A method for identifying and optimizing the aggregation characteristics of flexible loads
Through the combination of the characteristic identification model and the electricity price magnitude table, the time node and electricity node are divided, and the excitation electricity price is dynamically adjusted, which solves the problem of inaccurate incentive electricity price regulation in flexible load aggregation optimization in the existing technology, and achieves the maximization of the economy Benefits and load safety effects.
Patent Information
- Application Number
- CN202411139397.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-20
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-08-20
AI Technical Summary
The existing technology cannot accurately regulate the incentive electricity price during the optimization of flexible load aggregation, resulting in the lack of maximum economic benefits. Especially for industries with high power dependence or large energy storage users, it is impossible to achieve the best economy while ensuring load safety. benefit.
By obtaining characteristic identification models and electricity price magnitude tables for flexible loads, using incentive electricity prices and historical data to divide time and electricity nodes, generate optimization and control reports, and dynamically adjust the incentive electricity price strategy to ensure that under the premise of meeting system safety constraints, the system is safe and secure. Maximize economic benefits.
More precise optimization of incentive electricity price control is achieved, ensuring maximum economic benefits while ensuring load safety, and improving the stability and power efficiency of flexible loads.
Smart Images

Figure CN118676940B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power demand side response, and in particular to a method for identifying and optimizing flexible load aggregation characteristics. Background Art
[0002] New power systems urgently need to develop demand-side flexibility resources that have not yet been fully activated, fully grasp the opportunities for the construction of the electricity spot market, and develop electricity price incentive demand response technology, which has become an important means to improve demand-side flexibility. Demand-side flexibility resources mainly include a series of flexible loads, including electric vehicles, smart buildings, multi-energy microgrids, etc. These resources generally have massive heterogeneous characteristics, and the resources are dispersed, and they need to be efficiently aggregated and optimized to form large-scale controllable resources. Existing technologies rely on users to actively report operating parameters in modeling, so their modeling performance is deeply affected by the accuracy of the reported parameters. For example, when the operating parameters are distorted or there are malicious misreports, neither the centralized direct load control algorithm nor the distributed decomposition and coordination algorithm can obtain the true system optimal solution.
[0003] The Chinese invention patent with application number 202211128637.X provides a non-intrusive flexible load aggregation characteristic identification and optimization method, device and equipment, obtains a characteristic identification model and elasticity estimation model for flexible loads, obtains the incentive electricity price of the current round in real time, and inputs the incentive electricity price of the current round into the characteristic identification model and the elasticity estimation model respectively to output real-time response electricity consumption and real-time virtual elasticity matrix, and finally obtains the optimal incentive electricity price and the optimal response electricity consumption, realizing the aggregated optimization control of non-intrusive flexible loads, without relying on users to actively provide information reporting, and using statistical methods to establish an equivalent mapping relationship of external characteristics.
[0004] However, in actual applications, industries or industrial and commercial users with high electricity dependence, especially large-scale energy storage users, are extremely sensitive to fluctuations in electricity prices and have higher demands for electricity prices. Existing technologies cannot accurately obtain incentive electricity prices when performing aggregate optimization, and cannot achieve maximum economic benefits while ensuring load safety. Summary of the invention
[0005] This application solves the problem of the inability to accurately control the incentive electricity price and reduce economic benefits in the prior art by providing a method for identifying and optimizing flexible load aggregation characteristics, and achieves the technical effect of accurately controlling the incentive electricity price and maximizing economic benefits.
[0006] The present application provides a method for identifying and optimizing flexible load aggregation characteristics, the method comprising:
[0007] S100: Acquire a characteristic identification model for flexible loads, wherein the input of the characteristic identification model is the incentive electricity price, and the output of the characteristic identification model is the real-time response power consumption and the real-time load response characteristic vector; the real-time load response characteristic vector includes the real-time response power consumption change rate, load rate fluctuation, response sensitivity and response delay;
[0008] S200: Obtain an electricity price magnitude table for flexible loads, the electricity price magnitude table including incentive electricity prices and corresponding time nodes, select an initial incentive electricity price for first input into a characteristic identification model, select the remaining incentive electricity prices in turn according to the magnitude rule based on the initial incentive electricity price for input, obtain corresponding real-time response power consumption and real-time load response feature vectors respectively, and mark the time node corresponding to the incentive electricity price that meets the magnitude rule as a time node;
[0009] S300: Obtain the output real-time response power consumption, pre-set a power amplitude threshold according to the historical real-time response power consumption change rate, compare the real-time response power consumption change rate with the power amplitude threshold, and mark the time node corresponding to the real-time response power consumption that is not less than the power amplitude threshold as an electricity node; wherein the power amplitude threshold is set according to actual conditions and historical data, and is used to measure the change amplitude of the real-time response power consumption change rate.
[0010] S400: generating a time-degree curve according to different incentive electricity prices and time-degree nodes, generating an electric degree curve according to the real-time response power consumption and the electric degree node, determining a span node, obtaining the real-time load response characteristic vector of the span node, and generating an optimization control report; the span node is a node where a time-degree node and an electric degree node overlap;
[0011] S500: Aggregate and optimize the flexible loads according to the optimization and control report and update the electricity price level table.
[0012] Furthermore, the magnitude rule is based on the initial incentive electricity price, selects an incentive electricity price of a different magnitude from the initial incentive electricity price or not less than the measurement threshold as the incentive electricity price input subsequently; calculates the change range between incentive electricity prices of the same magnitude, and compares the change range with the measurement threshold. The measurement threshold is set according to actual conditions and experience, and is used to measure the change range of the incentive electricity price.
