Multi-element elastic load intelligent prediction method and platform for regional power grid
Through grid partition processing and power consumption database analysis, combined with predetermined load feedback coefficient and adjustable potential evaluation, the problem of load prediction results in the prior art is solved, achieving more accurate load prediction and grid elastic response, and improving the scheduling efficiency and safety of the regional power grid.
Patent Information
- Application Number
- CN202510607238.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-13
- Publication Date
- 2025-06-10
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing technology lacks the integration of adjustable potential analysis and dynamic regulation mechanisms, and the failure to deeply integrate the blocking characteristics of the power grid structure, resulting in the load prediction results only staying at the total level, and the regulation capabilities and response characteristics of user loads in different regions cannot be identified, which affects the scheduling accuracy and execution efficiency of the regional power grid in scenarios such as peak load reduction, real-time regulation and abnormal response.
By obtaining the administrative division information of the region, the power grid partition processing is performed, the historical power consumption records of the power grid blocks are extracted, the power consumption database is established, the power consumption trend is analyzed to predict the power consumption, and the load prediction is performed based on the predetermined load feedback coefficient, and the adjustable potential and historical regulation speed are evaluated, block sorting and elastic load partitioning are performed.
It improves load prediction accuracy, optimizes scheduling priority ranking, enhances the elastic response capability of the power grid, and improves the operating efficiency and safety of regional power systems.
Smart Images
Figure CN120127650A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of power grid technology, and particularly to a multi - element elastic load intelligent prediction method and platform for regional power grids. Background Art
[0002] With the continuous growth of regional power grid load and the increasingly complex electricity consumption structure, traditional power dispatching methods have been difficult to meet the ever - increasing requirements for supply - demand flexibility and regulation efficiency. To achieve the safe, economic, and stable operation of the power system, regional power grids increasingly rely on the prediction and regulation capabilities of elastic loads.
[0003] Currently, most existing load prediction technologies adopt static modeling methods, mainly relying on historical electricity consumption data for time - series modeling or using external variables such as meteorology and holidays to construct multivariate prediction models. These methods work well under the condition of stable load patterns, but when faced with the characteristics of regional loads with multiple types, multiple industries, and drastic changes in multiple time periods, problems such as low prediction accuracy and untimely response often occur. More importantly, existing technologies generally ignore the combination of the evaluation of adjustable potential of loads and dynamic regulation capabilities, only providing single - dimension electricity consumption predictions, which are difficult to support more complex load shedding scheduling strategies. In addition, most methods lack the adaptability to the internal structure division of the power grid and cannot perform elastic load analysis and prediction at the block level, resulting in lack of pertinence in regulation operations and inability to fully exploit the load elasticity differences of various users within the region.
[0004] In summary, there are technical problems in the existing technology that due to the lack of integration of adjustable potential analysis and dynamic regulation mechanisms, and the failure to deeply combine with the characteristics of power grid structure blockization, the load prediction results only stay at the total amount level, unable to identify the adjustment capabilities and response characteristics of loads of different regional users, further affecting the scheduling accuracy and execution efficiency of regional power grids in key scenarios such as peak load shedding, real - time regulation, and abnormal response. Summary of the Invention
[0005] The purpose of this application is to provide a multi - element elastic load intelligent prediction method and platform for regional power grids to solve the technical problems in the existing technology that due to the lack of integration of adjustable potential analysis and dynamic regulation mechanisms, and the failure to deeply combine with the characteristics of power grid structure blockization, the load prediction results only stay at the total amount level, unable to identify the adjustment capabilities and response characteristics of loads of different regional users, further affecting the scheduling accuracy and execution efficiency of regional power grids in key scenarios such as peak load shedding, real - time regulation, and abnormal response.
[0006] In view of the above problems, this application provides a multi - element elastic load intelligent prediction method and platform for regional power grids.
[0007] In a first aspect, the present application provides a multi - element elastic load intelligent prediction method for a regional power grid, which is implemented through a multi - element elastic load intelligent prediction platform for a regional power grid, and includes: obtaining administrative division information of a region, and performing power grid zoning processing on the region in combination with the administrative division information to obtain a power grid zoning result; extracting a first power grid block in the power grid zoning result, where the first power grid block corresponds to an identifier of a first block type; collecting historical electricity consumption records of the same - type blocks of the first block type, and forming an electricity consumption database with the first historical electricity consumption record of the first power grid block; analyzing the electricity consumption database to obtain a first predicted electricity consumption of the first power grid block at a first time, and combining a predetermined load feedback coefficient to obtain a first predicted load; establishing a power grid elastic load prediction list for the region according to a first correspondence relationship between the first power grid block and the first predicted load.
[0008] Preferably, the multi - element elastic load intelligent prediction method for a regional power grid further includes: obtaining a predetermined electricity consumption cycle; extracting first historical cycle electricity consumption data from the first historical electricity consumption record based on the predetermined electricity consumption cycle; performing an electricity consumption trend analysis on the first historical cycle electricity consumption data to obtain a first electricity consumption prediction value at the first time; randomly extracting any historical electricity consumption record of any power grid block from the historical electricity consumption records; extracting second historical cycle electricity consumption data from the any historical electricity consumption record based on the predetermined electricity consumption cycle; performing an electricity consumption trend analysis on the second historical cycle electricity consumption data to obtain a second electricity consumption prediction value at the first time; calculating an electricity consumption prediction mean value of the first electricity consumption prediction value and the second electricity consumption prediction value; forming the first predicted electricity consumption based on the first electricity consumption prediction value, the second electricity consumption prediction value, and the electricity consumption prediction mean value.
