An edge-computing-based power user demand control method

Abstract

The application belongs to the technical field of edge computing, and particularly relates to a power user demand control method based on edge computing, which comprises the following steps: obtaining user edge historical load data; performing load curve clustering based on the user historical load data to obtain a typical load curve; establishing a user side energy storage configuration model and a user side energy storage benefit model considering demand control; and calculating optimal configuration parameters in the user side energy storage configuration model based on the typical load curve and the user side energy storage benefit model. The application realizes real-time data acquisition, processing and optimal control by deploying a computing node at the edge of a power grid, effectively solves the problems caused by centralized control, and improves the real-time performance and accuracy of power grid management.

Description

Technical Field

[0001] This invention belongs to the field of edge computing technology, specifically relating to a power user demand control method based on edge computing. Background Technology

[0002] With the rapid development of smart grids, the requirements for refined grid management are increasing. Demand control, as an important means of power management, balances grid load and avoids grid overload by limiting user power consumption during specific time periods. Meanwhile, peak shaving and valley filling technologies use energy storage systems to discharge during peak loads and charge during off-peak periods to smooth the grid load curve and improve grid operating efficiency. However, existing technologies for the coordinated control of demand control and peak shaving and valley filling often rely on centralized computing platforms, which suffer from data transmission delays and limited computing resources. Furthermore, the load curves of each edge device are influenced by their local environment and exhibit different characteristics.

[0003] The importance of demand control for the refined management of the power grid is increasing. Demand control can achieve load balancing, prevent overload, and smooth load curves by peak shaving and valley filling, thereby improving grid operating efficiency, reducing demand-based electricity costs, and saving electricity expenses. Therefore, research on demand control is being conducted from various perspectives. However, in current mainstream technologies, the coordinated control of demand control and peak shaving / valley filling is completed by a cloud computing center. The cloud platform collects electricity consumption data from various devices, analyzes the grid load operation based on corresponding algorithms, derives relevant results, and then distributes the corresponding control strategies to each device to realize the demand control function. Based on this choice, existing demand control strategies suffer from problems such as data transmission delays, data loss, limited computing resources, and slow response. Summary of the Invention

[0004] The purpose of this invention is to provide a power user demand control method based on edge computing. By deploying computing nodes at the edge of the power grid, real-time data acquisition, processing, and optimized control are achieved, effectively solving the problems caused by centralized control and improving the real-time performance and accuracy of power grid management.

[0005] The specific technical solution adopted by this invention is as follows:

[0006] A power user demand control method based on edge computing includes:

[0007] Obtain historical load data at the user's edge;

[0008] Based on the user's historical load data, load curve clustering is performed to obtain typical load curves;

[0009] Establish a user-side energy storage configuration model and a user-side energy storage revenue model that take demand control into account;

[0010] Based on typical load curves and user-side energy storage revenue models, the optimal configuration parameters in the user-side energy storage configuration model are calculated.

[0011] In a preferred embodiment, the step of acquiring user edge historical load data includes:

[0012] Install edge computing devices and energy storage systems on the power user side, and ensure smooth communication between the two;

[0013] After the edge computing device is started, it automatically collects user load data, device status information and external environmental parameters, and uploads them to the local database.

[0014] In a preferred embodiment, the typical load curve is a load change curve generated during the operation of the distribution network system over a past period, determined by the average load at the same time over three months. Where ob represents the edge device number, · represents all working days, j∈[0~23] represents hours; Mean represents the average value.

[0015] In a preferred embodiment, the step of clustering load curves based on the user's historical load data to obtain typical load curves includes:

[0016] Obtain user load curves, and randomly select several load curves from the user load curves as the initial load cluster centers;

[0017] Calculate the Euclidean distance from each load curve to each cluster center;

[0018] For each load curve, compare its Euclidean distance to each cluster center, find the minimum value, and then classify it.

[0019] Calculate the average value of each type of load data, and use the averaged load curve as the new cluster center for each type.

[0020] Compare the distance between the new cluster center and the previous cluster center;

[0021] When the distance exceeds the initially set threshold, the Euclidean distance from each load curve to each cluster center is recalculated, and the distance between the new cluster center and the previous cluster center is compared.

[0022] When the distance between all cluster centers is less than a set threshold, the clustering curve and the final cluster centers are output.

[0023] The optimal number of clusters with the smallest DBI index among the final cluster centers is obtained by traversal method, and the clustering curve corresponding to the optimal number of clusters is the typical load curve.

