A photovoltaic area voltage over-limit control method and device based on regional coordinated U-shaped confluence strategy

Abstract

The present invention relates to the field of photovoltaic voltage control, and discloses a photovoltaic area voltage over-limit control method and device based on a regional collaborative U-shaped confluence strategy, which is used to solve the problem that when performing photovoltaic area voltage control, the collected data is affected by large equipment, resulting in misjudgment of power anomalies. The method and device include collecting power data within a time window and performing a first power anomaly judgment. If the current power is judged to be abnormal for the first time, the industrial load data within the time window is obtained, the industrial load mutation index is evaluated, and an industrial load mutation judgment is performed. If it is judged that an industrial load mutation occurs, the power data is corrected, and a second power anomaly judgment is performed based on the corrected power data. If the current power is judged to be abnormal for the second time, the U-shaped confluence strategy is executed, and the power data is continued to be collected in real time, and real-time dynamic adjustments are performed, thereby effectively improving the accuracy of the collected power data, reducing misjudgments of power anomalies, and improving the rationality of voltage control.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic voltage control, and more specifically to a method and device for controlling over-limit voltage in a photovoltaic area based on a regional collaborative U-shaped confluence strategy. Background Art

[0002] With the rapid development and widespread adoption of distributed photovoltaic power generation technology, rural power grid operations are facing new challenges, particularly in PV-connected power lines. During peak PV generation periods, reverse power feeds into the grid, causing voltages on related lines to rise, even exceeding the specified upper voltage limit. During periods without PV generation, line voltages fall back to normal or slightly lower levels. This long-term voltage fluctuation not only negatively impacts user experience but also increases the operational burden on grid equipment.

[0003] Some older power grid lines, due to their aging, small diameters, and long power supply radius, are prone to significant voltage drops and power losses when photovoltaic power is reversely fed back into the grid. An existing method for optimizing power operations involves adding a dedicated grid-connected line to serve photovoltaic users. This method effectively reduces the load on the main power supply line during peak daytime photovoltaic power generation, improves photovoltaic absorption capacity, and prevents voltage overloads in the substation area caused by photovoltaic reverse feeding.

[0004] However, in the process of implementing the technical solutions of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0005] In actual applications, the load situation in the substation area changes dynamically, and the load characteristics are different in different time periods and seasons. Simply adding new lines cannot flexibly adjust the operating status of the power grid. In addition, the collected data may be affected by large equipment, resulting in fluctuations and being misjudged as power anomalies. Power adjustments will waste resources and cause circuit instability. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides a photovoltaic area voltage over-limit control method and device based on a regional collaborative U-shaped confluence strategy to solve the problems existing in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A photovoltaic area voltage overlimit control method based on a regional collaborative U-shaped confluence strategy includes the following steps: Step 1: Using a sliding time window method, power data is collected in real time through a monitoring device, and the power data includes voltage data, current data, power factor and photovoltaic output data, and the power data within the time window is obtained; Step 2: Based on the power data within the time window, a power anomaly index is obtained, and the first power anomaly judgment is made based on the power anomaly index; Step 3: If the current power is judged to be abnormal for the first time, the industrial load data of each sampling time point in the time window is obtained, and the industrial load data includes voltage imbalance, current imbalance, harmonic content change rate and load curve mutation rate. The industrial load mutation index is obtained based on the industrial load data, and the industrial load mutation index is obtained based on the industrial load The load mutation index is used to judge the industrial load mutation; Step 4: If it is judged that an industrial load mutation occurs, the power data is corrected to obtain the corrected power data, and the power anomaly index is re-evaluated based on the corrected power data, and a second power anomaly judgment is performed; Step 5: If the current power is normal for the second time, the power data is continued to be collected. If the current power is abnormal for the second time, the U-type confluence strategy is implemented, and the reactive power of the photovoltaic inverter is adjusted through the reactive power compensation of the photovoltaic inverter to obtain the adjusted power data; Step 6: The voltage safety range and the adjusted voltage data are obtained, and the voltage anomaly judgment is made based on the adjusted voltage and the voltage safety range; Step 7: If the current voltage is abnormal, the intelligent route switching is performed, and the real-time power data collection and dynamic adjustment are continued.

[0009] Preferably, the power anomaly index acquisition step is: acquiring voltage data and rated voltage value within a time window, and calculating the voltage influence degree according to the voltage data and the rated voltage; acquiring current data within the time window, and calculating the current influence degree according to the current data; acquiring power factor within the time window, and calculating the power factor influence degree according to the power factor; acquiring photovoltaic output data within the time window, and calculating the photovoltaic processing influence degree according to the photovoltaic output data; normalizing the voltage influence degree, current influence degree, power factor influence degree, and photovoltaic processing influence degree, and evaluating the power anomaly index according to the normalized voltage influence degree, current influence degree, power factor influence degree, and photovoltaic processing influence degree. The specific acquisition steps are: Where, E total Expressed as power anomaly index, E U Expressed as the voltage influence degree, E I Expressed as the degree of current influence, E PF Expressed as the degree of power factor influence, E PV It is expressed as the degree of influence of photovoltaic treatment, using the sum of squares method to ensure that all abnormal factors can affect the final index.

[0010] Preferably, the first power anomaly judgment step based on the power anomaly index is: comparing the power anomaly index with the anomaly threshold; if the power anomaly index is greater than or equal to the anomaly threshold, the current power is judged to be abnormal; if the power anomaly index is less than the anomaly threshold, the current power is judged to be normal, and no adjustment is performed.