[0009] Preferably, the multi - element elastic load intelligent prediction method for a regional power grid further includes: performing stage segmentation on the first historical cycle electricity consumption data to obtain a first segmentation result, where the first segmentation result includes a first total electricity consumption corresponding to a first stage; establishing a first electricity consumption scatter plot according to a mapping relationship between the first stage and the first total electricity consumption; randomly obtaining a first scatter point group in the first electricity consumption scatter plot, and generating a first historical electricity consumption curve with the first scatter point group as a constraint; obtaining the first electricity consumption prediction value at the first time based on the first historical electricity consumption curve.
[0010] Preferably, the multi - element elastic load intelligent prediction method for a regional power grid further includes: extracting a first electricity consumption load corresponding to the first stage in the first segmentation result; using the ratio of the first total electricity consumption to the first electricity consumption load as the predetermined load feedback coefficient.
[0011] Preferably, the multi - element elastic load intelligent prediction method for the regional power grid further includes: comparing the first electricity consumption prediction value, the second electricity consumption prediction value and the average electricity consumption prediction value in the first predicted electricity consumption, and screening the maximum value, denoted as the first pessimistic estimated electricity consumption; comparing the first electricity consumption prediction value, the second electricity consumption prediction value and the average electricity consumption prediction value in the first predicted electricity consumption, and screening the minimum value, denoted as the first optimistic estimated electricity consumption; invoking the adjustable potential evaluation mechanism; according to the adjustable potential evaluation mechanism, taking the difference between the first pessimistic estimated electricity consumption and the first optimistic estimated electricity consumption, and performing normalization processing to obtain the first adjustable potential value; sorting the first power grid block by descending order of the first adjustable potential value to obtain a block sequence; and performing elastic load partition control of the region based on the block sequence.
[0012] Preferably, the multi - element elastic load intelligent prediction method for the regional power grid further includes: extracting the first historical regulation data from the first historical cycle electricity consumption data; analyzing the first historical regulation data to obtain the first historical regulation speed; and correcting the first adjustable potential value with the first historical regulation speed as the weight.
[0013] Preferably, the multi - element elastic load intelligent prediction method for the regional power grid further includes: when the first predicted load is at a predetermined load threshold, forming a first adjacent block set of the first power grid block; extracting the first block from the first adjacent block set, and obtaining the first load of the first block; determining whether the first load is at the predetermined load threshold; if not, using the first block as the load balance regulation block of the first power grid block.
[0014] Preferably, the multi - element elastic load intelligent prediction method for the regional power grid further includes: sorting the first adjacent block set in combination with the block sequence to obtain a first adjacent sequence; and denoting the first adjacent block at the head of the first adjacent sequence as the first block.
[0015] Preferably, the multi - element elastic load intelligent prediction method for the regional power grid further includes: if it is, forming a second adjacent block set of the first block, and adding the second adjacent block set to the first adjacent block set.
[0016] In a second aspect, the present application also provides a multi - element elastic load intelligent prediction platform for a regional power grid, which is used to execute the multi - element elastic load intelligent prediction method for a regional power grid as described in the first aspect, including: a power grid zoning module, configured to obtain the administrative division information of a region, and perform power grid zoning processing on the region in combination with the administrative division information to obtain a power grid zoning result; a block extraction module, configured to extract a first power grid block from the power grid zoning result, where the first power grid block corresponds to an identifier of a first block type; a database construction module, configured to collect the historical electricity consumption records of the same - type blocks of the first block type, and construct an electricity consumption database with the first historical electricity consumption record of the first power grid block; a load prediction module, configured to analyze the electricity consumption database to obtain a first predicted electricity consumption of the first power grid block at a first time, and obtain a first predicted load in combination with a predetermined load feedback coefficient; and a list establishment module, configured to establish a power grid elastic load prediction list of the region according to the first correspondence relationship between the first power grid block and the first predicted load.
[0017] The technical solutions provided in the present application have at least the following technical effects or advantages: By achieving the technical goal of multi - element elastic load intelligent prediction and zoning control based on block - level load behavior characteristics and adjustable potential evaluation, the technical effects of improving load prediction accuracy, optimizing dispatching priority sorting, enhancing the elastic response ability of the power grid, and improving the operation efficiency and safety of the regional power system are achieved.
[0018] The above description is only an overview of the technical solutions of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above - mentioned and other purposes, features, and advantages of the present application more obvious and understandable, the following specifically describes the embodiments of the present application. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings described below are only exemplary, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.
[0020] Figure 1 It is a flowchart of the multi - element elastic load intelligent prediction method for a regional power grid of the present application; Figure 2 It is a structural diagram of the multi - element elastic load intelligent prediction platform for a regional power grid of the present application.
[0021] Explanation of reference numerals: The power grid zoning module 11, the block extraction module 12, the database construction module 13, the load forecasting module 14, and the list establishment module 15. Detailed implementation manners
[0022] By providing a multi - elastic load intelligent prediction method and platform for regional power grids, this application solves the technical problems in the prior art. Due to the lack of integration of adjustable potential analysis and dynamic regulation mechanisms, and the failure to deeply combine the characteristics of power grid structure blockization, the load forecasting results only stay at the total amount level, unable to identify the regulation capabilities and response characteristics of users' loads in different regions, further affecting the dispatching accuracy and execution efficiency of regional power grids in key scenarios such as peak load reduction, real - time regulation, and abnormal response. It realizes the technical goal of multi - elastic load intelligent prediction and zoning regulation based on block - level load behavior characteristics and adjustable potential evaluation, achieving the technical effects of improving load forecasting accuracy, optimizing dispatching priority sorting, enhancing the elastic response ability of the power grid, and improving the operation efficiency and safety of the regional power system.