[0024] In a preferred embodiment, the load curve clustering calculation process is performed in real time on an edge computing device.

[0025] In a preferred embodiment, the energy storage configuration model includes an energy storage system device cost model, an energy storage system power balance model, an energy storage device charging and discharging power model, and a peak shaving and valley filling model.

[0026] In a preferred embodiment, the user-side energy storage revenue model incorporates a user-side energy storage revenue objective function, the formula of which is:

[0027] F max =D+C trans -W inv -C oper

[0028] In the formula, F max For user-side energy storage revenue; C trans The reduction in transformer cost due to maximum demand reduction; W inv For users' energy storage investment costs; C oper This refers to the operation and maintenance costs of energy storage throughout its entire lifecycle.

[0029] A second objective of this invention is to provide an edge computing-based power user demand control system. Using the aforementioned edge computing-based power user demand control method, the system includes:

[0030] The data acquisition module is used to acquire historical load data at the user edge.

[0031] The edge load curve clustering module is used to cluster load curves based on the user's historical load data to obtain typical load curves;

[0032] The energy storage configuration model building module is used to build a user-side energy storage configuration model and a user-side energy storage revenue model that take demand control into account.

[0033] Optimal configuration parameter calculation module: Based on typical load curves and user-side energy storage revenue models, calculates the optimal configuration parameters in the user-side energy storage configuration model.

[0034] In a preferred embodiment, the load curve clustering module first performs real-time calculation and response in the edge computing device, which stores historical iteration clustering results, and then obtains the latest clustering data based on the load data acquisition module.

[0035] The cloud control center comprehensively processes the calculation results of each edge device.

[0036] The cloud platform's control decision module adjusts the cloud-based demand control algorithm based on data uploaded from each edge device. The adjustment results generate a control policy and send corresponding adjustment instructions to the relevant edge devices.

[0037] The edge device re-sets its charging and discharging power based on the received control strategy instructions.

[0038] A third objective of this invention is to provide a control terminal, comprising:

[0039] The storage medium is used to store instructions and also for local data computation and receiving remote scheduling from the cloud control center.

[0040] The processor is configured to execute the edge computing-based power user demand control method according to the instructions.

[0041] The technical advantages achieved by this invention are as follows: By deploying computing resources on edge devices, real-time monitoring and prediction of power user loads and intelligent scheduling of energy storage devices are realized. This algorithm can not only effectively reduce demand-based electricity costs for power users and improve the economic efficiency of energy storage systems, but also enhance the stability and reliability of the power system, providing strong technical support for the development of smart grids.

[0042] The collaborative control method of this invention comprehensively considers the needs of demand control and energy storage optimization, and achieves effective collaboration between the two through the real-time data processing and decision-making capabilities of edge computing nodes. During the control process, the edge computing nodes dynamically adjust the demand control threshold and the charging and discharging strategy of the energy storage system according to the real-time changes in the grid load, ensuring that the grid load fluctuates within a safe and economical range.

[0043] The present invention. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating the present invention;

[0045] Figure 2 This is a schematic diagram of the modules of the present invention. Detailed Implementation

[0046] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0047] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0048] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that mutually excludes other embodiments.

[0049] Please see the appendix Figure 1 The image shows the first embodiment of the present invention, which provides a power user demand control method based on edge computing, including:

[0050] Obtain historical load data at the user's edge;

[0051] Based on the user's historical load data, load curves are clustered to obtain typical load curves. The clustering process is completed in real time in edge computing devices.

[0052] Establish a user-side energy storage configuration model and a user-side energy storage revenue model that take demand control into account;

[0053] Based on typical load curves and user-side energy storage revenue models, the optimal configuration parameters in the user-side energy storage configuration model are calculated.

[0054] In this embodiment, the energy storage configuration model includes an energy storage system device cost model, an energy storage system power balance model, an energy storage device charging and discharging power model, and a peak shaving and valley filling model. Based on these models...

[0055] The process of acquiring historical load data at the user edge involves installing edge computing devices and energy storage systems on the power user side.

[0056] And ensure unimpeded communication between the two;

[0057] After the edge computing device is started, it automatically collects user load data, device status information, and external environmental parameters, and uploads them to the local database.