[0013] Furthermore, two adjacent span nodes are obtained, and the time region between the two span nodes is divided into multiple auxiliary control regions according to the incentive electricity price, and the regions with the same incentive electricity price are the same auxiliary control region; in each of the auxiliary control regions, it is further divided into multiple segment regions according to the difference in incentive electricity prices; each segment region corresponds to a newly divided incentive electricity price, and the corresponding real-time response power consumption and the real-time load response characteristic vector are obtained according to the new incentive electricity price; the optimal incentive electricity price corresponding to the segment region and the corresponding time range are analyzed according to the real-time response power consumption and the real-time load response characteristic vector to generate an optimization control report.
[0014] Furthermore, the difference is the difference in incentive electricity prices between different auxiliary control areas, which is calculated by recording the ending electricity price and the starting electricity price of the auxiliary control area. The auxiliary control areas with a difference greater than 0, that is, the ending electricity price is greater than the starting electricity price, are divided; and the auxiliary control areas with a difference less than 0, that is, the ending electricity price is less than the starting electricity price, are set as pending areas.
[0015] Furthermore, the segment areas are divided according to the next order of magnitude of the difference, and the maximum number of the segment areas does not exceed 10. After the division is completed, a new incentive electricity price is set for each segment area, and the time length of the auxiliary control area is divided according to the number of the segment areas, and the initial time node and the end time node of each segment area are recorded.
[0016] Furthermore, the initial time node of the segment area is obtained, the time range is gradually increased according to the extension step, the real-time response power consumption change rate and the real-time load response characteristic vector are obtained according to the current time range and the new incentive electricity price, and it is determined whether the preset conditions are met. If the preset conditions are met or the maximum number of extensions is reached, the extension is stopped, and the time node at this time is recorded as the end time node.
[0017] Furthermore, the maximum extension times needs to be preset according to the time length and extension step of the auxiliary control area.
[0018] Furthermore, the preset conditions include a first threshold and a second threshold, the first threshold is set according to the real-time response power consumption change rate; the second threshold is set according to a comprehensive fluctuation value, and the comprehensive fluctuation value is generated comprehensively based on the load rate fluctuation, the response sensitivity and the response delay. If the real-time response power consumption change rate reaches the first threshold or the comprehensive fluctuation value reaches the second threshold, the extension is stopped.
[0019] Furthermore, the time range of the segment area and the incentive electricity price are obtained, a preliminary benchmark value and an initial floating range are set for each time range, and an optimization model with the electricity price floating range and magnitude as decision variables is constructed based on each time period. The optimal electricity price floating range and the best magnitude selection for each time period are recorded, and the electricity price magnitude table is updated.
[0020] Furthermore, the preliminary benchmark value is the optimal incentive electricity price within the segment area, and the initial floating range is the range between the incentive electricity price corresponding to the initial time node of the segment area and the incentive electricity price corresponding to the end time node; the optimization model evaluates the comprehensive benefits of the strategy based on the prediction results and optimization objectives, narrows the electricity price floating range, and optimizes the electricity price floating range and magnitude in each time period in combination with the prediction results of the characteristic identification model and the elasticity estimation model.
[0021] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0022] By using the electricity price level table input characteristic identification model and elasticity estimation model, the historical electricity price data is divided by level and associated with specific time nodes, and incentive electricity prices of various levels are used to achieve more accurate incentive electricity price control optimization effect; by combining electricity nodes with time nodes, the span nodes are determined to deeply analyze the load response characteristics, generate optimization and control reports, and propose targeted optimization and control strategies to achieve the effect of maximizing economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1 A schematic diagram of a flow chart of a flexible load aggregation characteristic identification and optimization method in an embodiment of the present invention;
[0024] Figure 2 A schematic diagram showing the magnitude of electricity price in a method for identifying and optimizing flexible load aggregation characteristics in an embodiment of the present invention;
[0025] Figure 3 It is a schematic diagram of a time period analysis curve in a flexible load aggregation characteristic identification and optimization method in an embodiment of the present invention;
[0026] Figure 4 Schematic diagram of segment area division of a flexible load aggregation characteristic identification and optimization method in an embodiment of the present invention. DETAILED DESCRIPTION
[0027] To facilitate the understanding of the present invention, the present application will be described more comprehensively below with reference to the relevant drawings; the drawings show preferred embodiments of the present invention, but the present invention can be implemented in many different forms and is not limited to the embodiments described herein; on the contrary, the purpose of providing these embodiments is to enable a more thorough and comprehensive understanding of the disclosed content of the present invention.
[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which the present invention belongs; the terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention; the term "and / or" used herein includes any and all combinations of one or more related listed items.
[0029] Embodiment 1: Figure 1 As shown, a method for identifying and optimizing flexible load aggregation characteristics, the method comprising:
[0030] S100: Acquire a characteristic identification model for flexible loads, wherein the input of the characteristic identification model is an incentive electricity price, and the output of the characteristic identification model is a real-time response power consumption and a real-time load response characteristic vector, wherein the real-time load response characteristic vector includes a real-time response power consumption change rate, load rate fluctuation, response sensitivity, and response delay.
