Load transfer line prediction method and system based on LSTM
By using an LSTM-based load transfer line prediction method, the differences in the time-series characteristics of line loads are accurately captured, solving the problems of low power grid resource utilization and unstable power supply in important power consumption scenarios, and realizing efficient utilization of power grid resources and stable power supply to important loads.
Patent Information
- Application Number
- CN202511410635.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-29
- Publication Date
- 2026-02-24
AI Technical Summary
Existing technologies are insufficient to improve the utilization rate of power grid resources during load transfer, and the power supply stability of important power consumption scenarios is difficult to guarantee, especially when substation overhaul coincides with peak load periods, resulting in serious line overload problems.
A load transfer line prediction method based on LSTM is adopted. The first and second fitting curves are obtained through the fitting algorithm. It is judged whether they meet the preset conditions, accurately captures the differences in the time sequence characteristics of line load, and obtains the necessary lines to be transferred and performs load transfer.
It significantly improved the utilization rate of power grid resources, ensured the power supply stability of important power consumption scenarios, reduced the resource input and equipment wear and tear of emergency measures, and improved the safety of power grid operation.
Smart Images
Figure CN121566474A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the technical field of power system automation, specifically to a load transfer line prediction method and system based on LSTM. Background Technology
[0002] In distribution network dispatching, load transfer is a crucial operation for ensuring the safe and stable operation of the power grid. Especially during special scenarios such as major overhauls of main transformers and busbars, it is necessary to centrally adjust the operating modes of more than ten or even twenty 10kV distribution network lines. During this process, dispatchers face two major decision-making challenges: first, accurately determining whether the connecting lines can withstand the increased load after load transfer without exceeding line operating limits; and second, ensuring that the main transformer receiving the transferred load does not exceed its rated capacity. Currently, when developing load transfer plans, dispatchers mainly refer to historical load data from the previous few days or the same period last year, using a method of directly adding the highest loads of the lines. While this method is simple to operate and can minimize overload risks, it has significant limitations. Because it does not consider the differences in load characteristics during actual operation, the calculation results often leave too much safety margin, leading to low utilization of power grid resources. Especially when substation overhauls overlap with peak load periods, many lines cannot achieve effective load transfer due to theoretically exceeding limits, further exacerbating the pressure on power grid operation. In addition, in scenarios where the load exceeds the limit, the original power supply is replaced by means such as arranging generator trucks or transferring loads on the low-voltage side, thereby reducing the load pressure on the lines to be transferred. However, the importance of electricity demand is not taken into account. By using an indiscriminate replacement mode, the power supply stability of important electricity demand scenarios cannot be prioritized. Summary of the Invention
[0003] To address the technical challenges of existing technologies in improving the utilization rate of power grid resources and ensuring power supply stability in critical power consumption scenarios, this invention provides a load transfer line prediction method and system based on LSTM. By fitting a first predicted load and a second predicted load using a fitting algorithm, first and second fitted curves are obtained. By determining whether the first and second fitted curves meet preset conditions—a curve feature-based matching mechanism—the differences in load timing characteristics of different lines can be accurately captured, significantly improving the utilization efficiency of power grid resources. When the preset conditions are not met, necessary transfer lines are identified based on the importance of the transfer lines obtained from the first fitted curve. Load transfer lines are then determined based on the first and second predicted loads corresponding to these necessary transfer lines, thereby ensuring power supply stability in critical power consumption scenarios. This solves the technical problem of existing technologies in improving the utilization rate of power grid resources and ensuring power supply stability in critical power consumption scenarios.
[0004] To address the aforementioned technical problems, this invention provides a load transfer line prediction method based on LSTM, comprising the following steps: S1: Construct a load forecasting model based on LSTM, and input the historical load of the line to be transferred into the load forecasting model to obtain the first forecast load; S2: The first predicted load is fitted using a fitting algorithm to obtain the first fitting curve, and the importance of the power transfer line is determined based on the first fitting curve. S3: Input the historical load of the candidate lines to be received into the load prediction model to obtain the second predicted load, and use the fitting algorithm to fit the second predicted load to obtain the second fitted curve; S4: Determine whether the first fitted curve and the second fitted curve meet the preset conditions. If they do, the line to be received is the load transfer line. If they do not meet the conditions, obtain the necessary lines to be transferred from the lines to be transferred based on their importance. Obtain the load transfer line based on the first predicted load and the second predicted load corresponding to the necessary lines to be transferred.
