Method, device and storage medium for estimating demand response baseline load

Abstract

The application discloses a demand response baseline load estimation method, equipment and storage medium, which comprises the following steps: dividing the load in the demand response into demand response users and non-demand response users, clustering a large number of non-demand response users, and mapping the load curve of the demand response users in the non-demand response period with the clustering curve of the non-demand response users, and then estimating the baseline load of the demand response users in the demand response period by the error of the values before and after the mapping and weighted estimation. Meanwhile, the demand response baseline estimation method using the traditional average method is also included, and the final load baseline is obtained according to the error weighting. The application has a wider application range and can be used in various load fluctuation and complex occasions such as residential load, commercial load and industrial load, and can significantly improve the estimation accuracy of the load baseline. The method considers more factors close to the actual operation, and the calculation result can significantly improve the technical and economic benefits of planning.

Description

Technical Field

[0001] This invention relates to the field of demand response technology for power system loads, and more specifically to a method for estimating baseline load for demand response. Background Technology

[0002] In new power systems dominated by new energy sources, the generation-side regulation capacity has significantly decreased, and there is an urgent need to fully tap the demand-side regulation capacity to compensate for the lack of flexibility caused by the grid connection of new energy sources.

[0003] Demand response treats demand-side resources as alternatives to supply-side electricity, reducing user load and thus improving the utilization rate of social resources. Demand response implementers incentivize users to change their electricity consumption behavior (load reduction or increase) during a specific period (the demand response window) by paying participants compensation. The participation compensation equals the product of the load response amount and its compensation unit price. The load response amount is defined as the absolute value of the difference between the load a user would have consumed without participating in demand response and the load actually consumed after participating in demand response; the latter represents the user's actual load during the demand response period, while the former is the user's baseline load. Once a user implements incentive-based demand response, their baseline load cannot be obtained through measurement in reality and must be estimated.

[0004] Accurate baseline load estimation is crucial for demand response implementation, as it directly impacts the economic interests of both implementers and participants. However, existing baseline load estimation methods, such as averaging, regression, and control group methods, perform well when demand response day load characteristics closely resemble typical daily load patterns, but their accuracy declines when these characteristics are not sufficiently similar. This deficiency is particularly pronounced for residential users with complex and variable load patterns. Furthermore, due to their heavy reliance on historical data, existing estimation methods suffer a sharp drop in accuracy or even complete failure when historical data is unavailable, rendering them unsuitable for newly added demand response projects with limited historical data. Summary of the Invention

[0005] To address the various drawbacks of baseline load estimation in demand response, this invention aims to provide a comprehensive baseline load estimation method based on multi-source data, which analyzes both historical and current data to improve the accuracy of baseline load estimation.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for estimating baseline load in demand response includes the following steps:

[0008] Step 1: Based on the electricity load data for the specified number of days prior to the demand response date, estimate the user baseline load for the demand response date using the averaging method;

[0009] Step 2: Based on the electricity load data of all users during the demand response day, estimate the user baseline load during the demand response day using the segmented control group method;

[0010] Step 3: During the non-demand response period within the demand response day, calculate the errors of the average method proposed in Step 1 and the segmented control group method in Step 2 respectively. Based on the errors of the two methods during the non-demand response period, fit the user baseline load during the demand response period within the demand response day.

[0011] Further, step 1 specifically includes,

[0012] Select historical electricity load data of users for Y days prior to the estimated demand response date, and remove unrepresentative days, namely historical demand response days and holidays, until Y days are added.

[0013] Then, select the X days with the highest daily load from these Y days, and calculate the average load at each time point on these X days as the user baseline load for the demand response day:

[0014]

[0015] In the formula, This represents the baseline user load within the demand response day estimated using the averaging method, and High(X,Y) represents the set of X days with the highest load levels among the Y days that meet the conditions before the demand response day. This represents the load value of the t-th time period on the d-th day in the High(X,Y) set.