[0011] Preferably, the steps for obtaining the industrial load mutation index are: obtaining three-phase voltage data within a time window, and obtaining the voltage unbalance coefficient by evaluating the three-phase voltage data through the norm method; obtaining three-phase current data within a time window, and obtaining the current unbalance coefficient by evaluating the three-phase current data through the norm method; obtaining harmonic content data within a time window, clustering the harmonic content data using the K-means clustering method, and calculating the harmonic content variation coefficient based on the clustering processing result; obtaining a load signal within a time window, decomposing the load signal into low-frequency and high-frequency components using discrete wavelet transform, extracting the high-frequency component, and using the energy of the high-frequency component as the load mutation coefficient; and using the voltage unbalance coefficient, current unbalance coefficient, and harmonic content as the coefficient of variation. The voltage unbalance coefficient, current unbalance coefficient, harmonic content variation coefficient, and load mutation coefficient are normalized, and the industrial load mutation index is evaluated based on the normalized voltage unbalance coefficient, current unbalance coefficient, harmonic content variation coefficient, and load mutation coefficient. The specific acquisition steps are: LM = a1 × VU + a2 × CU + a3 × HV + a4 × LM; where LM is the industrial load mutation index, VU is the voltage unbalance coefficient, CU is the current unbalance coefficient, HV is the harmonic content variation coefficient, LM is the load mutation coefficient, and a1, a2, a3, and a4 are the weight coefficients of the voltage unbalance coefficient, the current unbalance coefficient, the harmonic content variation coefficient, and the load mutation coefficient.

[0012] Preferably, the voltage imbalance coefficient acquisition step is as follows: constructing a voltage data matrix based on the three-phase voltage data in the time window; calculating the three-phase voltage mean at each sampling time point to obtain a voltage mean vector; performing difference calculation between the three-phase voltage at each sampling time point and the three-phase voltage mean, and constructing a voltage deviation matrix; using the norm method, calculating the norm of the voltage deviation matrix based on the voltage deviation matrix, which is used to measure the overall voltage deviation in the window. The specific acquisition steps are as follows: Where, ||ΔU|| F It is represented as the norm of the voltage deviation matrix, T is the number of sampling time points, 3 is the number of phases, ΔU ij It is represented as the deviation data of the i-th row and j-th column in the voltage deviation matrix; the rated voltage value is obtained, and the voltage unbalance coefficient is calculated by calculating the ratio of the norm of the voltage deviation matrix to the rated voltage value.

[0013] Preferably, the steps for obtaining the harmonic content variation coefficient are as follows: Step 3.1: taking the harmonic content data as clustering features, taking all the harmonic content data in the time window as a data set, and each harmonic content data in the data set as a data point; Step 3.2: using the silhouette coefficient method to determine the number of clusters K of the data set; Step 3.3: randomly selecting K data points in the data set as initial cluster centers, for each data point, calculating its Euclidean distance to each initial cluster center, for each data point, traversing the K initial cluster centers and assigning it to a distance The cluster corresponding to the nearest initial cluster center; Step 3.4: After traversing all data points, the initial clusters are obtained. For each initial cluster, the data points in it are averaged to obtain a new cluster center; Step 3.5: Repeat steps 3.3 and 3.4 until the cluster center no longer changes, and the final cluster and the final cluster center are obtained; Step 3.6: The mean of all final cluster centers is calculated to obtain the cluster center mean, and the harmonic content variation coefficient is calculated based on the final cluster center and the cluster center mean. The specific acquisition steps are: Where HV is the harmonic content variation coefficient, K is the number of clusters, and C k It represents the kth final cluster center, C represents the cluster center mean, and max(THD) represents the maximum harmonic content data in the time window.

[0014] Preferably, the step of judging the industrial load mutation based on the industrial load mutation index is: comparing the industrial load mutation index with the mutation threshold; if the industrial load mutation index is greater than or equal to the mutation threshold, it is judged that an industrial load mutation has occurred; if the industrial load mutation index is less than the mutation threshold, it is judged that no industrial load mutation has occurred.

[0015] Preferably, the steps for obtaining the corrected power data are: obtaining the power data within the time window, calculating the ratio of the industrial load mutation index to the mutation threshold and then subtracting 1 to obtain a correction factor; obtaining the rated value of the power data, subtracting the rated value of the power data from the power data within the time window and multiplying it by the correction factor to obtain the power data adjustment amount; calculating the difference between the voltage data and the power data adjustment amount to obtain the corrected power data.

[0016] Preferably, the step of judging voltage abnormality based on the adjusted voltage and the voltage safety range is: comparing the adjusted voltage with the voltage safety range; if the adjusted voltage is lower than the low voltage threshold, the current voltage is judged to be abnormal; if the adjusted voltage is higher than the high voltage threshold, the current voltage is judged to be abnormal; if the adjusted voltage is lower than the high voltage threshold and higher than the low voltage threshold, the current voltage is judged to be normal, and no intelligent route switching is performed.

[0017] Preferably, a photovoltaic area voltage overlimit control device based on a regional collaborative U-shaped confluence strategy, the device includes: a power data acquisition module, used to collect power data within a time window through a monitoring device, the power data including voltage data, current data, power factor and photovoltaic output data, and transmit the power data within the time window to a first power anomaly judgment module; the first power anomaly judgment module is used to obtain a power anomaly index based on the power data within the time window, and perform a first power anomaly judgment based on the power anomaly index; an industrial load mutation judgment module, if it is judged that the current power is abnormal, then obtain the industrial load data of each sampling time point within the time window, the industrial load data including voltage imbalance, current imbalance, harmonic content change rate and load curve mutation rate, obtain the industrial load mutation index based on the industrial load data evaluation, and perform an industrial load mutation based on the industrial load mutation index. Judgment; the second power anomaly judgment module, if it is judged that a sudden change in industrial load occurs, the power data is corrected to obtain the corrected power data, and the power anomaly index is re-evaluated based on the corrected power data, and a second power anomaly judgment is performed; the U-shaped confluence strategy execution module, if it is judged that the power is normal, continues to collect power data; if it is judged that the power is abnormal, executes the U-shaped confluence strategy, adjusts the reactive power of the photovoltaic inverter through the reactive compensation of the photovoltaic inverter, obtains the adjusted power data, and transmits the adjusted power data to the intelligent route switching module; the intelligent route switching module is used to obtain the voltage safety range, obtain the adjusted voltage data according to the adjusted power data, compare the adjusted voltage with the voltage safety range, and perform intelligent route switching if the adjusted voltage is not within the voltage safety range; the real-time dynamic adjustment module is used to continue to collect power data in real time and perform real-time dynamic adjustment.