[0031] In some embodiments, the input of the established characteristic identification model is the incentive electricity price, and the output of the characteristic identification model is the real-time response electricity consumption and the real-time load response feature vector. The established neural network model for characteristic identification is trained. Specifically, the incentive electricity price and the historical response electricity consumption constitute a first training data set, the loss function of the neural network model is set to the mean square error function, and the first training data set is used to train the neural network model for characteristic identification using a stochastic gradient descent algorithm. More than two hidden layers are set in the characteristic identification model, and two output layers are set. For the real-time response electricity consumption output, one neuron corresponds to the output of the predicted real-time response electricity consumption. ; For the real-time load response feature vector output, one neuron is added for each feature, and the output of each neuron corresponds to the predicted value of each feature.
[0032] In some embodiments, each time period corresponds to an incentive electricity price, and each input process selects incentive electricity prices for multiple time periods, where T is the total number of time periods, N is the number of features, the input layer receives T incentive electricity prices as input, the first hidden layer receives T incentive electricity prices, outputs n1 hidden units, uses the ReLU activation function, the second hidden layer receives n2 hidden units of the first hidden layer, and one neuron is used to output the predicted real-time response electricity consumption , The neurons are used to output the predicted value of each feature. , ,......, The loss function L is set to comprehensively consider the real-time response power consumption and the prediction error of all feature vectors:
[0033]
[0034] Among them, α and β are weight coefficients, which are used to balance the importance of the real-time response power consumption prediction error and the real-time load response characteristic vector prediction error; is the true value of the f-th feature in the t-th period, It is the actual real-time response power consumption in the tth period. It is measured and collected by setting multiple metering points in the power network using electric energy meters.
[0035] In some embodiments, the original data set is divided into a training set, a test set and a validation set. The validation set includes a variety of data samples, reflecting various situations that occur in actual applications. The original data set includes incentive electricity prices and historical response electricity consumption and historical load response feature vectors corresponding to the incentive electricity prices. The model is trained using the training set data, and the model parameters are adjusted. The input layer of the neural network model receives the original incentive electricity price data. The incentive electricity price data includes multiple incentive electricity price data for multiple time periods, indicating changes in electricity prices over a period of time. After the incentive electricity price data enters the first hidden layer, it undergoes a nonlinear transformation, that is, the input data is nonlinearly mapped through an activation function, thereby extracting the initial incentive electricity price data. The first step feature is the trend fluctuation of the incentive electricity price within the time period. In the second layer and subsequent hidden layers, the output results of the previous layer are further combined and transformed nonlinearly to obtain the relationship structure between the incentive electricity price and the load response feature vector, and generate high-level features. In the output layer, the neural network combines these high-level features to generate real-time load response feature vectors. These feature vectors include real-time response electricity consumption change rate, load rate fluctuation, response sensitivity and response delay. The output results are verified according to the verification set, and the internal parameters of the model are adjusted according to the verification results to achieve more accurate model prediction. The model is evaluated using an independent test set to verify the generalization ability on real-time data.
[0036] In some embodiments, the real-time load response characteristic vector includes the real-time response power consumption change rate, load rate fluctuation, response sensitivity and response delay. By acquiring and monitoring the real-time load response characteristic vector, the operating status of the flexible load under the incentive electricity price can be discovered in time, and potential supply and demand contradictions and risks can be accurately identified according to the real-time status and historical data, thereby enhancing the stability and reliability of the flexible load, ensuring the safety and continuity of power supply, and selecting the optimal incentive electricity price and the optimal real-time response power consumption while ensuring safe supply, thereby reducing costs and improving economic benefits.
[0037] Specifically, real-time response to power consumption change rate ( ):
[0038]
[0039] in, and The actual real-time response power consumption in the tth period and the t-1th period are respectively, and the real-time response power consumption monitored in the adjacent periods is subtracted and a ratio is generated to obtain the real-time response power consumption change rate; load rate fluctuation ( ): is the maximum possible load in period t, is the maximum possible load in the t-1 period, then
[0040]
[0041] The response delay (τ) needs to be obtained through experiments or long-term observations and can be adjusted according to actual conditions.
[0042] Response sensitivity ( ):
[0043]
[0044] in, , , represents the change in the response power consumption between the tth period and the t-1th period, represents the change in electricity price between the tth period and the t-1th period, refers to the incentive electricity price at time t, Refers to the incentive electricity price at time t-1.
[0045] S200: Obtain an electricity price level table for flexible loads, the electricity price level table including incentive electricity prices and corresponding time nodes, select an initial incentive electricity price for the first input into the characteristic identification model, select the remaining incentive electricity prices in turn according to the level rule based on the initial incentive electricity price for input, and obtain the corresponding real-time response electricity consumption and real-time load response characteristic vectors respectively, and mark the time nodes corresponding to the incentive electricity prices that meet the level rule as time nodes.
[0046] In some embodiments, Figure 2As shown, the electricity price level table includes all incentive electricity prices, and the incentive electricity prices are all historical electricity price data, dividing the electricity prices into levels such as yuan, jiao, fen, and cents. For example, ranging from 0.00 yuan to several yuan, each increase of 0.01 yuan or 0.1 yuan is regarded as an electricity price level; the incentive electricity prices have corresponding time nodes in the electricity price level table, and the time nodes are timed in 24 hours, that is, recorded from 0 o'clock to 24 o'clock, indicating the specific time of the incentive electricity price in a day; when the incentive electricity price is specifically selected for input, an incentive electricity price is randomly selected as the initial incentive electricity price, and the initial incentive electricity price is used as a benchmark to traverse the electricity price level table to query whether the gap with the initial incentive electricity price meets the level rule, and the incentive electricity price that meets the level rule is input into the characteristic identification model to generate the corresponding real-time response power consumption and real-time load response feature vector.