[0023] Next, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited by the example embodiments described here. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of this application. Additionally, it should be noted that for the sake of description, only the parts related to this application are shown in the accompanying drawings, rather than all of them. Embodiment 1
[0024] Please refer to the attached Figure 1 , this application provides a multi - elastic load intelligent prediction method for regional power grids, which is applied to a multi - elastic load intelligent prediction platform for regional power grids, and specifically includes the following steps: S1: Obtain the administrative division information of the region, and perform power grid zoning processing on the region in combination with the administrative division information to obtain a power grid zoning result.
[0025] Specifically, obtaining the administrative division information of the region means obtaining the specific administrative division levels of a region, such as provinces, cities, counties, streets, etc., through maps, geographic information systems, or relevant databases, which helps to divide the region into sub - regions with geographical and management boundaries, making subsequent analysis and operations more accurate.
[0026] Next, the power grid is partitioned for the region in combination with administrative division information, and then the power grid structure within the region can be re-divided according to the needs of power system operation. The power grid partitioning process not only considers geographical boundaries but may also be adjusted based on factors such as load density, power source distribution, and power supply paths, making the power grid division conform to administrative management logic and facilitating power system dispatching and management. The power grid partitioning result obtained through the power grid partitioning process is a specific power grid structure partition diagram or data, which is used for further load forecasting, power supply plan formulation, and equipment operation and maintenance arrangement.
[0027] S2: Extract the first power grid block from the power grid partitioning result, and the first power grid block corresponds to the identifier of the first block type.
[0028] Specifically, randomly extract a power grid block from the power grid partitioning result as the first power grid block. A power grid block refers to a partitioning result in the power grid partitioning result, which is a geographical unit with a relatively complete power grid structure. The first power grid block corresponds to the identifier of the first block type, and the first block type is used to characterize the function or characteristics of the first power grid block in the power grid system. For example, according to the main load composition or power consumption pattern within the first power grid block, it is classified into the first block type, including industrial type, commercial type, residential type, or comprehensive type, etc., which helps to more accurately apply different strategies for load forecasting, power supply regulation, and energy efficiency management during operation.
[0029] S3: Collect the historical power consumption records of the same-type blocks of the first block type and form a power consumption database with the first historical power consumption record of the first power grid block.
[0030] Specifically, the same-type blocks refer to the regions with the same functional attributes as the first power grid block. If the first block type is industrial type, the collected power consumption records should also be those of other industrial-type blocks. Obtain the power consumption data in the past period from all regions in the power grid that belong to the same type as the first power grid block. The historical power consumption record refers to the power consumption situation that occurred in the past period, usually including indicators such as hourly, daily, and monthly power consumption, maximum load, and minimum load.
[0031] Next, form a power consumption database by combining the historical power consumption records of the same-type blocks of the first block type with the first historical power consumption record of the first power grid block to establish a structured database. The first historical power consumption record refers to the past power consumption data of the first power grid block itself, including record items with different time granularities. The power consumption database is a storage system that centrally manages power consumption data, containing fields such as timestamp, block number, block type, actual power consumption, etc., and is used for subsequent data analysis, model training, and load forecasting tasks.
[0032] S4: Analyze the electricity consumption database to obtain a first predicted electricity consumption of the first power grid block at a first time, and obtain a first predicted load in combination with a predetermined load feedback coefficient.
[0033] Specifically, the power consumption database is analyzed to obtain the first predicted power consumption of the first power grid block at the first time, that is, the expected power demand of the first power grid block at a certain time in the future is inferred by performing data mining and trend analysis on the established power consumption database. The first time refers to the time point when the prediction needs to be made.
[0034] The expected load feedback coefficient is calculated by the ratio of historical load to actual load, which may be affected by external factors such as equipment status, adjustment delay, temperature change, etc. The first predicted load is obtained by combining the expected load feedback coefficient, that is, the load level of the block is not only judged based on the predicted power consumption itself, but also needs to be corrected by the expected load feedback coefficient. For example, by multiplying the first predicted power consumption with the expected feedback coefficient, it can be closer to the actual load situation and generate a more accurate first predicted load for subsequent scheduling strategies.
[0035] S5: Establishing a grid elastic load prediction list for the area according to a first corresponding relationship between the first grid block and the first predicted load.
[0036] Specifically, a regional grid elastic load forecast list is established based on the first corresponding relationship between the first grid block and the first forecast load, that is, the forecast load information of each first grid block at a specific future time is centrally sorted to form a list, which reflects the relative adjustability between each block, that is, whether each block can perform operations such as reduction, transfer, and postponement based on the forecast load, thereby constructing a panoramic view of the regional level regulation capability. Table 1 is the most recent forecast record of the grid elastic load forecast list.
