Electrical load release method based on clustering analysis

Abstract

The invention provides an electrical load release method based on clustering analysis, and relates to the technical field of electrical load adjustment. The electrical load release method based on clustering analysis comprises the steps of data collection, collection of historical data of electrical loads in a target area through a power grid load power monitoring system, data preprocessing, elimination of abnormal values in the collected electrical data in the target area, and extraction of static characteristics and dynamic characteristics of power data. And the data features are standardized. According to the method, the K-means clustering algorithm and the DBSCAN clustering algorithm are adopted, the convergence speed of the algorithms and the stability of the clustering result are effectively improved by combining the characteristics of the two algorithms, and the load release strategy which is high in pertinence and feasible is formulated based on different electrical load categories obtained through clustering analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of electric load adjustment, and in particular to an electric load release method based on cluster analysis. Background Art

[0002] With the rapid development of the global economy and the continuous improvement of people's living standards, society's demand for electricity continues to rise. As an indispensable energy source in modern society, the stable supply of electricity is crucial to the normal operation of economic development and social life. The intensification of the contradiction between electricity supply and demand will not only affect the development of the social economy, but may also have a negative impact on the quality of life of residents. In the case of insufficient electricity supply, industrial enterprises may be forced to reduce production or even stop production, resulting in economic losses, and residential users may face problems such as power outages and voltage instability, which affect the normal progress of daily life. In order to alleviate the contradiction between electricity supply and demand and improve the operating efficiency and reliability of the power system, the present invention proposes a method for releasing electricity load based on cluster analysis, thereby providing a scientific theoretical basis and practical technical means for the optimized operation of the power system, which plays an important role in alleviating the contradiction between electricity supply and demand, improving the operating efficiency and reliability of the power system, and promoting the sustainable development of the power industry. Summary of the Invention

[0003] Technical problems solved In response to the shortcomings of the existing technology, the present invention provides a power load release method based on cluster analysis. By cleaning the original data, missing values ​​and outliers in the data are effectively removed, ensuring the integrity and accuracy of the data. The Z-score standardization method is used to standardize the data to eliminate the dimensional differences between different features. The K-means clustering algorithm and the DBSCAN clustering algorithm are adopted. The combination of the two algorithm characteristics effectively improves the convergence speed of the algorithm and the stability of the clustering results. Based on the different power load categories obtained by cluster analysis, a targeted and practical load release strategy is formulated. According to the characteristics of different load categories, the charging and discharging strategies of the energy storage equipment are rationally planned. A configuration method combining distributed and centralized methods is adopted to give full play to the advantages of energy storage technology, realize peak shaving and valley filling of electricity, and improve the stability and reliability of the power system.

[0004] Technical Solution

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for releasing power load based on cluster analysis, the implementation steps of the method are as follows: Step 1: Data collection: Collect historical data on electricity consumption in the target area through the power grid load monitoring system; Step 2: Data preprocessing: remove outliers from the electricity consumption data collected in the target area, extract static and dynamic features of the power data, and standardize the data features; Step 3: Cluster analysis modeling, cluster analysis of data based on K-means clustering algorithm and DBSCAN clustering algorithm; Step 4: Analyze the cluster characteristics and count the central features of each cluster in the target area. Classify the clusters into high elasticity clusters, low elasticity clusters, and shiftable clusters based on their characteristics. Combined with the power demand of the power grid, set the power release priority level in the target area: Step 5: Design a load release strategy. Analyze the load cluster characteristics to determine the core objectives of the power release strategy within the target area and convert the objectives into quantifiable indicators. Step 6: Strategy implementation. During the implementation process, load aggregators are used to collect load status data in real time through Internet terminals and data sensors, and the data is compared with the strategy expectations for dynamic optimization.

[0006] Preferably, in the data collection step, the power data to be collected in the target area includes load operation data, user attribute data and environmental data; Through the above technical solution, load operation data includes power curves, voltage, current, etc., user attribute data includes user type and electrical equipment, etc., environmental data includes ambient temperature, holiday electricity consumption, etc., and data sources include sensor collection, smart meters and Internet of Things terminals.