[0005] Preferably, in S2, the step of fitting the first predicted load to obtain a first fitted curve using a fitting algorithm includes: The first predicted load is grouped according to the acquisition time of the historical load of the line to be transferred, and the first group load is obtained. The first daily average load curve is obtained based on the average load of the same first group load. The STL decomposition is used to extract features from the first daily average load curve, thereby obtaining the first typical daily fitted curve. The first typical daily fitted curve is then spliced together to obtain the first fitted curve.
[0006] Preferably, in S2, obtaining the importance of the line to be transferred based on the first fitted curve includes: The preset time period is obtained based on the electricity demand corresponding to the line to be transferred, and the matching degree between the peak and valley time periods of the first fitting curve and the preset time period is obtained. The importance of the power supply line to be transferred is determined based on the load stability and load amplitude stability of the first fitted curve and the matching degree.
[0007] Preferably, in S3, the step of fitting the second predicted load to obtain a second fitting curve using a fitting algorithm includes: The second predicted load is grouped according to the acquisition time of the historical load of the line to be received, and the second group load is obtained based on the average load of the same second group load. The second daily average load curve is decomposed using STL to extract features, thereby obtaining the fitting curve for the second typical day. The fitting curves for the second typical day are then spliced together to obtain the second fitting curve.
[0008] Preferably, in step S4, before determining whether the first fitted curve and the second fitted curve meet the preset conditions, the method further includes: The first and second fitted curves are superimposed to obtain the superimposed fitted curve. The degree of fit is obtained based on the electrical parameter constraints of the line to be received and the electrical parameter values of the superimposed fitted curve. When the degree of fit is less than the preset degree of fit, it means that the preset conditions are not met. When the degree of fit is greater than or equal to the preset degree of fit, it means that the preset conditions are met.
[0009] Preferably, when the degree of adaptation is greater than or equal to a preset degree of adaptation, it indicates that the preset condition is met, including: When the compatibility level is greater than or equal to the preset compatibility level, it is then determined whether the main transformer corresponding to the line to be received meets the electrical parameter constraints of the main transformer. If it does not meet the requirements, it means that the preset conditions are not met; otherwise, it means that the preset conditions are met.
[0010] Preferably, in S4, obtaining the load transfer lines based on the first predicted load and the second predicted load corresponding to the necessary transfer lines includes: Based on the first predicted load, the second predicted load, and the preset adaptability of the necessary transfer line, determine whether the receiving line meets the necessary power supply requirements corresponding to the necessary transfer line. If it does, the receiving line is a load transfer line. If it does not, the receiving line is removed from the candidate receiving line set, and S3 is executed again.
[0011] The beneficial effects of this plan are: By fitting the first predicted load and the second predicted load with a fitting algorithm to obtain the first fitting curve and the second fitting curve, respectively, and by judging whether the first fitting curve and the second fitting curve meet the preset conditions, that is, the matching mechanism based on curve features, the difference in load time sequence characteristics of different lines can be accurately captured, avoiding the safety margin problem caused by the direct superposition of maximum load, and significantly improving the utilization efficiency of power grid resources. When it is determined that the preset conditions are not met, the necessary lines to be transferred are obtained from the lines to be transferred based on the importance of the lines to be transferred obtained from the first fitting curve. The load transfer lines are obtained based on the first predicted load and the second predicted load corresponding to the necessary lines to be transferred. This solves the problem of unstable important loads caused by indiscriminate transfer and solves the technical problem that existing technologies are unable to improve the utilization rate of power grid resources and ensure the power supply stability of important power consumption scenarios. By accurately forecasting load and matching power transfer, the frequency of unnecessary generator dispatch and temporary line modifications is reduced, thus lowering resource input costs. At the same time, the scientific power transfer scheme effectively avoids the risk of line and main transformer overload, reduces equipment wear and tear and the probability of failure, and improves the overall operational safety of the power grid.