[0037] Table 1: The latest forecast record of the grid elastic load forecast list Power grid block Prediction time Predicted power consumption (kW) Predetermined load feedback coefficient Predicted load (kW) Adjustable potential value A01 14:00 on April 10 480 1.10 528 0.85 B02 14:00 on April 10 600 0.90 540 0.72 C03 14:00 on April 10 520 1.05 546 0.78 D04 14:00 on April 10 450 1.20 540 0.65 E05 14:00 on April 10 700 0.95 665 0.92 F06 14:00 on April 10 580 0.98 568 0.70 Further, this application also includes: obtaining a predetermined electricity consumption cycle; extracting first historical cycle electricity consumption data from the first historical electricity consumption record based on the predetermined electricity consumption cycle; performing an electricity consumption trend analysis on the first historical cycle electricity consumption data to obtain a first electricity consumption prediction value at the first time; randomly extracting any historical electricity consumption record of any grid block in the historical electricity consumption record; extracting second historical cycle electricity consumption data from the any historical electricity consumption record based on the predetermined electricity consumption cycle; performing an electricity consumption trend analysis on the second historical cycle electricity consumption data to obtain a second electricity consumption prediction value at the first time; calculating the average electricity consumption prediction of the first electricity consumption prediction value and the second electricity consumption prediction value; and forming the first predicted electricity consumption based on the first electricity consumption prediction value, the second electricity consumption prediction value, and the average electricity consumption prediction.
[0038] Specifically, to obtain a predetermined electricity consumption cycle, a time range is determined as the analysis time window. The predetermined electricity consumption cycle can be one day, one week, one month, or one quarter, specifically depending on the granularity and objective of the prediction task. For example, if analyzing the daily load fluctuation pattern in a residential area, one day may be selected as the electricity consumption cycle; if formulating a monthly electricity consumption plan, one month will be selected as the cycle. The predetermined electricity consumption cycle can ensure that subsequent processing steps are carried out on the same time scale, improving the consistency and comparability of the data.
[0039] In the historical electricity consumption data of the first grid block itself, data corresponding to the period range of the predetermined electricity consumption cycle is extracted to obtain the first historical cycle electricity consumption data. The first historical electricity consumption record is all the electricity consumption information of the first grid block in the past time period, while the first historical cycle electricity consumption data is the electricity consumption information of the predetermined electricity consumption cycle among them.
[0040] Perform an electricity consumption trend analysis on the first historical cycle electricity consumption data, judge the change trend of the electricity consumption data of the first grid block within a cycle, to identify the rule of the electricity consumption data increasing or decreasing over time, and then obtain a first electricity consumption prediction value at a specific future time point, that is, the first time, based on the trend speculation.
[0041] Next, the any grid block is a grid block of the same type as the first block type corresponding to the first grid block in the grid partition result. The any historical electricity consumption record is the historical electricity consumption record of the any grid block. Randomly extracting any historical electricity consumption record of any grid block in the historical electricity consumption record improves the adaptability of the prediction.
[0042] Then, the electricity consumption data of the predetermined electricity consumption cycle is extracted from the any historical electricity consumption record as the second historical cycle electricity consumption data for further analysis.
[0043] Perform an electricity consumption trend analysis on the electricity consumption data of the second historical period, model and analyze the electricity consumption trend of randomly selected power grid blocks within the set period, and obtain the second electricity consumption prediction value at the first time, thereby enhancing the universality and stability of the prediction.
[0044] Next, calculate the average electricity consumption prediction value of the first electricity consumption prediction value and the second electricity consumption prediction value, and then average process to obtain an intermediate value, which helps to reduce the deviation that may exist in a single data.
[0045] Finally, based on the first electricity consumption prediction value, the second electricity consumption prediction value, and the average electricity consumption prediction value, form the first predicted electricity consumption, and form a prediction value with higher stability and accuracy for the first time, which is used to guide dispatching and load response.
[0046] Furthermore, this application also includes: segmenting the electricity consumption data of the first historical period to obtain a first segmentation result, where the first segmentation result includes the total electricity consumption corresponding to the first stage; establishing a first electricity consumption scatter plot according to the mapping relationship between the first stage and the total electricity consumption; randomly obtaining a first scatter point group in the first electricity consumption scatter plot, and generating a first historical electricity consumption curve with the first scatter point group as a constraint; obtaining the first electricity consumption prediction value at the first time based on the first historical electricity consumption curve.
[0047] Specifically, segment the electricity consumption data of the first historical period, and cut it into several time periods with different characteristics according to the time sequence. For example, a complete cycle is one week, which can be divided into four stages: morning peak, daytime stable period, evening peak, and night valley. Each stage represents a relatively stable or representative time period of electricity consumption behavior characteristics. The first segmentation result is the output of the segmentation process, including the boundaries, durations, and corresponding electricity consumption information of each time period, which is used to support more fine-grained analysis.
[0048] Among them, in all the divided stages, calculate the total electricity consumption in the first segmentation result. The total electricity consumption reflects the intensity of the electricity demand in the time period within the first segmentation result and is used to measure the overall level of the load.
[0049] Then, establish a first electricity consumption scatter plot according to the mapping relationship between the first stage and the total electricity consumption, which means visualizing the corresponding relationship between multiple different stages and their respective total electricity consumption. In the first electricity consumption scatter plot, each scatter point represents a combination of a stage and its total electricity consumption. The horizontal axis can represent the time period number, and the vertical axis represents the corresponding total electricity consumption, which helps to discover the regularity or abnormal fluctuations of the load distribution.
[0050] Then, multiple scatter points in the first electricity consumption scatter plot are randomly obtained as the first scatter point group. A first historical electricity consumption curve is generated by connecting the scatter points of the first scatter point group, which is used to obtain the changing trend of electricity consumption over time within a predetermined electricity consumption cycle. Finally, the first time is the future electricity consumption time. Based on the first historical electricity consumption curve, a curve trend analysis is performed to obtain the first electricity consumption prediction value at the future electricity consumption time.
