Power distribution control method and system for smart substation based on power grid load

Abstract

The application relates to the technical field of power distribution control, in particular to a smart substation power distribution control method and system based on power grid load, which comprises the following steps: performing baseline correction on power grid load data of a current time period to obtain a load baseline part and a load correction part, and performing prediction to obtain a load baseline prediction value and a load correction prediction value; correcting the load correction prediction value; determining the reliability of each prediction method based on the difference between the prediction value and the actual monitoring value of the power grid load data of a historical time period, the numerical relationship between the prediction value and the actual monitoring value of the power grid load data obtained by combining the prediction correction value and the load baseline prediction value; and determining the power grid load data prediction value of the next moment to perform power distribution control. The application aims to improve the accuracy of power grid load prediction and accurately control and adjust the power distribution system of a substation.

Description

Technical Field

[0001] This application relates to the field of power distribution control technology, specifically to a smart substation power distribution control method and system based on grid load. Background Technology

[0002] Traditional substation distribution control strategies primarily rely on setpoint protection and preset operating modes, resulting in poor adaptability to load fluctuations. In actual operation, grid load is frequently affected by factors such as time, season, and the start-up and shutdown of electrical equipment, leading to problems such as short-term overload, three-phase imbalance, and voltage exceeding limits in transformers, feeders, and other equipment. Therefore, more accurate grid load forecasting is needed, and the substation operating mode should be adjusted in real time based on load change trends and forecast results to achieve dynamic and precise distribution control.

[0003] Existing power grid load forecasting algorithms typically only consider the historical data of the load itself, rely on human experience or simple threshold judgments, lack in-depth analysis of load data and trend prediction capabilities, and cannot fully reflect the influence of some external environmental factors during the forecasting process, resulting in insufficient reliability of the forecast results and difficulty in achieving dynamic and accurate power distribution control. Summary of the Invention

[0004] In view of the above, it is necessary to provide a smart substation power distribution control method and system based on grid load to solve the above problems.

[0005] The first aspect of this application provides a smart substation power distribution control method based on grid load, the method comprising:

[0006] For each node of the smart substation, the power grid load data and ambient temperature information for the current time period during power distribution are collected;

[0007] Baseline correction is performed on the power grid load data for the current time period to obtain the load baseline part and the load correction part, and then prediction is performed to obtain the load baseline prediction value and the load correction prediction value.

[0008] Based on the distribution characteristics of the load correction portions at the same time in the current time period and historical time periods, and combined with the changing characteristics of ambient temperature information, the first power grid load prediction correction value is determined; the similarity characteristics between the currently obtained load correction portion and the historically obtained load correction portion are analyzed to determine the second power grid load prediction correction value; the first power grid load prediction correction value and the second power grid load prediction correction value obtained by combining the load correction portions are used to correct the load correction prediction value to obtain the prediction correction value for the current time period.

[0009] The first reliability is determined based on the difference between the predicted results of the power grid load data obtained by the prediction algorithm and the actual power grid load monitoring value; the second reliability is determined based on the numerical relationship between the predicted results of the power grid load data obtained by the prediction correction value and the load baseline prediction value and the actual power grid load monitoring value; and each reliability is used as the weight of the prediction result corresponding to the current time period to determine the predicted value of the power grid load data at the next time moment.

[0010] Power distribution control is performed based on the predicted values ​​of the power grid load data.

[0011] Preferably, the first power grid load forecast correction value is specifically:

[0012] Select load correction data from time points with the same sequence number from a preset number of historical time periods corresponding to the current time period, and cluster them.

[0013] The degree of temperature influence is determined based on the degree of dispersion of the ambient temperature information corresponding to the elements in each cluster at the given time.

[0014] Cluster the ambient temperature information for the current time period, select the cluster where the ambient temperature information collected at the current moment belongs, and use the average value of the power grid load data of all elements in the cluster at the next moment as the first power grid load prediction correction value.

