Method for dividing typical operation modes of power system based on frequency resource supply-demand ratio

Abstract

The application discloses a kind of power system typical operating mode division methods based on frequency modulation resource supply-demand ratio, through the combination of local power system historical annual load, wind power, photovoltaic power curve obtains daily maximum power shortage, constructs the frequency safety constraint of frequency modulation resource demand amount evaluation model based on system dynamic frequency response characteristics and power system maximum frequency deviation limit value, simultaneously establishes the unit commitment optimization model considering thermal power, hydropower, energy storage primary frequency modulation capacity, accurately quantifies the supply-demand data of daily frequency modulation resource in whole year;And construct comprehensive clustering index, carry out daily frequency modulation resource clustering analysis in whole year with this as similarity measure, realize the accurate division of power system annual typical day, complete power system annual typical day clustering analysis, provide technical support for future efficient configuration of power system frequency modulation.

Description

Technical Field

[0001] This invention relates to the field of power system operation and planning technology, and in particular to a method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources. Background Technology

[0002] With the increasing proportion of wind and solar power generation in future new power systems, frequency issues are becoming increasingly prominent, necessitating attention to the supply and demand of frequency regulation resources across different time scales. Currently, under my country's power market reform environment, pilot provinces primarily use day-ahead methods to organize the frequency regulation market. However, when a high proportion of renewable energy generation leads to a shortage of frequency regulation resources, it is necessary to assess frequency regulation demand and allocate resources over a longer time scale to mitigate the risk of insufficient resources. Traditional methods based on typical operating modes of load and renewable energy output are insufficient to reflect the distribution of supply and demand for system frequency regulation resources, hindering optimal allocation and failing to meet the operational needs of new power systems. Existing clustering methods primarily focus on scenario generation based on load and wind / solar output characteristics, but struggle to capture the distribution of frequency regulation resources in the power market. Furthermore, existing clustering methods do not adequately consider the supply-demand ratio of frequency regulation resources during the clustering process, resulting in inaccurate reflection of resource sufficiency and limiting their practicality in frequency regulation market resource allocation. Summary of the Invention

[0003] This invention discloses a method for classifying typical operating modes of a power system based on the supply-demand ratio of frequency regulation resources, in order to overcome the above-mentioned technical problems.

[0004] To achieve the above objectives, the technical solution of the present invention is as follows: A method for classifying typical operating modes of a power system based on the supply-demand ratio of frequency regulation resources includes the following steps: S1: Obtain the historical annual load power curve, wind power power curve, and photovoltaic power curve of the local power system to obtain the daily maximum power deficit; S2: Based on the daily maximum power deficit, establish the swing equation of the dynamic frequency response of the power system, and establish a frequency regulation resource demand assessment model based on the frequency security constraint of the maximum frequency deviation limit of the power system to obtain the daily frequency regulation resource demand. S3: Establish a unit combination model that considers the primary frequency regulation capacity of multiple resources, including thermal power, hydropower, and energy storage, in order to obtain the daily frequency regulation resource supply. S4: Based on the daily frequency modulation resource demand and the daily frequency modulation resource supply, obtain the orientation angle of the two-dimensional data points composed of the daily frequency modulation resource demand and the daily frequency modulation resource supply to obtain the frequency modulation resource supply-demand ratio deviation angle, and then obtain the comprehensive clustering index. S5: Construct supply-demand sample pairs based on daily frequency regulation resource demand and daily frequency regulation resource supply, and cluster the supply-demand sample pairs using the K-means clustering method according to the comprehensive clustering index to obtain multiple clusters to determine typical days of the year; and obtain the average supply-demand ratio of the clusters to determine the frequency regulation resource pattern of the supply-demand sample pairs within the clusters, thereby providing technical support for the efficient allocation of frequency regulation in the future power system.

[0005] Furthermore, the formula used to obtain the daily maximum power deficit is as follows:

[0006] In the formula: The power deficit of the power system during time period t; The maximum upward fluctuation of the load during time period t; The maximum downward change in wind power output during time period t; This represents the maximum downward change in photovoltaic power output during time period t.

[0007] Furthermore, the swing equation for the dynamic frequency response of the power system is established as follows:

[0008] Where: H is the total inertia coefficient of the system; The system's rated frequency; This refers to the system frequency deviation. Index for time period; This is the load adjustment coefficient; The power deficit of the power system during time period t; It provides primary frequency regulation output for thermal power units.

