus voltage offset. Based on the imbalance deviation characteristics, the power imbalance level and causes of the off-grid wind-solar DC microgrid are determined.

[0014] Preferredly, the power imbalance level is determined by comparing the absolute value of the window statistical power difference with a preset imbalance threshold. The preset imbalance thresholds include a first imbalance threshold and a second imbalance threshold, with the first imbalance threshold being greater than the second imbalance threshold. When the absolute value of the window statistical power difference is less than the second imbalance threshold, it is determined to be a normal power condition; when the absolute value of the window statistical power difference is greater than or equal to the second imbalance threshold and less than the first imbalance threshold, it is determined to be at the second imbalance level; when the absolute value of the window statistical power difference is greater than or equal to the first imbalance threshold, it is determined to be at the first imbalance level. The preset imbalance threshold is obtained by multiplying the rated power of the off-grid wind-solar DC microgrid by a corresponding proportional coefficient, achieving adaptive configuration of the threshold according to the system capacity.

[0015] Prioritize determining the cause of power imbalance by comparing the total generated power and total load power after sliding statistical processing, and combining this with the offset direction of the DC bus voltage. When the total generated power is greater than the total load power and the DC bus voltage offsets positively, it is determined to be an imbalance due to overgeneration; when the total load power is greater than the total generated power and the DC bus voltage offsets negatively, it is determined to be an imbalance due to insufficient power supply. Prioritize performing power-voltage consistency verification to identify abnormal sampling conditions and instantaneous disturbances. Configure a consistency verification threshold. When the theoretical voltage offset direction corresponding to the comparison between the total generated power and the total load power is inconsistent with the actual offset direction of the DC bus voltage, it is determined to be a characteristic conflict condition, and the conflict deviation is calculated. When the conflict deviation does not exceed the consistency verification threshold, it is determined to be small noise and ignored; when the conflict deviation exceeds the consistency verification threshold, it is determined to be an abnormal sampling condition, the statistical results of this sliding evaluation window are discarded, and the statistics and judgment of the next window are repeated.

[0016] Prioritizing the generation of power imbalance at the first imbalance level, a set of severable branches and their operational characteristics are generated based on branch operation data. A swarm intelligence optimization model is constructed with the goal of minimizing microgrid operation cost. An ant colony optimization algorithm is used to iteratively optimize the model, solving for faulty branch isolation schemes, identifying faulty branches to be isolated, and performing faulty branch isolation operations. The set of severable branches includes all severable generating branches and controllable load branches; the operational characteristics include the power, switching status, and importance level of each branch.

[0017] Prioritizing the first imbalance level, where the off-grid wind-solar DC microgrid is nearing instability, the trial-and-error verification steps are skipped. Instead, the optimal isolation scheme is determined and executed directly through ant colony optimization to minimize processing delays and prioritize equipment hardware safety. According to the output isolation scheme, isolation control commands are issued to the corresponding branch's switching devices to complete the branch disconnection. After issuing the control commands, electrical feedback data from the corresponding branch's switching devices is collected to confirm that all switches have reliably disconnected and the branch has achieved electrical isolation.

[0018] Prior to this, when the power imbalance level is the second imbalance level, a set of candidate protection actions and their execution priorities are matched and screened from the preset action resource pool. The candidate protection actions are then tested and verified in order of execution priority.

[0019] Prioritizes the storage of multiple protection actions within the action resource pool. Each action is pre-bound with an appropriate imbalance level label and imbalance cause label, and is sorted according to a preset priority order. The preset priority ordering principle prioritizes power regulation actions, followed by load shedding actions. Power regulation actions include output power regulation of wind and solar power generation units and charging and discharging regulation of energy storage units, while load shedding actions include controllable load switching operations. Higher priority means that the action will be selected first during trial verification.

[0020] Prior to this, the specific steps of the trial verification include: obtaining the complete execution control quantity corresponding to the current candidate protection action; multiplying the complete execution control quantity by a preset trial ratio coefficient to generate a small trial control quantity; sending the small trial control quantity to the corresponding execution unit, starting a timer, and maintaining a preset observation and verification duration; and continuously collecting electrical feedback data of the microgrid within the preset observation and verification duration. The electrical feedback data includes time-series sampling data of DC bus voltage, total generating power, and total load power.

[0021] Prioritize determining whether the trial verification of the current candidate protection action passes based on electrical feedback data: if the electrical feedback data indicates that the absolute value of the window statistical power difference decreases, the bus voltage returns to its rated value, and the DC bus voltage remains within the allowable safety threshold range, then the trial verification is deemed to have passed; if the electrical feedback data indicates that the absolute value of the window statistical power difference increases, the bus voltage deviation further intensifies, the DC bus voltage exceeds the safety threshold, or power oscillation intensifies, then the trial verification is deemed to have failed. If the trial verification passes, the current candidate protection action is executed, i.e., a complete execution control quantity is sent to the corresponding execution unit; if the trial verification fails, the current candidate protection action is abandoned, and the next candidate protection action with the next execution priority is selected to re-execute the trial verification.

[0022] Prior to this, if all candidate protection actions in the candidate protection action set fail the test verification, the current operating condition is upgraded to the first imbalance level, and the fault branch isolation operation is performed.

[0023] Prior to this, after the fault branch isolation operation or candidate protection action is completed, the timing sampling data of the microgrid's operating electrical quantities and the branch operation data of each microgrid branch are collected again to enter the next round of power imbalance monitoring and judgment process.

[0024] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0025] This invention addresses the technical problem that random fluctuations in wind and solar power generation and frequent load switching in off-grid wind-solar DC microgrids can easily lead to power imbalance and cause drastic deviations in DC bus voltage. It provides a complete set of hierarchical self-healing protection and control schemes, from accurate identification and hierarchical classification of power imbalance to flexible execution of protection actions and closed-loop monitoring. By using a sliding evaluation window to perform sliding statistical processing on time-series sampled data, the system comprehensively analyzes multi-dimensional statistics to accurately output the power imbalance level and its causes. This overcomes the shortcomings of existing technologies that rely solely on instantaneous sampled data and are susceptible to noise interference, leading to false protection. Furthermore, for the first imbalance level, the system skips trial verification and directly uses an ant colony algorithm to iteratively optimize and quickly determine the optimal fault branch isolation scheme, prioritizing equipment hardware safety. For the second imbalance level, a closed-loop flexible adjustment mechanism of "trial-verification-execution" is used. Small trial control values ​​are used to observe the microgrid response, and only after successful verification is the full protection action executed. This avoids direct output of complete control commands that could cause severe bus voltage fluctuations and unnecessary load shedding, balancing power supply continuity and protection reliability. After the protection action is completed, the system returns to the data acquisition step, forming a closed-loop continuous monitoring system, enabling continuous tracking of the microgrid's dynamic operating conditions and multiple protection responses. Attached Figure Description

[0026] Figure 1 A flowchart of a hierarchical self-healing protection and control method for power imbalance in off-grid wind-solar DC microgrids;

[0027] Figure 2 A schematic diagram illustrating the classification of power imbalance levels;

[0028] Figure 3 Flowchart for selecting candidate protection actions for power imbalance;

[0029] Figure 4 This is a schematic diagram of the closed-loop monitoring process after the complete protection action is executed. Detailed Implementation

[0030] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof.

[0031] Please see Figure 1 This embodiment provides a hierarchical self-healing protection and control method for power imbalance in off-grid wind-solar DC microgrids, which includes the following steps:

[0032] S1: Collect time-series sampling data of the electrical quantities of the off-grid wind-solar DC microgrid, as well as the branch operation data of each microgrid branch; perform sliding statistical processing on the time-series sampling data of the electrical quantities of the microgrid using a pre-configured sliding evaluation window, extract imbalance deviation features, and determine the power imbalance level and cause of the off-grid wind-solar DC microgrid based on the imbalance deviation features.