[0018] Technical effects and advantages of the present invention:

[0019] The power data within the time window is collected and the first power anomaly judgment is made. If the current power is judged to be abnormal for the first time, the industrial load data within the time window is obtained, the industrial load mutation index is evaluated, and the industrial load mutation judgment is made. If it is judged that an industrial load mutation occurs, the power data is corrected and the second power anomaly judgment is made based on the corrected power data. If the current power is abnormal for the second time, the U-shaped confluence strategy is executed, and the real-time power data collection is continued, and real-time dynamic adjustments are made, which effectively improves the accuracy of the collected power data, reduces misjudgments of power anomalies, and improves the rationality of voltage control. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Flowchart of a photovoltaic area voltage over-limit control method based on a regional collaborative U-shaped confluence strategy provided in an embodiment of the present application

[0021] Figure 2 A structural diagram of a photovoltaic area voltage over-limit control device based on a regional collaborative U-shaped confluence strategy provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] The technical solutions of the present invention will be described clearly and completely below in conjunction with the drawings in the present invention. In addition, the forms of the various structures described in the following embodiments are merely examples. The method and device for controlling over-limit voltage in photovoltaic areas based on a regional collaborative U-shaped confluence strategy involved in the present invention are not limited to the various structures described in the following embodiments. All other implementations obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0023] The present invention provides a method for controlling over-limit voltage in photovoltaic areas based on regional coordinated U-shaped confluence strategy. Figure 1 As shown, the following steps are included:

[0024] Step 1: Use a sliding time window to collect real-time power data through the monitoring device. The power data includes voltage data, current data, power factor, and photovoltaic output data. The power data within the time window is obtained;

[0025] A sliding time window is a dynamic data processing method that continuously updates data within a fixed timeframe (e.g., 5 or 10 minutes). Over time, the window is moved at regular intervals (e.g., 1 or 30 seconds) to analyze the latest power data. This approach ensures that the system always makes decisions based on the latest power status, rather than relying on data sampling at fixed intervals, thereby improving the continuity and accuracy of grid status monitoring.

[0026] Using a sliding time window improves the system's real-time performance because it continuously updates data, avoiding the lag that can arise from fixed detection time periods. Furthermore, it smooths out short-term anomaly fluctuations, reducing misjudgments and enabling more precise triggering of U-shaped converging strategies and intelligent line switching. The sliding window also dynamically adapts to load fluctuations over time, ensuring more stable anomaly detection without unnecessary adjustments caused by a single sudden change in data.

[0027] Step 2: Evaluate the power data within the time window to obtain a power anomaly index, and make the first power anomaly judgment based on the power anomaly index;

[0028] In this embodiment, it should be specifically explained that the steps for obtaining the power anomaly index are:

[0029] Obtain the voltage data and rated voltage value within the time window, and calculate the voltage impact degree based on the voltage data and rated voltage. The specific acquisition steps are as follows:

[0030]

[0031] Where, E U It is expressed as the voltage influence degree, T is the number of sampling time points, U(t) is the voltage data at the t-th sampling time point, and U ref Expressed as rated voltage value;

[0032] Obtain the current data within the time window and calculate the current impact degree based on the current data. The specific acquisition steps are as follows:

[0033]

[0034] Where, E I It is expressed as the degree of current influence, T is the number of sampling time points, I(t) is the current data at the t-th sampling time point, I(t+1) is the current data at the next sampling time point adjacent to the t-th sampling time point, and I max Represented as the maximum current data within the time window;

[0035] Obtain the power factor within the time window and calculate the power factor impact based on the power factor. The specific acquisition steps are as follows:

[0036]

[0037] Where, E PF It is expressed as the degree of power factor influence, T is the number of sampling time points, PF(t) is the power factor at the t-th sampling time point, and the value range is between 0 and 1. 0.95 is the set reference power factor. It is generally believed that when the power factor is lower than 0.95, the reactive power has a greater impact. Taking the cube to amplify the impact of the low power factor on the abnormal index, but it will not excessively affect the power factor close to 1;

[0038] Obtain PV output data within the time window and calculate the impact of PV treatment based on the PV output data. The specific acquisition steps are as follows:

[0039]

[0040] Where, E PV It is expressed as the degree of influence of photovoltaic treatment, T is the number of sampling time points, P PV (t) represents the photovoltaic output data at the t-th sampling time point, P PV (t+1) represents the photovoltaic output data at the next sampling time point adjacent to the t-th sampling time point, P max Represented as the maximum photovoltaic output data within the time window;

[0041] The voltage impact, current impact, power factor impact, and photovoltaic treatment impact are normalized. Normalization is a data preprocessing method used to convert variables of different dimensions into the same numerical range for unified comparison and calculation. The power anomaly index is obtained based on the normalized voltage impact, current impact, power factor impact, and photovoltaic treatment impact. The specific acquisition steps are as follows:

[0042]

[0043] Where, E total Expressed as power anomaly index, E U Expressed as the voltage influence degree, E I Expressed as the degree of current influence, E PF Expressed as the degree of power factor influence, E PV It is expressed as the degree of influence of photovoltaic treatment, using the sum of squares method to ensure that all abnormal factors can affect the final index.