[0009] Furthermore, the frequency modulation resource demand assessment model is expressed as follows:

[0010] In the formula: The frequency modulation resource requirements needed to address the power deficit during time period t; The static characteristic coefficient of the power frequency of the thermal power unit; This is the overall time constant; This represents the maximum allowable frequency deviation of the system. The power deficit of the system during time period t; This is the load adjustment coefficient; The time it takes for the system frequency to reach its lowest point; The frequency safety constraint of the maximum frequency deviation limit of the power system is expressed as follows:

[0011] In the formula: This refers to the system frequency deviation. This represents the maximum allowable frequency deviation of the system. It is an absolute value.

[0012] Furthermore, the formula used to obtain the daily frequency regulation resource supply is as follows:

[0013] In the formula: The frequency regulation resource supply of the system during time period t; For the index of hydroelectric generator units; This represents the total number of hydropower units. i For the index number of the thermal power unit; N This represents the total number of thermal power units. This represents the primary frequency regulation capacity that thermal power unit i can provide during time period t. The primary frequency regulation capacity that hydropower unit h can provide during time period t; The primary frequency regulation capacity that energy storage can provide during time period t; in,

[0014] In the formula: Let i be the primary frequency regulation capacity coefficient of thermal power unit i; This represents the maximum output of thermal power unit i. This represents the output of the thermal power unit during time period t.

[0015] In the formula: The maximum output of the hydropower unit h; The output of the hydropower unit h during time period t; Let h be the power-frequency static characteristic coefficient of the hydroelectric generator unit;

[0016] In the formula: The charging power of energy storage during time period t; Let be the discharge power during time period t.

[0017] Furthermore, the method used to obtain the comprehensive clustering index is as follows: S41: Based on the daily frequency modulation resource demand and the daily frequency modulation resource supply, obtain the direction angle of the two-dimensional data point composed of the daily frequency modulation resource demand and the daily frequency modulation resource supply. S42: Obtain the frequency modulation resource supply-demand ratio deviation angle based on the direction angle of the two-dimensional data points; S43: Obtain the comprehensive clustering index based on the deviation angle of the supply and demand ratio of frequency modulation resources.

[0018] Furthermore, the formula used to obtain the direction angle of the two-dimensional data point composed of the frequency modulation resource demand and the frequency modulation resource supply on that day is as follows:

[0019] In the formula: The direction angle of a two-dimensional data point is formed by the frequency modulation resource demand and the frequency modulation resource supply on that day.

[0020] Furthermore, the formula used to obtain the frequency modulation resource supply-demand ratio deviation angle is as follows:

[0021] in, The deviation angle between the supply and demand ratio of frequency modulation resources; For the first j The orientation angle of each sample point; For the first k The orientation angle of each sample point; j , k All of these are indexes of sample points.

[0022] Furthermore, the formula used to obtain the comprehensive clustering index is as follows:

[0023] In the formula: For the first j The sample point and the first k A comprehensive clustering index among individual sample points; Euclidean distance weighting coefficient; The weighting coefficient for the supply-demand ratio deviation of frequency modulation resources; For the first j The sample point and the first k Euclidean distance between sample points; The maximum Euclidean distance between sample points; This represents the maximum angular deviation between sample points.