[0015] Preferably, determining the degree of temperature influence specifically involves:

[0016] Calculate the variance of the ambient temperature information of each element in each cluster at the corresponding time, and denote it as temperature variance;

[0017] The sum of the negative correlation mapping results of temperature variances obtained from all clusters at each time step is normalized to obtain the degree of temperature influence at each time step.

[0018] Preferably, determining the second power grid load forecast correction value specifically involves:

[0019] Curve fitting is performed on the load correction data for each time period to obtain the upper and lower envelopes of the fitted curves, and the similarity between the upper and lower envelope curves corresponding to the load correction portion in each time period is calculated.

[0020] The average similarity between the current time period and the load correction portion of each historical time period is used as the trend similarity between the current time period and the load correction portion of each historical time period.

[0021] The trend similarity between the current time period and its corresponding preset number of historical time periods is thresholded to obtain a similarity threshold. The load correction data of the next moment of the historical time period with a trend similarity greater than the similarity threshold is used as the predicted correction reference value of the load correction part of the current time period. The average of all the predicted correction reference values ​​is used as the second power grid load prediction correction value.

[0022] Preferably, obtaining the prediction correction value for the current time period specifically involves:

[0023] Based on the numerical value between the trend similarity obtained in the current time period and the corresponding similarity threshold, the trend similarity value of the current time period is determined and denoted as Hs.

[0024] In the formula, This represents the forecast correction value for the current time period, where Pc represents the load correction forecast value. Indicates the degree of influence of temperature at the current moment. This represents the first power grid load forecast correction value. This represents the revised value for the second power grid load forecast.

[0025] Preferably, the trend similarity value is specifically the average of all trend similarities greater than the corresponding similarity threshold obtained in the current time period.

[0026] Preferably, the specific process for obtaining the first reliability is as follows: the power grid load data for each historical time period corresponding to the current time period is predicted using a prediction algorithm to obtain the original power grid load prediction value for each historical time period; the average of the absolute values ​​of the differences between the original power grid load prediction values ​​obtained for all historical time periods and the actual power grid load data is normalized and the negative correlation mapping result is used as the first reliability.

[0027] The specific process for obtaining the second reliability is as follows: the sum of the load baseline prediction value and the prediction correction value for each historical time period is recorded as the final correction value for each historical time period. The average of the absolute values ​​of the differences between the final correction values ​​for all historical time periods and the actual power grid load data is normalized and the negative correlation mapping result is used as the second reliability.

[0028] Preferably, the formula for determining the predicted value of the power grid load data at the next moment is as follows: ;in, This represents the predicted value of the power grid load data at the next moment; Indicates primary reliability; Indicates second reliability; This represents the original power grid load forecast for the current time period; This represents the baseline load forecast for the current time period; This indicates the forecast correction value for the current time period.

[0029] Preferably, the specific steps for performing power distribution control are as follows:

[0030] If the predicted value of the power grid load data at the next moment is greater than or equal to the set threshold, the capacitor bank will be started in advance; otherwise, the redundant reactive power compensation device will be disconnected in advance.

[0031] Secondly, embodiments of this application also provide an intelligent substation power distribution control system based on grid load, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0032] This application has at least the following beneficial effects:

[0033] This application provides the most accurate and timely load information for subsequent analysis by collecting grid load data and ambient temperature information during the current time period of power distribution. Baseline correction is performed on the grid load data for the current time period, which removes abnormal factors and extracts stable load patterns. Based on the distribution characteristics of the load correction portion at the same time in historical time periods, and combined with changes in ambient temperature, the predicted correction value is determined. The current load data is corrected using the correction portion of historical data, and considering the influence of ambient temperature, the load forecast can be adjusted more accurately. The similarity characteristics between the currently obtained load correction portion and the historically obtained load correction portion are analyzed to determine the second grid load forecast correction value. Comparison of historical data can reveal the long-term trend and periodic fluctuations of load changes, identify similar patterns between the current load and historical loads, and thus adjust the predicted correction value, making the forecast more stable and reliable. The application integrates the first grid load forecast correction value and the second… The power grid load forecast correction value is used to obtain the forecast correction value for the current time period, making the obtained forecast correction value more reliable and better adaptable to different load change scenarios. Based on the difference between the forecast value and the actual monitoring value of the power grid load data in historical time periods, the first reliability is determined, which helps to evaluate the accuracy of the forecasting method and improve the system's adaptability. When the accuracy of historical forecasts is high, the system can provide more credible forecasts and reduce early warning errors. Based on the numerical relationship between the forecast value and the actual monitoring value of the power grid load data obtained from the forecast correction value and the load baseline forecast value, the second reliability is determined. By comparing the corrected forecast value with the actual value, the reliability of the split forecasting method is evaluated. The reliability is used as the weight of the forecast value to determine the forecast value of the power grid load data at the next moment. By weighted fusion of the reliability of different methods, the forecast value can be dynamically adjusted so that different forecasting methods have different impacts on the final forecast result according to their accuracy. This can effectively improve the accuracy of forecasts, enabling the system to cope with various load fluctuations. Based on the forecast values, power distribution control can be carried out, and the final forecast results can be used in the power distribution control system to help decision-makers take timely measures when the load surges or falls short, ensuring the stability and safety of the power grid operation. Timely adjustments can avoid problems such as voltage fluctuations and overloads, reducing the risk of power grid failures. Attached Figure Description

[0034] Figure 1 A flowchart illustrating the steps of a smart substation power distribution control method based on grid load, provided in one embodiment of this application;

[0035] Figure 2 This is a schematic diagram illustrating the acquisition of the final correction value provided in one embodiment of this application. Detailed Implementation

[0036] In the description of the embodiments in this application, the words "exemplary," "or," and "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary," "or," and "for example" is intended to present the relevant concepts in a specific manner.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this application's specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0038] It should also be noted that the terms "first" and "second" in this application and its accompanying drawings are used to distinguish similar objects, rather than to describe a specific order or sequence. The methods disclosed in the embodiments of this application or the methods shown in the flowcharts include one or more steps for implementing the method. Without departing from the scope of protection of this application, the execution order of multiple steps can be interchanged, and some steps can also be deleted.

[0039] The following description, in conjunction with the accompanying drawings, details the specific scheme of the intelligent substation power distribution control method and system based on grid load provided in this application.

[0040] Please see Figure 1 The diagram illustrates a flowchart of a smart substation power distribution control method based on grid load according to an embodiment of this application. The method includes the following steps:

[0041] The first step: For each node of the smart substation, collect the grid load data and ambient temperature information for the current time period during power distribution.

[0042] This application first obtains the grid load data of each node during power distribution in the smart substation. The grid load data is collected in real time. In this embodiment, the time interval for data collection is 1 minute.

[0043] Meanwhile, considering that the operating status of electrical equipment is directly affected by ambient temperature, especially in residential, commercial, and some industrial electricity sectors, incorporating temperature information into power grid load forecasting can significantly improve forecast accuracy. This is primarily because temperature is one of the important external factors affecting short-term power load fluctuations, especially under extreme weather conditions, where temperature plays a dominant role in load change trends. Therefore, considering temperature changes can more effectively capture the patterns of power grid load changes; hence, this application simultaneously collects ambient temperature information for the corresponding area of ​​each node in real time.

[0044] Since the analysis method for the power grid load data of each node is consistent, this application takes the power grid load data of one node as an example for subsequent analysis.

[0045] In this embodiment, when analyzing and regulating the power grid load, the power grid load data collected in real time at each node for a preset time length is processed as a set of data. The preset time length is 24 hours, that is, each moment corresponds to a time period of one day.

[0046] The second step is to perform baseline correction on the power grid load data for the current time period to obtain the load baseline part and the load correction part, and then make predictions to obtain the load baseline prediction value and the load correction prediction value.