[0012] This invention also provides an LSTM-based load transfer line prediction system, applicable to the aforementioned LSTM-based load transfer line prediction method, comprising: The first predicted load acquisition module is used to build a load prediction model based on LSTM and input the historical load of the line to be transferred into the load prediction model to obtain the first predicted load. The module for obtaining the importance of the line to be transferred is used to fit the first predicted load to obtain a first fitting curve through a fitting algorithm, and to obtain the importance of the line to be transferred based on the first fitting curve. The second fitting curve acquisition module is used to input the historical load of the candidate receiving lines into the load prediction model to obtain the second predicted load, and to fit the second predicted load to obtain the second fitting curve through the fitting algorithm. The load transfer line acquisition module is used to determine whether the first fitting curve and the second fitting curve meet the preset conditions. When the conditions are met, the line to be received is used as the load transfer line. When the conditions are not met, the necessary lines to be transferred are obtained from the lines to be transferred according to their importance, and the load transfer line is obtained according to the first predicted load and the second predicted load corresponding to the necessary lines to be transferred.
[0013] The beneficial effects of this plan are: By fitting the first and second predicted loads using a fitting algorithm, first and second fitted curves are obtained. By judging whether the first and second fitted curves meet preset conditions, i.e., a matching mechanism based on curve features, the differences in load timing characteristics of different lines can be accurately captured. This avoids the problem of excessive safety margin caused by the direct superposition of maximum loads, and significantly improves the utilization efficiency of power grid resources. When the preset conditions are not met, necessary transfer lines are obtained from the transfer lines based on the importance of the transfer lines obtained from the first fitted curve. The load transfer lines are then obtained based on the first and second predicted loads corresponding to the necessary transfer lines. This solves the problem of unstable important loads caused by indiscriminate transfer and addresses the technical problem that existing technologies cannot improve the utilization rate of power grid resources and ensure the power supply stability of important power consumption scenarios.
[0014] The present invention also provides a computer device, including: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the LSTM-based load transfer line prediction method.
[0015] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the LSTM-based load transfer line prediction method. Attached Figure Description
[0016] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. The drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings.
[0017] Figure 1 This is a flowchart illustrating the load transfer line prediction method based on LSTM of the present invention. Figure 2 This is a schematic diagram of the first fitting curve, the second fitting curve, and the superimposed fitting curve in the LSTM-based load transfer line prediction method of the present invention. Figure 3 This is a schematic diagram of the load rate curve in the LSTM-based load transfer line prediction method of this invention. Figure 4 This is an interface diagram of the LSTM-based load transfer line prediction system of the present invention. Detailed Implementation
[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only one preferred embodiment of this invention and are only used to explain this invention. They do not limit the scope of protection of this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.
[0019] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations (or steps) can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it may also have additional steps not included in the figures; the process may correspond to a method, function, procedure, subroutine, subroutine, etc.
[0020] Example 1: like Figure 1 As shown, the LSTM-based load transfer line forecasting method includes the following steps: S1: Construct a load forecasting model based on LSTM, and input the historical load of the line to be transferred into the load forecasting model to obtain the first forecast load.
[0021] LSTM excels at capturing long-term dependencies and periodic features in time-series data, effectively uncovering intraday peak-valley patterns, weekly / monthly / seasonal periodic fluctuations, and trend changes (such as load growth trends) hidden in the historical load of power transmission lines to be transferred. Compared to traditional simple forecasting methods such as mean analysis and linear regression, it has a stronger ability to fit dynamic load changes (such as load fluctuations caused by sudden weather events, holidays, etc.). Since there are long-term dependencies and periodic features among the historical loads of power transmission lines to be transferred, using LSTM to construct a load forecasting model improves the accuracy of the first-stage load forecast.
[0022] S2: The first predicted load is fitted using a fitting algorithm to obtain the first fitted curve, and the importance of the line to be transferred is determined based on the first fitted curve.
[0023] In some embodiments, S2, the step of fitting the first predicted load to obtain a first fitted curve using a fitting algorithm includes: The first predicted load is grouped according to the acquisition time of the historical load of the line to be transferred, and the first group load is obtained. The first daily average load curve is obtained based on the average load of the same first group load. The STL decomposition is used to extract features from the first daily average load curve, thereby obtaining the first typical daily fitted curve. The first typical daily fitted curve is then spliced together to obtain the first fitted curve.