[0047] In some embodiments, whether the system safety constraint is satisfied is determined based on the real-time response power consumption. The system safety constraint can be read from various parameters and performance requirements of the initial configuration of the aggregate optimization; the initial configuration generally includes checking the communication network status, importing the historical database, importing the historical experience model, and reading various parameters and performance requirements of the aggregate optimization. Based on the obtained real-time response power consumption, combined with the expression of the system safety constraint, the satisfaction of the system safety constraint is calculated and determined. For example, a common system safety constraint is a system capacity limit constraint. If the sum of all response power consumptions obtained based on the real-time response power consumption exceeds the given capacity limit value, the system safety constraint is not satisfied; otherwise, the system safety constraint is satisfied. If the system safety constraint has multiple constraints, such as capacity limit, voltage stability limit, frequency stability limit, etc., it is necessary to determine whether all system safety constraints are satisfied. If all system safety constraints are satisfied, it means that the flexible load state at this time will not cause system operation risks; for the incentive electricity price and real-time response power consumption that do not meet the safety constraints, it is necessary to continue to run the iterative optimization calculation to adjust the incentive electricity price, thereby changing the real-time response power consumption of the flexible load, and finally satisfying the safety constraint.
[0048] The magnitude rule is based on the initial incentive electricity price, selects an incentive electricity price of a different magnitude from the initial incentive electricity price or not less than the measurement threshold as the incentive electricity price for subsequent input; calculates the change range between incentive electricity prices of the same magnitude, and compares the change range with the measurement threshold. The measurement threshold is set according to actual conditions and experience, and is used to measure the change range of the incentive electricity price.
[0049] In some embodiments, the magnitude rule uses different magnitudes and measurement thresholds as division criteria and screening rules. For example, the initial incentive electricity price is 0.5 yuan, and the initial incentive electricity price corresponds to "0, 5, 0, 0" in the sequence of "yuan, jiao, fen, li" in the electricity price magnitude table. When traversing the electricity price magnitude table subsequently, first traverse the magnitude corresponding to 0 from left to right. When traversing to the sequence of "yuan, fen, li", the incentive electricity price corresponding to non-0 is the incentive electricity price of different magnitudes; the measurement threshold is experimentally measured, and the setting of the measurement threshold for different magnitudes is different. The larger the magnitude, the smaller the measurement threshold. In actual use, it is adjusted according to the sensitivity of different enterprises to electricity prices. Preferably, for "yuan, fen, li", the measurement threshold is adjusted according to the sensitivity of different enterprises to electricity prices. The measurement thresholds of "angle, minute, centimeter" are set to "1, 3, 5, 7" respectively, and the time nodes corresponding to the subsequent incentive electricity prices that meet the magnitude rules are marked as time nodes; when comparing the incentive electricity prices, there are two different standards. If the subsequent incentive electricity price and the initial incentive electricity price belong to different magnitudes, then it is used as the subsequent input incentive electricity price; if the subsequent incentive electricity price and the initial incentive electricity price belong to the same magnitude, then the change range is determined. The change range is the ratio of the absolute value of the difference between the subsequent incentive electricity price and the initial incentive electricity price. If the change range is not less than the measurement threshold, it is used as the subsequent input incentive electricity price; if it is less than the measurement threshold, it is not used as the subsequent input incentive electricity price.
[0050] S300: Obtain the output real-time response power consumption change rate, pre-set a power amplitude threshold according to the historical real-time response power consumption change rate, compare the real-time response power consumption change rate with the power amplitude threshold, and mark the time node corresponding to the real-time response power consumption that is not less than the power amplitude threshold as an electric degree node. The power amplitude threshold is set according to actual conditions and historical data, and is used to measure the change amplitude of the real-time response power consumption change rate.
[0051] In some embodiments, the real-time response electricity consumption generated according to the initial incentive electricity price is used as a benchmark, and compared with the subsequently generated real-time response electricity consumption to obtain the change amplitude. The amplitude by which the subsequent real-time response electricity consumption increases or decreases compared to the previous one is the change amplitude. The electricity amplitude threshold needs to be pre-set based on experimental measurements and historical data. Preferably, setting the electricity amplitude threshold to 10% can best reflect the trend of the change amplitude of the real-time response electricity consumption, and the time node where the real-time response electricity consumption is not less than the electricity amplitude threshold is marked as an electricity node.
[0052] S400: Generate a time curve according to different incentive electricity prices and time nodes, generate an electricity curve according to the real-time response power consumption and the electricity node, determine the span node, obtain the real-time load response characteristic vector of the span node, and generate an optimization control report; the span node is a node where the time node and the electricity node overlap.
[0053] In some embodiments, Figure 3 As shown, the horizontal axis is the node of each time period. The time degree curve and the electricity degree curve of the same period are superimposed in the time period analysis curve, and the span node is determined according to the intersection of the time degree curve and the electricity degree curve. The change of the load response characteristic vector on the span node is analyzed in detail, including the change amount and change rate of each characteristic and the comparison results with the adjacent nodes. The correlation analysis results between the change of the incentive electricity price and the change of the load response characteristic vector are displayed, including the correlation coefficient, regression equation, etc. The typical load response mode identified on the span node is described, including the mode characteristics, the load type involved, the existing impact, etc., and the corresponding optimization control report is generated. The optimization control report can be presented in a variety of forms.
[0054] S500: Aggregate and optimize the flexible loads according to the optimization and control report and update the electricity price level table.
