Photovoltaic power generation regulation method with power collection line and booster station cooperative operation

Abstract

The application discloses a photovoltaic power generation regulation and control method for coordinated operation of a power collection line and a booster station, relates to the technical field of photovoltaic power generation regulation and control, and comprises the following steps: constructing an active and reactive power ratio expression of each branch in the power collection line according to an identification result, extracting power ratio features by using a normalization processing mode, comparing the ratio differences between the branches, and generating ratio difference quantization results caused by control mode changes; comparing the ratio difference quantization results with a reactive power regulation target of the booster station, analyzing the offset relationship between bus voltage responses and reactive power outputs, and identifying the matching deviation between the reactive power regulation target of the booster station and the reactive power output capacity of each branch. The application solves the problem that the reactive power ratio differences of the branches caused by dynamic switching of inverter control strategies cannot be perceived and responded, and realizes accurate correction and coordinated optimization of reactive power regulation of the booster station.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation regulation, and particularly to a photovoltaic power generation regulation method for the coordinated operation of a collector line and a booster station. Background Art

[0002] The coordinated operation of a collector line and a booster station for photovoltaic power generation regulation refers to the real-time monitoring, data collection, intelligent analysis, and coordinated control of the operating states between the collector line (responsible for converging the electric energy generated by distributed photovoltaic modules) and the booster station (raising the voltage level to meet grid connection or power transmission requirements) in a photovoltaic power station system, so as to achieve the linkage dispatching of the two in terms of operation strategies, load distribution, fault response, energy optimization, etc., thereby improving the operating stability and power generation efficiency of the entire photovoltaic power generation system. The significance of this regulation method lies in that it can not only avoid system losses or shutdown risks caused by local overload, abnormal unilateral equipment, or uneven energy distribution, but also achieve intelligent response to dynamic meteorological conditions, grid load changes, and equipment health conditions, effectively improving the grid connection quality and operating safety of photovoltaic power, and is an important technical means for building a highly reliable and efficient intelligent photovoltaic power station.

[0003] The existing photovoltaic power generation regulation technology for the coordinated operation of a collector line and a booster station mainly relies on the integration of a SCADA (Supervisory Control and Data Acquisition) system, an EMS (Energy Management System), and protection and communication devices. By collecting the key operating parameters such as current, voltage, and power of each branch in the photovoltaic power station in real time, and combining the status information of equipment such as the main transformer, circuit breaker, and reactive power compensation device in the booster station, a centralized regulation system is formed. This system usually includes four core links: data collection and transmission, status monitoring and analysis, regulation strategy generation and command issuance, and execution feedback and optimization. Among them, data collection synchronizes the operating data of the collector line and the booster station to the regulation center through a communication interface; the status monitoring module judges abnormal situations according to set thresholds; the regulation strategy module formulates coordinated operation strategies based on goals such as load distribution, voltage quality, and system safety, such as adjusting the output of a certain inverter, switching to a standby line, starting reactive power compensation, etc.; the execution feedback link ensures that the strategy is accurately executed and makes adaptive adjustments in combination with the real-time operating status. The coherent coordination of this overall process ensures the stable, reliable, and efficient operation of the photovoltaic system in a complex operating environment.

[0004] The existing technology has the following deficiencies:

[0005] During the photovoltaic power generation regulation process involving the coordinated operation of the collection lines and the substation, some inverters, under the influence of solar irradiance fluctuations, automatically switch from constant reactive power control to power factor priority control based on their own control logic. This leads to differences in the active and reactive power output ratios between branches. Since this switching behavior is a local response mechanism of the inverter and is not incorporated into the unified regulation strategy plan of the centralized control system, the control center cannot perceive the dynamic changes in the control modes of each branch and still generates a station-wide reactive power compensation plan based on a preset unified target power factor. This results in a deviation between the actual reactive power response capability of the substation and the target voltage control command for the bus. When multiple branch control strategies drift, the substation cannot adjust the reactive power distribution in a timely manner, leading to increased local bus voltage fluctuations and some branches triggering power limiting operation due to voltage deviations. Existing photovoltaic power generation control technologies that coordinate the operation of collection lines and substations cannot dynamically correct the reactive power compensation plan of the substation based on the difference in active-reactive power ratio between branches caused by the dynamic switching of inverter power factor control strategies. This results in the continuous accumulation of control errors, leading to problems such as voltage control instability, output reduction, and abnormal system response, which seriously affects the overall control accuracy and operational reliability of photovoltaic power plants.

[0006] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] The purpose of this invention is to provide a photovoltaic power generation control method that coordinates the operation of the collector line and the booster station, so as to solve the problems in the background art mentioned above.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a photovoltaic power generation control method for coordinated operation of collection lines and step-up substations, specifically including the following steps:

[0009] S1. Collect the active power, reactive power, voltage, current and inverter operation control parameters of multiple branches in the collector line, perform time-series analysis on the active and reactive output change trends of each branch, and identify whether the inverter control mode has switched from constant reactive power control to power factor priority control.

[0010] S2. Based on the identification results, construct the active and reactive power ratio expressions for each branch in the collector line, extract the power ratio features using normalization, compare the ratio differences between each branch, and generate quantitative results of the ratio differences caused by changes in control mode.

[0011] S3. Compare the quantitative results of the ratio difference with the reactive power regulation target of the substation, analyze the offset relationship between the bus voltage response and reactive power output, and identify the matching deviation between the reactive power regulation target of the substation and the reactive power output capacity of each branch.

[0012] S4. Based on the matching deviation, and combined with the frequency of inverter control mode change and the power ratio change trend, the target voltage setpoint and reactive power output reference value of the booster station are corrected, and control commands for reactive power regulation are generated.

[0013] S5. Based on the corrected control commands, continuously collect the operating data of each branch in the collector line, track the changing trend of the control mode and the evolution trend of the ratio through the sliding time window, dynamically update the reactive power output adjustment parameters of the substation, and realize the coordinated operation and control between the collector line and the substation.

[0014] Preferably, S1 specifically includes the following steps:

[0015] The active power, reactive power, voltage, current, and inverter operation control parameters of multiple branches in the collector line are collected. The operation control parameters include reactive power target setpoint, active power setpoint, power factor target value, voltage setpoint, and control mode flag field.

[0016] The active and reactive power sampling data of each branch in multiple consecutive time segments are constructed into a two-dimensional time series matrix, and the matrix is ​​periodically truncated by a sliding window to extract the slope of instantaneous power change in each time segment.

[0017] Perform proportional analysis on the instantaneous power change slope for each time period to determine whether reactive power remains constant or changes synchronously with active power, and combine this with the inverter control mode flag field to determine the current control mode.

[0018] When it is detected that the reactive power changes proportionally with the active power and the power factor target value is enabled, it is determined that the control mode of the current branch inverter has been switched to power factor priority control.

[0019] Preferably, S2 specifically includes the following steps:

[0020] The active power and reactive power values ​​of each branch in the collector line that has switched from constant reactive power control to power factor priority control are extracted and identified in multiple consecutive time periods. The active power value and the reactive power value of the corresponding time period are formed into ordered point pairs and arranged in time series to form a two-dimensional power ratio expression with active power as the horizontal axis and reactive power as the vertical axis. This expression is used to express the power output ratio trajectory of the branch over time.

[0021] By using the vector magnitude normalization method, the active power value and reactive power value in each set of ratio expressions are normalized to obtain the power output direction vector;

[0022] A time window moving average is applied to the normalized vector of each branch over a continuous time period to generate a stable vector representation that expresses the power characteristics of the branch.

[0023] Calculate the angle deviation between the average vectors of all branches, and construct a ratio difference matrix based on the angle deviation to extract the branch groups with significant change characteristics;

[0024] The ratio difference matrix is ​​correlated and mapped with the control mode change results of the corresponding branch, and the ratio difference quantification result is output to quantify the degree of ratio change caused by the switching of each control strategy.

[0025] Preferably, a vector magnitude normalization method is used to normalize the active power and reactive power values ​​in each set of ratio expressions to obtain the power output direction vector. This specifically includes the following steps:

[0026] The active power value and reactive power value of each branch in each time period are combined into a two-dimensional vector to represent the power output status in that time period.