[0033] The off-grid wind-solar DC microgrid is an independent power system that does not rely on the main power grid for power supply. It uses distributed generation units such as wind and solar power and energy storage units as the core power source and supplies power to the load units through the DC bus. Off-grid wind-solar DC microgrids are widely used in scenarios without main power grid power supply, such as islands, remote rural areas, and field operation sites. Due to the strong random fluctuations in wind and solar power generation and frequent load switching, it is easy to cause an imbalance between the system's power generation and load power, resulting in a sharp deviation in the DC bus voltage. Power imbalance protection and control is a key technical issue for off-grid wind-solar DC microgrids.

[0034] The microgrid operating electrical quantities are a set of core electrical quantities characterizing the power-voltage balance state of off-grid wind-solar-DC microgrids; the microgrid operating electrical quantities include total generated power. Total load power The DC bus voltage, the total power generation is the sum of the output power of all power generation units in the off-grid wind-solar DC microgrid, the total load power is the sum of the power consumed by all load units in the off-grid wind-solar DC microgrid, and the DC bus voltage is the real-time voltage of the DC bus of the off-grid wind-solar DC microgrid.

[0035] The microgrid branch is an electrical branch in an off-grid wind-solar DC microgrid that can be independently connected or disconnected. For example, the microgrid branch includes a generation branch and a load branch; the generation branch connects to a generation unit, and the load branch connects to a load unit. The branch operation data is a set of data characterizing the real-time operating status of each microgrid branch; for example, the branch operation data includes branch power, branch switch status, and branch importance level, where the branch importance level is a preset parameter characterizing the impact of branch disconnection on microgrid operation.

[0036] Furthermore, S1 includes:

[0037] S11: According to the preset sampling period, continuously collect the microgrid operating electrical quantities at each sampling time to form the time-series sampling data of the microgrid operating electrical quantities;

[0038] The preset sampling period is a pre-configured time interval between two adjacent samples. It is set based on the sampling theorem's requirement for sampling frequency, that is, the sampling frequency corresponding to the preset sampling period is not less than twice the highest effective frequency component in the microgrid's operating electrical quantity, so as to ensure that the timing characteristics of microgrid power fluctuations can be fully captured and to avoid the loss of timing characteristics due to undersampling. For example, if the fastest effective change period of microgrid power fluctuations is about 1 second, then the preset sampling period should not be greater than 0.5 seconds. For example, the preset sampling period can be set to 0.1 seconds.

[0039] S12: According to the preset sampling period, synchronously collect the branch operation data of each microgrid branch, so that the branch operation data and the time-series sampling data of the microgrid operation electrical quantity are aligned in the time dimension, forming a data acquisition basis with the same source and the same sequence.

[0040] S13: Perform sliding statistical processing on the time-series sampling data of the microgrid's operating electrical quantities using a pre-configured sliding evaluation window to extract imbalance deviation features.

[0041] The sliding evaluation window is a pre-configured fixed-length time-series statistical window. As sampling data continuously flows in, the earliest set of sampling data within the sliding evaluation window is discarded, and the latest set of sampling data is included, thus realizing the sliding update of the time-series dataset. The window length of the sliding evaluation window must ensure that the number of sampling points within the window is not less than a preset sampling point threshold. The preset sampling point threshold is the lower limit of the number of sampling points required to ensure that the window statistical results have statistical significance. Its setting is based on the sample size requirement required to reduce the risk of misjudgment due to random sampling noise. For example, the preset sampling point threshold can be set to 20 sets of sampling points.

[0042] The imbalance deviation feature is a characteristic quantity used to characterize the degree of power imbalance deviation of the microgrid, obtained by statistical analysis of the time-series data of the microgrid's operating electrical quantities within the sliding evaluation window; the imbalance deviation feature includes the window statistical power difference. The bus voltage offset is the statistical power difference between the total power generation and the total load power at each sampling time within the sliding evaluation window, and the bus voltage offset is the offset of the average DC bus voltage relative to the rated reference voltage of the DC bus within the sliding evaluation window.

[0043] Furthermore, S13 includes:

[0044] S131: For each sampling time within the sliding evaluation window, calculate the difference between the total power generation and the total load power at that sampling time to obtain the instantaneous power difference at that sampling time; the instantaneous power difference is an instantaneous quantity characterizing the degree of deviation between the power generation and load supply and demand of the microgrid at that sampling time.

[0045] S132: Perform statistical calculations on the instantaneous power differences at each sampling time within the sliding evaluation window to obtain the window statistical power difference; for example, the statistical calculations include calculating the mean, fluctuation variance, and maximum offset of the instantaneous power differences within the sliding evaluation window; the window statistical power difference is a statistical value obtained after performing statistical calculations on the instantaneous power differences at multiple sampling times within the sliding evaluation window, used to represent the degree of deviation between microgrid generation and load supply and demand power.

[0046] S133: Extract the time-series data of the DC bus voltage within the sliding evaluation window, calculate the average value of the DC bus voltage within the sliding evaluation window, and use the difference between the average value of the DC bus voltage and the rated reference voltage of the DC bus as the bus voltage offset; the rated reference voltage of the DC bus is the factory-preset rated operating voltage of the DC bus of the off-grid wind-solar DC microgrid system; the bus voltage offset is used to characterize the degree to which the current DC bus deviates from the rated operating point.

[0047] S134: By combining the statistical power difference in the window with the bus voltage offset, the imbalance deviation characteristics are obtained for subsequent determination of the power imbalance level and the cause of the power imbalance.

[0048] S14: Compare the imbalance deviation characteristics with a preset imbalance threshold to determine the power imbalance level of the off-grid wind-solar DC microgrid. The preset imbalance threshold is a power deviation threshold based on the rated power configuration of the off-grid wind-solar DC microgrid, used to quantify the severity of the power imbalance; for example, the preset imbalance threshold includes a first imbalance threshold and a second imbalance threshold, and the first imbalance threshold is greater than the second imbalance threshold; the preset imbalance threshold is obtained by multiplying the rated power of the off-grid wind-solar DC microgrid by a corresponding proportional coefficient, so that the preset imbalance threshold adapts to the microgrid capacity. For example, the first imbalance threshold is obtained by multiplying the rated power of the off-grid wind-solar DC microgrid by a first proportional coefficient. The second imbalance threshold is obtained by multiplying the rated power of the off-grid wind-solar DC microgrid by a second proportional coefficient. We obtain, and satisfy 0 < < <1; for example, the first proportionality coefficient It can be set to 0.3, the second proportionality coefficient It can be set to 0.1. If the rated power of the off-grid wind-solar DC microgrid is 100kW, then the first imbalance threshold is 30kW and the second imbalance threshold is 10kW.

[0049] Furthermore, S14 includes:

[0050] S141: Based on the rated power of the off-grid wind-solar DC microgrid and the corresponding proportional coefficient, configure the first imbalance threshold. With the second imbalance threshold This allows the preset imbalance threshold to be adaptively configured to follow the microgrid capacity.

[0051] Please see Figure 2 , Figure 2 This is a schematic diagram illustrating the power imbalance level classification of the present invention; the horizontal axis represents the statistical power difference within the window. The absolute value of.

[0052] S142: Compare the absolute value of the window statistical power difference with the first imbalance threshold and the second imbalance threshold; when the absolute value of the window statistical power difference is less than the second imbalance threshold, the microgrid is determined to be in normal power operation; when the absolute value of the window statistical power difference is greater than or equal to the second imbalance threshold and less than the first imbalance threshold, it is determined to be a second imbalance level; when the absolute value of the window statistical power difference is greater than or equal to the first imbalance threshold, it is determined to be a first imbalance level.