[0047] Considering that users' basic electricity needs (such as those of household appliances like refrigerators) are generally predictable, these load demands are relatively easy to forecast. On the other hand, grid load is also affected by environmental factors such as temperature changes, especially during the use of equipment like air conditioners. Therefore, this application decomposes the real-time acquired grid load data over a specific time period to distinguish between basic load demand and load demand affected by environmental factors.

[0048] Therefore, this application employs the existing penalized least squares baseline correction method to process the acquired original power grid load curve, thereby effectively extracting the different characteristics of stable and fluctuating loads. In this application, the stable portion and the more volatile portion of the power grid load data are defined as the load baseline portion and the load correction portion, respectively. The curve representing stable load demand is the load baseline, representing the long-term stable electricity demand in the power grid. By applying the baseline correction algorithm, a relatively stable load curve can be obtained, reflecting the regular load demand of the power grid. In contrast, the corrected portion of the load curve reflects load fluctuations, especially when affected by changes in the external environment (such as temperature fluctuations). This application uses the penalized least squares (PLS) baseline correction method to process the original power grid load data, thereby effectively distinguishing between stable and fluctuating loads.

[0049] Existing time series forecasting algorithms are used to predict the original grid load, the load baseline portion, and the load correction portion, respectively, to obtain the original grid load forecast value Pa, the load baseline forecast value Pb, and the load correction forecast value Pc. In this embodiment, an exponential smoothing forecasting algorithm is used, which is a well-known existing technology and will not be described in detail here.

[0050] The third step is to determine the first power grid load forecast correction value based on the distribution characteristics of the load correction portion at the same time in the current time period and the historical time period, combined with the change characteristics of the ambient temperature information; analyze the similarity characteristics between the currently obtained load correction portion and the historically obtained load correction portion to determine the second power grid load forecast correction value; and correct the load correction forecast value by combining the first power grid load forecast correction value and the second power grid load forecast correction value obtained from the load correction portion to obtain the forecast correction value for the current time period.

[0051] For parts of the load that are relatively stable, time series forecasting algorithms can be used for predictive analysis; while for parts that are unstable, in addition to relying on time series forecasting algorithms, it is also necessary to combine real-time environmental characteristics for correction in order to improve the accuracy and adaptability of the forecast.

[0052] The load correction portion of the real-time acquired time period is analyzed to obtain the impact of temperature on the power grid load, thereby correcting the real-time acquired power grid load forecast results.

[0053] This application first analyzes the changes in user grid load over historical time periods under different temperature conditions, and compares these changes with historical data to determine the degree of influence of temperature variation on load changes under similar grid load conditions. The specific analysis method is as follows: Load correction data from time points with the same sequence number in a preset number of historical time periods corresponding to the current time period are selected. In this embodiment, the preset number is 50. The DBSCAN clustering algorithm is used to perform cluster analysis on the load correction data from the same time points. Taking the data at time i as an example, the number of clusters obtained is denoted as... Then, the variance of the ambient temperature information corresponding to each element in each cluster at any given time is calculated and denoted as the temperature variance. It should be noted that if the current time is 14:00, the current time period refers to the period from 14:00 yesterday to 14:00 today; the historical time period corresponding to the current time period is the data from the past 50 days calculated backward from the current time period for comparative analysis.

[0054] It should be understood that a larger temperature variance within a cluster indicates a smaller impact of temperature changes on grid load data; conversely, a smaller variance in ambient temperature information within a cluster indicates a larger impact of temperature changes on grid load data. This approach effectively assesses the strength of the effect of ambient temperature on grid load data changes and allows for optimization of load forecasting based on these analysis results.

[0055] The sum of the negative correlation mapping results of the temperature variances obtained from all clusters at each time point is normalized to obtain the degree of temperature influence of the power grid load data at each time point. In this embodiment, the formula for calculating the degree of temperature influence of the power grid load data at each time point is: ;in, This indicates the degree of temperature influence on the power grid load data at time i. This represents the number of clusters obtained at time i from the load correction portion of data over M time periods. Let represent the temperature variance of the j-th cluster, where norm() is the maximum-minimum normalization method. It should be understood that, within the same cluster containing the load correction data obtained at time i, a smaller temperature variance indicates a greater impact of temperature on the grid load at that time.