[0051] Furthermore, this application also includes: extracting the first electricity consumption load corresponding to the first stage in the first segmentation result; using the ratio of the first total electricity consumption to the first electricity consumption load as the predetermined load feedback coefficient.
[0052] Specifically, extracting the first electricity consumption load corresponding to the first stage in the first segmentation result, that is, the electricity consumption load data within the first stage, including power demand or load conditions, is used to evaluate the electricity consumption behavior characteristics of the first stage.
[0053] Using the ratio of the first total electricity consumption to the first electricity consumption load as the predetermined load feedback coefficient, that is, using the ratio between the total electricity consumption of the first stage and the average load of this stage to construct a coefficient, called the predetermined load feedback coefficient, which reflects the relationship between the total electricity consumption and the average load, and is used to evaluate the fit between the prediction and the actual load.
[0054] Through the above process, a mathematical connection between the total electricity consumption and the stage load can be established, providing a basis for subsequent model adjustment and prediction correction. Then, there is a logical sequence between extracting the load data and calculating the feedback coefficient. What is extracted are specific values, and what is calculated is a regular parameter. The feedback coefficient can not only be used for parameter tuning during the modeling process, but also be used to judge the consistency between the prediction result and the actual electricity consumption behavior after the prediction is completed.
[0055] Furthermore, this application also includes: comparing the first electricity consumption prediction value, the second electricity consumption prediction value and the average electricity consumption prediction value in the first predicted electricity consumption, and screening the maximum value, denoted as the first pessimistic estimated electricity consumption; comparing the first electricity consumption prediction value, the second electricity consumption prediction value and the average electricity consumption prediction value in the first predicted electricity consumption, and screening the minimum value, denoted as the first optimistic estimated electricity consumption; invoking the adjustable potential evaluation mechanism; according to the adjustable potential evaluation mechanism, taking the difference between the first pessimistic estimated electricity consumption and the first optimistic estimated electricity consumption, and performing normalization processing to obtain the first adjustable potential value; sorting the first power grid block by descending the first adjustable potential value to obtain a block sequence; based on the block sequence, performing elastic load zoning control on the area.
[0056] Specifically, compare the first electricity consumption prediction value, the second electricity consumption prediction value, and the average electricity consumption prediction in the first predicted electricity consumption, and select the maximum value, denoted as the first pessimistic estimated electricity consumption. This means making a horizontal comparison of the three calculated prediction values, which are respectively the prediction result based on the first power grid block, the prediction result of any power grid block, and the average of the two. The largest value among the three is selected as the estimate under the most unfavorable circumstances for the load demand at the first time, and is used as the first pessimistic estimated electricity consumption.
[0057] Next, compare the first electricity consumption prediction value, the second electricity consumption prediction value, and the average electricity consumption prediction in the first predicted electricity consumption, and select the minimum value, denoted as the first optimistic estimated electricity consumption. This indicates using the same three prediction values and selecting the smallest value among them as the estimation result under ideal or low-load conditions, thereby obtaining the first optimistic estimated electricity consumption, which is applicable to evaluating the possible operating state of the power grid in the most energy-efficient and lowest-demand scenarios.
[0058] After that, invoke the adjustable potential evaluation mechanism, which means calling a functional module to evaluate the adjustable resource potential in the current power grid system. Adjustable potential refers to the part of the electricity that can be adjusted through peak shaving and valley filling, load shifting, or demand response based on the load prediction. The adjustable potential evaluation mechanism may include parameters such as the user-side response ability, reserve capacity, or the participation ability of the energy storage system, etc., which are used to establish an adjustment space between the prediction result and the actual load.
[0059] According to the adjustable potential evaluation mechanism, take the difference between the first pessimistic estimated electricity consumption and the first optimistic estimated electricity consumption, and then obtain the theoretically load adjustment space at the first time, that is, the maximum range within which the power grid can fluctuate up and down. Normalize the difference to obtain the first adjustable potential value, which means converting the difference into a standardized proportional value between 0 and 1, facilitating subsequent comparison between different time periods and different regions.
[0060] After calculating the first adjustable potential values of multiple first power grid blocks, sort the first adjustable potential values in descending order, and then sort them corresponding to the first power grid blocks to obtain a block sequence, thereby preferentially identifying the power grid blocks with stronger adjustment capabilities, which reflects the theoretical and actual available load adjustment capabilities of each power grid block at the first time based on historical performance and current predictions.
[0061] Based on the block sequence, elastic load zoning regulation of the area is carried out, that is, according to the block sequence arranged in descending order of the adjustable potential value, the load regulation work within the entire area is guided. Among them, the block sequence reflects the strength of the potential role of each block in demand response, load transfer or reduction. Elastic load zoning regulation refers to dividing multiple blocks into several control units with different regulation capabilities according to the load elasticity level of each sub-area, and performing load balance regulation operations accordingly.
[0062] Furthermore, this application also includes: extracting the first historical regulation data from the first historical cycle power consumption data; analyzing the first historical regulation data to obtain the first historical regulation speed; and correcting the first adjustable potential value with the first historical regulation speed as the weight.
[0063] Specifically, extracting the first historical regulation data from the first historical cycle power consumption data means screening out the data related to load regulation, including the response behavior of the power grid during load mutation, the adjustment situation of electrical equipment, the intervention record of the energy storage system, etc.