[0053] S15: Compare the total power generation and total load power after sliding statistical processing, and combine this with the offset direction of the DC bus voltage to determine the cause of power imbalance in the off-grid wind-solar DC microgrid. The cause of power imbalance includes power generation excess imbalance and power supply deficiency imbalance.

[0054] Furthermore, S15 includes:

[0055] S151: Obtain statistical values ​​of total power generation and total load power after sliding statistical processing from the imbalance deviation characteristics, as well as the offset direction of DC bus voltage; the offset direction includes positive offset and negative offset. When the bus voltage offset is greater than zero, the DC bus voltage is positively offset, and when the bus voltage offset is less than zero, the DC bus voltage is negatively offset.

[0056] S152: Compare the total power generation after sliding statistical processing with the total load power, and combine the offset direction of the DC bus voltage to determine the cause of the power imbalance of the off-grid wind-solar DC microgrid; when the total power generation after sliding statistical processing is greater than the total load power and the DC bus voltage is offset in the positive direction, it is determined to be an imbalance caused by over-generation; when the total load power after sliding statistical processing is greater than the total power generation and the DC bus voltage is offset in the negative direction, it is determined to be an imbalance caused by insufficient power supply.

[0057] In another embodiment, step S15 further includes the identification and processing of abnormal operating conditions:

[0058] A consistency verification threshold is pre-configured. The consistency verification threshold is a pre-configured numerical threshold used to identify abnormal sampling conditions where the power statistics result and the voltage offset direction contradict each other. The threshold is set based on the measurement accuracy of the sensor. For example, the consistency verification threshold can be set to a value between 0.04 and 0.06.

[0059] When the theoretical voltage deviation direction corresponding to the comparison between the total power generation and the total load power is inconsistent with the actual deviation direction of the DC bus voltage, it is determined to be a characteristic conflict condition. For example, when the total power generation is greater than the total load power, the DC bus voltage should theoretically deviate in the positive direction. If the DC bus voltage actually deviates in the negative direction at this time, it is determined to be a characteristic conflict condition. When the total load power is greater than the total power generation, the DC bus voltage should theoretically deviate in the negative direction. If the DC bus voltage actually deviates in the positive direction at this time, it is determined to be a characteristic conflict condition.

[0060] The conflict deviation is a deviation value used to quantify the degree of contradiction between the power comparison relationship and the voltage offset direction when the characteristic conflict condition occurs; for example, the conflict deviation is taken as the absolute value of the bus voltage offset.

[0061] When the conflict deviation does not exceed the consistency verification threshold, it is determined to be small noise, the conflict is ignored, and subsequent protection control processing continues; when the conflict deviation exceeds the consistency verification threshold, it is determined to be an abnormal sampling condition, the statistical results of this sliding evaluation window are discarded, and the data statistics and power imbalance cause determination of the next sliding evaluation window are carried out again.

[0062] For example, if the consistency verification threshold is set to 0.05, and the total power generation after sliding statistical processing is 8kW and the total load power is 5kW, the total power generation is greater than the total load power, and theoretically the DC bus voltage should shift positively. When the DC bus voltage shift is -0.02, the DC bus voltage is actually shifted negatively, which is inconsistent with the theoretical positive shift direction. This is determined to be a characteristic conflict condition, and the conflict deviation is taken as the absolute value of the bus voltage shift of 0.02, which is less than the consistency verification threshold of 0.05. This is determined to be small noise, and the conflict is ignored, and subsequent protection control processing continues. When the DC bus voltage shift is -0.08, the DC bus voltage is actually shifted negatively, which is inconsistent with the theoretical positive shift direction. This is determined to be a characteristic conflict condition, and the conflict deviation is taken as the absolute value of the bus voltage shift of 0.08, which is greater than the consistency verification threshold of 0.05. This is determined to be a sampling abnormal condition, and the statistical results of this sliding evaluation window are discarded. The data statistics and power imbalance cause determination of the next sliding evaluation window are performed again.

[0063] When the power imbalance level is the first imbalance level, a set of removable branches and the operating characteristics of each removable branch are generated based on the branch operation data; a swarm intelligence optimization model is constructed with the goal of minimizing the microgrid operation cost; the ant colony algorithm is used to iteratively optimize the swarm intelligence optimization model to obtain the fault branch isolation scheme, determine the fault branches to be removed, and perform the fault branch isolation operation.

[0064] The removable branches refer to all removable power generation branches and controllable load branches. The set of removable branches is generated by filtering based on the switching status and controllability of each microgrid branch, traversing all microgrid branches, removing currently removable branches and uncontrollable branches, and forming the set of removable branches with the remaining controllable branches in operation. This set is used to limit the decision variable space for faulty branch isolation. The branch operation data is a set of data characterizing the real-time operation status of each microgrid branch. The operation features are extracted by reading the power, current switching status, and importance level of each branch in the set of removable branches from the branch operation data. This data is used to provide data support for the branch weighted loss calculation in the objective function. The swarm intelligence optimization model aims to minimize the comprehensive operation cost of the off-grid wind-solar DC microgrid after removal. Its objective function includes a weighted combination of three indicators: the weighted loss of the removable branch, the power deviation after removal, and the bus voltage offset. The faulty branch is the branch to be removed determined based on the optimal isolation scheme output by the ant colony algorithm. The controllable load refers to the load branch in the off-grid wind-solar DC microgrid that can be switched on or off by switching devices, and its disconnection will not have an unacceptable impact on the critical operating functions of the microgrid.

[0065] Furthermore, S2 includes:

[0066] S21: Select the corresponding processing branch based on the power imbalance level output by S1; if the current imbalance level is the first level, proceed to S221 to construct a swarm intelligence optimization model for the isolation scheme of the fault branch of the first imbalance level, use the ant colony algorithm to solve for the fault branch of the first imbalance level and perform the isolation operation; if the imbalance level is the second level, exit this S2 step, proceed to the subsequent S3 step to perform the trial verification process and select the corresponding protection operation based on the trial results.

[0067] Specifically, S21 includes: constructing a swarm intelligence optimization model for the isolation scheme of fault branches at the first imbalance level through a swarm intelligence algorithm, and using an ant colony algorithm to solve the swarm intelligence optimization model, outputting the optimal isolation scheme, determining the fault branch isolation scheme, and executing the fault branch isolation operation.

[0068] S221: Construct a swarm intelligence optimization model for isolating faulty branches at the first imbalance level. This model transforms the faulty branch isolation problem at the first imbalance level into a combinatorial optimization problem with the operational cost of the off-grid wind-solar DC microgrid after disconnection as the objective function. The decision variables are defined as the switching states of each disconnectable branch. Based on the operational characteristics of each branch extracted in S2, corresponding weight coefficients are assigned to each disconnectable branch. Based on the set of severable branches generated by S2, it is determined that there are N severable branches in the off-grid wind-solar DC microgrid, including generation branches and controllable load branches. The switch status of the branch circuit is , This indicates that the branch has been removed. If the branch is retained, then the isolation scheme vector is... .