[0024] Compared with the prior art, the present invention has the following beneficial effects: This invention discloses a method for classifying typical operating modes of a power system based on the supply-demand ratio of frequency regulation resources. It obtains the daily maximum power deficit by combining the load, wind power, and photovoltaic power curves of the local power system over historical years. A frequency regulation resource demand assessment model is constructed based on the system's dynamic frequency response characteristics and the frequency safety constraint of the power system's maximum frequency deviation limit. Simultaneously, a unit combination optimization model considering the primary frequency regulation capacity of thermal power, hydropower, and energy storage is established to accurately quantify the daily supply and demand data of frequency regulation resources throughout the year. Furthermore, a comprehensive clustering index is constructed and used as a similarity metric to conduct daily frequency regulation resource clustering analysis throughout the year. This achieves accurate classification of typical days in the power system for the year and completes the clustering analysis of typical days in the power system for the year, providing technical support for the efficient allocation of frequency regulation in the future power system. This invention addresses the frequency security requirements of new power systems with a high proportion of renewable energy access. It integrates the supply-demand ratio of frequency regulation resources as a core indicator into the clustering process, overcoming the shortcomings of traditional typical day classification methods that rely solely on load and renewable energy output and fail to reflect the sufficiency of frequency regulation resources. By conducting quantitative assessment of frequency regulation resource supply and demand and cluster analysis step by step, it reduces the difficulty of solving multi-dimensional parameter fusion analysis and enables the typical days to correspond to three frequency regulation resource modes: sufficient, insufficient, and critically balanced. Furthermore, the samples within the clusters exhibit good temporal continuity. The CH index, DB index, and average deviation angle index of the frequency regulation resource supply-demand ratio of the clustering results are all superior to traditional clustering methods, providing reliable technical support for the optimal allocation of frequency regulation resources in power systems. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a flowchart of the annual typical day division method for power systems that considers the supply and demand ratio of frequency regulation resources according to the present invention; Figure 2 This is a schematic diagram of the frequency response process in an embodiment of the present invention; Figure 3 This is a schematic diagram of scene clustering results in an embodiment of the present invention; Figure 4 This is a calendar diagram showing the annual frequency regulation supply and demand resource distribution in an embodiment of the present invention. Detailed Implementation

[0027] 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 only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] This embodiment introduces a method for classifying typical operating modes of a power system based on the supply-demand ratio of frequency regulation resources, such as... Figure 1 As shown, it includes the following steps: S1: Obtain the historical annual load power curve, wind power power curve, and photovoltaic power curve of the local power system to obtain the daily maximum power deficit; Specifically, the system collects historical power time-series curves of load, wind power, and photovoltaic power for the entire 365 days of the provincial power system, performs fluctuation analysis on the load data for each time period of each day, and extracts the maximum upward fluctuation of the load during time period t. Simultaneously, obtain the maximum downward change in wind power output and photovoltaic power output during time period t. , .

[0029] Preferably, the formula used to obtain the daily maximum power deficit is as follows: (1) In the formula: The power deficit of the power system during time period t; The maximum upward fluctuation of the load during time period t; The maximum downward change in wind power output during time period t; This represents the maximum downward change in photovoltaic power output during time period t.

[0030] Specifically, the power deficit for each time period of each day throughout the year is calculated, and the maximum value of the daily power deficit is selected as the daily maximum power deficit for that day, forming a dataset of daily maximum power deficit for 365 days throughout the year.

[0031] S2: Based on the stated daily maximum power deficit, establish the swing equation of the power system's dynamic frequency response. Using the frequency security constraint based on the power system's maximum frequency deviation limit, establish a frequency regulation resource demand assessment model to obtain the daily frequency regulation resource demand. ; This embodiment constructs a frequency regulation resource demand assessment model based on the system's dynamic frequency response characteristics and the frequency minimum point constraint to obtain the daily frequency regulation resource demand. Specifically, the system's dynamic frequency response characteristics refer to the dynamic process of system frequency change over time when a power deficit occurs in the power system, which is described in this embodiment using an oscillation equation. The frequency minimum point constraint means that the power system frequency deviation must not exceed the maximum allowable value, which is achieved in this embodiment through frequency safety constraints. These two characteristics together serve as the basis for constructing the frequency regulation resource demand assessment model.

[0032] Specifically, the total inertia coefficient of the power system is first determined based on its structural parameters. H Load regulation coefficient D, power frequency static characteristic coefficient of thermal power unit With the comprehensive time constant Basic parameters, combined with daily maximum power deficit The frequency change process under power deficit is described by the oscillation equation of the dynamic frequency response of the power system, such as... Figure 2 : (2) Where: H is the total inertia coefficient of the system; The system's rated frequency; This refers to the system frequency deviation. Index for time period; This is the load adjustment coefficient; The power deficit of the power system during time period t; For primary frequency regulation output of thermal power units; Preferably, the frequency security constraint of the maximum frequency deviation limit of the power system is expressed as follows: (3) In the formula: This refers to the system frequency deviation. This represents the maximum allowable frequency deviation of the system. It is the absolute value; Preferably, the frequency modulation resource demand assessment model is expressed as follows: (4) In the formula: The frequency modulation resource requirements needed to address the power deficit during time period t; The static characteristic coefficient of the power frequency of the thermal power unit; This is the overall time constant; This represents the maximum allowable frequency deviation of the system. The power deficit of the system during time period t; This is the load adjustment coefficient; The time it takes for the system frequency to reach its lowest point; Specifically, substitute the system's rated frequency. The required frequency regulation resources for each time period of each day throughout the year are calculated, and the maximum value of the daily frequency regulation resource requirements is selected as the frequency regulation resource requirements for that day. Complete the quantitative calculation of frequency regulation resource demand for 365 days a year.