[0044] In this embodiment, it should be specifically explained that the steps for performing the first power anomaly determination according to the power anomaly index are:

[0045] The power anomaly index is compared with the anomaly threshold. If the power anomaly index is greater than or equal to the anomaly threshold, the current power anomaly is determined. If the power anomaly index is less than the anomaly threshold, the current power is determined to be normal and no adjustments are made. The anomaly threshold is determined using the adaptive threshold method, which dynamically adjusts the anomaly detection threshold. The adaptive threshold method automatically calculates and adjusts the threshold based on factors such as historical power data, real-time grid status, and load fluctuation characteristics, adapting it to different operating conditions, rather than using a fixed static threshold. Compared to fixed thresholds, adaptive thresholds can adjust with changes in time, season, power load, photovoltaic output, and other factors, improving the accuracy of power anomaly detection.

[0046] Step 3: If the current power anomaly is determined for the first time, obtain the industrial load data for each sampling time point within the time window. The industrial load data includes voltage imbalance, current imbalance, harmonic content change rate, and load curve mutation rate. An industrial load mutation index is obtained based on the industrial load data, and the industrial load mutation judgment is performed based on the industrial load mutation index.

[0047] In this embodiment, it should be specifically explained that the steps for obtaining the industrial load mutation index are:

[0048] Three-phase voltage data is acquired within a time window. Based on this data, the voltage unbalance coefficient is evaluated using the norm method. Three-phase voltage data refers to the voltage values ​​of phases A, B, and C measured in a three-phase AC power grid. It is typically expressed in volts and represents the change in three-phase voltage at different points in time. Phases A, B, and C are the three power lines in a three-phase AC system, corresponding to the three independent components of the three-phase voltage. In a standard three-phase power supply system, the voltage amplitudes of phases A, B, and C are equal, with a phase difference of 120°, forming a balanced power system.

[0049] The norm method is a mathematical method used to measure the degree of deviation of vectors and matrices. In power system analysis, the norm method is often used to assess the degree of imbalance in three-phase voltage and current, quantifying anomalies by calculating the overall deviation of the data. For example, the norm can be used to measure the deviation of three-phase voltage or current within a time window, helping to identify problems such as sudden industrial load changes and unbalanced power supply. The core advantage of the norm method is that it simultaneously considers the overall trend of multiple data points, preventing single-point errors from affecting the calculation results. This has led to its widespread application in fields such as power grid anomaly detection, data analysis, and machine learning.

[0050] Acquire three-phase current data within the time window, and evaluate the current unbalance coefficient using the norm method based on the three-phase current data;

[0051] The harmonic content data is obtained within the time window, the K-means clustering method is used to cluster the harmonic content data, and the harmonic content variation coefficient is calculated based on the clustering results;

[0052] Obtain the load signal within the time window, select the wavelet basis function suitable for mutation detection, use discrete wavelet transform to decompose the load signal into low-frequency and high-frequency components, extract the high-frequency component, and use it to analyze the mutation situation. Discrete wavelet transform is a signal time-frequency analysis method. By decomposing the signal into low-frequency components and high-frequency components, the local characteristics of the signal are extracted. The energy of the high-frequency component is calculated based on the high-frequency component to quantify the degree of mutation, and the energy of the high-frequency component is used as the load mutation coefficient. The specific acquisition steps are as follows:

[0053]

[0054] Where LM is the energy of the high-frequency component, that is, the load mutation coefficient, W H (t) represents the high-frequency component at the t-th sampling time point, and T represents the number of sampling time points;

[0055] The voltage unbalance coefficient, current unbalance coefficient, harmonic content variation coefficient, and load sudden change coefficient are normalized. The industrial load sudden change index is obtained based on the normalized voltage unbalance coefficient, current unbalance coefficient, harmonic content variation coefficient, and load sudden change coefficient. The specific steps for obtaining the index are as follows:

[0056] LM=a1×VU+a2×CU+a3×HV+a4×LM;

[0057] In the formula, LM represents the industrial load mutation index, and VU represents the voltage imbalance coefficient. When certain industrial equipment suddenly starts or stops, it causes three-phase voltage deviations, increasing the voltage imbalance. Therefore, an increase in the voltage imbalance coefficient indicates that the system is experiencing a stronger industrial load shock, resulting in a larger industrial load mutation index. This relationship makes the voltage imbalance coefficient an important characteristic parameter for identifying industrial load mutations. CU represents the current imbalance coefficient. When industrial equipment suddenly starts or stops, or when the load fluctuates dramatically, the current of one phase suddenly increases or decreases, disrupting the three-phase current balance and increasing the current imbalance coefficient. Therefore, an increase in the current imbalance coefficient indicates that the sudden industrial load shock has a greater impact on the power grid, resulting in a corresponding increase in the industrial load mutation index. HV represents the harmonic content variation coefficient. When nonlinear industrial loads suddenly start or stop, or when their operating conditions change dramatically, the current harmonic content changes dramatically, leading to increased harmonic content fluctuations. Therefore, an increase in the harmonic content variation coefficient indicates that industrial load mutations have a greater impact on the power grid, resulting in a corresponding increase in the industrial load mutation index. LM represents the load mutation coefficient. When high-power equipment suddenly starts or shuts down, or the load changes rapidly, the power demand in the power grid changes dramatically in a short period of time, causing an increase in the load mutation coefficient. a1, a2, a3, and a4 represent the weight coefficients for the voltage imbalance coefficient, the current imbalance coefficient, the harmonic content variation coefficient, and the load mutation coefficient. a1 + a2 + a3 + a4 = 1. a1, a2, a3, and a4 are obtained using the Analytic Hierarchy Process (AHP). For example, a1, a2, a3, and a4 can be 0.2, 0.2, 0.3, and 0.3, respectively. The AHP is a multi-criteria decision-making method used to determine the relative importance of factors in complex problems. It constructs a hierarchical model to decompose the decision problem into the target layer, the criterion layer, and the indicator layer. Then, a judgment matrix is ​​established using a pairwise comparison method to calculate the weight coefficients of each factor.