[0016] Furthermore, step 2 specifically includes,

[0017] Step 2.1 On the demand response day, users are divided into demand response users and non-demand response users according to whether they participate in demand response, and the set of non-demand response users is used as the control group.

[0018] Step 2.2 Assuming multiple demand responses occur within a demand response day, then there are n non-demand response periods within that demand response day. And there are in: and These are the i-th non-demand response periods. The start and end times;

[0019] Step 2.3 adopts Figure 3 The K-Means algorithm clusters all non-demand response users within the control group on the demand response day, forming K load clustering curves.

[0020] Step 2.4 Obtain the demand response user's activity during n non-demand response periods within the demand response day. The load curve segment, with Indicate; will The non-demand response time periods corresponding to the K clustered load curves are respectively load curve segment Mapping is performed, that is, during a non-demand response time segment. Within, find the j-th load clustering curve that minimizes the error between the two. Specifically, as shown in formula (2):

[0021]

[0022] Simultaneously, load curve segments are calculated over n non-demand response periods. With the corresponding j-th load clustering curve The error between them is shown in the following formula:

[0023]

[0024] Step 2.5 set Map the load curve segment within the i-th non-demand response period to the j-th load cluster curve, and then use the segmentation error α obtained in step 2.4. i Clustering curves that map load curve segments of demand response users during n non-demand response periods within a demand response day. Weighted processing is performed on i = 1, 2, ..., n; j = 1, 2, ..., K to obtain the user baseline load within the demand response day under the segmented control group method:

[0025]

[0026] Furthermore, step 3 is further detailed as follows:

[0027] Step 3.1 Calculate the user baseline load obtained by the averaging method described in Step 1 within the demand response date. Error compared to actual data over n non-demand response periods:

[0028]

[0029] in: and Let represent the user baseline load estimated by the averaging method and the user actual load respectively during the i-th non-demand response period. Within the error;

[0030] Step 3.2 Calculate the user baseline load obtained during the demand response date using the segmented control group method described in Step 2. Error compared to actual data over n non-demand response periods:

[0031]

[0032] in: and Represent the baseline load and actual load of users estimated using the segmented control group method during the i-th non-demand response period, respectively. Error within;

[0033] Step 3.3 Based on the error magnitudes obtained in Steps 3.1 and 3.2, a weighted combination forecast is performed on the user baseline load to obtain the user baseline load P. t The details are as follows:

[0034]

[0035] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0036] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0037] As described above, this invention proposes a user baseline load estimation method considering a segmented control group approach. It divides the load within the demand response period into demand response users and non-demand response users, clusters a large number of non-demand response users, and maps the load curves of demand response users during non-demand response periods to the cluster curves of non-demand response users. Then, based on the error between the values ​​before and after mapping, a weighted estimate of the baseline load of demand response users during the demand response period is obtained. Simultaneously, to improve the accuracy of baseline load estimation, the traditional averaging method for demand response baseline estimation is also incorporated, and the final load baseline is obtained by weighting according to its error. Compared with existing load baseline estimation methods, the method proposed in this invention has a wider range of applications and can be used in various situations with large and complex load fluctuations, such as residential load, commercial load, and industrial load. It can significantly improve the accuracy of load baseline estimation. The factors considered in the method are closer to actual operation, and the calculation results can significantly improve the technical and economic benefits of planning.

[0038] This invention addresses the inherent "asynchronous matching" error mechanism in existing baseline load estimation methods, specifically the significant estimation errors caused by the temporal asynchrony between input data (historical load data) and output data (baseline load). The proposed method for estimating demand response baseline load uses a segmented control group method based on the synchronous mode matching principle to estimate the user baseline load within the demand response day, based on the electricity load data of all users. This solves the asynchronous matching problem of existing estimation methods, effectively improving estimation accuracy, and the estimation process does not require any historical data. Furthermore, it optimally combines the average method based on historical day data and asynchronous matching with the segmented control group method based on demand day data and synchronous matching, proposing a user baseline estimation method based on optimal weight combination. This fully integrates the advantages of both estimation methods, greatly improving estimation accuracy.