[0027] Calculate the magnitude of the two-dimensional vector as a normalization factor, and divide the two components of the vector by the magnitude to normalize it into a unit-length vector.

[0028] The normalized unit vector is used as the power output direction vector for that time period, representing the active and reactive power output ratio of that branch during that time period.

[0029] The power output direction vector of the same branch in multiple consecutive time periods is weighted and averaged to extract the representative output direction of the branch in the target period.

[0030] The obtained power output direction vector is used as the matching feature and input into the vector comparison process of matching differences to establish the matching direction comparison relationship between branches.

[0031] Preferably, S3 specifically includes the following steps:

[0032] The quantitative results of the ratio difference of each branch and the corresponding reactive power output value are collected within the same control cycle, and a branch reactive power output capacity expression vector is constructed based on the time series method.

[0033] Collect the target voltage value currently set at the substation and the real-time voltage response data of each bus node, and construct a voltage response vector using a unified time axis.

[0034] Align the branch reactive power output capability expression vector with the voltage response vector one by one, and establish the relationship between reactive power response and voltage change under the corresponding time slice through vector interpolation.

[0035] Calculate the reactive response offset of each branch under the target voltage command, and construct the matching difference vector between the expected output and the actual output of the booster station;

[0036] The matching difference vector is used to identify the set of target branches with response deviations within the control cycle, and the matching deviation between them and the reactive power regulation target of the booster station is output.

[0037] Preferably, the quantitative results of the ratio differences of each branch and the corresponding reactive power output values ​​are collected within the same control cycle, and a branch reactive power output capacity expression vector is constructed based on time series method, specifically:

[0038] Extract the quantitative results of the ratio difference of each branch identified as having switched control strategies within the target control cycle and the reactive power value of its corresponding time slice, and construct an ordered pair sequence composed of ratio difference and reactive power.

[0039] The ordered pair sequence is mapped to a two-dimensional coordinate plane, with the ratio difference quantification value as the abscissa and the reactive power value as the ordinate, to generate the ratio-reactive power relationship point set of the corresponding branch;

[0040] The point set is interpolated using curve fitting to construct the reactive power output variation curve of this branch within the target period;

[0041] The slope change, curvature inflection point, and extreme value distribution are extracted from the curve. These parameters are then combined in a fixed order into a vector structure, which serves as the various dimensional components of the branch reactive power output capability expression vector for subsequent deviation quantification and comparison.

[0042] Preferably, the target voltage value currently set at the substation and the real-time voltage response data of each bus node are collected, and a voltage response vector is constructed using a unified time axis, specifically as follows:

[0043] Within the target control cycle, the target voltage value set by the booster station and the real-time voltage value of each bus node are collected at a fixed time granularity, which constitute the target voltage sequence and the response voltage sequence, respectively.

[0044] A first-order filter is used to denoise the response voltage sequence, eliminating local spikes and ensuring that the voltage fluctuation trend is stable and comparable.

[0045] A linear interpolation method is used to synchronize and align the target voltage sequence and the response voltage sequence along the time axis, thereby constructing an hourly voltage difference sequence;

[0046] Based on the continuity and fluctuation amplitude of the voltage difference sequence on the time axis, three indicators are extracted: slope trend, frequency of change, and maximum deviation. These three indicators are then combined into a vector form in a preset order to serve as the voltage response vector for subsequent reactive power matching deviation identification.

[0047] Preferably, S4 is as follows:

[0048] Extract the matching difference vector of all branches within the target control cycle, construct the histogram of deviation intensity, and extract the mean deviation, maximum deviation, and standard deviation of deviation to quantify the degree of deviation of the current reactive power response of the booster station.

[0049] Based on the changes in the control mode marking field of each branch, the number of switching times and directions within the same period are counted. Combined with the continuous drift trajectory of the proportion direction vector of each branch, the control mode change frequency index and proportion change rate index are extracted.

[0050] The matching difference vector is fused with the control mode change frequency index and the ratio change rate index, and a weighted calculation method is used to construct the voltage target correction factor and the reactive power reference correction factor, wherein the weighting coefficient is dynamically assigned according to the stability of each index.

[0051] Adjust the original target voltage setting of the booster station according to the voltage target correction factor, adopt the downward adjustment method to alleviate the risk of bus overvoltage, and dynamically offset the reactive power output reference value according to the reactive power reference correction factor.

[0052] The corrected target voltage setpoint and reactive power output reference value are structurally combined to generate a set of control instructions containing specific execution timestamps, voltage command values, and reactive power target values, which serves as the scheduling basis for reactive power regulation of the booster station in the next cycle.

[0053] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0054] 1. The photovoltaic power generation control method for coordinated operation of collection lines and booster stations provided by this invention can achieve real-time identification and quantitative analysis of the dynamic switching behavior of inverter control strategies. By constructing active and reactive power ratio expressions for each branch and using vector normalization and comparison methods, it accurately extracts the output ratio differences caused by changes in control methods. Simultaneously, by combining the comparison relationship between the reactive power output capability expression vector of the branch and the voltage response vector of the booster station, a difference identification mechanism between the target command of the booster station and the actual response of the branch is established. This mechanism effectively compensates for the problem that traditional centralized control systems cannot detect inverter local strategy drift, achieving centralized insight into the distributed behavior of inverters and providing accurate basis for subsequent reactive power regulation strategy correction.

[0055] 2. This invention introduces quantitative indicators of the frequency and trend of control mode changes and ratio changes, integrates correction factors for voltage targets and reactive power benchmarks, and dynamically adjusts the reactive power output control commands of the booster station based on these factors, enabling the system to possess responsive feedback and adaptive control capabilities. Combined with a continuous data tracking mechanism using a sliding time window, it achieves dynamic updating and continuous optimization of control parameters, ensuring a high degree of matching between the control strategy and the on-site operating conditions. Compared to existing technologies, this invention significantly improves the reactive power response accuracy and bus voltage stability of photovoltaic power plants in multi-inverter heterogeneous control scenarios, reduces power limiting problems caused by voltage deviation, and effectively enhances the robustness and scheduling efficiency of system operation. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0057] Figure 1 This is a flowchart illustrating the photovoltaic power generation control method of the present invention, which coordinates the operation of the power collection line and the booster station. Detailed Implementation

[0058] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0059] This invention provides, for example Figure 1 The photovoltaic power generation control method shown, which coordinates the operation of the collector line and the step-up substation, specifically includes the following steps:

[0060] S1. Collect the active power, reactive power, voltage, current and inverter operation control parameters of multiple branches in the collector line, perform time-series analysis on the active and reactive output change trends of each branch, and identify whether the inverter control mode has switched from constant reactive power control to power factor priority control.

[0061] In this embodiment, S1 specifically includes the following steps:

[0062] The active power, reactive power, voltage, current, and inverter operation control parameters of multiple branches in the collector line are collected. The operation control parameters include reactive power target setpoint, active power setpoint, power factor target value, voltage setpoint, and control mode flag field.

[0063] By collecting data from multiple branches in the power collection line, the active power, reactive power, voltage, current, and inverter operating control parameters of each branch are obtained. This aims to provide timely and accurate foundational data for subsequent control strategy identification and power ratio analysis. Specifically, distributed intelligent acquisition terminals can be connected to the measurement nodes of each branch, combined with the existing measurement and control network of the photovoltaic power station, and high-frequency data transmission can be achieved using the IEC 61850 communication protocol or Modbus TCP protocol. During the data acquisition process, four types of electrical quantities—active power, reactive power, voltage, and current—are acquired through meters or sensors. Operating control parameters are extracted by accessing the inverter's internal registers or data interfaces, including the currently set reactive power output target, active power plan, target power factor value, expected voltage value, and control mode flag fields (such as whether the current mode is power factor priority mode or constant reactive power mode). All acquired data must be cached and forwarded under a unified timestamp to ensure that subsequent time-series analysis and control strategy identification have complete data dimensions and a consistent time base.

[0064] The active and reactive power sampling data of each branch in multiple consecutive time segments are constructed into a two-dimensional time series matrix, and the matrix is ​​periodically truncated by a sliding window to extract the slope of instantaneous power change in each time segment.