[0058] In this embodiment, it should be specifically explained that the steps for obtaining the voltage imbalance coefficient are:

[0059] According to the three-phase voltage data in the time window, the voltage data matrix is ​​constructed. The specific matrix is:

[0060]

[0061] Where U represents the voltage data matrix, the number of rows represents the number of sampling time points in the time window, and data is collected once at each time point t. A (t) represents the A-phase voltage at the t-th sampling time point, U B (t) represents the B-phase voltage at the t-th sampling time point, U C (t) represents the C-phase voltage at the t-th sampling time point;

[0062] Calculate the three-phase voltage mean at each sampling time point to obtain the voltage mean vector, which is specifically expressed as:

[0063]

[0064] Where, Expressed as the voltage mean vector, It is expressed as the mean of the three-phase voltage at the t-th sampling time point, which is used as a reference value to measure the degree of imbalance;

[0065] The difference between the three-phase voltage at each sampling time point and the three-phase voltage mean is calculated, and the voltage deviation matrix is ​​constructed. The specific matrix is:

[0066]

[0067] Where ΔU is the voltage deviation matrix, which represents the deviation of the three-phase voltage relative to the mean at each sampling time point. If the three phases of the power grid are balanced, all values ​​should be close to zero.

[0068] The norm method is used to calculate the norm of the voltage deviation matrix based on the voltage deviation matrix, which is used to measure the overall voltage deviation in the window. The specific acquisition steps are as follows:

[0069]

[0070] Where, ||ΔU|| F It is represented as the norm of the voltage deviation matrix, T is the number of sampling time points, ΔU ij It is represented as the deviation data of the i-th row and j-th column in the voltage deviation matrix, where the square root is taken after the sum of the squares to ensure that the deviations of all sampling time points and three-phase data are considered;

[0071] The rated voltage value is obtained, and the voltage unbalance coefficient is calculated by calculating the ratio of the norm of the voltage deviation matrix to the rated voltage value.

[0072] Using a matrix calculation method to calculate the voltage imbalance coefficient allows for comprehensive analysis of three-phase voltage variations within a sliding time window. Compared to traditional point-by-point calculation methods, this approach can account for voltage fluctuation trends throughout the entire time window, improving the accuracy of detecting sudden changes in industrial loads. Using the norm to calculate the overall voltage deviation effectively avoids misjudgments caused by short-term fluctuations at individual time points, making anomaly detection more stable. Furthermore, by processing multiple data points simultaneously through matrix operations, this method can more accurately reflect the voltage imbalance characteristics caused by industrial loads (such as large motors and welding machines), improving the ability to identify sudden changes in loads.

[0073] In this embodiment, it should be specifically explained that the steps for obtaining the current imbalance coefficient are:

[0074] According to the three-phase current data in the time window, the current data matrix is ​​constructed. The specific matrix is:

[0075]

[0076] Where I represents the current data matrix, the number of rows represents the number of sampling time points in the time window, and data is collected once at each time point t. A (t) represents the A-phase current at the t-th sampling time point, I B (t) represents the B-phase current at the t-th sampling time point, I C (t) represents the C-phase current at the t-th sampling time point;

[0077] Calculate the three-phase current mean at each sampling time point to obtain the current mean vector, which is specifically expressed as:

[0078]

[0079] Where, Expressed as the current mean vector, It is expressed as the mean value of the three-phase current at the t-th sampling time point, which is used as a reference value to measure the degree of imbalance;

[0080] The difference between the three-phase current at each sampling time point and the three-phase current mean is calculated, and the current deviation matrix is ​​constructed. The specific matrix is:

[0081]

[0082] Where ΔI is the current deviation matrix, which represents the deviation of the three-phase current relative to the mean at each sampling time point;

[0083] The norm method is used to calculate the norm of the current deviation matrix based on the current deviation matrix, which is used to measure the overall current deviation in the window. The specific acquisition steps are as follows:

[0084]

[0085] Where, ||ΔI|| F It is represented as the norm of the current deviation matrix, T is the number of sampling time points, ΔI ij It is represented as the deviation data of the i-th row and j-th column in the current deviation matrix, where the square root is taken after the sum of the squares to ensure that the deviations of all sampling time points and three-phase data are considered;

[0086] The rated current value is obtained, and the current unbalance coefficient is calculated by calculating the ratio of the norm of the current deviation matrix to the rated current value.

[0087] In this embodiment, it should be specifically explained that the steps for obtaining the harmonic content variation coefficient are:

[0088] Step 3.1: Use the harmonic content data as clustering features, all the harmonic content data in the time window as the data set, and each harmonic content data in the data set as a data point;

[0089] Step 3.2: Use the silhouette coefficient method to determine the number of clusters K of the data set;

[0090] The Silhouette Coefficient method is an unsupervised learning evaluation method used to measure the quality of data clustering and determine the number of clusters K. It calculates the Silhouette Coefficient for each data point to assess its closeness within the current cluster and its degree of separation from its nearest neighbor cluster. Generally, the optimal number of clusters is chosen to maximize the average Silhouette Coefficient. This ensures that data points are tightly clustered within each cluster while remaining separated from other clusters, thereby improving clustering accuracy and stability.

[0091] Step 3.3: Randomly select K data points in the data set as initial cluster centers. For each data point, calculate the Euclidean distance from each initial cluster center. The specific method is as follows:

[0092]

[0093] Where d(P,β) is the Euclidean distance from the data point to the cluster center, where P is the data point and β is the initial cluster center. For each data point, K initial cluster centers are traversed and assigned to the cluster corresponding to the nearest initial cluster center.