[0039] Compared with existing technologies, the improvements and beneficial effects of this invention are as follows:

[0040] (1) This invention improves and extends the traditional control group method for estimating user baseline load within the demand response day, fully considering multiple demand responses within the demand response day, and using load information before and after the DR period to synchronously match DR users and the control group. Moreover, this estimation method does not rely on any historical data and is still applicable to users who have newly joined demand responses and lack sufficient historical data.

[0041] (2) This invention combines the average value method based on historical daily data and asynchronous matching with the segmented control group method based on demand daily data and synchronous matching to propose a user baseline estimation method based on the optimal weight combination. This fully integrates the advantages of the two estimation methods, effectively prevents the impact of excessively long duration of multiple demand segments and the rebound effect of demand response on the performance of the proposed method, and greatly improves the estimation accuracy. Attached Figure Description

[0042] Figure 1 This is a flowchart of a method according to an embodiment of the present invention;

[0043] Figure 2 This is a diagram illustrating the demand response period within a demand response day;

[0044] Figure 3 The calculation process of the K-Means algorithm. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0046] Demand response treats demand-side resources as substitutes for supply-side electricity, altering electricity consumption behavior through price signals and incentive mechanisms. This allows it to fully leverage its role in improving power system supply-demand balance, peak shaving and valley filling, and promoting renewable energy integration. A key factor in evaluating the effectiveness of demand response implementation is determining the user baseline load, as it directly impacts the economic interests of both implementers and participants. However, once a user implements demand response, their baseline load cannot be measured in reality and must be estimated. Existing baseline load estimation methods, such as averaging, regression, and control group methods, perform well when the demand response day load characteristics are similar to typical daily load patterns, but their accuracy declines when these characteristics are not sufficiently similar. This deficiency is even more pronounced for residential users with complex and variable load patterns. Furthermore, due to their heavy reliance on historical data, existing estimation methods suffer a sharp drop in accuracy or even complete failure when historical data is lacking, making them unsuitable for users newly added to demand response projects who lack historical data records.

[0047] To address this, this invention proposes a user baseline load estimation method considering a segmented control group approach. This method divides the load within the demand response period into demand response users and non-demand response users. A large number of non-demand response users are clustered, and the load curves of demand response users during non-demand response periods are mapped to the cluster curves of non-demand response users. Then, the baseline load of demand response users during the demand response period is weighted and estimated based on the error between the values ​​before and after mapping. Simultaneously, to improve the accuracy of baseline load estimation, the traditional averaging method for demand response baseline estimation is also incorporated, and the final load baseline is obtained by weighting according to its error. Compared with existing load baseline estimation methods, the method proposed in this invention has a wider range of applications and can be used in various situations with large and complex load fluctuations, such as residential load, commercial load, and industrial load. It can significantly improve the accuracy of load baseline estimation. The factors considered in the method are closer to actual operation, and the calculation results can significantly improve the technical and economic benefits of planning.

[0048] like Figure 1 As shown, the method for estimating the baseline load of demand response described in this embodiment includes the following steps:

[0049] Step 1: Based on the electricity load data for a certain number of days prior to the demand response date, estimate the user baseline load for the demand response date using the averaging method. The specific method is as follows:

[0050] Select historical electricity load data for users from Y days prior to the demand response date to be estimated, and remove unrepresentative days (historical demand response dates and holidays) until Y days are reached. Then, select X days with the highest daily load levels from these Y days, and calculate the average load at each time point on these X days as the baseline load for users within the demand response date.

[0051]

[0052] In the formula, This represents the baseline user load within the demand response day estimated using the averaging method, and High(X,Y) represents the set of X days with the highest load levels among the Y days that meet the conditions before the demand response day. This represents the load value of the t-th time period on the d-th day in the High(X,Y) set.