[0064] Analyzing the first historical regulation data to obtain the first historical regulation speed means extracting the response speed information therein, and further obtaining the time required for the power grid to achieve load change after receiving the regulation signal. The first historical regulation speed can be measured by the amount of load regulated per unit time, reflecting the sensitivity and efficiency of the load adjustment ability of the first power grid block.
[0065] Finally, correcting the first adjustable potential value with the first historical regulation speed as the weight. If the first historical regulation speed is faster, it means that the first power grid block has strong regulation ability and the adjustable potential value is highly credible, and a higher weight can be given; otherwise, the weight is low, and the actually available potential value after correction will decrease.
[0066] Furthermore, this application also includes: when the first predicted load is at the predetermined load threshold, forming the first adjacent block set of the first power grid block; extracting the first block in the first adjacent block set, and obtaining the first load of the first block; judging whether the first load is at the predetermined load threshold; if not, using the first block as the load balance regulation block of the first power grid block.
[0067] Specifically, the predetermined load threshold is the boundary for judging whether to trigger regulation. For example, when the power consumption of the block exceeds 9 million kilowatts and reaches the set threshold of 9 million kilowatts, the adjacent block identification mechanism is started, and the predetermined load threshold is set by those skilled in the art according to the actual situation. When the first predicted load is at the predetermined load threshold, forming the first adjacent block set of the first power grid block means identifying the adjacent areas that are physically or electrically connected closely to it and forming an adjacent block set.
[0068] Next, extract the first block from the first adjacent block set and obtain the first load of the first block, that is, select the block with the strongest regulation ability and most likely to play an effective role in load regulation from the identified adjacent block set, and obtain the real-time load data of the first block. The first load refers to the current power consumption value of the first block, which is used to compare with a predetermined load threshold to determine whether the first block has the regulation ability.
[0069] Subsequently, determine whether the first load is at the predetermined load threshold, that is, determine whether it is in a critical state. If the current load of the first block has not reached the threshold, it means that the first block still has surplus power supply or regulation ability at this time and has the possibility of taking on part of the load of other areas.
[0070] If it is not in that state, use the first block as the load balance regulation block of the first power grid block, that is, select this adjacent block as the load regulation auxiliary area of the first power grid block with a high current load. The load balance regulation block refers to an area that can help the target block reduce the load pressure through technical means or dispatching strategies, such as energy transfer, load transfer, or demand response measures.
[0071] Furthermore, this application also includes: sorting the first adjacent block set in combination with the block sequence to obtain a first adjacent sequence; and denoting the first adjacent block in the first adjacent sequence as the first block.
[0072] Specifically, sorting the first adjacent block set in combination with the block sequence to obtain a first adjacent sequence means reordering the multiple blocks in the adjacent block set. For example, taking the relative position of the adjacent blocks in the overall sorting as a reference, arranging the blocks with stronger regulation ability in front according to the block sequence, so as to form a new adjacent block order with optimized priority.
[0073] Next, denote the first adjacent block in the first adjacent sequence as the first block, which means selecting the first block with the highest ranking in the sorted adjacent block sequence as the target for the next regulation judgment or operation. Since the first block has the strongest comprehensive regulation ability in the adjacent set, it is most likely to play an effective role in load regulation.
[0074] Furthermore, this application also includes: if it is in that state, form a second adjacent block set of the first block and add the second adjacent block set to the first adjacent block set.
[0075] Specifically, it is determined whether the first load is at a predetermined load threshold. If so, it indicates that the first block itself is also at a high load and temporarily does not have the ability to provide regulation support for the original target block. Then, by forming a second adjacent block set of the first block and adding the second adjacent block set to the first adjacent block set, that is, not directly abandoning the first block, but continuing to expand outward, searching for power grid blocks adjacent to the first block again with the first block as the center. The newly identified set of adjacent blocks is called the second adjacent block set, and thus recursive expansion can be carried out on the original adjacency relationship, that is, searching for more potential regulation resource areas starting from the first block. The newly added adjacent blocks may be more suitable in terms of load level or regulation ability to support the currently overloaded original power grid block. Adding the second adjacent block set to the first adjacent block set means adding the newly expanded blocks to the original list of adjacent blocks, expanding the range of candidate blocks, and continuing with subsequent screening and sorting to ensure finding the optimal regulation path.
[0076] In summary, the multi - elastic load intelligent prediction method for regional power grids provided by this application has the following technical effects: By achieving the technical goal of multi - elastic load intelligent prediction and zoning regulation based on block - level load behavior characteristics and adjustable potential evaluation, it achieves the technical effects of improving load prediction accuracy, optimizing the scheduling priority ranking, enhancing the elastic response ability of the power grid, and improving the operation efficiency and security of the regional power system. Embodiment 2
[0077] Based on the same inventive concept as the multi - elastic load intelligent prediction method for regional power grids in the foregoing embodiment, this application also provides a multi - elastic load intelligent prediction platform for regional power grids. Please refer to the appendix Figure 2 which includes: a power grid zoning module 11 for obtaining the administrative division information of the region and performing power grid zoning processing on the region in combination with the administrative division information to obtain a power grid zoning result; a block extraction module 12 for extracting the first power grid block in the power grid zoning result, where the first power grid block corresponds to the identifier of the first block type; a database construction module 13 for collecting the historical electricity consumption records of the same - type blocks of the first block type and constructing an electricity consumption database with the first historical electricity consumption record of the first power grid block; a load prediction module 14 for analyzing the electricity consumption database to obtain the first predicted electricity consumption of the first power grid block at the first time and obtaining the first predicted load in combination with a predetermined load feedback coefficient; and a list establishment module 15 for establishing a power grid elastic load prediction list of the region according to the first corresponding relationship between the first power grid block and the first predicted load.