[0058] The aforementioned control method includes edge computing, energy storage systems, smart meters, and a power grid control center. Edge computing nodes are deployed at key nodes of the power grid, responsible for collecting power grid operation data, executing demand control algorithms, monitoring load curves, and sending control commands to energy storage systems and smart meters. Edge computing refers to an open platform that integrates network, computing, storage, and application capabilities at the network edge, close to the source of data or objects, providing edge intelligent services locally to meet the key needs of industry digitalization in areas such as agile connectivity, real-time business, data optimization, application intelligence, and security and privacy protection. The demand control method in this invention dynamically adjusts the user's power usage limits based on the real-time data processing capabilities of edge computing nodes. First, based on historical data and real-time load forecasts, demand thresholds are set for each time period (e.g., exceeding 90% of the transformer load, or other settings depending on the actual situation). Then, the actual power consumption of users is monitored in real time. When the actual power consumption exceeds the demand threshold, a control signal is sent to the user through the smart meter to limit their power usage. Leveraging the computing power of edge computing devices and combining historical load data, user production plan information, and weather forecasts, a precise load forecasting method is constructed to predict the power load curve for a future period. This model employs advanced machine learning algorithms, such as Long Short-Term Memory (LSTM) networks, to improve the accuracy and robustness of the predictions. Based on the predicted load curve and the state information of the energy storage system, an intelligent energy storage scheduling strategy is executed. This strategy comprehensively considers the charging and discharging efficiency of the energy storage devices, remaining capacity, lifecycle costs, and grid constraints on the user side, dynamically adjusting the charging and discharging strategies of the energy storage devices to maximize the reduction of demand charges and improve the economic efficiency of the energy storage system while meeting user electricity demand.

[0059] To further understand and illustrate the typical load curve, it will be discussed in detail below. The typical load curve is the load change curve generated during the operation of the distribution network system over a past period, determined by the average load at the same time over three months. Where ob represents the edge device number, · represents all working days, j∈[0~23] represents hours; Mean represents the average value.

[0060] In a preferred embodiment, referring to Figure X, the steps of clustering load curves based on the user's historical load data to obtain typical load curves include:

[0061] Obtain user load curves, and randomly select several load curves from the user load curves as the initial load cluster centers;

[0062] Calculate the Euclidean distance from each load curve to each cluster center;

[0063] For each load curve, compare its Euclidean distance to each cluster center, find the minimum value, and then classify it.

[0064] Calculate the average value of each type of load data, and use the averaged load curve as the new cluster center for each type.

[0065] Compare the distance between the new cluster center and the previous cluster center;

[0066] When the distance exceeds the initially set threshold, the Euclidean distance from each load curve to each cluster center is recalculated, and the distance between the new cluster center and the previous cluster center is compared.

[0067] When the distance between all cluster centers is less than a set threshold, the clustering curve and the final cluster centers are output.

[0068] The optimal number of clusters with the smallest DBI index among the final cluster centers is obtained by traversal method, and the clustering curve corresponding to the optimal number of clusters is the typical load curve.

[0069] To better understand the energy storage configuration model, it is further analyzed below. The energy storage configuration model includes the energy storage system device cost model, the energy storage system power balance model, the energy storage equipment charging and discharging power model, and the peak shaving and valley filling model.

[0070] Furthermore, the cost model for the energy storage system device is as follows:

[0071] The investment in energy storage configuration mainly includes fixed investment costs and subsequent operation and maintenance costs, where W = W1 + W2 + W3, where W is the total fixed investment cost; W1 represents the investment cost of battery cells, etc.; W2 represents the investment cost of PCS equipment; and W3 represents the investment cost of EMS control module.

[0072] The operation and maintenance cost is C = fQ, where C represents the annual operation and maintenance cost of the system; f represents the actual annual operating input per unit of power; and Q represents the rated power of the system.

[0073] Furthermore, the power balance model of the energy storage system is as follows:

[0074] T load (t)=P load (t)+P EES (t)

[0075] T peak,m =(1-δ m )P peak,m

[0076] In the formula, P load (t) represents the actual user load, P EES(t) represents the power of the energy storage device, P EES (t)>0 indicates a charging state, P EES (t)<0 indicates the discharge state, T load (t) represents the actual power at the user's grid connection point, δ m To achieve peak shaving efficiency after configuring energy storage devices, P peak,m T represents the peak power (calendar month) at the user's grid connection point when no energy storage device is installed. peak,m Peak power (calendar month) at the user's grid connection point after the installation of energy storage devices;

[0077] Furthermore, the charging and discharging power model of the energy storage device is as follows:

[0078] P EES (t)=P c (t)B c (t)-P dc (t)B dc (t)

[0079] 0≤P c (t)≤B c (t)P cmax

[0080] 0≤P dc (t)≤B dc (t)P dcmax

[0081] B c (t)+B dc (t)≤1

[0082] In the formula, P EES (t) represents the power of the energy storage device, P c (t) represents the energy storage charging power, P dc (t) represents the energy storage discharge power, P cmax For the rated charging power, P dcmax For rated discharge power, B c (t) and B dc (t) is a Boolean variable (corresponding to the charging and discharging state);

[0083] Furthermore, the peak shaving and valley filling model is as follows:

[0084] P load (t)+P EES (t)≤(1-δ m )P peak,m .

[0085] It should also be further explained that, based on typical load curves and user-side energy storage revenue objective functions, the optimal configuration parameters in the user-side energy storage configuration model are calculated.

[0086] The peak shaving and valley filling effects achievable using energy storage devices can generate revenue from the peak-valley price difference. D represents the user's electricity purchase cost benefit value; P1 and P2 represent the discharge and charging power values ​​for a certain period of time, respectively; c represents the actual electricity price data for a certain period of time; and x represents time.

[0087] The user-side energy storage revenue model incorporates a user-side energy storage revenue objective function, the formula of which is:

[0088] F max =D+C trans -W inv -C oper

[0089] In the formula, F max For user-side energy storage revenue; C trans The reduction in transformer costs due to the reduction in maximum demand (peak load); W inv For users' energy storage investment costs; C oper This refers to the operation and maintenance costs of energy storage throughout its entire lifecycle.

[0090] like Figure 2 As shown, a second objective of this invention is to provide an edge computing-based power user demand control system. Using the aforementioned edge computing-based power user demand control method, the system includes:

[0091] The data acquisition module is used to acquire historical load data at the user edge.

[0092] The edge load curve clustering module is used to cluster load curves based on the user's historical load data to obtain typical load curves;

[0093] The energy storage configuration model building module is used to build a user-side energy storage configuration model and a user-side energy storage revenue model that take demand control into account.

[0094] Optimal configuration parameter calculation module: Based on typical load curves and user-side energy storage revenue models, calculates the optimal configuration parameters in the user-side energy storage configuration model.

[0095] When the data acquisition module is executed, edge computing devices and energy storage systems need to be installed on the power user side, and communication between the two must be ensured. After the edge computing device is started, it automatically collects user load data, equipment status information and external environmental parameters, and uploads them to the local database.

[0096] The load curve clustering module first performs real-time calculation and response in the edge computing device during the load curve clustering calculation process. The edge computing device stores historical iteration clustering results and obtains the latest clustering data based on the load data acquisition module.

[0097] The cloud control center comprehensively processes the calculation results of each edge device.

[0098] The cloud platform's control decision module adjusts the cloud-based demand control algorithm based on data uploaded from each edge device. The adjustment results generate a control policy and send corresponding adjustment instructions to the relevant edge devices.

[0099] The edge device re-sets its charging and discharging power based on the received control strategy instructions.

[0100] In the aforementioned system, edge computing devices and energy storage systems are installed on the power user side, ensuring smooth communication between them. Once activated, the edge computing devices automatically collect user load data, device status information, and external environmental parameters, uploading them to a local database. A load forecasting model (using the average load over three months at the same time) predicts the power load curve for a future period, storing the prediction results in the edge computing devices. Based on the predicted load curve and energy storage system status information, an energy storage scheduling strategy is formulated, and the charging and discharging strategies of the energy storage devices are adjusted in real time. The edge computing devices monitor the actual electricity demand of users in real time and compare it with the predicted load curve. Based on a default threshold (e.g., exceeding 90% of the transformer load), a dispatch command is triggered to suspend the charging and discharging of the corresponding edge devices. A comprehensive benefit evaluation module monitors and evaluates the algorithm's performance in real time, generating visual reports and charts for users to view and analyze.

[0101] A third objective of this invention is to provide a control terminal, comprising:

[0102] The storage medium is used to store instructions and also for local data computation and receiving remote scheduling from the cloud control center.

[0103] A processor, configured to execute the edge computing-based power user demand control method according to the instructions.

[0104] The above description is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained in this invention are implemented according to conventional methods in the art unless otherwise specified or limited.