[0094] Euclidean distance is a geometric distance calculation method used to measure the straight-line distance between two points. It is defined as the square root of the sum of the squares of the differences in the corresponding dimensions of the coordinates of the two points in multidimensional space.

[0095] Step 3.4: After traversing all data points, we get the initial clusters. For each initial cluster, we calculate the mean of the data points in it and get the new cluster center.

[0096] Step 3.5: Repeat steps 3.3 and 3.4 until the cluster center no longer changes, and obtain the final cluster and the final cluster center;

[0097] Step 3.6: Calculate the mean of all final cluster centers to obtain the cluster center mean. Calculate the harmonic content variation coefficient based on the final cluster center and the cluster center mean. The specific steps are as follows:

[0098]

[0099] Where HV is the harmonic content variation coefficient, K is the number of clusters, and C k Represented as the kth final cluster center, It is expressed as the cluster center mean, and max(THD) is expressed as the maximum harmonic content data in the time window.

[0100] In this embodiment, it should be specifically explained that the steps for determining industrial load mutation according to the industrial load mutation index are as follows:

[0101] The industrial load mutation index is compared with the mutation threshold. If the industrial load mutation index is greater than or equal to the mutation threshold, it is judged that an industrial load mutation has occurred; if the industrial load mutation index is less than the mutation threshold, it is judged that no industrial load mutation has occurred. The mutation threshold is obtained by the adaptive threshold method.

[0102] Step 4: If a sudden change in industrial load is detected, the power data is corrected to mask the impact of the industrial load on the power data within the time window. Corrected power data is obtained, and the power anomaly index is re-evaluated based on the corrected power data, followed by a second power anomaly determination.

[0103] In this embodiment, it should be specifically explained that the steps for obtaining the corrected power data are:

[0104] Obtain power data within the time window, calculate the ratio of the industrial load mutation index to the mutation threshold, and then subtract 1 to obtain the correction factor;

[0105] Obtain the rated value of power data, which includes the rated voltage, rated current, rated power factor, and rated photovoltaic output. Subtract the rated value from the power data within the time window and multiply it by the correction factor to obtain the power data adjustment. The power data adjustment includes the voltage data adjustment, current data adjustment, power factor adjustment, and photovoltaic output data adjustment. For example, to calculate the voltage data adjustment, the specific steps are as follows:

[0106] TU(t)=(U(t)-U ref )×λ;

[0107] Where TU(t) represents the voltage data adjustment amount at the t-th sampling time point, U(t) represents the voltage data at the t-th sampling time point, and U ref It is expressed as the rated voltage value, and λ is expressed as the correction factor;

[0108] The voltage data and the power data adjustment amount are calculated to obtain the corrected power data. The corrected power data includes corrected voltage data, corrected current data, corrected power factor, and corrected photovoltaic output data. For example, the corrected voltage data is calculated in the following steps:

[0109] JU(t)=U(t)-TU(t);

[0110] JU(t) represents the corrected voltage data at the t-th sampling time point, U(t) represents the voltage data at the t-th sampling time point, and TU(t) represents the voltage data adjustment amount at the t-th sampling time point.

[0111] Step 5: If the current power is normal for the second time, return to step 1 to continue collecting power data. If the current power is abnormal for the second time, execute the U-shaped confluence strategy, adjust the reactive power of the photovoltaic inverter through the reactive power compensation of the photovoltaic inverter, and obtain the adjusted power data;

[0112] The U-shaped converging strategy is a grid voltage regulation method based on regional collaborative optimization, used to address voltage overshoots in photovoltaic (PV) substations. The core of the U-shaped converging strategy is to dynamically adjust the reactive power of multiple PV inverters, enabling them to coordinate when grid voltage anomalies occur, thereby improving the stability of the PV substation grid.

[0113] PV inverter reactive power compensation utilizes PV inverters to regulate reactive power to optimize grid voltage, power factor, and power quality. When voltage is too high, the inverter absorbs reactive power to reduce it; when voltage is too low, the inverter provides reactive power to increase it. This method dynamically adjusts reactive power by monitoring grid conditions in real time, enabling PV grid-connected systems to maintain normal power quality while reducing grid fluctuations and losses.

[0114] Step 6: Obtain the voltage safety range. Obtain the adjusted voltage data based on the adjusted power data, and determine voltage anomaly based on the adjusted voltage and the voltage safety range.

[0115] In this embodiment, it should be specifically explained that the steps for determining voltage abnormality based on the adjusted voltage and the voltage safety range are as follows:

[0116] The adjusted voltage is compared with the voltage safety range. If the adjusted voltage is lower than the low voltage threshold, the current voltage is judged to be abnormal; if the adjusted voltage is higher than the high voltage threshold, the current voltage is judged to be abnormal; if the adjusted voltage is lower than the high voltage threshold and higher than the low voltage threshold, the current voltage is judged to be normal and no intelligent route switching is performed.

[0117] Step 7: If the current voltage is judged to be abnormal, intelligent route switching is performed, and real-time power data collection continues, and real-time dynamic adjustments are made.

[0118] In this embodiment, it should be specifically explained that, Figure 2 As shown, a photovoltaic area voltage over-limit control device based on a regional coordinated U-shaped confluence strategy includes:

[0119] The power data acquisition module is used to collect power data within a time window through a monitoring device. The power data includes voltage data, current data, power factor, and photovoltaic output data, and transmit the power data within the time window to the first power anomaly judgment module;

[0120] A first power anomaly judgment module is used to evaluate the power data within the time window to obtain a power anomaly index and perform a first power anomaly judgment based on the power anomaly index;

[0121] The industrial load mutation judgment module, if it determines that the current power is abnormal, obtains the industrial load data for each sampling time point within the time window. The industrial load data includes voltage imbalance, current imbalance, harmonic content change rate, and load curve mutation rate. The industrial load mutation index is evaluated based on the industrial load data, and the industrial load mutation judgment is performed based on the industrial load mutation index.