[0007] Preferably, in the data preprocessing step, the interquartile range of the power load is calculated based on the interquartile range method to eliminate data outliers. The first quartile Q1 indicates that 25% of the data is less than this value, and the third quartile Q3 indicates that 75% of the data is less than this value. According to IQR=Q3-Q1, data points smaller than Q1-1.5IQR or greater than Q3+1.5IQR are determined as outliers. Through the above technical solution, during the data preprocessing process, by visually comparing and statistically analyzing the data before and after processing, after data cleaning, outliers are effectively processed and the data distribution is more reasonable.

[0008] Preferably, in the data preprocessing step, the static feature data of the target area are extracted, including the classification of electricity users, the rated power of electricity equipment, and the upper limit of the adjustable capacity of the power load; the dynamic features of the target area are extracted, including the peak and valley period distribution of the power load, the power load adjustment rate, and the delayed response time. The data are normalized based on the Z-score standard deviation. By converting the mean and standard deviation of the original data, the data is mapped to a standard normal distribution with a mean of 0 and a standard deviation of 1. The calculation formula for X of the concentrated power load is:

[0009] x is the original data, x 1 is the standardized data, mean( x ) is the original data x The mean, std( x ) is the original data x The standard deviation of Through the above technical solution, the Z-score does not have strict requirements on the distribution of data and has a certain robustness to outliers. Even if there are a small number of outliers in the data set, normalization by calculating the mean and standard deviation can reduce the impact of outliers on the overall data distribution to a certain extent.

[0010] Preferably, in the cluster analysis modeling step, the algorithm indicators of the K-means clustering algorithm are data values ​​of closeness and separation in the range of [-1, 1], and the calculation formula for calculating the average distance between each data point and other data points in the same cluster and the average distance between each data point and data points in other clusters is:

[0011] S(i) Represents each data point, Indicates the closeness between clusters, which is the average distance between each data point and other data points in the same cluster. The minimum distance between each data point and one of the points represents the degree of separation between clusters; Through the above technical solution, the K-means clustering algorithm divides the data set into K clusters so that the sum of the squared distances between the data points in each cluster and the center point of the cluster is minimized, and the DBSCAN clustering algorithm identifies clusters by defining the density connection area of ​​the data points.

[0012] Preferably, in the cluster characteristic analysis step, the cluster center characteristics include the average adjustable power of the power load, the main power consumption period of high power load and the average response delay, the high elasticity cluster characteristics are fast response and large adjustment range, the low elasticity cluster characteristics are non-adjustable, the shiftable cluster characteristics are flexible adjustment of the power consumption period but low adjustment rate, and the high elasticity cluster is determined to have the highest priority for power load release adjustment.

[0013] Preferably, in the actual steps of the load release strategy, a real-time segmented electricity price strategy is adopted for high-elasticity clusters, a subsidy-based control is adopted for shiftable clusters, and a contract-reserved adjustable capacity is adopted for low-elasticity clusters; Through the above technical solution, the time-of-use electricity price strategy formulates different electricity prices according to the electricity supply and demand situation and cost in different time periods. The electricity price is moderate during normal periods and low during valley periods. For industrial users in high-elasticity clusters, production activities during valley periods can reduce electricity costs. For commercial users in shiftable clusters, commercial users are guided to increase electricity consumption during valley periods and reduce electricity consumption during peak periods. At the same time, certain economic compensation or service discounts are provided to users to increase their enthusiasm for participation. For residential users in low-elasticity clusters, part of the power load needs to be reserved to ensure the adjustable capacity of user electricity consumption.