[0024] The first predicted load is grouped according to the acquisition time of the historical load of the lines to be transferred. Specifically, if the acquisition time of historical load A for the lines to be transferred is Monday, the acquisition time of historical load B is Tuesday, the acquisition time of historical load C is Tuesday, and the acquisition time of historical load D is Monday, then historical load A and historical load D form one group, and historical load B and historical load C form another group. The average load of the historical loads acquired at each time point on Monday constitutes the first daily average load curve. It is understandable that, in addition to weekly units, monthly, quarterly, and yearly units can also be used.
[0025] STL decomposition is a time series data decomposition method based on locally weighted regression, specifically designed to decompose time series data into three core components: a trend term, an intraday periodic term, and a residual term. Its core principle is to separate long-term trends, periodic fluctuations, and random noise in the data through local fitting of a sliding window, thereby clearly extracting the intrinsic characteristics of the time series data. Using STL decomposition to extract features from the first daily average load curve aims to extract the trend term, intraday periodic term, and residual term. Spline interpolation smoothing is applied to the intraday periodic term, moving average processing is applied to the trend term, and noise filtering is applied to the residual term (retaining values within ±1.5 standard deviations). Finally, the trend term, intraday periodic term, and residual term are reconstructed to obtain the fitted curve for the first typical day.
[0026] In another embodiment, S2, obtaining the importance of the line to be transferred based on the first fitted curve includes: The preset time period is obtained based on the electricity demand corresponding to the line to be transferred, and the matching degree between the peak and valley time periods of the first fitting curve and the preset time period is obtained. The importance of the power supply line to be transferred is determined based on the load stability and load amplitude stability of the first fitted curve and the matching degree.
[0027] Load sustainability stability measures the ability of a load to persist over time, focusing on whether the load can exist for extended periods without prolonged interruptions (zero load or extremely low load), reflecting the continuity of power supply requirements of the line's users. Load amplitude stability refers to the degree to which the amplitude of the power load remains relatively stable during operation, reflecting the load's fluctuations over a certain period. Higher load sustainability stability, higher load amplitude stability, and a higher degree of matching between the peak and valley periods of the first fitted curve and the preset time period indicate greater importance of the line to be supplied. Obtaining the importance of the line to be supplied from multiple dimensions improves the reliability of its importance assessment.
[0028] S3: Input the historical load of the candidate receiving lines into the load prediction model to obtain the second predicted load, and use the fitting algorithm to fit the second predicted load to obtain the second fitted curve.
[0029] Specifically, in S3, the step of fitting the second predicted load to obtain a second fitted curve using a fitting algorithm includes: The second predicted load is grouped according to the acquisition time of the historical load of the line to be received, and the second group load is obtained based on the average load of the same second group load. The second daily average load curve is decomposed using STL to extract features, thereby obtaining the fitting curve for the second typical day. The fitting curves for the second typical day are then spliced together to obtain the second fitting curve.
[0030] The second predicted load is obtained by grouping the historical loads of the lines to be received into groups. Specifically, if the historical load 1 of the lines to be received was obtained on Monday, the historical load 2 on Tuesday, the historical load 3 on Tuesday, and the historical load 4 on Monday, then historical load 1 and historical load 4 form one group, and historical load 2 and historical load 3 form another group. The average load of the historical loads obtained at each time point on Monday constitutes the second daily average load curve. It is understandable that, in addition to weekly units, monthly, quarterly, and yearly units can also be used.
[0031] S4: Determine whether the first fitted curve and the second fitted curve meet the preset conditions. If they do, the line to be received is the load transfer line. If they do not meet the conditions, obtain the necessary lines to be transferred from the lines to be transferred based on their importance. Obtain the load transfer line based on the first predicted load and the second predicted load corresponding to the necessary lines to be transferred.
[0032] In some embodiments, before determining whether the first fitted curve and the second fitted curve meet the preset conditions in S4, the method further includes: The first and second fitted curves are superimposed to obtain the superimposed fitted curve. The degree of fit is obtained based on the electrical parameter constraints of the line to be received and the electrical parameter values of the superimposed fitted curve. When the degree of fit is less than the preset degree of fit, it means that the preset conditions are not met. When the degree of fit is greater than or equal to the preset degree of fit, it means that the preset conditions are met.