[0033] S3: With the goal of minimizing system operating costs, establish a unit combination model that considers the primary frequency regulation capacity of multiple resources such as thermal power, hydropower, and energy storage to obtain the daily frequency regulation resource supply. Specifically, the first step is to construct a unit combination model that considers the primary frequency regulation capacity of multiple resources, including thermal power, hydropower, and energy storage, with the objective of minimizing the total power generation cost of the power system. (5) In the formula: C This represents the total power generation cost of the power system. T This represents the total number of time periods per day. , , All are thermal power units i The power generation cost coefficient; This represents the output of the thermal power unit during time period t. For thermal power units i The startup cost during time period t; For thermal power units i The start / stop status; The charging and discharging power of energy storage units and their operation and maintenance costs; The output of the energy storage unit during time period t; It is the absolute value; Secondly, constraints on unit combination are set, including system power balance constraints, upper and lower limits of thermal power unit output, ramp rate, start-up and shutdown time constraints, output-related constraints of hydropower units, and energy storage charging and discharging power and energy status constraints, etc.; then, frequency regulation capacity evaluation models are established for thermal power, hydropower, and energy storage respectively. The frequency regulation capacity of thermal power units must meet the constraints of remaining available capacity and primary frequency regulation capacity coefficient. (6) In the formula: Let i be the primary frequency regulation capacity coefficient of thermal power unit i; This represents the maximum output of thermal power unit i. This represents the output of the thermal power unit during time period t. The frequency regulation capacity of hydropower units must meet the constraints of remaining output capacity and power frequency static characteristic coefficient: (7) In the formula: The maximum output of the hydropower unit h; The output of the hydropower unit h during time period t; Let h be the power-frequency static characteristic coefficient of the hydroelectric generator unit; Energy storage frequency regulation capacity is constrained by charge / discharge state and maximum charge / discharge power: (8) In the formula: The charging power of energy storage during time period t; The discharge power during time period t; Finally, the unit combination optimization model is solved to obtain the output and frequency regulation capacity data of thermal power, hydropower, and energy storage for each time period of the year. The frequency regulation capacity of each type of resource in each time period is summed to obtain the system frequency regulation resource supply for each time period of the day. (9) In the formula: The frequency regulation resource supply of the system during time period t; For the index of hydroelectric generator units; This represents the total number of hydropower units. i For the index number of the thermal power unit; N This represents the total number of thermal power units. This represents the primary frequency regulation capacity that thermal power unit i can provide during time period t. The primary frequency regulation capacity that hydropower unit h can provide during time period t; The primary frequency regulation capacity that energy storage can provide during time period t; Specifically, the maximum value among the daily frequency regulation resource supply is selected as the frequency regulation resource supply for that day. This yields a dataset of frequency modulation resource supply for all 365 days of the year.