[0014] Preferably, in the strategy implementation step, the power usage in the target area is optimized in a rolling manner using the K-means clustering algorithm and the DBSCAN clustering algorithm, and a closed-loop system of "data collection - cluster analysis - strategy generation - execution feedback" is constructed to perform real-time strategy update optimization; Through the above technical solutions, complex load systems are transformed into manageable cluster groups, providing a scientific basis for the intelligent regulation of power systems. Beneficial effects

[0015] The present invention provides a method for releasing power load based on cluster analysis. It has the following beneficial effects: The present invention provides an electricity load release method based on cluster analysis. By cleaning the original data, missing values ​​and outliers in the data are effectively removed, ensuring the integrity and accuracy of the data. The Z-score standardization method is used to standardize the data to eliminate the dimensional differences between different features. The K-means clustering algorithm and the DBSCAN clustering algorithm are adopted. The characteristics of the two algorithms are combined to effectively improve the convergence speed of the algorithm and the stability of the clustering results. Based on the different electricity load categories obtained by cluster analysis, a highly targeted and practical load release strategy is formulated. According to the characteristics of different load categories, the charging and discharging strategies of the energy storage equipment are rationally planned. A configuration method combining decentralized and centralized methods is adopted to give full play to the advantages of energy storage technology, realize peak shaving and valley filling of electricity, and improve the stability and reliability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of the power load release method based on cluster analysis of the present invention. DETAILED DESCRIPTION

[0017] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the accompanying drawings. It should be noted that the described embodiments are only some of the embodiments of the present invention, and are not intended to be exhaustive. Based on the embodiments of the present invention, all other embodiments derived by persons of ordinary skill in the art without inventive effort are within the scope of protection of the present invention. It should be noted that the terms used herein are intended solely to describe specific embodiments and are not intended to limit the exemplary embodiments of the present invention. For ease of description, the dimensions of the various parts shown in the drawings are not drawn to scale. Technologies, methods, and devices known to persons of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such technologies, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely illustrative and not limiting. Therefore, other examples of the exemplary embodiments may have different values. It should be noted that similar reference numerals and letters denote similar items in the following figures. Therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures. Example 1:

[0018] like Figure 1 As shown, an embodiment of the present invention provides a method for releasing power load based on cluster analysis, and the method is implemented in the following steps: Step 1: Data collection: Collect historical data on electricity consumption in the target area through the power grid load monitoring system; Step 2: Data preprocessing: remove outliers from the electricity consumption data collected in the target area, extract static and dynamic features of the power data, and standardize the data features; Step 3: Cluster analysis modeling, cluster analysis of data based on K-means clustering algorithm and DBSCAN clustering algorithm; Step 4: Analyze the cluster characteristics and count the central features of each cluster in the target area. Classify the clusters into high elasticity clusters, low elasticity clusters, and shiftable clusters based on their characteristics. Combined with the power demand of the power grid, set the power release priority level in the target area: Step 5: Design a load release strategy. Analyze the load cluster characteristics to determine the core objectives of the power release strategy within the target area and convert the objectives into quantifiable indicators. Step 6: Strategy implementation. During the implementation process, load aggregators are used to collect load status data in real time through Internet terminals and data sensors, and the data is compared with the strategy expectations for dynamic optimization.

[0019] In the data collection step, the power data that needs to be collected in the target area include load operation data, user attribute data and environmental data. The load operation data includes power curve, voltage, current, etc. The user attribute data includes user type and electrical equipment, and the environmental data includes ambient temperature, holiday electricity consumption, etc. The data sources include sensor collection, smart meters and Internet of Things terminals. In the data preprocessing step, the interquartile range of power load is calculated based on the interquartile range method to eliminate data outliers. The first quartile Q1 means that 25% of the data is less than this value, and the third quartile Q3 means that 75% of the data is less than this value. According to IQR=Q3-Q1, data points with a value less than Q1-1.5IQR or greater than Q3+1.5IQR are determined to be outliers. In the process of data preprocessing, by visual comparison and statistical analysis of the data before and after processing, after data cleaning, the outliers are effectively processed and the data distribution is more reasonable.