[0055] In some embodiments, by analyzing the span nodes and real-time load response characteristic vectors, it is possible to clearly determine the electricity price corresponding to the time node and the changing trend of the real-time response electricity consumption, and then aggregate and optimize the flexible load according to the changing trend to achieve more precise regulation and control, thereby achieving maximum economic benefits while ensuring load safety.
[0056] In this embodiment, by using the electricity price level table to input the characteristic identification model, the historical electricity price data is divided by level and associated with specific time nodes, and incentive electricity prices of multiple levels are used to achieve more accurate incentive electricity price control optimization; by comparing the real-time response electricity consumption under different incentive electricity prices, the electricity nodes with significant electricity changes are identified; combined with the time node in the electricity price level table, the span node (that is, the point where the electricity node and the time node overlap) is further determined, and the time curve and electricity curve are generated based on these nodes, and the real-time load response characteristic vector is deeply analyzed; based on the analysis results of the real-time load response characteristic vector of the span node, an optimization control report is generated, and targeted optimization control strategies are proposed; according to the electricity price level table and the real-time system status, the incentive electricity price strategy is dynamically adjusted to ensure that the economic benefits are maximized while meeting the system safety constraints.
[0057] The technical solutions in the above embodiments of the present application have at least the following technical effects or advantages:
[0058] This application uses the electricity price level table as input into the characteristic identification model, divides the historical electricity price data by level, and associates it with specific time nodes, and uses incentive electricity prices of various levels to achieve more accurate incentive electricity price control optimization effects; by combining electricity nodes with time nodes, the span nodes are determined to deeply analyze the real-time load response characteristic vectors, generate optimization and control reports, and propose targeted optimization and control strategies, thereby achieving the effect of maximizing economic benefits.
[0059] Embodiment 2: In embodiment 1, an electricity price level table is set and the time node and electricity node are obtained, and the span node is obtained for analysis to obtain an optimized control report for precise control. This embodiment makes further improvements on the basis of the above embodiment.
[0060] Obtain two adjacent span nodes, and divide the time area between the two span nodes into multiple auxiliary control areas according to the incentive electricity price, and the areas with the same incentive electricity price are the same auxiliary control area; in each of the auxiliary control areas, further divide it into multiple segment areas according to the difference in incentive electricity prices; each segment area corresponds to a newly divided incentive electricity price, and the corresponding real-time response power consumption and the real-time load response characteristic vector are obtained according to the new incentive electricity price; according to the real-time response power consumption and the real-time load response characteristic vector, the optimal incentive electricity price corresponding to the segment area and the corresponding time range are analyzed to generate an optimization control report.
[0061] The difference is the difference in incentive electricity prices between different auxiliary control areas, which is calculated by recording the ending electricity price and the starting electricity price of the auxiliary control area. The auxiliary control areas with a difference greater than 0, that is, the ending electricity price is greater than the starting electricity price, are divided; and the auxiliary control areas with a difference less than 0, that is, the ending electricity price is less than the starting electricity price, are set as pending areas.
[0062] In some embodiments, according to two adjacent span nodes, the time area between the two span nodes is divided into multiple auxiliary control areas according to the incentive electricity price, the same incentive electricity price is the same auxiliary control area, and the number of auxiliary control areas is consistent with the number of different incentive electricity prices; the auxiliary control area is further divided into multiple segment areas according to the difference in incentive electricity prices, and the difference is the difference in incentive electricity prices between different auxiliary control areas. The end electricity price and the start electricity price of the auxiliary control area are used to perform a difference calculation, and the auxiliary control areas with a difference greater than 0, that is, the end electricity price is greater than the start electricity price, are divided; for the auxiliary control areas with a difference less than 0, that is, the end electricity price is less than the start electricity price, no further division is performed. Preferably, the auxiliary control areas that have not been divided are set as pending areas, and the pending areas are re-divided after waiting for the remaining auxiliary control areas to be processed.
[0063] The segment areas are divided according to the next order of magnitude of the difference, and the maximum number of the segment areas does not exceed 10. After the division is completed, a new incentive electricity price is set for each segment area. The time length of the auxiliary control area is divided according to the number of the segment areas, and the initial time node and the end time node of each segment area are recorded.
[0064] In some embodiments, the auxiliary control area is divided according to the next magnitude of the difference, the maximum number of the segment areas does not exceed 10, the incentive electricity price corresponding to the first segment area is the starting electricity price of the auxiliary control area, and the incentive electricity price corresponding to the last segment area is the ending electricity price of the auxiliary control area. For each segment area, starting from the starting electricity price, the incentive electricity price is gradually increased according to the set growth rate. For example, the starting electricity price of the auxiliary control area 1 is 0.5, the ending electricity price is 0.6, and the magnitude is "jiao". When dividing the auxiliary control area, the next magnitude, that is, "fen", is used. The auxiliary control area 1 is divided according to the new magnitude into ten segment areas, which are 0.5, 0.51, 0. 52, ..., 0.6. After the division is completed, a new incentive electricity price is set for each segment area and the time length of the auxiliary control area is divided according to the number of the segment areas. For example, the time length of the auxiliary control area is one hour, and the auxiliary control area is divided into 10 segment areas. Then the time range of each segment area is 6 minutes. The initial time node and the end time node of the segment area are generated according to the set time range. According to the new incentive electricity price, the corresponding real-time response power consumption and real-time load response characteristic vector are generated. According to the real-time response power consumption and the real-time load response characteristic vector, the optimal incentive electricity price corresponding to the segment area and the corresponding time range are obtained.