[0065] This step aims to structure and extract features from the changes in active and reactive power over time for each branch, in order to identify subsequent changes in control strategies. Specifically, the active and reactive power data collected from each branch within multiple equally spaced time segments are first arranged chronologically to construct a two-dimensional time series matrix. One dimension represents the time index, and the other corresponds to the active and reactive power values, respectively. Based on this, a fixed-length sliding time window (e.g., five consecutive sampling points constitute one window) is introduced, progressively extracting local data segments from the matrix with a sliding step size. For the data within each time window, the changes in active and reactive power are calculated using the first-order difference method. Furthermore, the power change trend within this time period is fitted using linear regression or least squares methods to obtain slope values ​​representing the rate of power change, denoted as the active power change slope and reactive power change slope, respectively. These slopes are used to measure the dynamic characteristics of the branch's output within the corresponding time period, providing a quantitative basis for subsequently determining whether its control mode has changed.

[0066] Perform proportional analysis on the instantaneous power change slope for each time period to determine whether reactive power remains constant or changes synchronously with active power, and combine this with the inverter control mode flag field to determine the current control mode.

[0067] After extracting the instantaneous power change slope, a ratio analysis can be performed on the active power change slope and the reactive power change slope to determine whether the reactive power remains constant or changes synchronously with the active power. Specifically, the active power change slope in each time period is denoted as ΔP, and the reactive power change slope is denoted as ΔQ. Their ratio ΔQ / ΔP is calculated, and a set of reference threshold intervals is set. If the ratio approaches zero, it indicates that the reactive power does not change significantly with the active power during that time period, meaning it may be in a constant reactive power control state. If the ratio remains stable and significantly non-zero, and is consistent with historical power factor characteristics, it indicates that the reactive power adjusts synchronously with the active power, suggesting that the inverter may be in a power factor priority control state. Furthermore, to improve the accuracy of the judgment, the control mode flag field in the inverter's internal operating data can be read simultaneously. This field generally represents the control strategy in the form of bit flags (e.g., 0 indicates constant reactive power mode, 1 indicates power factor priority mode). The proportional analysis results are checked for consistency with the flag field. When the trend of ΔQ / ΔP changes is consistent with the control flag and conforms to the same logic in multiple consecutive time windows, the control mode of the current branch inverter can be determined, thereby enhancing the robustness of identifying the strategy drift process.

[0068] When it is detected that the reactive power changes proportionally with the active power and the power factor target value is enabled, it is determined that the control mode of the current branch inverter has been switched to power factor priority control.

[0069] After analyzing the power change trends across multiple time windows, if the reactive and active power change directions are consistent and the change ratios are stable over consecutive time periods, and the ratio of the reactive power change slope ΔQ to the active power change slope ΔP is within the empirically set linear response range, it indicates that the inverter in this branch no longer maintains a fixed reactive power output in the current period, but rather changes dynamically with the active power output. In this case, it is necessary to further read the inverter's power factor target value parameter. If this parameter is not zero and is active, it indicates that the inverter possesses the prerequisite for reactive power adjustment according to power factor priority logic. After comprehensive judgment, it can be confirmed that the inverter in the current branch has completed the mode switch from constant reactive power control to power factor priority control. This identification method uses dual logical cross-verification of data behavior patterns and control parameter states. This avoids misjudging the control strategy due to short-term disturbances and enables highly reliable control state identification through output characteristic behavior in inverter scenarios lacking direct mode feedback, providing crucial preliminary basis for subsequent corrections to the reactive power compensation plan.

[0070] The identification result will be used as the basis for the change of the branch control strategy, and will be used in the subsequent steps of ratio feature extraction and substation adjustment parameter correction.

[0071] S2. Based on the identification results, construct the active and reactive power ratio expressions for each branch in the collector line, extract the power ratio features using normalization, compare the ratio differences between each branch, and generate quantitative results of the ratio differences caused by changes in control mode.

[0072] In this embodiment, S2 specifically includes the following steps:

[0073] The active power and reactive power values ​​of each branch in the collector line that has switched from constant reactive power control to power factor priority control are extracted and identified in multiple consecutive time periods. The active power value and the reactive power value of the corresponding time period are formed into ordered point pairs and arranged in time series to form a two-dimensional power ratio expression with active power as the horizontal axis and reactive power as the vertical axis. This expression is used to express the power output ratio trajectory of the branch over time.

[0074] When modeling the power distribution of a branch whose control mode has switched from constant reactive power control to power factor priority control, the first step is to extract the active and reactive power values ​​of the branch over multiple consecutive time periods using a monitoring platform with consistent sampling frequencies. The active power value and its corresponding reactive power value for each time period are constructed as ordered pairs of points according to their temporal correspondence, and these points are arranged sequentially in chronological order. Subsequently, using a coordinate system with active power as the horizontal axis and reactive power as the vertical axis, these point pairs are mapped into a set of trajectory points in a two-dimensional plane, forming a continuously distributed power distribution trajectory. This representation clearly reflects the changes in the active and reactive power output ratio of the branch over different time periods. This two-dimensional power distribution expression not only preserves numerical amplitude information but can also be further used in subsequent processing through geometric features, angle analysis, etc., to extract the differences in the distribution trend before and after the control strategy change. It can be efficiently constructed and analyzed using a time series database combined with vector mathematical operations.

[0075] By using the vector magnitude normalization method, the active power value and reactive power value in each set of ratio expressions are normalized to obtain the power output direction vector;

[0076] A time window moving average is applied to the normalized vector of each branch over a continuous time period to generate a stable vector representation that expresses the power characteristics of the branch.

[0077] To generate a stable vector representation of branch power characteristics, a sliding time window process is needed for the normalized vector sequence obtained for each branch over multiple consecutive time periods. This process can be achieved by weighted averaging of the normalized power output direction vectors within a fixed-length time window. For example, a sliding window containing five time slices can be set up. Each time a time slice is moved, different weights are assigned to the five normalized vectors within the window (e.g., the weights are higher closer to the current time slice). The vectors are then weighted and summed, and the resulting vector is normalized again to obtain the smooth vector characteristics at that time point. By continuously applying this window and sliding it along the time axis, a set of stationary vector sequences reflecting the changing trend of power output direction can be generated. For example, if a branch has direction vectors of (0.8, 0.6), (0.82, 0.58), (0.85, 0.52), (0.88, 0.47), and (0.90, 0.44) over five time periods, a stable direction vector representing the recent power output characteristics of that branch can be obtained by weighting the vectors with weights of 0.1, 0.15, 0.3, 0.25, and 0.2, and then normalizing the result. This method not only eliminates the interference of abnormal fluctuations at a single time point but also enhances the ability to express the trend of the power distribution under changes in control strategy.

[0078] Calculate the angle deviation between the average vectors of all branches, and construct a ratio difference matrix based on the angle deviation to extract the branch groups with significant change characteristics;

[0079] To identify branch groups with significant changes in power distribution characteristics, it is necessary to first compare the angles of the stable vectors formed after moving averages for each branch. Specifically, this is achieved by calculating the cosine angle between each pair of average power output direction vectors of all branches. The formula for the angle θ is arccos((A·B) / (|A||B|)), where A and B are the average vectors of any two branches, · represents the vector dot product, and || represents the magnitude. This method constructs a symmetric power distribution difference matrix, where each element represents the angle of difference in power output direction between two branches. For example, if the average vectors of branch X and branch Y are (0.86, 0.51) and (0.75, 0.66) respectively, then their angle deviation is arccos((0.86×0.75+0.51×0.66) / (√(0.86)). 2 +0.51 2 )×√(0.75 2 +0.66 2The angle is approximately 7.2 degrees. After constructing the matrix, an angle threshold can be set. A deviation greater than 10 degrees is considered significant. By statistically analyzing the set of branches in the matrix whose angles with multiple branches simultaneously exceed the threshold, the branch groups whose output direction changes drastically after the control strategy is switched can be extracted. These branches deviate significantly from other branches in terms of power ratio characteristics, often accompanied by abnormal local control responses, and possess important dynamic adjustment priorities.