[0122] The second power anomaly judgment module, if it is determined that a sudden change in industrial load has occurred, corrects the power data to obtain corrected power data, re-evaluates the power anomaly index based on the corrected power data, and performs a second power anomaly judgment;

[0123] The U-shaped confluence strategy execution module continues to collect power data if it determines that the power is normal. If it determines that the power is abnormal, it executes the U-shaped confluence strategy and adjusts the reactive power of the photovoltaic inverter through reactive compensation to obtain the adjusted power data, which is then transmitted to the intelligent route switching module.

[0124] The intelligent route switching module is used to obtain the voltage safety range, obtain the adjusted voltage data based on the adjusted power data, compare the adjusted voltage with the voltage safety range, and perform intelligent route switching if the adjusted voltage is not within the voltage safety range;

[0125] The real-time dynamic adjustment module is used to continue to collect power data in real time and perform real-time dynamic adjustments.

[0126] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0127] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for controlling over-limit voltage in photovoltaic areas based on a regional coordinated U-shaped confluence strategy, characterized in that: The following steps are involved: Step 1: Use the sliding time window method to collect real-time power data through the monitoring device. The power data includes voltage data, current data, power factor, and photovoltaic output data, and obtain the power data within the time window; Step 2: Evaluate the power data within the time window to obtain a power anomaly index, and make the first power anomaly judgment based on the power anomaly index; Step 3: If the current power anomaly is determined for the first time, obtain the industrial load data for each sampling time point within the time window. The industrial load data includes voltage imbalance, current imbalance, harmonic content change rate, and load curve mutation rate. An industrial load mutation index is obtained based on the industrial load data, and the industrial load mutation judgment is performed based on the industrial load mutation index. Step 4: If it is determined that a sudden change in industrial load has occurred, the power data is corrected to obtain corrected power data, and the power anomaly index is re-evaluated based on the corrected power data, and a second power anomaly judgment is performed; Step 5: If the current power is normal for the second time, continue to collect power data. If the current power is abnormal for the second time, execute the U-shaped confluence strategy, adjust the reactive power of the photovoltaic inverter through the reactive power compensation of the photovoltaic inverter, and obtain the adjusted power data; Step 6: Obtain the voltage safety range and the adjusted voltage data, and determine voltage anomaly based on the adjusted voltage and the voltage safety range; Step 7: If the current voltage is judged to be abnormal, intelligent route switching is performed, and real-time power data collection continues, and real-time dynamic adjustments are made.

2. The method for controlling over-limit voltage in a photovoltaic area based on a regional coordinated U-shaped confluence strategy according to claim 1, characterized in that: The steps for obtaining the power anomaly index are: Obtain voltage data and rated voltage values ​​within a time window, and calculate the voltage impact degree based on the voltage data and rated voltage; Obtain current data within the time window, and calculate the current impact degree based on the current data; Obtain the power factor within the time window and calculate the power factor impact based on the power factor; Obtain PV output data within the time window and calculate the impact of PV treatment based on the PV output data; The voltage impact, current impact, power factor impact, and photovoltaic treatment impact are normalized, and the power anomaly index is obtained based on the normalized voltage impact, current impact, power factor impact, and photovoltaic treatment impact. The specific acquisition steps are as follows: Where, E total Expressed as power anomaly index, E U Expressed as the voltage influence degree, E I Expressed as the degree of current influence, E PF Expressed as the degree of power factor influence, E PV It is expressed as the degree of influence of photovoltaic treatment, using the sum of squares method to ensure that all abnormal factors can affect the final index.

3. The method for controlling over-limit voltage in photovoltaic areas based on regional coordinated U-shaped confluence strategy according to claim 1, characterized in that: The steps of performing the first power anomaly judgment according to the power anomaly index are as follows: Compare the power anomaly index with the anomaly threshold. If the power anomaly index is greater than or equal to the anomaly threshold, it is determined that the current power anomaly is present. If the power anomaly index is less than the anomaly threshold, it is determined that the current power is normal and no adjustment is performed.

4. The method for controlling over-limit voltage in photovoltaic areas based on regional coordinated U-shaped confluence strategy according to claim 1, characterized in that: The steps for obtaining the industrial load mutation index are: Acquire three-phase voltage data within the time window, and evaluate the voltage unbalance coefficient using the norm method based on the three-phase voltage data; Acquire three-phase current data within the time window, and evaluate the current unbalance coefficient using the norm method based on the three-phase current data; The harmonic content data is obtained within the time window, the K-means clustering method is used to cluster the harmonic content data, and the harmonic content variation coefficient is calculated based on the clustering results; Obtain the load signal within the time window, use discrete wavelet transform to decompose the load signal into low-frequency and high-frequency components, extract the high-frequency component, and use the energy of the high-frequency component as the load mutation coefficient; The voltage unbalance coefficient, current unbalance coefficient, harmonic content variation coefficient, and load sudden change coefficient are normalized. The industrial load sudden change index is obtained based on the normalized voltage unbalance coefficient, current unbalance coefficient, harmonic content variation coefficient, and load sudden change coefficient. The specific steps for obtaining the index are as follows: LM=a1×VU+a2×CU+a3×HV+a4×LM; Wherein, LM represents the industrial load mutation index, VU represents the voltage unbalance coefficient, CU represents the current unbalance coefficient, HV represents the harmonic content variation coefficient, LM represents the load mutation coefficient, and a1, a2, a3, and a4 represent the weight coefficients of the voltage unbalance coefficient, the current unbalance coefficient, the harmonic content variation coefficient, and the load mutation coefficient.