[0069] The combined optimization problem aims to minimize the overall operational cost after branch disconnection, while satisfying power balance constraints, bus voltage safety constraints, and branch disconnection quantity constraints. It seeks the optimal isolation scheme from all possible branch disconnection combinations. The decision variables are the switching states of each disconnectable branch. , This indicates that the branch has been removed. This indicates that the branch will be retained; the isolation scheme is a vector determined by the combination of values ​​of all decision variables, and the set of faulty branches to be removed consists of branches with a value of 1. The objective function is to minimize the integrated operation cost of the off-grid wind-solar DC microgrid after removal, and its mathematical expression is as follows:

[0070]

[0071] in, For the first The weighting coefficients for each branch are set according to the branch type: For load branches, the weighting coefficient is determined based on the importance level and load capacity of the load; the higher the importance level and the larger the load capacity, the larger the weighting coefficient. For power generation branches, the weighting coefficient is determined based on the proportion of the power generation capacity of the branch; the larger the power generation capacity, the larger the weighting coefficient. For example, the weighting coefficient can be in the range of 0. 1; This represents the total remaining power generation capacity after the removal. This represents the total remaining load power after the load is cut off; among which, It equals the total generating power minus the power of the disconnected generating branches. The total load power minus the power of the disconnected load branch, the , , These are the weighting coefficients for each indicator. Corresponding to the weighted loss term of the cut-off branch, Corresponding power deviation term, For the corresponding voltage deviation term, the weighting coefficient satisfies The configuration can be tailored to the specific off-grid wind-solar DC microgrid's emphasis on power supply continuity, power balance, and voltage stability. Constraints include: power balance constraints, bus voltage safety constraints, and branch disconnection quantity constraints. Specifically, the power balance constraint requires that the absolute value of the difference between the remaining generating power and the remaining load power after disconnection be less than a preset allowable deviation. It satisfies the following formula:

[0072]

[0073] The preset allowable deviation The threshold for determining whether the supply and demand power of an off-grid wind-solar DC microgrid reaches an acceptable balance after disconnection is physically defined as: the maximum tolerable difference between the generated power and the load power within the allowable range of normal operation of the off-grid wind-solar DC microgrid. When the absolute value of the difference between the remaining generated power and the remaining load power after disconnection is less than or equal to... At this time, the off-grid wind-solar DC microgrid can automatically suppress the residual deviation through its own control methods such as energy storage buffering and voltage regulation, without the need for further branch disconnection or power regulation operations. The preset allowable deviation... The value is set comprehensively based on the rated power, load fluctuation characteristics, energy storage regulation capability, and measurement accuracy of the off-grid wind-solar DC microgrid. For example, for an off-grid wind-solar DC microgrid with a rated power of 100kW, small load fluctuation, and fast energy storage response speed, the preset allowable deviation is... It can be set to 2kW.

[0074] Specifically, the bus voltage safety constraint is the DC bus voltage after disconnection. Not lower than the safety lower limit And not higher than the safety limit It satisfies the following formula:

[0075]

[0076] Specifically, the constraint on the number of branches to be removed is that the number of branches removed in a single isolation operation does not exceed a preset maximum value. This is to avoid excessive disconnection that could lead to a severe decrease in the system's power supply capacity.

[0077] S222: Solve the swarm intelligence optimization model S221 using the ant colony optimization algorithm. The specific solution steps are as follows:

[0078] Specifically, based on the first imbalance level information, the current set of optional branches S is determined. This set of optional branches S includes all severable power generation branches and controllable load branches. Ant colony algorithm parameters are set as follows: number of ants. Pheromones importance factor Heuristic function importance factor pheromone volatility coefficient Maximum number of iterations Initialize the pheromone concentration of each branch. The The constant indicates that each branch has the same probability of being selected at the initial moment. For example, It can be set to 1.

[0079] The number of ants is the total number of ants used to search for isolation schemes in each iteration, used to control the parallel search capability of the algorithm; the pheromone importance factor is a weighted index of pheromone concentration in path selection, used to adjust the influence of historical search experience on the current decision; the heuristic function importance factor is a weighted index of heuristic information in path selection, used to adjust the influence of prior information such as branch importance on the current decision; the pheromone evaporation coefficient is the evaporation rate of pheromone during the iteration process, used to control the balance between new and old search experience in the algorithm; the maximum number of iterations is one of the algorithm termination criteria, used to limit the maximum number of iterations the algorithm can run; the pheromone concentration is the accumulated search experience value of the branch during the iteration process, used to record the success rate of each branch in being selected to build a high-quality isolation scheme in the historical search, and the higher the concentration of the branch, the greater the probability of it being selected in subsequent iterations. The settings of the number of ants, pheromone importance factor, heuristic function importance factor, pheromone evaporation coefficient, maximum number of iterations, and pheromone concentration are based on the size of the set of removable branches and the optimization convergence stability requirements. For example, the pheromone importance factor can be set to 1, the heuristic function importance factor can be set to 2, the pheromone evaporation coefficient can be set to 0.1, the number of ants can be set to 20, and the preset maximum number of iterations can be set to 50. The pheromone importance factor and the heuristic function importance factor together determine the relative influence of historical experience and current branch features during the optimization process, and the pheromone evaporation coefficient controls the forgetting rate of historical pheromones to avoid the algorithm getting trapped in local optima.

[0080] Specifically, Each ant is placed at the starting node. Following the rules of roulette wheel selection, each ant chooses a path from the set of available paths S to construct a complete isolation scheme. In the Select a branch in the next iteration. The probability is:

[0081]

[0082] in, For the first Branches during the next iteration The concentration of pheromones on the surface; branch road The heuristic function value is the reciprocal of the branch importance. This indicates that the branch with the lower cost of removal is preferred; For ants The current set of available branches.

[0083] The starting point is the first node in the ant colony algorithm where each ant begins constructing its isolation scheme. Each ant starts from the first node and proceeds sequentially from the set of available paths. The process continues until a complete set of cutoff branches is constructed. The roulette wheel rule is a probabilistic random selection mechanism in the ant colony algorithm for choosing branches: for ants... Calculate the probability of each of the currently available branches being selected. The higher the probability of a branch being selected, the larger the sector area it occupies on the roulette wheel. Ants select the next branch according to this probability distribution by randomly spinning the roulette wheel, making branches with higher probabilities more likely to be selected, while branches with lower probabilities still have a chance of being selected, thus maintaining the algorithm's exploration capability. Once an ant selects a branch, that branch is removed from the current set of available branches. If a path is removed from the list, the ant continues to select the next path from the updated set of available paths using the same probability selection mechanism, until a complete isolation scheme vector is constructed. The heuristic function value represents the... The prior preference value for the priority of cutting off a branch is determined based on the branch weight coefficient. The smaller the branch weight coefficient, the larger the heuristic function value, and the higher the probability of being selected by ants during path construction. This coefficient is calculated once during algorithm initialization and remains unchanged throughout the iteration process. It works in conjunction with dynamically updated pheromone concentrations during probabilistic selection to guide ants to prioritize branches with lower cut costs. For example, suppose an off-grid wind-solar DC microgrid contains three cuttable branches: the photovoltaic branch PV1 accounts for 25% of the power generation capacity. , PV2 power generation capacity accounts for 12.5% ​​of the total capacity of photovoltaic branches. , The core load capacity of PV3 in the load branch is 15kW. , Therefore, PV2 has the largest heuristic function value and is preferentially selected for cutting during path construction, which aligns with the optimization direction of minimizing cutting costs.

[0084] Specifically, after each ant completes its path selection, it follows the objective function in S221. Calculate the fitness value of the isolation scheme and record the optimal scheme in the current iteration round. .

[0085] The fitness value is a numerical indicator used to quantify the quality of the isolation scheme. The smaller the value, the lower the overall system operating cost after the removal of the scheme, and the better the scheme. In each iteration, by comparing the fitness values ​​of all the schemes constructed by the ants, the scheme with the smallest fitness value is selected as the isolation scheme for the current iteration round. This is used to enhance the pheromone of the high-quality scheme during subsequent pheromone updates, guiding subsequent iterations to converge toward a better solution.