[0080] The ratio difference matrix is ​​correlated and mapped with the control mode change results of the corresponding branch, and the ratio difference quantification result is output to quantify the degree of ratio change caused by the switching of each control strategy.

[0081] To quantify the degree of power ratio change caused by control strategy switching, it is necessary to establish a one-to-one correlation between the constructed power ratio difference matrix and the control mode change status of each branch. Specifically, each branch is first labeled to indicate whether it switched from constant reactive power control to power factor priority control during the analysis period. Then, by comparing the angle deviation values ​​between this branch and other branches in the power ratio difference matrix, the branch groups whose mean or extreme angle values ​​exceed a set threshold among the branches where the control strategy has switched are identified as significantly affected. The average angle difference between these branches and the branches that have not switched is further calculated, and this is used to construct a control strategy switching impact score. For example, if branches A and B switch control strategies, and their average angles with branches C, D, and E are 15° and 18° respectively, while the average angle between branches that have not switched is only 5°, this indicates that the control strategy change directly caused a significant shift in the power ratio direction. Finally, a vector result can be output, with the branch as the index and the average angle as the quantitative indicator. This serves as a quantitative expression of the degree of deviation of the ratio of the branch after the control strategy is switched, providing a reliable basis for subsequent dynamic correction.

[0082] In this embodiment, the vector magnitude normalization method is used to normalize the active power value and reactive power value in each set of ratio expressions to obtain the power output direction vector. The specific steps include:

[0083] The active power value and reactive power value of each branch in each time period are combined into a two-dimensional vector to represent the power output status in that time period.

[0084] To represent the power output state of each branch within a specific time period, a two-dimensional vector can be constructed. Specifically, within each sampling period, the active power value P and reactive power value Q of a branch are collected and combined into an ordered vector pair (P, Q), reflecting the power distribution characteristics within that time period in two-dimensional coordinates. The direction of this vector represents the ratio of active to reactive power output, while the magnitude represents the overall power level within that time period. For example, if the active power of branch X is 320 kW and the reactive power is 240 kV in a certain time period, its power output vector is (320, 240), indicating that its output direction is biased towards a lower power factor operating state. Using this method, a continuous vector sequence can be constructed for each branch on the time axis, providing a standardized data structure for subsequent normalization, ratio analysis, and vector angle calculation, ensuring that the power output behavior before and after control strategy switching can be accurately modeled and compared. This method has good linear computability and is suitable for batch processing and sliding window processing scenarios, making it a key step in realizing the basic feature modeling of dynamic regulation.

[0085] Calculate the magnitude of the two-dimensional vector as a normalization factor, and divide the two components of the vector by the magnitude to normalize it into a unit-length vector.

[0086] When standardizing the power output state of each branch, it is necessary to normalize the magnitude of the constructed two-dimensional vector to eliminate the interference of power magnitude on the distribution direction. Specifically, the Euclidean norm can be used to calculate the magnitude of each vector (P, Q), and the calculation formula is √(P / Q). 2 + Q 2 This value represents the total combined active and reactive power within that time period. After calculation, the active component P and reactive component Q of the vector are divided by the magnitude to obtain the unit length vector (P / √(P)). 2 + Q 2 ), Q / √(P 2 + Q 2 Taking branch Y as an example within a certain time period, its active power is 300 kW and reactive power is 400 kV, corresponding to a modulus of 500. After normalization, the resulting unit vector is (0.6, 0.8). This vector has the same direction as the original vector, but its length is normalized to 1, retaining only the power ratio characteristics and eliminating the influence of the total power scale. This processing method can significantly improve the accuracy of the power ratio angle analysis and is a core prerequisite for subsequent calculation of the direction vector angle and measurement of the difference in power ratio between branches. Through this method, a standardized power output representation can be effectively constructed, which is conducive to the unified analysis of high-dimensional, multi-time period data.

[0087] The normalized unit vector is used as the power output direction vector for that time period, representing the active and reactive power output ratio of that branch during that time period.

[0088] Using the normalized unit vector as the power output direction vector for each time period aims to accurately express the active and reactive power output ratio of a branch within that time period, without being affected by the total power output. Specifically, after normalizing the vector magnitude, a unit vector of the form (P′, Q′) is obtained, where P′ and Q′ are the normalized results of the original active and reactive power values, respectively. The direction of this unit vector reflects the characteristics of the power output. If P′ is much larger than Q′, it indicates that active power output is dominant during that time period; if they are similar, it indicates that the output is relatively balanced; if Q′ is larger than P′, it indicates that the branch is primarily responsible for reactive power support. Taking the output of branch Z in two time periods as an example, the normalized vector for the first time period is (0.87, 0.5), and for the second time period it is (0.6, 0.8). The change in direction indicates that its control strategy may be shifting from active power dominance to enhanced reactive power support. By constructing such directional vectors in each time period, a time-series power ratio trajectory can be formed, providing a clear and stable data foundation for subsequent extraction of ratio trend features, identification of the impact of control switching, and realization of dynamic regulation.

[0089] The power output direction vector of the same branch in multiple consecutive time periods is weighted and averaged to extract the representative output direction of the branch in the target period.

[0090] Weighted averaging of the power output direction vectors of the same branch over multiple consecutive time periods aims to extract representative directional features from time-series data, accurately reflecting the overall power distribution trend of the branch within a control cycle. In implementation, unit vectors for each time period can be arranged chronologically to construct a vector sequence. A weighting function can then be used to assign weights to vectors from different time periods; for example, more recent time periods can be given higher weights to enhance sensitivity to the latest trends. Specifically, assuming the normalized vectors of a branch over five time periods are V1 to V5, with corresponding weights w1 to w5, the weighted average direction can be calculated using the formula Vavg = (w1×V1 + w2×V2 + ... + w5×V5) / (w1+ w2 + ... + w5). For example, if vectors V1 to V5 are (0.6, 0.8), (0.65, 0.76), (0.7, 0.71), (0.72, 0.69), and (0.74, 0.67) respectively, with progressively increasing weights, then the representative vector obtained after weighted averaging might be (0.7, 0.71), representing the dominant direction of the power distribution of this branch within this period. This approach not only smooths out the impact of short-term disturbances but also enhances the stability of the analysis, serving as an important foundation for constructing difference measurement and control strategies.

[0091] The obtained power output direction vector is used as the matching feature and input into the vector comparison process of matching differences to establish the matching direction comparison relationship between branches.

[0092] The obtained power output direction vector is used as a feature input to the vector comparison process of the ratio difference. The purpose is to establish a clear quantitative relationship of the ratio difference among multiple branches, thereby identifying the output direction change caused by the switching of control strategies. In implementation, the representative power output direction vector of each branch in the target period can be used as a unit vector. The angle between different branches is calculated in a vector space to measure the directional difference of their active and reactive power output ratio. For example, if the direction vectors corresponding to two branches are (0.8, 0.6) and (0.6, 0.8) respectively, the difference value is obtained by calculating the cosine angle. 1 The angle ((0.8×0.6 + 0.6×0.8) / 1), approximately 18.43 degrees, indicates a significant deviation in the output ratio of the two branches. This method can systematically reveal the consistency of operating states between branches, providing a quantitative basis for subsequent identification of strategy drift trends and accurate correction of reactive power regulation plans at booster stations. It avoids voltage control response lag due to undetected ratio changes, ensuring regulation accuracy and grid stability.

[0093] S3. Compare the quantitative results of the ratio difference with the reactive power regulation target of the substation, analyze the offset relationship between the bus voltage response and reactive power output, and identify the matching deviation between the reactive power regulation target of the substation and the reactive power output capacity of each branch.

[0094] In this embodiment, S3 specifically includes the following steps:

[0095] The quantitative results of the ratio difference of each branch and the corresponding reactive power output value are collected within the same control cycle, and a branch reactive power output capacity expression vector is constructed based on the time series method.

[0096] Collect the target voltage value currently set at the substation and the real-time voltage response data of each bus node, and construct a voltage response vector using a unified time axis.