[0065] In some embodiments, according to the real-time response power consumption and real-time load response characteristic vector obtained for each segment area, according to the changes in incentive power prices and real-time response power consumption, the load response situations under different incentive power prices are recorded, the differences in load responses between adjacent segment areas are compared, the impact of the growth rate on the load response is evaluated, and compared with the real-time response power consumption between span nodes, the load response law under different incentive power price growth rates is analyzed, and based on the test results, an aggregated optimization control strategy for flexible loads is proposed, the style response characteristics of different auxiliary control areas and segment areas are analyzed, and differentiated control suggestions are proposed. The control suggestions should include specific control strategies, expected effects and risk assessments, etc. According to the optimized control suggestions, the incentive power price settings in the power price scale table are adjusted, and the effective incentive power price growth rate found in the test is included in the power price scale table to optimize the load response effect.
[0066] In some embodiments, the optimal incentive electricity price for the segment area and the time range are obtained to set an optimization control strategy, and different optimal incentive electricity prices are used in different time ranges to obtain the optimal real-time response electricity consumption. Under the premise of ensuring load safety, precise control of the incentive electricity price is achieved to obtain the maximum economic benefit.
[0067] In this embodiment, a method for dividing auxiliary control areas and segment areas is introduced. The time area between two adjacent span nodes is divided into multiple auxiliary control areas according to the incentive electricity price, and each auxiliary control area is further subdivided into segment areas according to the electricity price difference, thereby achieving more refined regulation of the time area. In each segment area, a dynamic electricity price test is implemented by setting a new incentive electricity price and obtaining the corresponding real-time response electricity consumption and real-time load response characteristic vector, observing the load response at different electricity price levels, and finding the optimal incentive electricity price in the segment area.
[0068] The technical solutions in the above embodiments of the present application have at least the following technical effects or advantages:
[0069] This application achieves a more refined control effect on flexible loads by dividing the auxiliary control area and the segment area, and analyzing the load response under different incentive electricity prices in each segment area; the optimal incentive electricity price and corresponding time range are obtained based on data analysis, which can more effectively guide the electricity consumption behavior of flexible loads, optimize the allocation and utilization efficiency of power resources, and improve the overall economic benefits.
[0070] Embodiment three: In the above embodiment, the time length between two adjacent span nodes is divided into multiple auxiliary control areas and segment areas, but the time range of each segment area is evenly divided, which affects the precise control. This embodiment makes further limitations on the basis of the above embodiment.
[0071] Obtain the initial time node of the segment area, gradually increase the time range according to the extension step, obtain the real-time response power consumption change rate and the real-time load response characteristic vector according to the current time range and the new incentive electricity price, and judge whether the preset conditions are met. If the preset conditions are met or the maximum extension times are reached, stop the extension and record the time node at this time as the end time node.
[0072] The maximum extension times needs to be preset according to the time length and extension step of the auxiliary control area.
[0073] The preset conditions include a first threshold and a second threshold. The first threshold is set according to the real-time response power consumption change rate; the second threshold is set according to the comprehensive fluctuation value, and the comprehensive fluctuation value is generated comprehensively based on the load rate fluctuation, the response sensitivity and the response delay. If the real-time response power consumption change rate reaches the first threshold or the comprehensive fluctuation value reaches the second threshold, the extension is stopped.
[0074] In some embodiments, the extension step is pre-set, and the extension step is the length of time added each time the extension is performed, such as three minutes, five minutes, etc.; the setting of the extension step needs to be set according to the time length of the auxiliary control area, and according to the current electricity price level information of the auxiliary control area, different extension steps are set according to different levels. The larger the level, the smaller the extension step. It is adjusted according to the actual situation, and the maximum number of extensions is set according to the extension step and the time length of the auxiliary control area, which should not exceed the time length of the auxiliary control area; each time a new step is added to obtain a new time range, the real-time response power consumption and the real-time load response characteristic vector are predicted based on the new incentive electricity price using the characteristic identification model, and it is determined whether the preset conditions are met. If not, the extension continues to expand the time range. If the preset conditions are met or the maximum number of extensions is reached, the extension is stopped, and the current time node is determined as the end time node. The time range is generated from the initial time node to the end time node.
[0075] In some embodiments, Figure 4 As shown, the horizontal axis is the time period node. The first threshold and the second threshold are set between the span nodes, that is, between T1 and T2. The actual curve represents the specific changes in the extension process. When the first threshold is reached at t1, the extension stops. The T1-t1 time period is the time range of a segment area. The incentive electricity price corresponding to t1 is the optimal incentive electricity price for the time range. t1-t2 and t2-T2 are the time ranges of the second and third segment areas generated. Figure 4It is only a schematic diagram; the preset condition sets a first threshold and a second threshold. The first threshold is set according to the real-time response power consumption change rate. The real-time response power consumption change rate can quickly capture the real-time changes of the load, which is convenient for horizontal and vertical comparisons between different time periods and different load types. When the real-time response power consumption change rate reaches the first threshold, it means that there is an obvious change trend in the time range at this time; the second threshold is set according to the real-time load response characteristic vector. There are multiple key indicators in the real-time load response characteristic vector. The real-time response power consumption change rate is used as the first threshold, and the comprehensive fluctuation value is set as the second threshold. The comprehensive fluctuation value is generated based on the load rate fluctuation, the response sensitivity and the response delay. It measures the safety of the power load and is used to ensure the stability of the load when the step length is extended. When setting the second threshold, it is necessary to analyze the historical data information to obtain a reasonable change range for each key indicator. For load rate fluctuations, a reasonable load rate fluctuation range needs to be set. For example, a maximum allowable fluctuation percentage is set, such as 5%. If the real-time load rate fluctuation is found to exceed the maximum allowable fluctuation percentage during real-time monitoring, it is judged that the load stability is threatened at this time; response sensitivity and response delay are two relative key indicators. The sensitivity of different load types to changes in electricity prices is analyzed, and an upper limit of sensitivity is set to avoid drastic changes in load due to excessive sensitivity; response delay cannot directly reflect stability, but too high or too short response delay causes the system to be unable to accurately identify changes in the current incentive electricity price. The minimum delay time is set to ensure that changes in the incentive electricity price can be accurately identified; the second threshold is set in combination with the three indicators of load rate fluctuation, response sensitivity and response delay, specifically the three indicator data of load rate fluctuation range, sensitivity upper limit and minimum delay time. The second threshold is obtained by normalizing the three indicator data and dividing them into different weight values for weighted summation; according to the initial time node of the segment area, the time range is automatically divided according to the set extension step and preset conditions.