[0053] Step 2: Based on the electricity load data of all users during the demand response day, estimate the user baseline load during the demand response day using the segmented control group method;

[0054] Step 2.1 On the demand response day, users are divided into demand response users and non-demand response users according to whether they participate in demand response, and the full-day load curves of k non-demand response users are obtained. Furthermore, the set of all non-demand response users was used as a control group.

[0055] Step 2.2 Assuming multiple demand responses occur within a demand response day, then there are n non-demand response periods within that demand response day. Specifically, such as Figure 2 As shown in the figure. It can be seen from the figure that... in: and These are the i-th non-demand response periods. The start and end times.

[0056] Step 2.3 Cluster all non-demand response users in the control group within the demand response day to form K load clustering curves. Specifically, it includes the following 6 steps.

[0057] Step 2.3.1 Calculate the load curves of k users who did not participate in demand response. Perform per-unit processing, i.e., ppNDR.it = PNDR.itmax(PNDR.it)i = 1,...k;

[0058] Step 2.3.2 will As a collection of sample data, K (K is much smaller than k) samples are randomly selected as cluster centers;

[0059] Step 2.3.3 Calculate the distances of other samples in the sample to these K cluster centers, and assign each of these samples to the category of its nearest cluster center;

[0060] Step 2.3.4: Calculate the average value for each category of the samples that have been classified above, and solve for the new cluster centroids.

[0061] Step 2.3.5 compares the K cluster centroids obtained in the previous calculation. If the cluster centroids have changed, proceed to step 2.3.3; otherwise, proceed to step 2.3.6.

[0062] Step 2.3.6 When the centroid does not change (when a centroid is found, the samples assigned to this centroid in each iteration are consistent, that is, each newly generated cluster is consistent, and all sample points will not move from one cluster to another, so the centroid will not change), stop and output the clustering results. The K-Means algorithm calculation process is as follows: Figure 3 As shown:

[0063] Step 2.4 Obtain the demand response user's activity during n non-demand response periods within the demand response day. The load curve segment, with Indicate. Will The non-demand response time periods corresponding to the K clustered load curves are respectively load curve segment Mapping is performed, that is, during a non-demand response time segment. Within, find the j-th load clustering curve that minimizes the error between the two. Specifically, as shown in formula (2):

[0064]

[0065] Simultaneously, load curve segments are calculated over n non-demand response periods. With the corresponding j-th load clustering curve The error between them is shown in the following formula:

[0066]

[0067] Step 2.5 set Map the load curve segment within the i-th non-demand response period to the j-th load cluster curve, and then use the segmentation error α obtained in step 2.4. i Clustering curves that map load curve segments of demand response users during n non-demand response periods within a demand response day. Weighted processing is performed to obtain the user baseline load within the demand response day under the segmented control group method, and...

[0068]

[0069] Step 3: During the non-demand response period within the demand response day, calculate the errors of the average method proposed in Step 1 and the segmented control group method in Step 2 respectively. Based on the errors of the two methods during the non-demand response period, fit the user baseline load during the demand response period within the demand response day.

[0070] Step 3.1 Calculate the user baseline load obtained by the averaging method described in Step 1 within the demand response date. Error compared to actual data over n non-demand response periods:

[0071]

[0072] in: and Let represent the user baseline load estimated by the averaging method and the user actual load respectively during the i-th non-demand response period. Error within.

[0073] Step 3.2 Calculate the user baseline load obtained during the demand response date using the segmented control group method described in Step 2. Error compared to actual data over n non-demand response periods:

[0074]

[0075] in: and Represent the baseline load and actual load of users estimated using the segmented control group method during the i-th non-demand response period, respectively. Error within.

[0076] Step 3.3 Based on the error magnitudes obtained in Steps 3.1 and 3.2, a weighted combination forecast is performed on the user baseline load to obtain the user baseline load P. t The details are as follows:

[0077]

[0078] The following are examples:

[0079] The simulation dataset used in this invention comes from measured data from a certain location. This invention selects the load data of 763 residential users throughout 2020 for simulation analysis, with a data sampling resolution of 1 hour.