[0056] The ambient temperature information for the current time period is clustered. In this embodiment, the DBSCAN clustering algorithm is used to select the cluster where the ambient temperature information collected at the current time is located. The average value of the power grid load data of all elements in the cluster at the next time is used as the first power grid load prediction correction value, denoted as Dn.

[0057] Although short-term fluctuations may still exist in the load correction data, load correction data obtained from different time periods may still show certain trend correlations. Therefore, this application further analyzes the trend changes of the load correction data to provide a basis for correcting the current grid load forecast, thereby improving the accuracy of the forecast.

[0058] Because grid load changes are affected by factors such as volatility, directly performing similarity analysis on the load correction portions of different time periods may result in low similarity between the two curves. In other words, pairwise comparisons of grid load curves across all time periods generally yield low similarity, which is detrimental to extracting accurate and effective reference grid load data. Therefore, this application performs curve fitting on the load correction portion data for each time period, obtaining the upper and lower envelopes of the fitted curves. By calculating the similarity between the upper and lower envelope curves corresponding to the load correction portion in each time period, minor differences can be better ignored while better analyzing the similarity of the changing trends of the two curves. The average similarity obtained between the current time period and the load correction portions of each historical time period is used as the trend similarity between the current time period and the load correction portions of each historical time period. The trend similarity obtained between the current time period and its corresponding preset number of historical time periods is thresholded to obtain a similarity threshold. The load correction portion data of the next moment of the historical time period with a trend similarity greater than the similarity threshold is used as the predicted correction reference value for the load correction portion of the current time period. The average of all obtained predicted correction reference values ​​is used as the second grid load predicted correction value, denoted as De.

[0059] In this embodiment, similarity is obtained by calculating the reciprocal of the DTW distance between curves and normalizing it. The normalization method selected is the maximum-minimum normalization method.

[0060] The first grid load forecast correction value Dn and the second grid load forecast correction value De obtained from the integrated load correction section are used to correct the load correction forecast value Pc to obtain the forecast correction value for the current time period:

[0061]

[0062] in, This represents the forecast correction value for the current time period, where Pc represents the load correction forecast value. Indicates the degree of influence of temperature at the current moment. This represents the first power grid load forecast correction value, and Hs represents the trend similarity value for the current time period, specifically the mean of all trend similarities greater than the corresponding similarity threshold. This represents the revised value for the second power grid load forecast.

[0063] The fourth step is to determine the first reliability based on the difference between the predicted results of the power grid load data obtained by the prediction algorithm and the actual power grid load monitoring value; to determine the second reliability based on the numerical relationship between the predicted results of the power grid load data obtained by the prediction correction value and the load baseline prediction value and the actual power grid load monitoring value; and to determine the predicted value of the power grid load data for the next time moment by using each reliability as the weight of the prediction result corresponding to the current time period.

[0064] Furthermore, this application uses the same method to split and predict historical data, and obtains the reliability of each prediction result by analyzing the difference between the split prediction results and the actual monitoring results. Specifically, the above three steps are used to analyze the power grid load data of a preset number of historical time periods corresponding to the current time period to obtain the original power grid load prediction value for each historical time period. Baseline load forecast The predicted correction values ​​obtained for each historical time period .

[0065] Raw power grid load forecasts obtained from all historical time periods The negative correlation mapping result, normalized from the average absolute value of the difference between the actual power grid load data and the actual power grid load data, is used as the first reliability. In this embodiment, the normalization method adopts the arctangent transform normalization method, and the negative correlation mapping result of the variable is calculated by the difference between the natural number 1 and the variable.

[0066] It should be understood that the smaller the absolute value of the difference between the prediction results obtained by the existing prediction algorithm and the actual power grid load data, the stronger the reliability of the prediction results of the power grid load data for the current time period using the existing prediction algorithm.