[0097] Align the branch reactive power output capability expression vector with the voltage response vector one by one, and establish the relationship between reactive power response and voltage change under the corresponding time slice through vector interpolation.

[0098] Aligning the reactive power output capability vectors of branches with their voltage response vectors, and establishing the relationship between reactive power response and voltage changes at corresponding time slices through vector interpolation, aims to achieve multi-source information fusion analysis on a unified time scale. This allows for a clear numerical correspondence between the response capability of each branch and the dynamic behavior of the bus voltage. In implementation, it is crucial to ensure the consistency of the time granularity of the two vector sequences. If there are missing data or time axis misalignments, vector interpolation is used for correction. Vector interpolation involves using numerical fitting techniques (such as piecewise linear interpolation, polynomial interpolation, or cubic spline interpolation) to estimate the vector components at missing time points, given the existence of vector data at known time points. This generates a complete, continuous, and time-aligned vector sequence. At each time slice, the interpolated branch vector represents its reactive power output capability characteristics at that time, while the voltage response vector reflects the actual voltage state of the bus at that time. Pairing the two establishes a set of "input-output" relationship samples, providing fundamental data support for subsequent matching deviation quantification. This process improves data integrity and temporal consistency through numerical interpolation, avoiding interference from incomplete data or time mismatches in the analysis results, and helps to achieve higher accuracy in identifying matching differences.

[0099] Calculate the reactive response offset of each branch under the target voltage command, and construct the matching difference vector between the expected output and the actual output of the booster station;

[0100] The reactive power response offset of each branch under the target voltage command is calculated, and a matching difference vector is constructed between the expected and actual outputs of the booster station. This aims to quantify the response consistency during the execution of the booster station's control commands. This process is achieved by performing a one-to-one time-point differential operation on the target response value and the actual response value of each branch. The target response value originates from the unified voltage command issued by the booster station. Based on the historical power ratio, reactive power output capacity expression vector, and voltage response vector of each branch, the theoretically achievable reactive power output value is predicted. The actual response value is obtained from the reactive power measurement data of each branch at the target time point. By performing point-by-point difference calculations on the two over a unified time series, the degree of response offset of each branch under a given target voltage can be obtained. By assembling the response offset values ​​of all branches into a vector, a matching difference vector can be constructed for subsequent deviation identification and control correction.

[0101] For example, suppose the target voltage of the booster station is set to 500V in a certain time slot. The reactive power output capability vector of branch A predicts that it should output 300kVar under this voltage, but the actual measured output of this branch is only 240kVar. Then its reactive power response offset is 60kVar. Similarly, the same operation is performed on branches B, C, etc. After calculating the corresponding offset values, they are arranged into a vector [60, 20, -15, …] according to the branch order. This vector is the matching difference vector of each branch of the booster station under the current voltage command. Positive values ​​represent low response, and negative values ​​represent over-response. This vector can intuitively reflect the degree of coordination between the response of each branch and the booster station command, providing data support for the accurate correction of the control strategy.

[0102] The matching difference vector is used to identify the set of target branches with response deviations within the control cycle, and the matching deviation between them and the reactive power regulation target of the booster station is output.

[0103] Identifying the set of target branches with response deviations within the control cycle and outputting their matching deviations from the reactive power control target of the booster station is to discover the mismatch areas generated during the execution of control commands in each branch, thus providing a basis for subsequent compensation adjustments. This process is achieved through numerical analysis of the constructed matching difference vector. First, upper and lower limits of the matching deviation can be set based on the statistical threshold method, for example, setting a tolerance band of ±ΔkVar centered at 0. When the deviation value of a branch exceeds this range, its response is considered to have an abnormal offset. In addition, a clustering algorithm can be used to cluster the difference vector, thereby grouping branches with significant deviations into one class and marking it as the deviation set. Outputting this set not only indicates which branches have deviated in response but also further quantifies the degree of deviation, serving as an auxiliary input for the dynamic correction of the booster station's subsequent reactive power control plan.

[0104] For example, within a certain control cycle, assuming the matching difference vector is [8, -3, 0, 22, -18, 5], and the matching tolerance is set to ±10kVar, then branches 4 and 5 deviate by 22kVar and -18kVar respectively, both exceeding the set range, and are therefore identified as target branches with response deviations. This set is output as {4,5}, and the matching deviation value between it and the reactive power regulation target of the booster station is calculated, which is the absolute value of their respective deviations [22, 18]. This process can be regarded as a precise screening of the branch response effectiveness under the reactive power regulation target of the booster station, providing data support and strategic basis for dynamic correction in the control process.

[0105] In this embodiment, the quantitative results of the ratio differences of each branch and the corresponding reactive power output values ​​are collected within the same control cycle, and a branch reactive power output capability expression vector is constructed based on time series method, specifically:

[0106] Extract the quantitative results of the ratio difference of each branch identified as having switched control strategies within the target control cycle and the reactive power value of its corresponding time slice, and construct an ordered pair sequence composed of ratio difference and reactive power.

[0107] Within the target control cycle, branches that have switched from constant reactive power control to power factor priority control can be identified first. These branches' control strategy changes within the cycle directly affect their active and reactive power ratios. For each identified branch, the quantified ratio difference value is extracted at each discrete time segment within the control cycle, and the corresponding reactive power value is simultaneously obtained. Then, the quantified ratio difference value is used as the x-axis, and the reactive power value as the y-axis, corresponding to each other in chronological order, forming an ordered pair sequence of ratio-reactive power output for that branch within the control cycle. This sequence describes the functional relationship between the degree of ratio change and the reactive power response, providing structured input for subsequent curve fitting and capacity modeling. This method can achieve efficient data extraction through database indexing or a multidimensional array structure cached in memory, ensuring temporal consistency and data integrity.

[0108] The ordered pair sequence is mapped to a two-dimensional coordinate plane, with the ratio difference quantification value as the abscissa and the reactive power value as the ordinate, to generate the ratio-reactive power relationship point set of the corresponding branch;

[0109] When constructing the point set of ratio-reactive power relationship, the previously formed ordered pair sequence consisting of the quantified ratio difference values ​​and corresponding reactive power values ​​can be used as the basic data input. Following coordinate mapping rules, each ordered pair is plotted on a two-dimensional coordinate plane, with the quantified ratio difference values ​​as the abscissa and the reactive power values ​​as the ordinate, reflecting the reactive power response capability of the branch under different ratio states. This process can be implemented using a matrix coordinate mapping function, or in practical applications, using visualization tools or graphical computing interfaces for batch vector mapping to form a continuously distributed point set trajectory. To ensure the representativeness of the point set distribution and the usability of the calculation, a ratio difference segment with representative changes within the control cycle should be selected, and the consistency of the sampling time interval should be ensured. The final point set is not only used for subsequent curve fitting but also serves as a structural expression of the branch response characteristics after control strategy switching, reflecting the actual impact trend of ratio fluctuations on reactive power output.

[0110] The point set is interpolated using curve fitting to construct the reactive power output variation curve of this branch within the target period;

[0111] To construct the reactive power output variation curve of a branch within the target period, a two-dimensional relationship point set consisting of the mapped ratio difference quantification value and reactive power value can be interpolated using curve fitting methods such as polynomial regression, spline interpolation, or Gaussian process fitting. Polynomial regression is suitable for cases where the point set exhibits a continuous trend with small fluctuations, capturing the overall trend by fitting a polynomial function of a certain order. Spline interpolation is suitable for cases where the point set has uneven spacing or inflection points, smoothly connecting data points through piecewise low-order functions while preserving local characteristics. Gaussian process fitting is suitable for scenarios where modeling requires assessing fitting uncertainty or where the point set exhibits irregular fluctuations, generating variation curves with confidence intervals. In implementation, an appropriate fitting strategy is selected based on the point set distribution characteristics, expressing the interpolated continuous function as a reactive power output variation curve. This curve characterizes the reactive power response evolution path of the branch within the target period under control strategy switching, thus providing a structural foundation for subsequent output capacity modeling and matching analysis.

[0112] The slope change, curvature inflection point, and extreme value distribution are extracted from the curve. These parameters are then combined in a fixed order into a vector structure, which serves as the various dimensional components of the branch reactive power output capability expression vector for subsequent deviation quantification and comparison.