[0080] After a demand response (DR) is executed, the true baseline load no longer exists. Therefore, to evaluate the performance of baseline load estimation methods, it is usually necessary to test on "DR-like days" (i.e., days with conditions similar to actual demand response days but where no DR event was actually executed). On "DR-like days," since no DR was actually executed, the load value measured is the true baseline load. This invention uses three evaluation metrics to evaluate the baseline load estimation method: accuracy, bias, and robustness.

[0081] (1) Accuracy. Accuracy can be measured by the mean absolute error (MAE) between the estimated value and the true value, as shown in the following formula.

[0082]

[0083] In the formula, and Let $\mathbf{n}$ and $\mathbf{t}$ represent the estimated and actual values ​​of the load baseline for the nth DR user during the tth time period on the dth DR day, respectively; $N$ represents the number of DR users to be estimated; $D$ represents the number of DR days to be estimated; and $|δ|$ represents the duration of the DR (in hours). Obviously, the smaller the MAE, the more accurate the CBL estimation.

[0084] (2) Deviation.

[0085] Bias can be represented by the algebraic difference between the estimated and actual load baseline values, as shown in the following formula.

[0086]

[0087] Unlike MAE, Bias is calculated using the algebraic difference between the actual and estimated values, rather than the absolute value. This means Bias can be positive or negative. A positive Bias indicates that the estimated load baseline is greater than its actual value, meaning the project provider needs to pay more compensation to project participants; conversely, project participants receive less compensation than they should have received. Compared to MAE, Bias is more relevant to the interests of DR project providers and participants. Clearly, the closer Bias is to 0, the smaller the deviation in the load baseline estimate, and the better the load baseline estimation method.

[0088] (3) Robustness. Robustness refers to the adaptability of the CBL estimation method under different scenarios (including different users, different DR days and different DR time periods). It can be measured by the relative error ratio (RER). The RER value of the nth DR user can be calculated by the following formula.

[0089]

[0090] In the formula, std(·) represents the standard deviation; avg(·) represents the average. and B n,d (δ)∈R |δ| The difference represents the estimated and actual load baseline values ​​within the DR event window δ.

[0091] The final RER value is obtained by averaging the RERs of all users. The smaller the RER value, the better the robustness of the load baseline estimation method.

[0092] Table 1 lists the comparison results (average performance) of several different baseline load estimation methods in different scenarios.

[0093] Table 1 Comparison of Indicators of Different Baseline Load Estimation Methods under Different Scenarios

[0094]

[0095]

[0096] Specifically, among the five CBL estimation methods mentioned above, the method proposed in this invention exhibits the best performance in terms of bias and robustness, and the estimation results have high accuracy.

[0097] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0098] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0099] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to perform any of the demand response baseline load estimation methods described in the above embodiments.

[0100] It is understood that the system provided in the embodiments of the present invention corresponds to the method provided in the embodiments of the present invention, and the explanation, examples and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

[0101] This application also provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, communication interface, and memory communicate with each other via the communication bus.

[0102] Memory, used to store computer programs;

[0103] When the processor executes the program stored in memory, it implements the above-mentioned method for estimating the baseline load of the demand response.

[0104] The communication bus mentioned in the aforementioned electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc.