[0033] In this embodiment, the electrical parameter constraints of the receiving line specifically refer to the current limit of the receiving line, and the electrical parameter values of the superimposed fitting curve specifically refer to the current values of the superimposed fitting curve. Obtaining the fit degree based on the electrical parameter constraints of the receiving line and the electrical parameter values of the superimposed fitting curve specifically refers to dividing the current value of the superimposed fitting curve by the current limit of the receiving line to obtain the load rate of the receiving line. A higher load rate indicates a lower fit degree. Specifically, a load rate exceeding 80% is considered heavy load, and exceeding 100% is considered overload. Users can flexibly set the preset fit degree according to actual needs. Figure 2 As shown, the Donglian A411 line is the line to be transferred, and the Dongqian 472 line is the line to be received. The first fitted curve of the line to be transferred is shown as the blue curve, the second fitted curve of the line to be received is shown as the green curve, and the superimposed fitted curve is shown as the orange curve. Figure 3 As shown in the figure, the load rate of the receiving line obtained by dividing the current value of the superimposed fitted curve by the current limit of the receiving line is shown in the blue curve. At this time, since some of them exceed 80%, they are overloaded.
[0034] Specifically, when the adaptation degree is greater than or equal to the preset adaptation degree, it indicates that the preset condition is met, including: When the compatibility level is greater than or equal to the preset compatibility level, it is then determined whether the main transformer corresponding to the line to be received meets the electrical parameter constraints of the main transformer. If it does not meet the requirements, it means that the preset conditions are not met; otherwise, it means that the preset conditions are met.
[0035] The electrical parameter constraints of the main transformer specifically refer to the current limit of the main transformer. When the total current of the receiving line and other lines connected to the main transformer exceeds the current limit of the main transformer, it indicates that the preset conditions are not met.
[0036] In some preferred embodiments, in S4, obtaining the load transfer line based on the first predicted load and the second predicted load corresponding to the necessary transfer line includes: Based on the first predicted load, the second predicted load, and the preset adaptability of the necessary transfer line, determine whether the receiving line meets the necessary power supply requirements corresponding to the necessary transfer line. If it does, the receiving line is a load transfer line. If it does not, the receiving line is removed from the candidate receiving line set, and S3 is executed again.
[0037] The determination of whether the receiving line meets the necessary power supply requirements of the necessary receiving line based on the first predicted load, the second predicted load, and the preset adaptation level specifically refers to the following: If the first predicted load of the necessary receiving line is 'a', the second predicted load is 'b', and the preset adaptation level is 'c', then the sum of 'a' and 'b' divided by the current limit of the receiving line is used to obtain the actual load rate. When the actual load rate is less than or equal to the preset adaptation level, the sum of the total current of the remaining lines connected to the main transformer corresponding to the receiving line, the first predicted load, and the second predicted load is obtained. When the sum of the total current of the remaining lines connected to the main transformer corresponding to the receiving line, the first predicted load, and the second predicted load is less than or equal to the current limit of the main transformer, it indicates that the necessary power supply requirements of the necessary receiving line are met. When the actual load rate is greater than the preset adaptation level, or when the sum of the total current of the remaining lines connected to the main transformer corresponding to the receiving line, the first predicted load, and the second predicted load is greater than the current limit of the main transformer, it indicates that the necessary power supply requirements of the necessary receiving line are not met.
[0038] Example 2: This embodiment also provides an LSTM-based load transfer line prediction system, applicable to the aforementioned LSTM-based load transfer line prediction method, including: The first predicted load acquisition module is used to build a load prediction model based on LSTM and input the historical load of the line to be transferred into the load prediction model to obtain the first predicted load. The module for obtaining the importance of the line to be transferred is used to fit the first predicted load to obtain a first fitting curve through a fitting algorithm, and to obtain the importance of the line to be transferred based on the first fitting curve. The second fitting curve acquisition module is used to input the historical load of the candidate receiving lines into the load prediction model to obtain the second predicted load, and to fit the second predicted load to obtain the second fitting curve through the fitting algorithm. The load transfer line acquisition module is used to determine whether the first fitting curve and the second fitting curve meet the preset conditions. When the conditions are met, the line to be received is used as the load transfer line. When the conditions are not met, the necessary lines to be transferred are obtained from the lines to be transferred according to their importance, and the load transfer line is obtained according to the first predicted load and the second predicted load corresponding to the necessary lines to be transferred.