[0113] To extract three parameters—slope change, curvature inflection point, and extreme value distribution—from the reactive power output change curve, a segmented analysis based on the derivative information of the curve can be performed first. By calculating the first derivative, the slope value sequence at each time point can be obtained, and then the fluctuation frequency, average slope, and slope direction change points on the time axis can be analyzed to quantify the stability and trend of the output rate. Next, the curvature information of the curve is extracted by calculating the second derivative, locating curvature abrupt change points to identify the response-sensitive intervals and inertial lag regions of output changes. Simultaneously, the curve is scanned, local maxima and minima are marked, and the frequency of extreme value occurrences, the interval length between extreme values, and the peak-to-valley amplitude difference are extracted to reflect the fluctuation amplitude and adjustment elasticity of the output capability. After standardization, the above three parameter sets are arranged in a predetermined order to form a fixed-structure vector representation. This vector serves as the digital expression basis for the branch output capability in subsequent comparisons, facilitating quantitative comparison and matching deviation analysis between multiple branches.

[0114] In this embodiment, the target voltage value currently set at the booster station and the real-time voltage response data of each bus node are collected, and a voltage response vector is constructed using a unified time axis, specifically:

[0115] Within the target control cycle, the target voltage value set by the booster station and the real-time voltage value of each bus node are collected at a fixed time granularity, which constitute the target voltage sequence and the response voltage sequence, respectively.

[0116] Within the target control cycle, the target voltage value set by the substation and the real-time voltage values ​​of each bus node are collected at a fixed time granularity. This can be achieved by setting a uniform sampling period in the dispatch control platform, such as collecting data every 1 second or 5 seconds. The target voltage value is usually derived from the voltage setting parameters at the control command generation end and can be directly extracted from the setting interface. The real-time voltage values ​​of each bus node are continuously fed back to the data acquisition layer by voltage sampling devices installed at the bus measuring points, and are time-aligned with the target voltage value through timestamp synchronization. The collected target voltage values ​​form a target voltage sequence in time series form, while the real-time voltage values ​​are recorded according to bus number, synchronously forming multiple response voltage sequences. In this process, it is necessary to ensure that the target voltage sequence and each response voltage sequence are completely consistent in time granularity so as to conduct point-by-point comparison and fluctuation trend analysis, thereby providing an accurate data basis for identifying the dynamic deviation of reactive power response.

[0117] A first-order filter is used to denoise the response voltage sequence, eliminating local spikes and ensuring that the voltage fluctuation trend is stable and comparable.

[0118] A first-order filter is used to denoise the response voltage sequence, primarily to eliminate local spikes caused by electromagnetic interference, measurement accuracy fluctuations, or instantaneous load disturbances, thus preventing these abnormal data from misinterpreting subsequent voltage trend analysis. A first-order filter is a linear low-pass filter that smooths the input sequence recursively. Its basic implementation involves weighting and superimposing the original value at the current time step with the filtered value from the previous time step according to a certain ratio. The calculation formula is usually expressed as: Y n = α × X n + (1 − α) × Y n ₋1, where Y n X is the current filter output. n Y is the current sampled value. n ₋1 represents the previous filtered output, and α is the smoothing coefficient, ranging from 0 to 1. By appropriately selecting the value of α, the overall voltage fluctuation trend can be preserved while effectively suppressing abrupt noise, making the processed voltage response sequence more continuous and comparable, which is beneficial for subsequent construction of the voltage response vector and reactive power deviation identification.

[0119] A linear interpolation method is used to synchronize and align the target voltage sequence and the response voltage sequence along the time axis, thereby constructing an hourly voltage difference sequence;

[0120] The use of linear interpolation to synchronize the target voltage sequence and the response voltage sequence along the time axis addresses the issue of inconsistencies in the sampling time points between the two sets of voltage data, ensuring a strict one-to-one correspondence in subsequent comparative analysis. Linear interpolation is a numerical calculation method based on connecting known points with line segments. It is suitable for estimating the approximate value at any intermediate time point from the values ​​of a known function at two adjacent time points. Its basic principle is to construct a linear function y = y1 + (y2 − y1) × (x − x1) / (x2 − x1) between the known points (x1, y1) and (x2, y2) to calculate the approximate value at any x∈(x1, x2). In specific implementation, the target voltage sequence and the response voltage sequence are first mapped onto a unified time axis. Then, for the missing target time point in the response voltage, the approximate voltage value is calculated using linear interpolation based on the adjacent sampled points before and after it, thus obtaining a complete hourly voltage difference sequence. This method is simple to implement and has stable accuracy, which can effectively improve the time consistency and trend continuity of voltage response comparison, laying the foundation for subsequent deviation index extraction.

[0121] Based on the continuity and fluctuation amplitude of the voltage difference sequence on the time axis, three indicators are extracted: slope trend, frequency of change, and maximum deviation. These three indicators are then combined into a vector form in a preset order to serve as the voltage response vector for subsequent reactive power matching deviation identification.

[0122] Extracting three indicators—slope trend, frequency of change, and maximum deviation—from the continuity and fluctuation amplitude of the voltage difference sequence along the time axis is to quantify the dynamic response characteristics of the bus voltage to the reactive power regulation commands of the substation into a comparable vector form, thereby enabling accurate identification of regulation deviations. The slope trend can be obtained by calculating the linear rate of change of the difference sequence over each time period, reflecting whether the voltage response gradually deviates from or gradually approaches the target. The frequency of change is determined by setting an amplitude change threshold and counting the number of changes exceeding that threshold, characterizing the activity level of voltage fluctuations. The maximum deviation is the extreme value in the difference sequence, used to measure the most severe voltage control error. After uniformly normalizing the three indicators, they are combined in a fixed order to form a three-dimensional voltage response vector, which is used for difference matching analysis with the reactive power output capability vectors of each branch. This method, by compressing the time series response characteristics into a standardized vector, not only enhances the comparability of voltage behavior data but also improves the algorithm efficiency and accuracy of matching deviation identification, avoiding the problems of high dimensionality and high noise encountered when directly processing raw voltage data. The above processing relies on a pre-linear interpolation method, which is used to construct a complete time-aligned sequence to ensure that the extracted metrics have strict temporal consistency and logical continuity.

[0123] S4. Based on the matching deviation, and combined with the frequency of inverter control mode change and the power ratio change trend, the target voltage setpoint and reactive power output reference value of the booster station are corrected, and control commands for reactive power regulation are generated.

[0124] In this embodiment, S4 specifically refers to:

[0125] Extract the matching difference vector of all branches within the target control cycle, construct the histogram of deviation intensity, and extract the mean deviation, maximum deviation, and standard deviation of deviation to quantify the degree of deviation of the current reactive power response of the booster station.

[0126] Within the target control cycle, the matching difference vectors of all branches can be extracted by calculating the hourly difference between the expected and actual reactive power output values ​​of each branch during that cycle. After summing the difference vectors of all branches and grouping them by amplitude, a histogram of deviation intensity can be constructed. This histogram reflects the degree of reactive power response deviation and its frequency distribution characteristics at different time points for different branches. After construction, statistical analysis of the histogram can extract key parameters such as the mean deviation, maximum deviation, and standard deviation. The mean is used to assess the overall response deviation trend, the maximum deviation reflects extreme deviations, and the standard deviation measures the overall deviation fluctuation range. These indicators are crucial in the control decisions of the booster station, determining whether the current control strategy is mismatched and serving as an important quantitative basis for subsequent adjustments to the target voltage setpoint and reactive power benchmark, thereby achieving more robust and accurate reactive power regulation strategy updates.

[0127] Based on the changes in the control mode marking field of each branch, the number of switching times and directions within the same period are counted. Combined with the continuous drift trajectory of the proportion direction vector of each branch, the control mode change frequency index and proportion change rate index are extracted.