[0086] Specifically, after all ants have completed path construction, the pheromone concentration of each branch is updated:

[0087]

[0088] in, For ants Remaining on the branch in this iteration The increase in pheromones on the ants; if ants The branch was removed in the isolation plan. ,but ,otherwise ; The pheromone constant in the pheromone update formula controls the magnitude of the total amount of pheromone released by the ant on the selected branch in each iteration. For example, in this example, it can be taken as... ,or . For ants The objective function value of the constructed isolation scheme. The smaller the objective function value, the greater the corresponding pheromone increment, thus guiding subsequent ants to tend to choose the branch combinations within that isolation scheme. This is the pheromone evaporation coefficient in the pheromone update formula, representing the proportion of the original pheromone concentration retained in each branch after each iteration. Its value ranges from [value missing]. .

[0089] The objective function value is the objective function. The value of is used to quantify and evaluate the overall operational cost of the isolation scheme. A smaller objective function value indicates a lower overall operational cost of the system after the isolation scheme is removed, and a better isolation scheme. During the pheromone update process, the objective function value... Used to calculate pheromone increment The smaller the objective function value, the greater the pheromone increment of the isolation scheme, which makes the branch corresponding to the isolation scheme more likely to be selected in subsequent iterations, guiding the algorithm to converge toward the global optimal solution.

[0090] S223: The optimal isolation scheme output by the ant colony algorithm. Perform faulty branch isolation. The optimal isolation scheme, after convergence of the ant colony algorithm iterations, outputs the objective function. The isolation scheme with the smallest value; [This is followed by a seemingly unrelated sentence about an isolation scheme.] The corresponding branch switch device issues an isolation control command to complete the disconnection of the faulty branch. After issuing the control command, electrical feedback data of each corresponding branch switch device is collected to confirm that all switches have been reliably disconnected and the branch has achieved electrical isolation. After the isolation action is confirmed, the process returns to S1 and re-enters the loop of continuously collecting microgrid electrical data to reassess the microgrid power imbalance status and achieve closed-loop monitoring.

[0091] The electrical feedback data is a set of data characterizing the actual operating response status of the microgrid after receiving control commands, including total power generation sampling values, total load power sampling values, DC bus voltage, and status signals and current sampling values ​​of switching devices. The statement that all switches have been reliably disconnected includes: judging based on the collected electrical feedback data; if the collected status signal changes from "on" to "off", then it is determined that the branch switch has been reliably disconnected.

[0092] For example, the rated power of a certain off-grid wind-solar DC microgrid First Imbalance Threshold Second imbalance threshold Based on time-series sampling data within the sliding evaluation period, the window statistical power difference is calculated. ,because The current imbalance level is determined to be the first level, and the total power generation is read. Total load power The total generating power exceeds the total load power, and the bus voltage deviates positively, indicating that the imbalance is caused by over-generation. Proceeding to step S2, skipping the trial-and-error verification phase, a fault branch isolation optimization model is constructed. The objective function is to minimize the system operating cost after fault removal, and an ant colony algorithm is used for iterative solution. The ant colony algorithm parameters are set as follows: number of ants... Pheromones importance factor Heuristic function importance factor pheromone volatility coefficient Maximum number of iterations The optimal isolation scheme is output after 100 iterations. Assuming the solution is and branch road ,in For photovoltaic branch circuits, For controllable load branches, disconnect and Branch circuit, remaining total power generation after disconnection Remaining total load power Power deviation The bus voltage drops back to near the rated value, meeting the system's safe operation requirements. The solution involves disconnecting the generating branch and controllable load branch; issuing a disconnection control command to the DC circuit breaker of that branch; collecting feedback signals from the circuit breaker to confirm that the switch disconnection is complete and the branch is isolated; after execution, returning to S1, re-collecting the total generating power, total load power, and DC bus voltage timing data, and re-determining the power imbalance level.

[0093] S3: In response to the power imbalance level being the second imbalance level, a set of candidate protection actions and their execution priorities are matched and filtered from the preset action resource pool. According to the order of the execution priorities, trial control quantities are output to the candidate protection actions in sequence and trial verification is performed: electrical feedback data is collected within the preset verification time window to determine whether the trial verification of the candidate protection action passes; if it passes, the current candidate protection action is executed; if it fails, the candidate protection action with the next execution priority is selected and the trial verification is re-executed.

[0094] The action resource pool is a pre-built database or data structure storing various protection actions and their corresponding matching conditions, including power imbalance level tags and power imbalance cause tags. The protection action set is a group of protection actions selected from the action resource pool that are applicable to the current second imbalance level, and are called sequentially according to priority in subsequent trial and verification stages. For example, the tags for improving energy storage charging power matching are the second imbalance level and the overcapacity generation type. The preset verification time window is a pre-configured fixed duration for collecting microgrid electrical feedback data after outputting the trial control quantity. Its value is configured according to the microgrid electrical inertia, energy storage response speed, and power regulation unit response characteristics to ensure that the microgrid power and voltage complete a dynamic response under the action of the trial control quantity. The specific method of matching and filtering is as follows: each protection action is pre-bound with an appropriate power imbalance level and power imbalance cause, and selected according to a preset priority order.

[0095] Furthermore, S3 includes:

[0096] S31: Combining the power imbalance level output by S1 with the determined cause of the power imbalance, candidate protection actions suitable for the current operating condition are selected from the preset action resource pool, providing candidate action input for the subsequent small-scale trial verification stage. The small-scale trial verification is a closed-loop verification process of "trial-verification" for candidate protection actions under the second imbalance level, specifically including: generating an output amplitude of 20% of the complete execution control quantity for the current candidate protection action. A small-amplitude trial control input of 50% continuously collects electrical feedback data from the microgrid within a preset verification time window. Based on this data, it determines whether the microgrid power imbalance has been alleviated and whether the bus voltage has returned to its rated value, thereby verifying whether the current candidate protection action is suitable for the current operating conditions. If the trial verification passes, the protection action is executed in its entirety. If the trial verification fails, the next priority candidate protection action is switched to for re-execution of the trial verification. This process verifies the suitability of the protection action under limited disturbance conditions, avoiding direct output of complete control commands that could lead to severe bus voltage fluctuations or unnecessary load shedding.

[0097] The output amplitude is the value of the small-amplitude probing control quantity, and its value is 20% of the full-scale execution control quantity. 50%; The complete execution control quantity is the control command that is fully output after the candidate protection action passes the trial verification, and is used to finally execute the protection action to eliminate the power imbalance of the microgrid.

[0098] Specifically, S31 includes:

[0099] S311: Obtain the power imbalance level output by S1 and confirm the current second imbalance level; obtain the cause of the power imbalance output by S1, retrieve all protection actions that match the level label and cause label from the preset action resource pool, sort them according to the preset priority from high to low, and output the candidate protection action set; this step is only executed for the second imbalance level, the first imbalance level directly jumps to S2 and does not enter this step.

[0100] For details, please refer to Figure 3 , Figure 3 This is a flowchart illustrating the selection of candidate protection actions for power imbalance in an embodiment of the present invention. After determining the cause of power imbalance, it is divided into two types of operating conditions: overcapacity generation and undercapacity supply.