[0039] In this embodiment, as Figure 4 As shown, when the receiving line cannot meet the user demand of the transfer line, an early warning needs to be issued. The system monitors the status of each substation and line in real time, and by identifying risks in advance, the spread of hidden dangers is prevented, thus protecting the user's electricity rights and the safety of the power grid.
[0040] Before determining whether the first fitted curve and the second fitted curve meet the preset conditions, the method further includes: The first and second fitted curves are superimposed to obtain a superimposed fitted curve. The degree of fit is determined based on the electrical parameter constraints of the receiving line and the electrical parameter values of the superimposed fitted curve. If the degree of fit is less than a preset degree of fit, it indicates that the preset conditions are not met; if the degree of fit is greater than or equal to the preset degree of fit, it indicates that the preset conditions are met. The statement that the preset conditions are met when the degree of fit is greater than or equal to the preset degree of fit includes: if the degree of fit is greater than or equal to the preset degree of fit, then it is further determined whether the main transformer corresponding to the receiving line meets the main transformer's electrical parameter constraints. If it does not meet these constraints, it indicates that the preset conditions are not met; otherwise, it indicates that the preset conditions are met.
[0041] Example 3: This embodiment also provides a computer device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the LSTM-based load transfer line prediction method.
[0042] Example 4: This embodiment also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the LSTM-based load transfer line prediction method.
[0043] The beneficial effects of this invention are: By fitting the first predicted load and the second predicted load with a fitting algorithm to obtain the first fitting curve and the second fitting curve, respectively, and by judging whether the first fitting curve and the second fitting curve meet the preset conditions, that is, the matching mechanism based on curve features, the difference in load time sequence characteristics of different lines can be accurately captured, avoiding the safety margin problem caused by the direct superposition of maximum load, and significantly improving the utilization efficiency of power grid resources. When it is determined that the preset conditions are not met, the necessary lines to be transferred are obtained from the lines to be transferred based on the importance of the lines to be transferred obtained from the first fitting curve. The load transfer lines are obtained based on the first predicted load and the second predicted load corresponding to the necessary lines to be transferred. This solves the problem of unstable important loads caused by indiscriminate transfer and solves the technical problem that existing technologies are unable to improve the utilization rate of power grid resources and ensure the power supply stability of important power consumption scenarios. By accurately forecasting load and matching power transfer, the frequency of unnecessary generator dispatch and temporary line modifications is reduced, thus lowering resource input costs. At the same time, the scientific power transfer scheme effectively avoids the risk of line and main transformer overload, reduces equipment wear and tear and the probability of failure, and improves the overall operational safety of the power grid.
[0044] The specific embodiments described above are preferred embodiments of the LSTM-based load transfer line prediction method and system of the present invention, and are not intended to limit the specific scope of the present invention. The scope of the present invention includes, but is not limited to, these specific embodiments. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.
Claims
1. A load transfer line prediction method based on LSTM, characterized in that, Includes the following steps: S1: Construct a load forecasting model based on LSTM, and input the historical load of the line to be transferred into the load forecasting model to obtain the first forecast load; S2: The first predicted load is fitted using a fitting algorithm to obtain the first fitting curve, and the importance of the power transfer line is determined based on the first fitting curve. S3: Input the historical load of the candidate lines to be received into the load prediction model to obtain the second predicted load, and use the fitting algorithm to fit the second predicted load to obtain the second fitted curve; S4: Determine whether the first fitted curve and the second fitted curve meet the preset conditions. If they do, the line to be received is the load transfer line. If they do not meet the conditions, obtain the necessary lines to be transferred from the lines to be transferred based on their importance. Obtain the load transfer line based on the first predicted load and the second predicted load corresponding to the necessary lines to be transferred.