[0076] The technical solutions in the above embodiments of the present application have at least the following technical effects or advantages:
[0077] The present application avoids the inaccuracy that may be caused by the equal division of the time range through dynamic extension, making the division of the time range more flexible and better able to adapt to changes in load response characteristics; by monitoring the real-time response power consumption change rate and the real-time load response characteristic vector, and stopping the extension when the preset conditions are met, it ensures that each segment area can accurately reflect the response characteristics of the load within the time period, thereby achieving precise optimization and regulation.
[0078] Embodiment 4: In the above embodiment, the time range of the segment area is automatically obtained according to the preset conditions and the set extension step size to achieve more accurate optimization and control. This embodiment makes further limitations on the basis of the above embodiment.
[0079] Obtain the time range of the segment area and the incentive electricity price, set a preliminary benchmark value and an initial floating range for each time range, take each time period as the basic unit, construct an optimization model with the electricity price floating range and magnitude as decision variables, record the optimal electricity price floating range and the best magnitude selection for each time period, and update the electricity price magnitude table.
[0080] The preliminary benchmark value is the optimal incentive electricity price in the segment area, and the initial floating range is the range between the incentive electricity price corresponding to the initial time node of the segment area and the incentive electricity price corresponding to the end time node; the optimization model evaluates the comprehensive benefits of the strategy based on the prediction results and optimization objectives, narrows the electricity price floating range, and optimizes the electricity price floating range and magnitude for each time period based on the prediction results of the characteristic identification model.
[0081] In some embodiments, the maintenance length of the segment area and the optimal incentive electricity price can be accurately obtained according to the time range of the segment area and the incentive electricity price. Each time range includes multiple time periods. A preliminary benchmark value and an initial floating range are set within the time range. The initial floating range refers to the floating range of the incentive electricity price, such as Figure 4 As shown, the T1-t1 time period is the time range of a segment area, the incentive electricity price corresponding to t1 is the optimal incentive electricity price in this time range, and the incentive electricity price corresponding to T1 and the incentive electricity price corresponding to t1 are the initial floating range; taking each time period as the basic unit, an optimization model with the electricity price floating range and magnitude as decision variables is constructed, and the optimization model adopts heuristic algorithms such as genetic algorithm and particle swarm optimization algorithm. According to the prediction results of the characteristic identification model, the electricity price floating range and magnitude of each time period are optimized. In each time period, a series of candidate electricity price strategies are generated according to the initial electricity price benchmark value and floating range, and the characteristic identification model is used to predict different The load response under the candidate electricity price strategy, based on the prediction results and optimization objectives, evaluates the comprehensive benefits of each candidate strategy, especially the economic benefits, and gradually narrows the electricity price fluctuation range through iterative optimization until the optimal solution is found. The optimal electricity price fluctuation range and the best level selection for each time period are recorded, and the electricity price level table is updated; the optimal electricity price fluctuation range and the best level selection for each time period are integrated, and a new electricity price level table is compiled based on the integrated optimization results to clarify the electricity price fluctuation range and the best level for each time period. The new electricity price level table is checked for safety constraints to ensure that it meets system safety standards; the new electricity price strategy is implemented, and the incentive electricity price for each time period is adjusted according to the new electricity price level table.
[0082] In this embodiment, based on the dynamic time range division method in the previous embodiment, the time range of the segment area and the corresponding incentive electricity price are obtained, and a preliminary benchmark value (that is, the optimal incentive electricity price for the area) and an initial floating range (determined by the incentive electricity prices at the start and end time nodes of the area) are set for the time range of each segment area. Taking each time period as the basic unit, an optimization model with the electricity price floating range and magnitude as decision variables is constructed, the optimal electricity price floating range and the optimal magnitude selection of each time period are integrated, the electricity price magnitude table is updated, and the electricity price floating range and the optimal magnitude of each time period are clarified.
[0083] The technical solutions in the above embodiments of the present application have at least the following technical effects or advantages:
[0084] This application achieves precise electricity price regulation by setting the optimal electricity price floating range and optimal level for each time period, so that the electricity price strategy can be adjusted in real time according to load characteristics and system requirements, enhancing the flexibility of the strategy, finding the optimal electricity price strategy for each time period, and achieving the effect of improving economic benefits.