[0078] Further, the intelligent prediction platform for diversified flexible loads of the regional power grid is also used for: obtaining a predetermined power consumption cycle; extracting first historical cycle power consumption data from the first historical power consumption records based on the predetermined power consumption cycle; performing power consumption trend analysis on the first historical cycle power consumption data to obtain a first power consumption prediction value at the first time; randomly extracting any historical power consumption record of any power grid block in the historical power consumption records; extracting second historical cycle power consumption data from the any historical power consumption record based on the predetermined power consumption cycle; performing power consumption trend analysis on the second historical cycle power consumption data to obtain a second power consumption prediction value at the first time; calculating the average power consumption prediction value of the first power consumption prediction value and the second power consumption prediction value; and forming the first predicted power consumption based on the first power consumption prediction value, the second power consumption prediction value, and the average power consumption prediction value.
[0079] Further, the intelligent prediction platform for diversified flexible loads of the regional power grid is also used for: performing stage segmentation on the first historical cycle power consumption data to obtain a first segmentation result, where the first segmentation result includes a first total power consumption corresponding to a first stage; establishing a first power consumption scatter plot according to the mapping relationship between the first stage and the first total power consumption; randomly obtaining a first scatter plot group in the first power consumption scatter plot, and generating a first historical power consumption curve with the first scatter plot group as a constraint; and obtaining the first power consumption prediction value at the first time based on the first historical power consumption curve.
[0080] Further, the intelligent prediction platform for diversified flexible loads of the regional power grid is also used for: extracting the first power consumption load corresponding to the first stage in the first segmentation result; and using the ratio of the first total power consumption to the first power consumption load as the predetermined load feedback coefficient.
[0081] Further, the intelligent prediction platform for diversified flexible loads of the regional power grid is also used for: comparing the first power consumption prediction value, the second power consumption prediction value, and the average power consumption prediction value in the first predicted power consumption, and screening the maximum value, denoted as the first pessimistic estimated power consumption; comparing the first power consumption prediction value, the second power consumption prediction value, and the average power consumption prediction value in the first predicted power consumption, and screening the minimum value, denoted as the first optimistic estimated power consumption; invoking an adjustable potential evaluation mechanism; according to the adjustable potential evaluation mechanism, taking the difference between the first pessimistic estimated power consumption and the first optimistic estimated power consumption, and performing normalization processing to obtain a first adjustable potential value; sorting the first power grid block in descending order of the first adjustable potential value to obtain a block sequence; and performing elastic load zoning control of the region based on the block sequence.
[0082] Further, the multi - elastic load intelligent prediction platform for regional power grids is also configured to: extract first historical regulation data from the first - cycle power consumption data; analyze the first historical regulation data to obtain a first historical regulation speed; and correct the first adjustable potential value with the first historical regulation speed as a weight.
[0083] Further, the multi - elastic load intelligent prediction platform for regional power grids is also configured to: when the first predicted load is at a predetermined load threshold, form a first adjacent block set of the first power grid block; extract a first block from the first adjacent block set and obtain the first load of the first block; determine whether the first load is at the predetermined load threshold; if not, use the first block as the load - balancing regulation block of the first power grid block.
[0084] Further, the multi - elastic load intelligent prediction platform for regional power grids is also configured to: sort the first adjacent block set according to the block sequence to obtain a first adjacent sequence; and denote the first adjacent block at the head of the first adjacent sequence as the first block.
[0085] Further, the multi - elastic load intelligent prediction platform for regional power grids is also configured to: if it is at the threshold, form a second adjacent block set of the first block and add the second adjacent block set to the first adjacent block set.
[0086] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The multi - elastic load intelligent prediction method and specific examples for regional power grids in the first - mentioned embodiment 1 are equally applicable to the multi - elastic load intelligent prediction platform for regional power grids in this embodiment. Through the detailed description of the multi - elastic load intelligent prediction method for regional power grids above, those skilled in the art can clearly know the multi - elastic load intelligent prediction platform for regional power grids in this embodiment. Therefore, for the sake of simplicity of the specification, it will not be elaborated here.
[0087] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.
[0088] Obviously, those skilled in the art can make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of this application and its equivalent technologies, this application also intends to include these changes and modifications.
Claims
1. A multivariate elastic load intelligent prediction method for a regional power grid, characterized in that: include: Obtaining administrative division information of a region, and performing power grid division processing on the region in combination with the administrative division information to obtain a power grid division result; Extracting a first power grid block from the power grid partitioning result, where the first power grid block corresponds to an identifier of a first block type; Collecting historical power consumption records of blocks of the same type as the first block type, and establishing a power consumption database together with the first historical power consumption records of the first power grid block; Analyze the power consumption database to obtain a first predicted power consumption of the first power grid block at a first time, and obtain a first predicted load in combination with a predetermined load feedback coefficient; A grid elastic load prediction list for the area is established based on a first corresponding relationship between the first grid block and the first predicted load.