[0101] For example, regarding imbalances caused by power generation overcapacity:

[0102] In the event of power generation overcapacity and imbalance, the first priority candidate protection action is to increase the charging power of energy storage. When the energy storage unit has not yet reached its maximum charging power, selecting this candidate protection action generates the corresponding complete execution control quantity. The small-scale trial control quantity is used to collect electrical feedback data within the observation window for trial verification within the preset observation and verification period. If the energy storage unit has reached its maximum charging power, or if the DC bus voltage exceeds the limit during the trial verification process, the verification is deemed unsuccessful, the current candidate protection action is abandoned, the next priority action is selected, and the power output of wind and solar power generation is switched to reduce the power output. The power reduction adjustment target value is determined in combination with the power surplus of the power generation imbalance. Based on the power reduction adjustment target value, a corresponding small-scale trial control quantity is generated, and the trial verification continues within the preset observation and verification period.

[0103] Specifically, before outputting the test control quantity, the current charging power value of the energy storage unit is read and compared with the maximum allowable charging power of the energy storage unit. If the current charging power has reached or exceeded the maximum allowable charging power, it means that the energy storage unit has no remaining charging capacity and cannot execute the protection action of "increasing the energy storage charging power". In this case, there is no need to issue a test control quantity, and the verification is directly determined to fail.

[0104] Specifically, after outputting the test control quantity, the DC bus voltage is continuously sampled over a preset observation and verification period. The real-time bus voltage value at each sampling moment is compared point by point with the pre-configured voltage safety threshold range. If the bus voltage exceeds the upper safety limit or falls below the lower safety limit at any sampling moment within the observation and verification period, it is determined that the DC bus voltage exceeds the limit. A bus voltage exceeding the limit indicates that the test control quantity has caused an excessive voltage surge to the system. The candidate protection action is not suitable for the current operating conditions, and the calibration test should be stopped immediately and determined to fail. If the bus voltage at all sampling moments within the observation and verification period is within the safety threshold range, it is determined that the voltage does not exceed the limit.

[0105] Specifically, the power difference obtained from the window statistics of the sliding evaluation cycle is used as the power surplus, and the power reduction adjustment target value is set as the power surplus, so that the total power generation of the wind and solar power generation units after the power reduction adjustment is performed can reach a balance with the total load power.

[0106] For example, regarding power shortage-type imbalance:

[0107] In the event of insufficient or unbalanced power supply, the first priority candidate protection action is to increase the energy storage discharge power. When the energy storage unit has not yet reached its maximum discharge power, selecting this candidate protection action generates the corresponding complete execution control quantity. A small-scale trial control quantity is used to collect electrical feedback data within the observation window for trial verification within a preset observation and verification period. If the energy storage unit has reached its maximum discharge power, or if the DC bus voltage exceeds the limit during the trial verification process, the verification is deemed unsuccessful, the current candidate protection action is abandoned, and the controllable load branch is switched to be disconnected. The total power of the load to be disconnected is determined in conjunction with the power deficit, and a corresponding small-scale trial control quantity is generated based on the total power of the controllable load to be disconnected. The trial verification continues within the preset observation and verification period.

[0108] Specifically, before outputting the test control quantity, the current discharge power value of the energy storage unit is read and compared with the maximum allowable discharge power of the energy storage unit. If the current discharge power has reached or exceeded the maximum allowable discharge power, it means that the energy storage unit has no remaining discharge capacity and cannot perform the protection action of "increasing the energy storage discharge power". In this case, there is no need to issue a test control quantity, and the verification is directly determined to fail.

[0109] Specifically, after outputting the test control quantity, the system continuously collects time-series sampling data of the DC bus voltage within a preset observation and verification period. The real-time bus voltage values ​​collected at each moment are compared point-by-point with a pre-configured voltage safety threshold range. If the bus voltage at any sampling moment within the observation and verification period exceeds the upper safety limit or falls below the lower safety limit, it is determined that the DC bus voltage is out of limit. An out-of-limit bus voltage indicates that the test control quantity has caused an excessive voltage surge to the system; the candidate protection action is not suitable for the current operating conditions, and the calibration test is immediately stopped and determined to fail. If the bus voltage at all sampling moments within the observation and verification period is within the safety threshold range, it is determined that the voltage is not out of limit.

[0110] Specifically, the absolute value of the window statistical power difference obtained from the sliding evaluation cycle is used as the power deficit. This power deficit represents the portion of the current load power that exceeds the power generation power, i.e., the power value that the system needs to reduce from the load side.

[0111] Specifically, the steps for disconnecting the controllable load branch are as follows: issuing a disconnection control command to the corresponding branch's switching devices, collecting electrical feedback data from each corresponding branch's switching devices, and confirming that all switches have reliably disconnected. If all candidate protection actions in the preset action resource pool fail the verification, the operating condition is directly upgraded to the first imbalance level, skipping the trial verification and performing the faulty branch isolation operation.

[0112] S312: Based on the candidate protection action, a small-amplitude trial control quantity is generated. The calculation formula for the small-amplitude trial control quantity is as follows:

[0113]

[0114] in, This is a small-scale trial to determine the control quantity; The complete execution control quantity corresponding to the candidate protection action; To test the proportionality coefficient, the range of values ​​is: The proposed trial ratio is a pre-set parameter that can be configured based on the microgrid system capacity, device tolerance, and load importance level. In this example, it can be set to [value missing]. ,when At that time, the trial control quantity is only 30% of the output quantity of the complete protection action.

[0115] S32: Output a small trial control value, start the timer, and maintain the preset observation and verification duration.

[0116] The timer is a timing unit used to measure the duration of the microgrid's electrical response after the output of the test control quantity. Its timing duration is equal to the preset observation and verification duration. After the timer is started, if the timing has not reached the preset observation and verification duration, the electrical feedback data of the microgrid will be continuously collected. When the timing reaches the preset observation and verification duration, the collection will stop and the test and verification result determination stage will begin.

[0117] Specifically, S32 includes:

[0118] S321: Issue a command to the corresponding execution unit of the microgrid. A small trial control quantity is made; at the same time, a timer is started, with the timing duration equal to the preset observation and verification duration.

[0119] The preset observation and verification duration is a pre-configured time parameter, used to allow time for the microgrid's power and voltage to complete a dynamic response. This parameter is configured based on the microgrid's electrical inertia, energy storage response speed, and power regulation unit response characteristics. As an implementation example, the preset observation and verification duration can be set to a range of values. If the observation and verification time is set too short, the microgrid electrical quantities have not yet completed the dynamic transition, and the collected data cannot truly reflect the system effect after the trial control is applied; if the observation and verification time is set too long, it will lengthen the total protection response time and reduce the real-time performance of fault handling.

[0120] The electrical inertia of the grid in the off-grid wind-solar DC microgrid refers to the rate of change of electrical quantities determined by the electromagnetic transient processes within the DC bus capacitor energy storage and power generation units and load units. This reflects the dynamic response speed of the bus voltage and system power after the microgrid is subjected to power disturbances. The greater the grid electrical inertia, the slower the voltage and power changes, and the longer the required observation and verification time. The energy storage response speed is the time required for the actual output power of the energy storage system to change from the current value to the target value after receiving the charging and discharging power adjustment command. This reflects the dynamic adjustment capability of the energy storage converter. The faster the energy storage response speed, the faster the power imbalance of the microgrid can be suppressed, and the shorter the required observation and verification time. The power adjustment unit response characteristics are the dynamic response characteristics of the output power of the wind and solar power generation unit converter after receiving the power derating or scaling command, including response delay and power change rate. The better the response characteristics, the shorter the required observation and verification time.

[0121] The grid electrical inertia, the energy storage response speed, and the power regulation unit response characteristics jointly determine the configuration value of the preset observation and verification time. In actual configuration, the slowest response component is taken as the dominant factor to ensure that the observation and verification time is sufficient to cover the longest time required for the system to complete the dynamic transition.