[0085] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for identifying and optimizing flexible load aggregation characteristics, characterized in that: The method comprises: S100: Acquire a characteristic identification model for flexible loads, wherein the input of the characteristic identification model is the incentive electricity price, and the output of the characteristic identification model is the real-time response power consumption and the real-time load response characteristic vector; the real-time load response characteristic vector includes the real-time response power consumption change rate, load rate fluctuation, response sensitivity and response delay; S200: Obtain an electricity price magnitude table for flexible loads, the electricity price magnitude table including incentive electricity prices and corresponding time nodes, select an initial incentive electricity price for first input into a characteristic identification model, select the remaining incentive electricity prices in turn according to the magnitude rule based on the initial incentive electricity price for input, obtain corresponding real-time response power consumption and real-time load response feature vectors respectively, and mark the time node corresponding to the incentive electricity price that meets the magnitude rule as a time node; S300: Acquire the output real-time response power consumption change rate, pre-set a power amplitude threshold according to the historical real-time response power consumption change rate, compare the real-time response power consumption change rate with the power amplitude threshold, and mark the time node corresponding to the real-time response power consumption that is not less than the power amplitude threshold as a power node; S400: generating a time-degree curve according to different incentive electricity prices and time-degree nodes, generating an electric degree curve according to the real-time response power consumption and the electric degree node, determining a span node, obtaining the real-time load response characteristic vector of the span node, and generating an optimization control report; the span node is a node where a time-degree node and an electric degree node overlap; S500: performing aggregation optimization on the flexible loads according to the optimization control report and updating the electricity price level table; Among them, two adjacent span nodes are obtained, and the time area between the two span nodes is divided into multiple auxiliary control areas according to the incentive electricity price, and the areas with the same incentive electricity price are the same auxiliary control area; in each of the auxiliary control areas, it is further divided into multiple segment areas according to the difference in incentive electricity prices; each segment area corresponds to a newly divided incentive electricity price, and the corresponding real-time response power consumption and the real-time load response characteristic vector are obtained according to the new incentive electricity price; according to the real-time response power consumption and the real-time load response characteristic vector, the optimal incentive electricity price corresponding to the segment area and the corresponding time range are analyzed to generate an optimization control report; The magnitude rule is to use the initial incentive electricity price as a benchmark, select an incentive electricity price of a different magnitude from the initial incentive electricity price or not less than the measurement threshold as the incentive electricity price for subsequent input; calculate the change range between incentive electricity prices of the same magnitude, and compare the change range with the measurement threshold. The change range is the ratio of the absolute value of the difference between the subsequent incentive electricity price and the initial incentive electricity price. If the change range is not less than the measurement threshold, it is used as the incentive electricity price for subsequent input; if it is less than the measurement threshold, it is not used as the incentive electricity price for subsequent input; The difference is the difference in incentive electricity prices between different auxiliary control areas, which is calculated by recording the ending electricity price and the starting electricity price of the auxiliary control area. The auxiliary control areas with a difference greater than 0, that is, the ending electricity price is greater than the starting electricity price, are divided; the auxiliary control areas with a difference less than 0, that is, the ending electricity price is less than the starting electricity price, are set as pending areas; The segment areas are divided according to the next order of magnitude of the difference, and the maximum number of the segment areas does not exceed 10. After the division is completed, a new incentive electricity price is set for each segment area. The time length of the auxiliary control area is divided according to the number of the segment areas, and the initial time node and the end time node of each segment area are recorded.
2. A method for identifying and optimizing flexible load aggregation characteristics as claimed in claim 1, characterized in that: Obtain the initial time node of the segment area, gradually increase the time range according to the extension step, obtain the real-time response power consumption change rate and the real-time load response characteristic vector according to the current time range and the new incentive electricity price, and judge whether the preset conditions are met. If the preset conditions are met or the maximum extension times are reached, stop the extension and record the time node at this time as the end time node.
3. A method for identifying and optimizing flexible load aggregation characteristics as claimed in claim 2, characterized in that: The maximum extension times needs to be preset according to the time length and extension step of the auxiliary control area.
4. A method for identifying and optimizing flexible load aggregation characteristics as claimed in claim 2, characterized in that: The preset conditions include a first threshold and a second threshold. The first threshold is set according to the real-time response power consumption change rate; the second threshold is set according to the comprehensive fluctuation value, and the comprehensive fluctuation value is generated comprehensively based on the load rate fluctuation, the response sensitivity and the response delay. If the real-time response power consumption change rate reaches the first threshold or the comprehensive fluctuation value reaches the second threshold, the extension is stopped.
5. A method for identifying and optimizing flexible load aggregation characteristics as claimed in claim 1, characterized in that: Obtain the time range of the segment area and the incentive electricity price, set a preliminary benchmark value and an initial floating range for each time range, take each time period as the basic unit, construct an optimization model with the electricity price floating range and magnitude as decision variables, record the optimal electricity price floating range and the best magnitude selection for each time period, and update the electricity price magnitude table.
6. A method for identifying and optimizing flexible load aggregation characteristics as claimed in claim 5, characterized in that: The preliminary benchmark value is the optimal incentive electricity price in the segment area, and the initial floating range is the range between the incentive electricity price corresponding to the initial time node of the segment area and the incentive electricity price corresponding to the end time node; the optimization model evaluates the comprehensive benefits of the strategy based on the prediction results and optimization objectives, narrows the electricity price floating range, and optimizes the electricity price floating range and magnitude for each time period in combination with the prediction results of the characteristic identification model and the elasticity estimation model.