5. The method for controlling over-limit voltage in a photovoltaic area based on a regional coordinated U-shaped confluence strategy according to claim 4, characterized in that: The steps for obtaining the voltage unbalance coefficient are: Construct a voltage data matrix based on the three-phase voltage data within the time window; Calculate the three-phase voltage mean at each sampling time point to obtain the voltage mean vector; Calculate the difference between the three-phase voltage at each sampling time point and the three-phase voltage mean, and construct a voltage deviation matrix; The norm method is used to calculate the norm of the voltage deviation matrix based on the voltage deviation matrix, which is used to measure the overall voltage deviation in the window. The specific acquisition steps are as follows: Where, ||ΔU|| F It is represented as the norm of the voltage deviation matrix, T is the number of sampling time points, 3 is the number of phases, ΔU ij Represented as the deviation data of row i and column j in the voltage deviation matrix; The rated voltage value is obtained, and the voltage unbalance coefficient is calculated by calculating the ratio of the norm of the voltage deviation matrix to the rated voltage value.

6. The method for controlling over-limit voltage in a photovoltaic area based on a regional coordinated U-shaped confluence strategy according to claim 4, characterized in that: The steps for obtaining the harmonic content variation coefficient are as follows: Step 3.1: Use the harmonic content data as clustering features, all the harmonic content data in the time window as the data set, and each harmonic content data in the data set as a data point; Step 3.2: Use the silhouette coefficient method to determine the number of clusters K of the data set; Step 3.3: Randomly select K data points in the data set as initial cluster centers. For each data point, calculate its Euclidean distance to each initial cluster center. For each data point, traverse the K initial cluster centers and assign it to the cluster corresponding to the initial cluster center closest to it. Step 3.4: After traversing all data points, we get the initial clusters. For each initial cluster, we calculate the mean of the data points in it and get the new cluster center. Step 3.5: Repeat steps 3.3 and 3.4 until the cluster center no longer changes, and obtain the final cluster and the final cluster center; Step 3.6: Calculate the mean of all final cluster centers to obtain the cluster center mean. Calculate the harmonic content variation coefficient based on the final cluster center and the cluster center mean. The specific steps are as follows: Where HV is the harmonic content variation coefficient, K is the number of clusters, and C k Represented as the kth final cluster center, It is expressed as the cluster center mean, and max(THD) is expressed as the maximum harmonic content data in the time window.

7. The method for controlling over-voltage in photovoltaic areas based on a regional coordinated U-shaped confluence strategy according to claim 1, characterized in that: The steps for judging industrial load mutation according to the industrial load mutation index are as follows: The industrial load mutation index is compared with the mutation threshold. If the industrial load mutation index is greater than or equal to the mutation threshold, it is determined that an industrial load mutation has occurred; if the industrial load mutation index is less than the mutation threshold, it is determined that no industrial load mutation has occurred.

8. The method for controlling over-voltage in photovoltaic areas based on a regional coordinated U-shaped confluence strategy according to claim 1, characterized in that: The steps for obtaining the corrected power data are as follows: Obtain power data within the time window, calculate the ratio of the industrial load mutation index to the mutation threshold, and then subtract 1 to obtain the correction factor; Obtaining a rated value of power data, subtracting the rated value from the power data within the time window and multiplying the result by a correction factor to obtain a power data adjustment amount; The difference between the voltage data and the power data adjustment amount is calculated to obtain the corrected power data.

9. The method for controlling over-limit voltage in a photovoltaic area based on a regional coordinated U-shaped confluence strategy according to claim 1, characterized in that: The steps of determining voltage abnormality based on the adjusted voltage and the voltage safety range are as follows: Compare the adjusted voltage with the voltage safety range. If the adjusted voltage is lower than the low voltage threshold, the current voltage is judged to be abnormal; if the adjusted voltage is higher than the high voltage threshold, the current voltage is judged to be abnormal. If the adjusted voltage is lower than the high voltage threshold and higher than the low voltage threshold, the current voltage is determined to be normal and no intelligent route switching is performed.

10. A photovoltaic area voltage over-limit control device based on a regional coordinated U-shaped confluence strategy, used to implement a photovoltaic area voltage over-limit control method based on a regional coordinated U-shaped confluence strategy according to any one of claims 1 to 9, characterized in that: The device comprises: The power data acquisition module is used to collect power data within a time window through a monitoring device. The power data includes voltage data, current data, power factor, and photovoltaic output data, and transmit the power data within the time window to the first power anomaly judgment module; A first power anomaly judgment module is used to evaluate the power data within the time window to obtain a power anomaly index and perform a first power anomaly judgment based on the power anomaly index; The industrial load mutation judgment module, if it determines that the current power is abnormal, obtains the industrial load data for each sampling time point within the time window. The industrial load data includes voltage imbalance, current imbalance, harmonic content change rate, and load curve mutation rate. The industrial load mutation index is evaluated based on the industrial load data, and the industrial load mutation judgment is performed based on the industrial load mutation index. The second power anomaly judgment module, if it is determined that a sudden change in industrial load has occurred, corrects the power data to obtain corrected power data, re-evaluates the power anomaly index based on the corrected power data, and performs a second power anomaly judgment; The U-shaped confluence strategy execution module continues to collect power data if it determines that the power is normal. If it determines that the power is abnormal, it executes the U-shaped confluence strategy and adjusts the reactive power of the photovoltaic inverter through reactive compensation to obtain the adjusted power data, which is then transmitted to the intelligent route switching module. The intelligent route switching module is used to obtain the voltage safety range, obtain the adjusted voltage data based on the adjusted power data, compare the adjusted voltage with the voltage safety range, and perform intelligent route switching if the adjusted voltage is not within the voltage safety range; The real-time dynamic adjustment module is used to continue to collect power data in real time and perform real-time dynamic adjustments.

Images