[0105] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0106] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0107] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0108] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0109] It should be noted that in this invention, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0110] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0111] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for estimating baseline load in demand response, characterized in that, Includes the following steps, Step 1: Based on the electricity load data for the specified number of days prior to the demand response date, estimate the user baseline load for the demand response date using the averaging method; Step 2: Based on the electricity load data of all users during the demand response day, estimate the user baseline load during the demand response day using the segmented control group method; Step 3: During the non-demand response period within the demand response day, calculate the errors of the average method proposed in Step 1 and the segmented control group method in Step 2, respectively. Based on the errors of the two methods during the non-demand response period, fit the user baseline load during the demand response period within the demand response day; Step 2 specifically includes, Step 2.1 On the demand response day, users are divided into demand response users and non-demand response users according to whether they participate in demand response, and the set of non-demand response users is used as the control group. Step 2.2 Assuming multiple demand responses occurred within the demand response day, then the demand response day contained: n Non-demand response period And there are ,in: and The first i Non-demand response period The start and end times; Step 2.3: Using the K-Means algorithm shown in Figure 3, all non-demand response users in the control group are clustered within the demand response day to form K load clustering curves. ; Step 2.4 Obtain the demand response from the user within the specified timeframe. n Non-demand response period The load curve segment, with Indicate; will The non-demand response time periods corresponding to the K clustered load curves are respectively load curve segment Mapping is performed, that is, during a non-demand response time segment. Within, find the first that satisfies the requirement of minimizing the error between the two. j Clustering curves of bar load Specifically, as shown in formula (2): （2） Simultaneously, calculation n Load curve segments during non-demand response periods With the corresponding number j Clustering curves of bar load The error between them is shown in the following formula: （3） Step 2.5 Set For the first i The load curve segment during the non-demand response period is mapped to the first... j The load clustering curve is then used to determine the segmentation error obtained in step 2.

4. Within the demand response day, the demand response user will... n Clustering curves mapped to load curve segments during each non-demand response period After weighting, the baseline user load within the demand response day under the segmented control group method is obtained: （4）； The specific steps of step 3 are as follows: Step 3.1 Calculate the user baseline load obtained by the averaging method described in Step 1 within the demand response date. exist n Errors between actual data and data during non-demand response periods: （5） in: and These represent the user baseline load estimated by the averaging method and the user actual load on the [number]th [month]. i Non-demand response period Within the error; Step 3.2 Calculate the user baseline load obtained in Step 2 using the segmented control group method during the demand response date. exist n Errors between actual data and data during non-demand response periods: （6） in: and These represent the user baseline load and the actual user load estimated using the segmented control group method on the [number]th [year]. i Non-demand response period Error within; Step 3.3 Based on the error magnitudes obtained in Steps 3.1 and 3.2, perform weighted combination forecasting on the user baseline load to obtain the user baseline load. The details are as follows: （7）。 2. The method for estimating the baseline load of demand response according to claim 1, characterized in that: Step 1 specifically includes, Select historical electricity load data of users for Y days prior to the estimated demand response date, and remove unrepresentative days, namely historical demand response days and holidays, until Y days are added. Then, select the X days with the highest daily load from these Y days, and calculate the average load at each time point on these X days as the user baseline load for the demand response day: （1） In the formula, This represents the intraday user baseline load estimated using the average method for demand response. High ( X , Y () represents the set of X days with the highest load levels among the Y days that meet the conditions before the demand response date. express High ( X , Y The set of d Heavenly t Load values ​​for a given time period.

3. The method for estimating the baseline load of demand response according to claim 2, characterized in that: Step 2.3 specifically includes, Step 2.3.1: Calculate the load curves of k users who did not participate in demand response. Perform per-unit processing, that is ; Step 2.3.2 will As a collection of sample data, K samples are randomly selected as cluster centers, where K is much smaller than k. Step 2.3.3 Calculate the distances of other samples in the sample to these K cluster centers, and assign each of these samples to the category of its nearest cluster center; Step 2.3.4 Calculate the average value for each category of the samples that have been classified above, and solve for the new cluster centroids; Step 2.3.5 compares the K cluster centroids obtained from the previous calculation. If the cluster centroids have changed, proceed to step 2.3.3; otherwise, proceed to step 2.3.

6. Step 2.3.6 When the centroid does not change, stop and output the clustering results.

4. A computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of the method as described in any one of claims 1 to 3.

5. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method as claimed in any one of claims 1 to 3.

Images