[0020] In the data preprocessing step, static feature data of the target area are extracted, including the classification of electricity users, the rated power of electricity equipment, and the upper limit of the adjustable capacity of the power load. Dynamic features of the target area are extracted, including the peak and valley period distribution of electricity load, the power load adjustment rate, and the delayed response time. Normalization is performed based on the Z-score standard deviation. By converting the original data into a standard normal distribution with a mean of 0 and a standard deviation of 1, the calculation formula for X with concentrated power load is:

[0021] x is the original data, x 1 is the standardized data, mean( x ) is the original data x The mean, std( x ) is the original data x Through the above technical solution, the Z-score has no strict requirements on the distribution of data and has a certain robustness to outliers. Even if there are a small number of outliers in the data set, normalization by calculating the mean and standard deviation can reduce the impact of outliers on the overall data distribution to a certain extent.

[0022] In the cluster analysis modeling step, the algorithm indicators of the K-means clustering algorithm are the data values ​​of the closeness and separation in the range of [-1, 1]. For each data point, the average distance between it and other data points in the same cluster and the average distance between it and data points in other clusters is calculated as follows:

[0023] S(i) Represents each data point, Indicates the closeness between clusters, which is the average distance between each data point and other data points in the same cluster. The minimum distance between each data point and one of the points represents the degree of separation between clusters. The K-means clustering algorithm divides the data set into K clusters so that the sum of the squares of the distances between the data points in each cluster and the center point (centroid) of the cluster is minimized. The DBSCAN clustering algorithm identifies clusters by defining the density connection area of ​​the data points. The DBSCAN clustering algorithm defines the neighborhood radius. ϵ And the minimum number of points MinPoint for each data point in the dataset p , calculated with p is the center and the radius is ϵ The number of data points in the neighborhood of a core point that has not been visited is calculated. Starting from an unvisited core point, the data points that are directly density-reachable from it are added to the same cluster. Then, these newly added data points are checked. If they are also core points, the data points that are directly density-reachable from them are added to the cluster, continuously expanding the range of the cluster. This process is repeated until no new data points can be added to the cluster. If a data point is neither a core point nor density-reachable from any core point, it is marked as a noise point.

[0024] In the cluster characteristic analysis step, the cluster center characteristics include the average adjustable power of the power load, the main power consumption period of high power load and the average response delay. The high elasticity cluster is characterized by rapid response and large adjustment range, the low elasticity cluster is characterized by non-adjustability, and the shiftable cluster is characterized by flexible adjustment of the power consumption period but low adjustment rate. The high elasticity cluster is determined to be the highest priority for power load release adjustment. In the actual steps of the load release strategy, a real-time segmented electricity price strategy is used for the high elasticity cluster, a subsidy control is used for the shiftable cluster, and a contract reserved adjustable capacity is used for the low elasticity cluster. The time-of-use electricity price strategy is to formulate different electricity prices according to the electricity supply and demand situation and cost in different time periods. The electricity price in normal period is moderate, and the electricity price in valley period is low. For industrial users in high-elasticity clusters, production activities in valley period can reduce electricity costs. For commercial users in shiftable clusters, commercial users are guided to increase electricity consumption in valley period and reduce electricity consumption in peak period. At the same time, certain economic compensation or service discounts are provided to users to increase their enthusiasm for participation. For residential users in low-elasticity clusters, part of the power load needs to be reserved to ensure the adjustable capacity of user electricity consumption.

[0025] During the strategy implementation step, the K-means clustering algorithm and the DBSCAN clustering algorithm are used to perform rolling optimization on the power usage in the target area. A closed-loop system of "data collection - cluster analysis - strategy generation - execution feedback" is constructed to perform real-time strategy update optimization, transforming the complex load system into a manageable cluster group, providing a scientific basis for the intelligent regulation of the power system.