[0034] S4: Based on daily frequency regulation resource requirements and daily frequency regulation resource supply The direction angle of two-dimensional data points (sample points) consisting of daily frequency modulation resource demand and daily frequency modulation resource supply is obtained, and the difference between the direction angles of two sample points is defined as the frequency modulation resource supply-demand ratio deviation angle. The frequency modulation resource supply-demand ratio deviation angle is obtained, and then combined with Euclidean distance to obtain a comprehensive clustering index. Preferably, the method used to obtain the comprehensive clustering index is as follows: S41: Based on daily frequency regulation resource requirements and daily frequency regulation resource supply Obtain the direction angle of a two-dimensional data point composed of the frequency modulation resource demand and the frequency modulation resource supply for the day; Specifically, the daily frequency regulation resource demand for 365 days a year. As the horizontal axis, supply volume Using the vertical axis, a two-dimensional sample space is constructed. Each two-dimensional data point (sample point) corresponds to a day's supply and demand data, calculated according to the direction angle formula: (10) In the formula: The directional angle of a two-dimensional data point is formed by the frequency modulation resource demand and the frequency modulation resource supply on that day. S42: Obtain the frequency modulation resource supply-demand ratio deviation angle based on the direction angle of the two-dimensional data points; Specifically, the angle between the line connecting each sample point to the origin and the horizontal axis (i.e., the direction angle) is calculated to obtain the direction angles of all sample points; the absolute value of the difference between the direction angles of any two sample points is defined as the frequency modulation resource supply-demand ratio deviation angle as follows: (11) in, The deviation angle between the supply and demand ratio of frequency modulation resources; For the first j The orientation angle of each sample point; For the first k The orientation angle of each sample point; j , k All are indexes of sample points; S43: Obtain the comprehensive clustering index based on the deviation angle of the supply and demand ratio of frequency modulation resources; Simultaneously calculate the Euclidean distance between sample points. Introducing Euclidean distance weighting coefficients And frequency modulation resource supply and demand ratio deviation angle weighting coefficient The two indicators are weighted and fused to construct a comprehensive clustering indicator: (12) In the formula: For the first j The sample point and the first k A comprehensive clustering index among individual sample points; Euclidean distance weighting coefficient; The weighting coefficient for the supply-demand ratio deviation of frequency modulation resources; For the first j The sample point and the first k Euclidean distance between sample points; The maximum Euclidean distance between sample points; This represents the maximum angular deviation between sample points; Specifically, the comprehensive clustering index serves as the basis for measuring sample similarity in subsequent clustering analysis, and the weighting coefficients can be calibrated according to the actual needs of power system frequency regulation resource supply and demand analysis.

[0035] S5: Construct supply-demand sample pairs based on the daily demand and supply of frequency modulation (FM) resources, and cluster the supply-demand sample pairs using the K-means clustering method according to the comprehensive clustering index to obtain multiple clusters to determine typical days of the year; and obtain the average supply-demand ratio of the clusters to determine the FM resource mode of the supply-demand sample pairs within the clusters, including three modes: sufficient FM resources, insufficient FM resources, and critical balance of FM resources.

[0036] Among them, the supply and demand sample pairs corresponding to the cluster centers of the clusters are the process sample pairs of the typical days of the year, which can be used to determine the typical days of the year.

[0037] Using the frequency regulation resource demand required to address the daily power deficit throughout the year and the daily frequency regulation resource supply throughout the year as sample characteristics, and using a comprehensive clustering index as a similarity measure, we conduct cluster analysis on the daily frequency regulation resources throughout the year to obtain typical days of the year and promote the optimal allocation of frequency regulation resources in the new power system.

[0038] Specifically, the daily frequency regulation resource supply and daily frequency regulation resource demand Using the comprehensive clustering index as a similarity measure, a K-means clustering method incorporating supply and demand ratios was employed to conduct cluster analysis on 365 sample points throughout the year. First, k cluster centers were initialized. Through multiple iterative calculations, the cluster centers were updated with the goal of minimizing the comprehensive clustering index until the cluster centers converged, determining the optimal number of clusters and the final cluster centers. Sample points with high similarity were clustered into one class to obtain the clustering results for the supply and demand of frequency regulation resources throughout the year. The supply and demand data corresponding to the cluster centers of each cluster were used as the supply and demand data for typical days of that cluster throughout the year. The number of samples contained in each cluster was counted, and its proportion probability throughout the year was calculated. Simultaneously, the average supply and demand ratio of each cluster was analyzed. (13) In the formula: The average supply-demand ratio for the g-th cluster; This is the index of the samples contained in the g-th cluster; The total number of samples contained in the g-th cluster; For the g-th cluster, the th Frequency modulation resource requirements for each sample; For the g-th cluster, the th Frequency modulation resource supply per sample; Based on the average supply-demand ratio, each cluster is divided into three modes: sufficient frequency regulation resources, insufficient frequency regulation resources, and critical balance of frequency regulation resources, providing data support and decision-making basis for the annual optimization allocation of frequency regulation resources in the power system.

[0039] A specific implementation example of the present invention is as follows: This analysis uses a provincial power grid in my country as an example. The example includes a total installed capacity of 32,680 MW for thermal power units, 14,286 MW for wind power, 9,577 MW for photovoltaic power, 3,505 MW for hydropower, and 1,350 MW for energy storage. Historical data is used for both renewable energy output and load power (including wind and photovoltaic). The following parameters are selected for system frequency analysis: primary frequency regulation capability coefficient of thermal power units is 10%, inertia time constant of motor load is 0.5, system rated frequency is 50 Hz, maximum frequency deviation limit is 0.2 Hz, and frequency dead zone is set to 0.033 Hz.