[0122] S322: If the timer has not reached the preset observation and verification duration, the electrical feedback data of the microgrid within the observation window will be continuously collected; the electrical feedback data includes the total power generation sample value, the total load power sample value, and the DC bus voltage. .

[0123] Electrical feedback data is collected in real time by sensors in each branch and is used to characterize the actual operating response of the entire microgrid system after the application of trial control. The sampling frequency is consistent with the data acquisition frequency in step S1 to ensure that the statistical caliber of the data before and after is consistent and to avoid judgment deviation caused by inconsistent sampling frequencies.

[0124] S33: Based on the electrical feedback data in the observation window, observe the response status of the microgrid and complete the verification of the small-scale trial control.

[0125] Specifically, S33 includes:

[0126] S331: Perform statistical processing on multiple sets of electrical feedback data collected during the observation and verification period to calculate the window statistical power difference and DC bus voltage, thereby obtaining the actual response characteristics of the microgrid after the application of the complete trial control quantity. The window statistical processing here uses the same sliding statistical calculation logic as step S1 to ensure a unified imbalance judgment benchmark. Through window statistical processing, instantaneous sensor noise and random sampling disturbances are filtered out, avoiding misjudgments of verification results caused by single instantaneous sampling.

[0127] Specifically, the method for obtaining the statistical power difference of the window is as follows: for each sampling time within the observation and verification period, calculate the difference between the total power generation and the total load power at that sampling time to obtain the instantaneous power difference at that sampling time; perform statistical calculations on the instantaneous power differences of all sampling times within the observation and verification period to obtain the statistical power difference of the observation window.

[0128] The actual response characteristics are statistical characteristics obtained after statistical processing of the electrical quantities of the microgrid within the observation window after the application of the trial control quantity. The window statistical processing involves extracting statistical features from multiple sets of electrical feedback data collected during the observation and verification period using the same sliding statistical calculation logic as step S1. Specifically, for each sampling moment within the observation and verification period, the difference between the total generating power and the total load power is calculated to obtain the instantaneous power difference. Then, statistical calculations are performed on all instantaneous power differences to obtain the window statistical power difference. Simultaneously, all DC bus voltage values ​​within the observation and verification period are extracted. These two statistical quantities together constitute the actual response characteristics of the microgrid after the application of the trial control quantity, used for comparison and verification with the theoretically expected response characteristics.

[0129] S332: Compare and verify the actual response characteristics after the trial with the theoretically expected response characteristics;

[0130] The theoretical expected response characteristics are predetermined by the current cause and level of imbalance. Specifically, under an imbalance of excess power generation, the total power generation is greater than the total load power, the window statistical power difference is positive, and the absolute value of the expected power difference gradually decreases and approaches zero. Under an imbalance of insufficient power supply, the total load power is greater than the total power generation, the window statistical power difference is negative, and the expected power difference should gradually increase. If the electrical feedback data shows that the absolute value of the window statistical power difference decreases, the bus voltage returns to its rated value, and the DC bus voltage remains within the system's allowable safety threshold range, then the trial verification is considered successful. Successful verification means that the current candidate protection action is adapted to the current microgrid operating conditions and a complete execution control quantity can be applied.

[0131] If the electrical feedback data indicates that the absolute value of the statistical power difference in the window increases, the bus voltage deviation worsens, or the DC bus voltage exceeds the safety threshold, then the test verification is deemed unsuccessful. Failure to verify means that the candidate protection action is unsuitable for the current operating conditions and cannot be executed directly as a complete protection action. The current candidate protection action must be abandoned, the next priority candidate protection action switched, and the process returned to S32 to regenerate a small-amplitude test control quantity, repeating the test verification process.

[0132] S34. For the second imbalance level, after the trial verification is completed, the candidate protection action is fully executed.

[0133] Further step S34 includes: outputting the complete execution control quantity corresponding to the candidate protection action. The execution unit corresponding to the microgrid.

[0134] Specifically, S341 includes:

[0135] S341: Candidate protection actions obtained through step S31, and the complete execution control quantity corresponding to the candidate protection action. Its goal is to offset the power deviation obtained from the sliding evaluation cycle statistics, so that the microgrid power returns to a state of supply and demand balance. All parameters are taken from the statistical results of the same imbalance judgment cycle, ensuring that the operating conditions and control commands match each other, and avoiding the problem of mismatch of control quantities caused by inconsistent sampling data across cycles.

[0136] For details, please refer to Figure 4 , Figure 4 This is a schematic diagram of the closed-loop monitoring process after a complete protection action is executed. After issuing the complete control quantity to complete the protection action, the process returns to S1 for resampling and a new round of sliding evaluation. If the judgment returns to normal, monitoring continues; if power imbalance still exists, the protection process is re-entered to achieve closed-loop continuous control of dynamic operating conditions.

[0137] S342: The microgrid's corresponding execution unit issues a complete execution control statement. Perform complete protection actions;

[0138] For imbalances caused by excess power generation, the complete execution control quantity corresponds to the adjustment of energy storage charging power or the adjustment of power reduction for wind and solar power generation units. For imbalances caused by insufficient power supply, the complete execution control quantity corresponds to the adjustment of energy storage discharging power or the power cut-off of secondary controllable loads. During the issuance of the complete execution control quantity, the control output is output smoothly according to the preset rate of change to avoid abrupt changes in the control quantity from impacting the DC bus and preventing sudden voltage rises and falls. After the complete protection action is executed, it is used to eliminate power surplus or power deficit within the microgrid, suppress DC bus voltage deviation, and promote the return of the bus voltage to the normal operating range.

[0139] The preset rate of change is a pre-defined maximum allowable range of change in the control quantity per unit time when a complete protection action is performed. It is used to limit the rate of change of control commands such as the adjustment of energy storage charging and discharging power or the adjustment of power reduction of wind and solar power generation units. For example, the preset rate of change can be set to 10% of the rated power per second. For a microgrid with a rated power of 100kW, the rate of change of the control quantity shall not exceed 10kW per second.

[0140] S35: After the complete protection action is issued, return to step S1, re-collect the microgrid electrical timing sampling data, and enter the next round of power imbalance monitoring and judgment process.

[0141] Specifically, S35 includes:

[0142] S351: After the protection action is completed, the control flow is not terminated. Instead, the process jumps to S1 and continues to collect the total power generation, total load power, and DC bus voltage timing data of the off-grid wind-solar DC microgrid.

[0143] S352: Initiate a new round of sliding evaluation cycle statistics and imbalance level determination, and continuously monitor the microgrid's operating status; if the current protection action is effective, the new round of sliding evaluation cycle is determined to be normal operating condition, and the microgrid maintains the current operating status and continues to be monitored; if power imbalance still exists, the imbalance level and imbalance cause determination will be carried out again, and the subsequent protection logic will be entered again.

[0144] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A hierarchical self-healing protection and control method for power imbalance in off-grid wind-solar DC microgrids, characterized in that, Includes the following steps: Collect time-series sampling data of the electrical quantities of the off-grid wind-solar DC microgrid, as well as the branch operation data of each microgrid branch; perform sliding statistical processing on the time-series sampling data of the electrical quantities of the microgrid using a pre-configured sliding evaluation window, extract imbalance deviation features, and determine the power imbalance level and cause of the off-grid wind-solar DC microgrid based on the imbalance deviation features; In response to the power imbalance level being the first imbalance level, a set of detachable branches and the operating characteristics of each detachable branch are generated based on the branch operation data; a swarm intelligence optimization model is constructed with the goal of minimizing the microgrid operation cost; the ant colony algorithm is used to iteratively optimize the swarm intelligence optimization model to obtain the fault branch isolation scheme, determine the fault branches to be disconnected, and perform the fault branch isolation operation; When the power imbalance level is the second imbalance level, a set of candidate protection actions and their execution priorities are matched and filtered from a preset action resource pool. According to the order of the execution priorities, trial control quantities are output to the candidate protection actions in sequence and trial verification is performed: electrical feedback data is collected within a preset verification time window to determine whether the trial verification of the candidate protection action passes; if it passes, the current candidate protection action is executed; if it fails, the candidate protection action with the next execution priority is selected and the trial verification is re-executed.