[0067] Meanwhile, the sum of the load baseline forecast and the forecast correction for each historical time period is recorded as the final correction value for that historical time period. The average of the absolute values ​​of the differences between the final correction values ​​for all historical time periods and the actual grid load data is normalized and the resulting negative correlation mapping is used as the second reliability. In this embodiment, the normalization method adopts the arctangent transform normalization method, and the negative correlation mapping result of the variable is calculated by the difference between the natural number 1 and the variable.

[0068] The diagram illustrating the acquisition of the final correction value is shown below. Figure 2 As shown.

[0069] It should be understood that the smaller the difference between the actual collected value and the sum of the baseline predicted value and the baseline corrected predicted value, the stronger the reliability of the predicted value obtained by using the split prediction method.

[0070] Finally, the predicted values ​​are weighted using the first reliability and the second reliability respectively to obtain the predicted grid load value for the next time step. The specific formula is as follows: ;in, Indicates primary reliability; Indicates second reliability; This represents the original power grid load forecast for the current time period; This represents the baseline load forecast for the current time period; This indicates the forecast correction value for the current time period.

[0071] The aforementioned correction method synthesizes the reliability of two different forecasting approaches, weighting them to arrive at a final, more accurate power grid load forecast. The smaller the forecast error and the more stable the historical performance of the corresponding forecasting methods, the higher the reliability of the forecast. By fusing the two forecasting methods, the overfitting or underfitting problems of a single forecasting method under non-stationary load scenarios can be effectively mitigated, forecast variance can be reduced, and the system's adaptability to abnormal operating conditions can be enhanced.

[0072] The fifth step: Perform power distribution control based on the predicted values ​​of the power grid load data.

[0073] If the current grid load forecast reaches or exceeds the set threshold, it indicates a potential short-term load surge (e.g., concentrated air conditioning startup). In this case, capacitor banks should be started in advance or the taps of on-load tap-changing transformers adjusted to prevent a sudden voltage drop. Conversely, if the forecast is below the threshold, redundant reactive power compensation devices should be disconnected in advance to avoid voltage exceeding the upper limit or reactive power backflow. In this embodiment, the threshold is set to 4MW, but implementers can adjust it according to actual conditions.

[0074] Based on the same inventive concept as the above methods, this application also provides an intelligent substation power distribution control system based on grid load, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the above methods.

[0075] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than disclosed in the description; sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0076] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from its essential characteristics. Therefore, the embodiments described above should be considered exemplary and non-limiting in all respects; modifications to the technical solutions described in the foregoing embodiments, or equivalent substitutions of some technical features, without causing the essence of the corresponding technical solutions to deviate from the scope of the technical solutions in the embodiments of this application, should all be included within the protection scope of this application.

Claims

1. A smart substation power distribution control method based on grid load, characterized in that, The method includes the following steps: For each node of the smart substation, the power grid load data and ambient temperature information for the current time period during power distribution are collected; Baseline correction is performed on the power grid load data for the current time period to obtain the load baseline part and the load correction part, and then prediction is performed to obtain the load baseline prediction value and the load correction prediction value. Based on the distribution characteristics of the load correction portions at the same time in the current time period and historical time periods, and combined with the changing characteristics of ambient temperature information, the first power grid load prediction correction value is determined; the similarity characteristics between the currently obtained load correction portion and the historically obtained load correction portion are analyzed to determine the second power grid load prediction correction value; the first power grid load prediction correction value and the second power grid load prediction correction value obtained by combining the load correction portions are used to correct the load correction prediction value to obtain the prediction correction value for the current time period. The first reliability is determined based on the difference between the predicted results of the power grid load data obtained by the prediction algorithm and the actual power grid load monitoring value; the second reliability is determined based on the numerical relationship between the predicted results of the power grid load data obtained by the prediction correction value and the load baseline prediction value and the actual power grid load monitoring value; and each reliability is used as the weight of the prediction result corresponding to the current time period to determine the predicted value of the power grid load data at the next time moment. Power distribution control is performed based on the predicted values ​​of the power grid load data.