[0040] A comprehensive clustering index was constructed based on Euclidean distance and the supply-demand ratio of frequency modulation resources. A K-means clustering method incorporating the supply-demand ratio was used to conduct cluster analysis on a yearly sample set of 365 samples. The K-means clustering method was used to determine the cluster centers and the number of clusters. All sample points were clustered, and the clustering results and the cluster centers of each cluster were obtained as follows: Figure 3 As shown in Table 1, the cluster centers of each cluster are used as typical days, and the typical days and their probabilities in the clusters are divided based on the characteristics of frequency modulation resource supply and demand as sample points.

[0041] Based on the proposed K-means clustering method incorporating supply-demand ratios, the 365 days of the year are clustered into 8 clusters. These clusters can then be further categorized into three modes according to their supply-demand ratios: FM resource sufficiency mode, FM resource insufficiency mode, and FM resource critical balance mode. A FM resource supply-demand ratio greater than 1 within a cluster indicates FM resource sufficiency; less than 1 indicates FM resource insufficiency; and close to 1 indicates FM resource critical balance. "Close to 1" means that sample points within a cluster are distributed close to both sides of a line where the supply-demand ratio is 1. For example... Figure 3 Cluster 2 is a middle cluster, and the average supply-demand ratio of all sample points in this cluster is 1.075.

[0042] The basis for classifying each cluster into three modes is: classification based on the degree of matching between the supply and demand of frequency regulation resources. Different modes reflect the sufficiency of frequency regulation resources in the power system.

[0043] (1) Frequency modulation resource sufficient mode The frequency regulation (FM) resource sufficiency mode refers to a mode where the supply of FM resources significantly exceeds the demand, corresponding to clusters 3, 4, 5, 6, 7, and 8. The sample dates included in this mode are mainly distributed in spring and autumn, when load levels are low and relatively stable. The FM resource supply significantly exceeds demand, and this mode applies to the vast majority of days throughout the year, indicating a sufficient FM supply for the system.

[0044] Table 1. Clustering Results and Probabilities of Frequency Modulation Resource Supply and Demand for 365 Days of the Year

[0045] (2) Frequency modulation resource shortage mode Cluster 1 falls under the frequency regulation resource shortage mode. Due to significant heating demand during winter cold waves in the province, the supply of frequency regulation resources is scarce, and the grid's frequency regulation demand is also high. Under this mode, the supply of frequency regulation resources is consistently significantly lower than the demand. In late July, due to rising summer temperatures, the load on air conditioning and other appliances increases sharply, and wind power occasionally experiences extreme heat and no wind, leading to a significant increase in the demand for frequency regulation resources. Under this mode, the frequency regulation resources of new energy sources and energy storage should be fully utilized to increase the supply of frequency regulation resources and prevent the system frequency from falling below the limit. It is necessary to allocate frequency regulation resources such as energy storage and new energy sources in advance at the annual level to increase the configuration capacity of frequency regulation resources under this mode.

[0046] (3) Frequency modulation resource critical balance mode Cluster 2 represents a critical balance mode for frequency regulation resources. This mode is mainly distributed in winter, where the supply and demand of frequency regulation resources are very close due to the randomness of load and renewable energy fluctuations. Under this mode, the power grid needs to flexibly allocate various resources and make full use of frequency regulation resources with fast response speeds, such as energy storage, to cope with frequency fluctuations caused by factors such as changes in power sources and loads, and to avoid the system frequency falling below the limit.

[0047] Table 2 shows a comparison of the CH and DB indices for different numbers of clusters. When the number of clusters is 8, the CH index reaches its maximum value of 463.1 and the DB index reaches its minimum value of 0.994. Therefore, the optimal number of clusters for the sample set composed of the supply and demand of frequency regulation resources for 365 days throughout the year is 8.