2. The off-grid wind-solar DC microgrid power imbalance graded self-healing protection and control method according to claim 1, characterized in that, The steps for extracting imbalance and deviation features include: According to a preset sampling period, the electrical quantities of the microgrid are collected to form time-series sampling data of the electrical quantities of the microgrid operation; the electrical quantities of the microgrid operation include total power generation, total load power and DC bus voltage. Synchronously collect branch operation data of each microgrid branch; Using a pre-configured sliding evaluation window, the time-series sampling data of the microgrid's operating electrical quantities are subjected to sliding statistical processing to extract imbalance deviation features, including: For each sampling time within the sliding evaluation window, calculate the difference between the total power generation and the total load power at that sampling time to obtain the instantaneous power difference at that sampling time; The instantaneous power difference at each sampling time within the sliding evaluation window is statistically calculated to obtain the window statistical power difference. Extract the time-series data of the DC bus voltage within the sliding evaluation window, calculate the average value of the DC bus voltage within the sliding evaluation window, and use the difference between the average value of the DC bus voltage and the rated reference voltage of the DC bus as the bus voltage offset.

3. The off-grid wind-solar DC microgrid power imbalance graded self-healing protection and control method according to claim 2, characterized in that, The steps for determining the power imbalance level of the off-grid wind-solar DC microgrid include: Obtain the window statistical power difference; Configure a preset imbalance threshold, which includes a first imbalance threshold and a second imbalance threshold, wherein the first imbalance threshold is greater than the second imbalance threshold; The absolute value of the window statistical power difference is compared with a first imbalance threshold and a second imbalance threshold; when the absolute value of the window statistical power difference is less than the second imbalance threshold, it is determined to be normal power. When the absolute value of the window statistical power difference is greater than or equal to the second imbalance threshold and less than the first imbalance threshold, it is determined to be the second imbalance level; When the absolute value of the statistical power difference in the window is greater than or equal to the first imbalance threshold, it is determined to be at the first imbalance level.

4. The off-grid wind-solar DC microgrid power imbalance graded self-healing protection and control method according to claim 1, characterized in that, The steps for determining the cause of power imbalance in the off-grid wind-solar DC microgrid include: Obtain the total power generation, total load power, and DC bus voltage offset direction after sliding statistical processing; the offset direction includes positive offset and negative offset, when the bus voltage offset is greater than zero, it is determined to be positive offset, and when the bus voltage offset is less than zero, it is determined to be negative offset; The imbalance is determined by comparing the total power generation with the total load power and considering the direction of the DC bus voltage deviation: when the total power generation is greater than the total load power and the DC bus voltage deviates in a positive direction, it is determined to be an imbalance caused by power generation excess; when the total load power is greater than the total power generation and the DC bus voltage deviates in a negative direction, it is determined to be an imbalance caused by power supply deficiency.

5. The off-grid wind-solar DC microgrid power imbalance graded self-healing protection and control method according to claim 1, characterized in that, The preset action resource pool contains candidate protection actions. Each candidate protection action is pre-bound with a power imbalance level label and a power imbalance cause label. The combination of the same power imbalance level and the same power imbalance cause uniquely corresponds to a set of candidate protection actions. The candidate protection action corresponding to the current operating condition is obtained by matching and filtering the power imbalance level label and the power imbalance cause label.

6. The off-grid wind-solar DC microgrid power imbalance graded self-healing protection and control method according to claim 1, characterized in that, The execution methods of the candidate protection actions include: For imbalances caused by overcapacity in power generation, priority is given to issuing charging power commands to energy storage converters to increase the charging power of energy storage; when the energy storage unit has reached its maximum charging power, the operation of issuing power reduction commands to wind and solar power generation unit converters is switched to reduce the output power of wind and solar power generation. For imbalances caused by insufficient power supply, priority is given to issuing discharge power commands to the energy storage converter to increase the energy storage discharge power; when the energy storage unit has reached its maximum discharge power, the operation of issuing disconnection control commands to the switching devices of the corresponding branch is switched to cut off the controllable load. The process of disconnecting controllable loads includes issuing disconnection control commands to the switching devices of the corresponding branches, collecting electrical feedback data from the switching devices of each corresponding branch, and confirming that all switches have been reliably disconnected.

7. The off-grid wind-solar DC microgrid power imbalance graded self-healing protection and control method according to claim 1, characterized in that, Also includes: Obtain the complete execution control quantity corresponding to the current candidate protection action; Multiply the complete execution control quantity by a preset trial ratio coefficient to generate a small trial control quantity; The small-amplitude trial control quantity is sent to the corresponding execution unit; Within a preset observation and verification period, continuously collect microgrid electrical feedback data, and determine whether candidate protection actions meet the execution conditions based on the electrical feedback data; The determination of whether a candidate protection action meets the execution conditions includes: Within a preset observation and verification period, electrical feedback data including window statistical power difference and bus voltage offset is collected. At the end of the preset verification time window, the absolute value of the statistical power difference and the bus voltage offset are used to determine whether the candidate protection action meets the execution conditions.

8. The off-grid wind-solar DC microgrid power imbalance graded self-healing protection and control method according to claim 1, characterized in that, It also includes steps for identifying and handling abnormal sampling conditions: Configure a consistency verification threshold, which is a numerical threshold used to identify abnormal sampling conditions where the power statistics results and voltage offset direction contradict each other; When the theoretical voltage offset direction corresponding to the comparison between the total power generation and the total load power is inconsistent with the actual offset direction of the DC bus voltage, it is determined to be a characteristic conflict condition, and the conflict deviation is calculated. When the conflict deviation does not exceed the consistency verification threshold, it is determined to be small noise and ignored; When the conflict deviation exceeds the consistency verification threshold, it is determined to be an abnormal sampling condition. The statistical results of this sliding evaluation window are discarded, and the data statistics and power imbalance cause determination of the next sliding evaluation window are carried out again.

9. The off-grid wind-solar DC microgrid power imbalance graded self-healing protection and control method according to claim 1, characterized in that, The step of using the ant colony algorithm to iteratively optimize the swarm intelligence optimization model and solve for the fault branch isolation scheme includes: A set of selectable branches is determined based on the first imbalance level and the branch operation data, the set of selectable branches including disconnectable power generation branches and controllable load branches; Initialize the pheromone concentration of each branch; Multiple ants are used to construct isolation schemes. Each ant selects a branch from the set of available branches based on the branch pheromone concentration and the branch importance, and generates an isolation scheme vector for cutting off the branch. The fitness value of each isolation scheme is calculated according to a preset objective function, wherein the objective function is to optimize the overall operating cost of the system after the isolation is removed. Update the pheromone concentration of each branch based on the fitness value of each isolation scheme; Determine whether the iteration termination condition is met. If so, output the isolation scheme with fitness value as the fault branch isolation scheme. The fault branch isolation scheme includes the number information of at least one branch to be cut.

10. A hierarchical self-healing protection and control method for power imbalance in an off-grid wind-solar DC microgrid according to claim 9, characterized in that, Under the second imbalance level, if all candidate protection actions in the preset action resource pool fail the test and verification, the current imbalance is upgraded to the first imbalance level, and the fault branch isolation operation is performed.

Images