2. The method for intelligent prediction of multivariate elastic loads for a regional power grid according to claim 1, characterized in that: Analyzing the power consumption database to obtain a first predicted power consumption of the first power grid block at a first time includes: Obtaining a scheduled electricity usage cycle; Extracting first historical period electricity consumption data from the first historical electricity consumption record based on the predetermined electricity consumption period; Performing a power consumption trend analysis on the power consumption data of the first historical period to obtain a first power consumption forecast value at the first time; Randomly extract any historical electricity consumption record of any power grid block in the historical electricity consumption record; Extracting second historical period electricity consumption data from any historical electricity consumption record based on the predetermined electricity consumption period; Performing a power consumption trend analysis on the power consumption data of the second historical period to obtain a second power consumption forecast value under the first time period; Calculating a predicted power consumption average of the first power consumption prediction value and the second power consumption prediction value; The first predicted power consumption is constructed based on the first predicted power consumption value, the second predicted power consumption value and the predicted power consumption average.
3. The method for intelligent prediction of multivariate elastic loads for a regional power grid according to claim 2, characterized in that: Performing a power consumption trend analysis on the first historical period power consumption data to obtain a first power consumption forecast value at the first time includes: Performing stage segmentation on the first historical period electricity consumption data to obtain a first segmentation result, wherein the first segmentation result includes a first total electricity consumption corresponding to the first stage; Establishing a first electricity consumption scatter plot according to a mapping relationship between the first stage and the first total electricity consumption; Randomly obtain a first scatter point group in the first power consumption scatter plot, and generate a first historical power consumption curve using the first scatter point group as a constraint; The first power consumption forecast value at the first time is obtained based on the first historical power consumption curve.
4. The method for intelligent prediction of multivariate elastic loads for a regional power grid according to claim 3, characterized in that: Before analyzing the power consumption database to obtain a first predicted power consumption of the first power grid block at a first time and obtaining a first predicted load in combination with a predetermined load feedback coefficient, the method includes: Extracting a first power load corresponding to the first stage from the first segmentation result; The ratio of the first total power consumption to the first power load is used as the predetermined load feedback coefficient.
5. The method for intelligent prediction of multivariate elastic loads for a regional power grid according to claim 2, characterized in that: After analyzing the power consumption database to obtain a first predicted power consumption of the first power grid block at a first time, the method further includes: Compare the first power consumption forecast value, the second power consumption forecast value and the power consumption forecast mean in the first forecast power consumption, and select the maximum value, which is recorded as the first pessimistic estimated power consumption; Compare the first power consumption forecast value, the second power consumption forecast value and the power consumption forecast mean in the first forecast power consumption, and select the minimum value, which is recorded as the first optimistic estimated power consumption; Retrieve adjustable potential assessment mechanism; According to the adjustable potential evaluation mechanism, taking the difference between the first pessimistic estimated power consumption and the first optimistic estimated power consumption, and normalizing them to obtain a first adjustable potential value; Sort the first power grid blocks by the first adjustable potential values in descending order to obtain a block sequence; The elastic load partitioning control of the area is performed based on the block sequence.
6. The method for intelligent prediction of multivariate elastic loads for a regional power grid according to claim 5, characterized in that: After obtaining the first adjustable potential value by taking the difference between the first pessimistic estimated power consumption and the first optimistic estimated power consumption according to the adjustable potential evaluation mechanism and normalizing the difference to obtain the first adjustable potential value, the method further includes: Extracting first historical regulation data from the first historical period electricity consumption data; Analyzing the first historical control data to obtain a first historical control speed; The first adjustable potential value is corrected using the first historical control speed as a weight.
7. The method for intelligent prediction of multivariate elastic loads for a regional power grid according to claim 5, characterized in that: The elastic load partitioning control of the region based on the block sequence includes: When the first predicted load is at a predetermined load threshold, forming a first adjacent block set of the first power grid block; Extracting a first block from the first set of adjacent blocks, and obtaining a first load of the first block; determining whether the first load is within the predetermined load threshold; If not, the first block is used as the load balancing control block of the first power grid block.
8. The method for intelligent prediction of multivariate elastic loads for a regional power grid according to claim 7, characterized in that: Extracting the first block from the first set of adjacent blocks includes: sorting the first adjacent block set in combination with the block sequence to obtain a first adjacent sequence; The first adjacent block in the first adjacent sequence is recorded as the first block.
9. The method for intelligent prediction of multivariate elastic loads for a regional power grid according to claim 7, characterized in that: Determining whether the first load is within the predetermined load threshold further includes: if so, forming a second adjacent block set of the first block, and adding the second adjacent block set to the first adjacent block set.
10. A multi-element elastic load intelligent prediction platform for regional power grids, characterized in that: The steps for implementing the method for intelligent prediction of multivariate elastic loads for a regional power grid as described in any one of claims 1 to 9 include: A power grid partitioning module is used to obtain administrative division information of a region, and perform power grid partitioning processing on the region in combination with the administrative division information to obtain a power grid partitioning result; A block extraction module, configured to extract a first power grid block from the power grid partitioning result, wherein the first power grid block corresponds to an identifier of a first block type; a database building module, configured to collect historical power consumption records of blocks of the same type as the first block type, and to build a power consumption database together with the first historical power consumption records of the first power grid block; A load prediction module, configured to analyze the power consumption database to obtain a first predicted power consumption of the first power grid block at a first time, and to obtain a first predicted load in combination with a predetermined load feedback coefficient; A list establishing module is used to establish a grid elastic load prediction list of the area according to a first corresponding relationship between the first grid block and the first predicted load.