[0048] Based on the clustering effectiveness index, the proposed K-means clustering algorithm with supply-demand ratio is compared with the K-means algorithm. When the number of clusters is 8, the CH index, DB index, and average deviation angle index of the frequency modulation resource supply-demand ratio of each cluster obtained by different clustering methods are shown in Table 3. It can be seen from the three indices that the proposed K-means clustering algorithm with supply-demand ratio is superior to other methods in all aspects, proving that the clustering method proposed in this embodiment has certain advantages.

[0049] Table 2 Relationship between effectiveness indicators and cluster number

[0050] Table 3 Comparison of effectiveness metrics for different clustering algorithms

[0051] To address the issue of defining typical days for frequency security in new power systems, a method for assessing the demand for frequency regulation resources based on the dynamic frequency response characteristics of the power system is proposed. Furthermore, a modeling and assessment of the supply of frequency regulation resources is conducted based on unit combination simulation. A K-means clustering method incorporating the supply-demand ratio deviation angle and Euclidean distance of frequency regulation resources is then proposed. Using the supply and demand of frequency regulation resources as sample point features, a clustering analysis of typical days throughout the year in the power system is completed. Based on the supply and demand of frequency regulation resources, the sample set of the power system for the entire 365 days of the year is clustered into 8 clusters, with the cluster center of each cluster serving as a typical day. Figure 4 Analysis revealed that the sample points within each cluster exhibited good temporal continuity. Based on the clustering method using the supply-demand ratio of frequency modulation (FM) resources, the sufficiency of FM resources in each cluster can be reflected, including sufficiency, insufficiency, and critical balance modes. The proposed clustering method fully considers the deviation angle of the FM resource supply-demand ratio during the clustering process. Compared with existing clustering methods, the proposed K-means clustering method with supply-demand ratio yields the best CH index, DB index, and average deviation angle index of the FM resource supply-demand ratio, verifying the superiority of the proposed clustering method in terms of clustering quality and effectiveness, and demonstrating significant practical value.

[0052] This embodiment presents a method for classifying typical operating modes of a power system based on the supply-demand ratio of frequency regulation resources. First, it obtains the historical annual load, wind power, and photovoltaic power curves of the power system to determine the daily maximum power deficit. Then, based on the system's dynamic frequency response characteristics and the constraint of the lowest frequency point, it constructs a frequency regulation resource demand assessment model. Simultaneously, with the goal of minimizing system operating costs, it establishes a unit combination model containing primary frequency regulation capacity for various resources, calculating the daily demand and supply of frequency regulation resources throughout the year. Subsequently, it calculates the directional angle based on supply and demand and defines the frequency regulation resource supply-demand ratio deviation angle, constructing a comprehensive clustering index using Euclidean distance. Finally, using frequency regulation resource supply and demand as sample features and the comprehensive clustering index as a similarity measure, it employs a K-means clustering method incorporating the supply-demand ratio to conduct annual daily frequency regulation resource clustering analysis, obtaining typical days for the year. This invention integrates the frequency regulation resource supply-demand ratio into the clustering process, overcoming the deficiency of traditional methods in reflecting the sufficiency of frequency regulation resources. The classified typical days correspond to three modes: sufficient, insufficient, and critically balanced frequency regulation resources. The clustering index performs better, providing reliable technical support for the optimized allocation of frequency regulation resources in new power systems.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; 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 or all of the technical features in the formula; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for classifying typical operating modes of a power system based on the supply-demand ratio of frequency regulation resources, characterized in that, Includes the following steps: S1: Obtain the historical annual load power curve, wind power power curve, and photovoltaic power curve of the local power system to obtain the daily maximum power deficit; S2: Based on the daily maximum power deficit, establish the swing equation of the dynamic frequency response of the power system, and establish a frequency regulation resource demand assessment model based on the frequency security constraint of the maximum frequency deviation limit of the power system to obtain the daily frequency regulation resource demand. S3: Establish a unit combination model that considers the primary frequency regulation capacity of multiple resources, including thermal power, hydropower, and energy storage, in order to obtain the daily frequency regulation resource supply. S4: Based on the daily frequency modulation resource demand and the daily frequency modulation resource supply, obtain the orientation angle of the two-dimensional data points composed of the daily frequency modulation resource demand and the daily frequency modulation resource supply to obtain the frequency modulation resource supply-demand ratio deviation angle, and then obtain the comprehensive clustering index. S5: Construct supply and demand sample pairs based on the daily frequency regulation resource demand and the daily frequency regulation resource supply, and use the K-means clustering method to cluster the supply and demand sample pairs according to the comprehensive clustering index to obtain multiple clusters to determine the typical days of the year; The average supply-demand ratio of the clusters is obtained to determine the frequency regulation resource pattern of the supply-demand sample pairs within the clusters, thereby providing technical support for the efficient allocation of frequency regulation in the future power system.