[0128] To extract the frequency and rate of change of control mode indicators and the ratio change indicators, the following steps are taken: First, the control mode marker field of each branch within the target control cycle is collected to identify the specific switching state of the control strategy within that cycle. Then, the number and timing of switching from constant reactive power control to power factor priority control and reverse switching are statistically analyzed using timestamp information, thus determining the control mode switching frequency and its directionality. Next, the continuously generated power output direction vectors of each branch within that cycle are sorted to construct a time-series vector trajectory, and the rate of angle change between consecutive vectors is calculated to form a ratio direction change rate sequence. This rate sequence can be used to extract the degree of drift through moving average or regression trend analysis to reflect the dynamics of the branch power output characteristics. The purpose of this process is to quantify the instability of the inverter operating strategy and the corresponding fluctuation trend of the output power state, providing a dynamic compensation reference for the changing trends when the booster station generates new target voltage setpoints and reactive power output reference values, thereby improving the adaptability and accuracy of the overall control strategy.

[0129] The matching difference vector is fused with the control mode change frequency index and the ratio change rate index. A weighted calculation method is used to construct the voltage target correction factor and the reactive power reference correction factor. The weighting coefficients are dynamically assigned based on the stability of each index to enhance the control robustness.

[0130] To effectively integrate the matching difference vector with the control mode change frequency index and the ratio change rate index, the three types of parameters are first normalized to ensure numerical consistency. Then, a fusion vector is constructed for each branch, mapping the matching difference domain to the static error dimension, the control mode change frequency to the strategy stability dimension, and the ratio change rate to the dynamic volatility dimension. To comprehensively measure the impact of different indicators on control accuracy, the weight coefficient of each indicator is dynamically determined using an entropy weighting method or a weighted variance method. The weight is automatically adjusted based on its stability over recent periods; smaller fluctuations result in higher weights, thus strengthening the dependence on stable data sources. Based on the fused weighted vector, a three-dimensional function model of error-frequency-rate is established to generate a voltage target correction factor and a reactive power reference correction factor for the booster station. The former is used to adjust the voltage setpoint, and the latter to adjust the reactive power output target. This method enhances the robustness and responsiveness of the control strategy to system disturbances and strategy switching by integrating static error and dynamic variability information, thereby achieving stable and accurate target value correction for the booster station.

[0131] The original target voltage setting of the booster station is adjusted according to the voltage target correction factor, and the bus overvoltage risk is mitigated by adjusting downward. The reactive power output reference value is dynamically offset according to the reactive power reference correction factor to improve the alignment of response capability.

[0132] To achieve target voltage adjustment based on the voltage target correction factor and dynamic correction of the reactive power output reference value based on the reactive power reference correction factor, the voltage target correction factor must first be applied as an adjustable amplitude coefficient to the original target voltage setpoint of the booster station. This is done by adjusting downwards, i.e., decreasing the original setpoint by multiplying the correction factor by a certain proportion to form a new voltage control target. This reduces the risk of bus overvoltage, which is particularly critical when frequent control strategy switching causes distribution imbalances or severe fluctuations in sunlight. Simultaneously, the reactive power reference correction factor must be used as a control parameter for the adjustment amplitude, superimposed or reduced to the current reactive power target value of the booster station, achieving quantitative dynamic compensation based on the actual branch response capability. To ensure the continuity and stability of the adjustment process, a correction factor filtering mechanism with hysteresis can be introduced to limit and smooth the correction amplitude, avoiding over-adjustment that could cause system oscillations. This approach, by embedding the correction factor as a dynamic feedback variable into the control strategy, enables dual adaptive compensation for bus voltage risk and reactive power response deviation, thereby improving the steady-state adjustment accuracy and dynamic control response capability of the photovoltaic power plant's booster stage.

[0133] The corrected target voltage setpoint and reactive power output reference value are structurally combined to generate a set of control instructions containing specific execution timestamps, voltage command values ​​and reactive power target values. This set serves as the scheduling basis for reactive power regulation of the booster station in the next cycle, achieving precise control driven by strategy evolution characteristics.

[0134] To achieve a structured combination of the corrected target voltage setpoint and reactive power output reference value to form a complete control instruction set, a command unit with a fixed field structure must first be constructed based on the target voltage setpoint and reactive power reference value output at the end of the current control cycle. This command unit includes an execution timestamp field (indicating the effective time of the control instruction), a target voltage field (corresponding to the corrected voltage setpoint), and a reactive power field (corresponding to the corrected reactive power output reference value). To ensure that the control instruction can be accurately scheduled and responded to in real time in the next cycle, the above fields need to be encapsulated into a data packet with a standard format and transmitted to the substation control device to trigger reactive power output control adjustments for each branch and bus node. Furthermore, to balance the dynamics of strategy evolution and the predictability of execution, the generation logic of this instruction set should be updated in each control cycle to ensure consistency with the frequency of control mode changes, the trend of ratio changes, and the feedback results of matching deviations. This method, through timestamp marking and structured combination, achieves efficient mapping of strategy correction parameters to actual control actions, thereby improving the substation's response speed and instruction accuracy to changes in the photovoltaic system's operating status.

[0135] S5. Based on the corrected control commands, continuously collect the operating data of each branch in the collector line, track the changing trend of the control mode and the evolution trend of the ratio through the sliding time window, dynamically update the reactive power output adjustment parameters of the substation, and realize the coordinated operation and control between the collector line and the substation.

[0136] In this embodiment, S5 specifically refers to:

[0137] Based on the control command set generated in the previous control cycle, the system collects the target voltage setpoint and reactive power output reference value transmitted from the booster station and establishes a control plan execution queue corresponding to the timestamp. On this execution queue, the system continuously collects active power, reactive power, voltage, current, and inverter control mode fields for each branch in the collector line at fixed time intervals, and registers them with a unified time axis. All collected data is encapsulated into time-series samples, serving as the basic data source for sliding time window tracking analysis. The sampling frequency is preferably between 1 and 5 seconds to ensure sufficient capture of transient changes during control strategy switching while avoiding redundant delays in data processing.

[0138] Based on a sliding time window, a two-dimensional power ratio trajectory vector sequence is constructed according to the active and reactive power sampling points of each branch over multiple consecutive time periods. The trajectory of the change in the power ratio direction vector is tracked, and dynamic indicators reflecting the evolution trend of the branch power ratio are extracted by analyzing the angle offset amplitude, vector drift rate, and direction fluctuation frequency. Simultaneously, changes in the inverter control mode field are monitored, recording the number of control strategy switches and the switching direction (constant reactive power control switching to power factor priority control or reverse switching) for each branch within the time window. Based on these two types of data, a control mode change trend map is generated by setting dynamic thresholds to identify which branches are experiencing frequent strategy changes and generating significant power ratio disturbances.

[0139] The system compares the control mode change trend map of each branch in the current cycle with the control strategy trajectory in historical control cycles, calculates the change value of its deviation rate, and constructs a "control stability weighting factor" and a "power ratio fluctuation factor" by combining the stability index of the power ratio vector sequence. Based on this, the system dynamically adjusts the current reactive power output regulation parameters of the substation according to the real-time reactive power response capability requirements of the substation and the bus voltage fluctuation trend. Regulation methods include slightly increasing or decreasing the reactive power baseline command value of the substation, delaying the timestamp of certain response commands, and temporarily freezing the reactive power dispatch participation level of some branches. The update mechanism of these regulation parameters adopts an exponential weighted average method to avoid causing drastic disturbances to the system.

[0140] Before the start of a new control cycle, the updated reactive power output regulation parameters of the substation are re-encoded into the control instruction set, along with an execution time stamp and a branch matching mapping table. This control instruction set will be used to guide the reactive power control behavior of each branch in the next cycle. Simultaneously, to ensure the sustainability of dynamic correction, the system also needs to perform closed-loop backtracking on the actual response results after execution, comparing the actual response curves of each branch in the collector line with the preset control curves, identifying control errors, and feeding them back into the parameter calculation process for the next cycle, thus achieving rolling strategy correction. Through this method, the system achieves data-driven dynamic evolutionary control in each control cycle, effectively improving the coordination and response consistency of reactive power regulation between the collector line and the substation.