2. The load transfer line prediction method based on LSTM according to claim 1, characterized in that, In S2, the step of fitting the first predicted load to obtain the first fitted curve using a fitting algorithm includes: The first predicted load is grouped according to the acquisition time of the historical load of the line to be transferred, and the first group load is obtained. The first daily average load curve is obtained based on the average load of the same first group load. The STL decomposition is used to extract features from the first daily average load curve, thereby obtaining the first typical daily fitted curve. The first typical daily fitted curve is then spliced together to obtain the first fitted curve.
3. The load transfer line prediction method based on LSTM according to claim 1, characterized in that, In S2, obtaining the importance of the line to be transferred based on the first fitted curve includes: The preset time period is obtained based on the electricity demand corresponding to the line to be transferred, and the matching degree between the peak and valley time periods of the first fitting curve and the preset time period is obtained. The importance of the power supply line to be transferred is determined based on the load stability and load amplitude stability of the first fitted curve and the matching degree.
4. The load transfer line prediction method based on LSTM according to claim 2, characterized in that, In S3, the step of fitting the second predicted load to obtain a second fitted curve using a fitting algorithm includes: The second predicted load is grouped according to the acquisition time of the historical load of the line to be received, and the second group load is obtained based on the average load of the same second group load. The second daily average load curve is decomposed using STL to extract features, thereby obtaining the fitting curve for the second typical day. The fitting curves for the second typical day are then spliced together to obtain the second fitting curve.
5. The load transfer line prediction method based on LSTM according to claim 1, characterized in that, In step S4, before determining whether the first fitted curve and the second fitted curve meet the preset conditions, the method further includes: The first and second fitted curves are superimposed to obtain the superimposed fitted curve. The degree of fit is obtained based on the electrical parameter constraints of the line to be received and the electrical parameter values of the superimposed fitted curve. When the degree of fit is less than the preset degree of fit, it means that the preset conditions are not met. When the degree of fit is greater than or equal to the preset degree of fit, it means that the preset conditions are met.
6. The load transfer line prediction method based on LSTM according to claim 5, characterized in that, The condition that the adaptation degree is greater than or equal to the preset adaptation degree indicates that the preset condition is met includes: When the compatibility level is greater than or equal to the preset compatibility level, it is then determined whether the main transformer corresponding to the line to be received meets the electrical parameter constraints of the main transformer. If it does not meet the requirements, it means that the preset conditions are not met; otherwise, it means that the preset conditions are met.
7. The LSTM-based load transfer line prediction method according to claim 5, characterized in that, In S4, the step of obtaining the load transfer line based on the first predicted load and the second predicted load corresponding to the necessary transfer line includes: Based on the first predicted load, the second predicted load, and the preset adaptability of the necessary transfer line, determine whether the receiving line meets the necessary power supply requirements corresponding to the necessary transfer line. If it does, the receiving line is a load transfer line. If it does not, the receiving line is removed from the candidate receiving line set, and S3 is executed again.
8. A load transfer line prediction system based on LSTM, applicable to the load transfer line prediction method based on LSTM as described in any one of claims 1-7, characterized in that, include: The first predicted load acquisition module is used to build a load prediction model based on LSTM and input the historical load of the line to be transferred into the load prediction model to obtain the first predicted load. The module for obtaining the importance of the line to be transferred is used to fit the first predicted load to obtain a first fitting curve through a fitting algorithm, and to obtain the importance of the line to be transferred based on the first fitting curve. The second fitting curve acquisition module is used to input the historical load of the candidate receiving lines into the load prediction model to obtain the second predicted load, and to fit the second predicted load to obtain the second fitting curve through the fitting algorithm. The load transfer line acquisition module is used to determine whether the first fitting curve and the second fitting curve meet the preset conditions. When the conditions are met, the line to be received is used as the load transfer line. When the conditions are not met, the necessary lines to be transferred are obtained from the lines to be transferred according to their importance, and the load transfer line is obtained according to the first predicted load and the second predicted load corresponding to the necessary lines to be transferred.
9. A computer device, characterized in that: include: The computer device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the LSTM-based load transfer line prediction method as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the LSTM-based load transfer line prediction method as described in any one of claims 1-7.