2. The method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources according to claim 1, characterized in that, The formula used to obtain the daily maximum power deficit is as follows: In the formula: The power deficit of the power system during time period t; The maximum upward fluctuation of the load during time period t; The maximum downward change in wind power output during time period t; This represents the maximum downward change in photovoltaic power output during time period t.

3. The method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources according to claim 1, characterized in that, The swing equation for the dynamic frequency response of the power system is established as follows: Where: H is the total inertia coefficient of the system; The system's rated frequency; This refers to the system frequency deviation. Index for the time period; This is the load adjustment coefficient; The power deficit of the power system during time period t; It provides primary frequency regulation output for thermal power units.

4. The method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources according to claim 1, characterized in that, The frequency modulation resource demand assessment model is expressed as follows: In the formula: The frequency modulation resource requirements needed to address the power deficit during time period t; The static characteristic coefficient of the power frequency of the thermal power unit; This is the overall time constant; This represents the maximum allowable frequency deviation of the system. The power deficit of the system during time period t; This is the load adjustment coefficient; The time it takes for the system frequency to reach its lowest point; The frequency safety constraint of the maximum frequency deviation limit of the power system is expressed as follows: In the formula: This refers to the system frequency deviation. This represents the maximum allowable frequency deviation of the system. It is an absolute value.

5. The method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources according to claim 1, characterized in that, The formula used to obtain the daily frequency regulation resource supply is as follows: In the formula: The frequency regulation resource supply of the system during time period t; For the index of hydroelectric generator units; This represents the total number of hydropower units. i For the index number of the thermal power unit; N This represents the total number of thermal power units. The primary frequency regulation capacity that thermal power unit i can provide during time period t; The primary frequency regulation capacity that hydropower unit h can provide during time period t; The primary frequency regulation capacity that energy storage can provide during time period t; in, In the formula: Let i be the primary frequency regulation capacity coefficient of thermal power unit i; This represents the maximum output of thermal power unit i. This represents the output of the thermal power unit during time period t. In the formula: The maximum output of the hydropower unit h; The output of the hydropower unit h during time period t; Let h be the power-frequency static characteristic coefficient of the hydropower unit; In the formula: The charging power of energy storage during time period t; Let be the discharge power during time period t.

6. The method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources according to claim 1, characterized in that, The method used to obtain the comprehensive clustering index is as follows: S41: Based on the daily frequency modulation resource demand and the daily frequency modulation resource supply, obtain the direction angle of the two-dimensional data point composed of the daily frequency modulation resource demand and the daily frequency modulation resource supply. S42: Obtain the frequency modulation resource supply-demand ratio deviation angle based on the direction angle of the two-dimensional data points; S43: Obtain the comprehensive clustering index based on the deviation angle of the supply and demand ratio of frequency modulation resources.

7. The method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources according to claim 6, characterized in that, The formula used to obtain the direction angle of the two-dimensional data point composed of the frequency modulation resource demand and the frequency modulation resource supply on the same day is as follows: In the formula: The direction angle of a two-dimensional data point is formed by the frequency modulation resource demand and the frequency modulation resource supply on that day.

8. The method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources according to claim 7, characterized in that, The formula used to obtain the frequency modulation resource supply-demand ratio deviation angle is as follows: in, The deviation angle between the supply and demand ratio of frequency modulation resources; For the first j The orientation angle of each sample point; For the first k The orientation angle of each sample point; j , k All of these are indexes of sample points.

9. The method for classifying typical operation modes of a power system based on the supply-demand ratio of frequency regulation resources according to claim 8, characterized in that, The formula used to obtain the comprehensive clustering index is as follows: In the formula: For the first j The sample point and the first k A comprehensive clustering index among individual sample points; Euclidean distance weighting coefficient; The bias angle weighting coefficient for the supply-demand ratio of frequency modulation resources; For the first j The sample point and the first k Euclidean distance between sample points; The maximum Euclidean distance between sample points; This represents the maximum angular deviation between sample points.

Images