Claims

1. A power user demand control method based on edge computing, characterized in that: include Obtain historical load data at the user's edge; Based on the user's historical load data, load curve clustering is performed to obtain typical load curves; Establish a user-side energy storage configuration model and a user-side energy storage revenue model that take demand control into account; Based on typical load curves and user-side energy storage revenue models, the optimal configuration parameters in the user-side energy storage configuration model are calculated. The load curve clustering calculation process is completed in real time in the edge computing device. The edge computing device stores historical iteration clustering results and obtains the latest clustering data based on the load data acquisition module. The cloud control center comprehensively processes the calculation results of each edge device; the cloud platform's control decision module adjusts the cloud demand control algorithm based on the data uploaded by each edge device, and the adjustment results generate a control strategy and send corresponding adjustment instructions to the relevant edge devices; the edge devices then set their own charging and discharging power again according to the received control strategy instructions. The energy storage configuration model includes an energy storage system device cost model, an energy storage system power balance model, an energy storage device charging and discharging power model, and a peak shaving and valley filling model. The energy storage system device cost model is as follows: Investment in energy storage configurations mainly includes fixed investment costs and subsequent operation and maintenance costs, among which , Total fixed investment cost; This is expressed as the cost of battery cell investment; This represents the investment cost of some PCS equipment. This is represented as the investment cost of the EMS control module; The operation and maintenance cost is , Expressed as the annual system maintenance cost; Expressed as annual operating input per unit power; This is expressed as the system's rated power. The peak shaving and valley filling effects achievable using energy storage devices can generate revenue from the peak-valley price difference. , This is expressed as the benefit value of the user's electricity purchase cost; as well as These represent the discharge and charging power values ​​for a specific time period, respectively. This represents the actual electricity price data for a specific period. Represented as time; The user-side energy storage revenue model incorporates a user-side energy storage revenue objective function, the formula of which is: ； In the formula, For user-side energy storage revenue; The reduction in transformer costs due to maximizing demand; Reduce user energy storage investment costs; This refers to the operation and maintenance costs of energy storage throughout its entire lifecycle.

2. The power user demand control method based on edge computing according to claim 1, characterized in that: The step of obtaining historical load data at the user edge includes: Install edge computing devices and energy storage systems on the power user side, and ensure smooth communication between the two; After the edge computing device is started, it automatically collects user load data, device status information and external environmental parameters, and uploads them to the local database.

3. The power user demand control method based on edge computing according to claim 1, characterized in that: The typical load curve is a load change curve generated during the operation of the distribution network system over a past period, determined by the average load at the same time over three months. ,in, Represents the edge device number. Represents all working days within a 3-month period. , indicating hours; This indicates taking the average value.

4. The power user demand control method based on edge computing according to claim 1, characterized in that: The step of clustering load curves based on the user's historical load data to obtain typical load curves includes: Obtain user load curves, and randomly select several load curves from the user load curves as the initial load cluster centers; Calculate the Euclidean distance from each load curve to each cluster center; For each load curve, compare its Euclidean distance to each cluster center, find the minimum value, and then classify it. Calculate the average value of each type of load data, and use the averaged load curve as the new cluster center for each type. Compare the distance between the new cluster center and the previous cluster center; When the distance exceeds the initially set threshold, the Euclidean distance from each load curve to each cluster center is recalculated, and the distance between the new cluster center and the previous cluster center is compared. When the distance between all cluster centers is less than a set threshold, the clustering curve and the final cluster centers are output. The optimal number of clusters with the smallest DBI index among the final cluster centers is obtained by traversal method, and the clustering curve corresponding to the optimal number of clusters is the typical load curve.

5. A power user demand control system based on edge computing, characterized in that: The method for demand control of electricity users based on edge computing, as described in any one of claims 1 to 4, comprises: The data acquisition module is used to acquire historical load data at the user edge. The edge load curve clustering module is used to cluster load curves based on the user's historical load data to obtain typical load curves; The energy storage configuration model building module is used to build a user-side energy storage configuration model and a user-side energy storage revenue model that take demand control into account. Optimal configuration parameter calculation module: Based on typical load curves and user-side energy storage revenue models, calculates the optimal configuration parameters in the user-side energy storage configuration model.

6. A control terminal, characterized in that, include: The storage medium is used to store instructions and also for local data computation and receiving remote scheduling from the cloud control center. A processor, the processor being configured to execute, according to the instructions, the power user demand control method based on edge computing as described in any one of claims 1-4.

Images