2. The intelligent substation power distribution control method based on grid load as described in claim 1, characterized in that, The first power grid load forecast correction value is as follows: Select load correction data from time points with the same sequence number from a preset number of historical time periods corresponding to the current time period, and cluster them. The degree of temperature influence is determined based on the degree of dispersion of the ambient temperature information corresponding to the elements in each cluster at the given time. Cluster the ambient temperature information for the current time period, select the cluster where the ambient temperature information collected at the current moment belongs, and use the average value of the power grid load data of all elements in the cluster at the next moment as the first power grid load prediction correction value.

3. The intelligent substation power distribution control method based on grid load as described in claim 2, characterized in that, The determination of the degree of influence of temperature specifically includes: Calculate the variance of the ambient temperature information of each element in each cluster at the corresponding time, and denote it as temperature variance; The sum of the negative correlation mapping results of temperature variances obtained from all clusters at each time step is normalized to obtain the degree of temperature influence at each time step.

4. The intelligent substation power distribution control method based on grid load as described in claim 3, characterized in that, The determination of the second power grid load forecast correction value specifically involves: Curve fitting is performed on the load correction data for each time period to obtain the upper and lower envelopes of the fitted curves, and the similarity between the upper and lower envelope curves corresponding to the load correction portion in each time period is calculated. The average similarity between the current time period and the load correction portion of each historical time period is used as the trend similarity between the current time period and the load correction portion of each historical time period. The trend similarity between the current time period and its corresponding preset number of historical time periods is thresholded to obtain a similarity threshold. The load correction data of the next moment of the historical time period with a trend similarity greater than the similarity threshold is used as the predicted correction reference value of the load correction part of the current time period. The average of all the predicted correction reference values ​​is used as the second power grid load prediction correction value.

5. The intelligent substation power distribution control method based on grid load as described in claim 4, characterized in that, The process of obtaining the prediction correction value for the current time period is as follows: Based on the numerical value between the trend similarity obtained in the current time period and the corresponding similarity threshold, the trend similarity value of the current time period is determined and denoted as Hs. In the formula, This represents the forecast correction value for the current time period, where Pc represents the load correction forecast value. Indicates the degree of influence of temperature at the current moment. This represents the first power grid load forecast correction value. This represents the revised value for the second power grid load forecast.

6. The intelligent substation power distribution control method based on grid load as described in claim 5, characterized in that, The trend similarity value is specifically the average of all trend similarities greater than the corresponding similarity threshold obtained in the current time period.

7. The intelligent substation power distribution control method based on grid load as described in claim 1, characterized in that, The specific process for obtaining the first reliability is as follows: the power grid load data for each historical time period corresponding to the current time period is predicted using a prediction algorithm to obtain the original power grid load prediction value for each historical time period; the average of the absolute values ​​of the differences between the original power grid load prediction values ​​obtained from all historical time periods and the actual power grid load data is normalized and the negative correlation mapping result is used as the first reliability. The specific process for obtaining the second reliability is as follows: the sum of the load baseline forecast value and the forecast correction value for each historical time period is recorded as the final correction value for each historical time period. The average of the absolute values ​​of the differences between the final correction values ​​for all historical time periods and the actual power grid load data is normalized and the negative correlation mapping result is used as the second reliability.

8. The intelligent substation power distribution control method based on grid load as described in claim 7, characterized in that, The formula for determining the predicted value of the power grid load data at the next moment is as follows: ;in, This represents the predicted value of the power grid load data at the next moment; Indicates primary reliability; Indicates second reliability; This represents the original power grid load forecast for the current time period; This represents the baseline load forecast for the current time period; This indicates the forecast correction value for the current time period.

9. The intelligent substation power distribution control method based on grid load as described in claim 1, characterized in that, The specific steps for performing power distribution control are as follows: If the predicted value of the power grid load data at the next moment is greater than or equal to the set threshold, the capacitor bank will be started in advance; otherwise, the redundant reactive power compensation device will be disconnected in advance.

10. A smart substation power distribution control system based on grid load, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-9.

Images