[0026] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for releasing power load based on cluster analysis, characterized by: The implementation steps of the method are: Step 1: Data collection: Collect historical data on electricity consumption in the target area through the power grid load monitoring system; Step 2: Data preprocessing: remove outliers from the electricity consumption data collected in the target area, extract static and dynamic features of the power data, and standardize the data features; Step 3: Cluster analysis modeling, cluster analysis of data based on K-means clustering algorithm and DBSCAN clustering algorithm; Step 4: Analyze the cluster characteristics and count the central features of each cluster in the target area. Classify the clusters into high elasticity clusters, low elasticity clusters, and shiftable clusters based on their characteristics. Combined with the power demand of the power grid, set the power release priority level in the target area: Step 5: Design a load release strategy. Analyze the load cluster characteristics to determine the core objectives of the power release strategy within the target area and convert the objectives into quantifiable indicators. Step 6: Strategy implementation. During the implementation process, load aggregators are used to collect load status data in real time through Internet terminals and data sensors, and the data is compared with the strategy expectations for dynamic optimization.

2. The method for releasing power load based on cluster analysis according to claim 1, characterized in that: In the data collection step, the power data to be collected in the target area includes load operation data, user attribute data and environmental data.

3. The method for releasing power load based on cluster analysis according to claim 1, characterized in that: In the data preprocessing step, the interquartile range of the power load is calculated based on the interquartile range method to eliminate data outliers. The first quartile Q1 indicates that 25% of the data is less than this value, and the third quartile Q3 indicates that 75% of the data is less than this value. According to IQR=Q3-Q1, data points smaller than Q1-1.5IQR or greater than Q3+1.5IQR are determined as outliers.

4. The method for releasing power load based on cluster analysis according to claim 1, characterized in that: In the data preprocessing step, static feature data of the target area are extracted, including the classification of electricity users, the rated power of electricity equipment, and the upper limit of the adjustable capacity of the power load. Dynamic features of the target area are extracted, including the peak and valley period distribution of the power load, the power load adjustment rate, and the delayed response time. Normalization is performed based on the Z-score standard deviation. By converting the original data into a mean and standard deviation, the data is mapped to a standard normal distribution with a mean of 0 and a standard deviation of 1. The calculation formula for X of the concentrated power load is: ， x is the original data, x 1 is the standardized data, mean( x ) is the original data x The mean, std( x ) is the original data x The standard deviation of .

5. The method for releasing power load based on cluster analysis according to claim 1, characterized in that: In the cluster analysis modeling step, the algorithm indicators of the K-means clustering algorithm are the data values ​​of the closeness and separation in the range of [-1, 1]. For each data point, the average distance between it and other data points in the same cluster and the average distance between it and data points in other clusters is calculated as follows: ， S(i) represents each data point, Indicates the closeness between clusters, which is the average distance between each data point and other data points in the same cluster. The minimum distance between each data point and one of the points represents the degree of separation between clusters.

6. The method for releasing power load based on cluster analysis according to claim 1, characterized in that: In the load cluster characteristic analysis step, the cluster center characteristics include the average adjustable power of the electric load, the main power consumption period of high power load and the response average value. The high elasticity cluster characteristics are fast response and large adjustment range, the low elasticity cluster characteristics are unadjustable, and the shiftable cluster characteristics are flexible adjustment of the power consumption period but low adjustment rate, and the high elasticity cluster is determined to have the highest priority for power load release adjustment.

7. The method for releasing power load based on cluster analysis according to claim 1, characterized in that: In the actual steps of the load release strategy, a real-time segmented electricity price strategy is adopted for high-elasticity clusters, a subsidy-based control is adopted for shiftable clusters, and a contract-reserved adjustable capacity is adopted for low-elasticity clusters.

8. The method for releasing power load based on cluster analysis according to claim 1, characterized in that: In the strategy implementation step, the K-means clustering algorithm and the DBSCAN clustering algorithm are used to perform rolling optimization on the power usage in the target area, and a closed-loop system of "data collection - cluster analysis - strategy generation - execution feedback" is constructed to perform real-time strategy update optimization.

Images