[0141] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means (e.g., infrared, wireless, microwave, etc.). A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0142] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0143] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0144] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0145] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0146] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0147] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A photovoltaic power generation control method for coordinated operation of collection lines and step-up substations, characterized in that, Specifically, the following steps are included: S1. Collect the active power, reactive power, voltage, current and inverter operation control parameters of multiple branches in the collector line, perform time-series analysis on the active and reactive output change trends of each branch, and identify whether the inverter control mode has switched from constant reactive power control to power factor priority control. S2. Based on the identification results, construct the active and reactive power ratio expressions for each branch in the collector line, extract the power ratio features using normalization, compare the ratio differences between each branch, and generate quantitative results of the ratio differences caused by changes in control mode. S2 specifically includes the following steps: The active power and reactive power values ​​of each branch in the collector line that has switched from constant reactive power control to power factor priority control are extracted and identified in multiple consecutive time periods. The active power value and the reactive power value of the corresponding time period are formed into ordered point pairs and arranged in time series to form a two-dimensional power ratio expression with active power as the horizontal axis and reactive power as the vertical axis. This expression is used to express the power output ratio trajectory of the branch over time. By using the vector magnitude normalization method, the active power value and reactive power value in each set of ratio expressions are normalized to obtain the power output direction vector; A time window moving average is applied to the normalized vector of each branch over a continuous time period to generate a stable vector representation that expresses the power characteristics of the branch. Calculate the angle deviation between the average vectors of all branches, and construct a ratio difference matrix based on the angle deviation to extract the branch groups with significant change characteristics; The ratio difference matrix is ​​correlated and mapped with the control mode change results of the corresponding branch, and the ratio difference quantification result is output to quantify the degree of ratio change caused by the switching of each control strategy. S3. Compare the quantitative results of the ratio difference with the reactive power regulation target of the substation, analyze the offset relationship between the bus voltage response and reactive power output, and identify the matching deviation between the reactive power regulation target of the substation and the reactive power output capacity of each branch. S4. Based on the matching deviation, and combined with the frequency of inverter control mode change and the power ratio change trend, the target voltage setpoint and reactive power output reference value of the booster station are corrected, and control commands for reactive power regulation are generated. S5. Based on the corrected control commands, continuously collect the operating data of each branch in the collector line, track the changing trend of the control mode and the evolution trend of the ratio through the sliding time window, dynamically update the reactive power output adjustment parameters of the substation, and realize the coordinated operation and control between the collector line and the substation.

2. The photovoltaic power generation control method for coordinated operation of the collector line and the booster station according to claim 1, characterized in that, S1 specifically includes the following steps: The active power, reactive power, voltage, current, and inverter operation control parameters of multiple branches in the collector line are collected. The operation control parameters include reactive power target setpoint, active power setpoint, power factor target value, voltage setpoint, and control mode flag field. The active and reactive power sampling data of each branch in multiple consecutive time segments are constructed into a two-dimensional time series matrix, and the matrix is ​​periodically truncated by a sliding window to extract the slope of instantaneous power change in each time segment. Perform proportional analysis on the instantaneous power change slope for each time period to determine whether reactive power remains constant or changes synchronously with active power, and combine this with the inverter control mode flag field to determine the current control mode. When it is detected that the reactive power changes proportionally with the active power and the power factor target value is enabled, it is determined that the control mode of the current branch inverter has been switched to power factor priority control.

3. The photovoltaic power generation control method for coordinated operation of collector lines and substations according to claim 1, characterized in that, The active power and reactive power values ​​in each set of ratio expressions are normalized using a vector magnitude normalization method to obtain the power output direction vector. This process includes the following steps: The active power value and reactive power value of each branch in each time period are combined into a two-dimensional vector to represent the power output status in that time period. Calculate the magnitude of the two-dimensional vector as a normalization factor, and divide the two components of the vector by the magnitude to normalize it into a unit-length vector. The normalized unit vector is used as the power output direction vector for that time period, representing the active and reactive power output ratio of that branch during that time period. The power output direction vector of the same branch in multiple consecutive time periods is weighted and averaged to extract the representative output direction of the branch in the target period. The obtained power output direction vector is used as the matching feature and input into the vector comparison process of matching differences to establish the matching direction comparison relationship between branches.

4. The photovoltaic power generation control method for coordinated operation of the collector line and the booster station according to claim 1, characterized in that, S3 specifically includes the following steps: The quantitative results of the ratio difference of each branch and the corresponding reactive power output value are collected within the same control cycle, and a branch reactive power output capacity expression vector is constructed based on the time series method. Collect the target voltage value currently set at the booster station and the real-time voltage response data of each bus node, and construct a voltage response vector using a unified time axis. Align the branch reactive power output capability expression vector with the voltage response vector one by one, and establish the relationship between reactive power response and voltage change under the corresponding time slice through vector interpolation. Calculate the reactive response offset of each branch under the target voltage command, and construct the matching difference vector between the expected output and the actual output of the booster station; The matching difference vector is used to identify the set of target branches with response deviations within the control cycle, and the matching deviation between them and the reactive power regulation target of the booster station is output.

5. The photovoltaic power generation control method for coordinated operation of the collector line and the booster station according to claim 4, characterized in that, The quantitative results of the ratio differences of each branch and the corresponding reactive power output values ​​are collected within the same control cycle. A branch reactive power output capacity expression vector is constructed based on a time series approach, specifically: Extract the quantitative results of the ratio difference of each branch identified as having switched control strategies within the target control cycle and the reactive power value of its corresponding time slice, and construct an ordered pair sequence composed of ratio difference and reactive power. The ordered pair sequence is mapped to a two-dimensional coordinate plane, with the ratio difference quantification value as the abscissa and the reactive power value as the ordinate, to generate the ratio-reactive power relationship point set of the corresponding branch; The point set is interpolated using curve fitting to construct the reactive power output variation curve of this branch within the target period; The slope change, curvature inflection point, and extreme value distribution are extracted from the curve. These parameters are then combined in a fixed order into a vector structure, which serves as the various dimensional components of the branch reactive power output capability expression vector for subsequent deviation quantification and comparison.

6. The photovoltaic power generation control method for coordinated operation of the collector line and the booster station according to claim 4, characterized in that, The target voltage value currently set at the substation and the real-time voltage response data of each bus node are collected, and a voltage response vector is constructed using a unified time axis. Specifically: Within the target control cycle, the target voltage value set by the booster station and the real-time voltage value of each bus node are collected at a fixed time granularity, which constitute the target voltage sequence and the response voltage sequence, respectively. A first-order filter is used to denoise the response voltage sequence, eliminating local spikes and ensuring that the voltage fluctuation trend is stable and comparable. A linear interpolation method is used to synchronize and align the target voltage sequence and the response voltage sequence along the time axis, thereby constructing an hourly voltage difference sequence; Based on the continuity and fluctuation amplitude of the voltage difference sequence on the time axis, three indicators are extracted: slope trend, frequency of change, and maximum deviation. These three indicators are then combined into a vector form in a preset order to serve as the voltage response vector for subsequent reactive power matching deviation identification.

7. The photovoltaic power generation control method for coordinated operation of collector lines and substations according to claim 1, characterized in that, S4 specifically refers to: Extract the matching difference vector of all branches within the target control cycle, construct the deviation intensity histogram, and extract the mean deviation, maximum deviation and standard deviation of deviation to quantify the degree of current reactive power response deviation of the booster station. Based on the changes in the control mode marking field of each branch, the number of switching times and directions within the same period are counted. Combined with the continuous drift trajectory of the proportion direction vector of each branch, the control mode change frequency index and proportion change rate index are extracted. The matching difference vector is fused with the control mode change frequency index and the ratio change rate index, and a weighted calculation method is used to construct the voltage target correction factor and the reactive power reference correction factor, wherein the weighting coefficient is dynamically assigned according to the stability of each index. Adjust the original target voltage setting of the booster station according to the voltage target correction factor, adopt the downward adjustment method to alleviate the risk of bus overvoltage, and dynamically offset the reactive power output reference value according to the reactive power reference correction factor. The corrected target voltage setpoint and reactive power output reference value are structurally combined to generate a set of control instructions containing specific execution timestamps, voltage command values, and reactive power target values, which serves as the scheduling basis for reactive power regulation of the booster station in the next cycle.

Images