A flexible resource adaptive allocation method based on multi-scenario fusion

Abstract

This invention relates to a flexible resource adaptive allocation method based on multi-scenario fusion, comprising: obtaining predicted values ​​of wind power, photovoltaic power, and load power of a regional power grid; constructing a scenario identification objective function with the goal of minimizing the weighted sum of flexibility characteristic indicators for power balance, peak shaving, and frequency regulation scenarios; solving the scenario identification objective function based on scenario identification constraints to obtain a flexibility characteristic indicator vector; determining the scenario identification result based on the flexibility characteristic indicator vector; if the scenario is one or more of power balance, peak shaving, and frequency regulation, constructing a two-stage adaptive allocation model with the goal of minimizing the capacity allocation of flexible resources on each side of the power source, grid, load, and storage, as well as the control losses of resources on each side; solving for the flexible resource allocation scale based on the scenario identification result and the first-stage constraints; and solving for the flexible resource operation plan based on the predicted values ​​of wind power, photovoltaic power, and load power, the newly added flexible resource allocation scale, and the second-stage deterministic and uncertain constraints.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and in particular to a flexible resource adaptive configuration method based on multi-scenario fusion. Background Technology

[0002] As the proportion of new energy sources in the power system increases year by year, the demand for stability and flexibility of the power system is also rising. New energy power generation exhibits significant volatility and uncontrollability, posing a greater challenge to the reliability of power supply. Improving the flexibility and regulation capabilities of the power system to adapt to the derivative characteristics brought about by the volatility of new energy sources is an urgent problem to be solved. The medium- and long-term seasonal fluctuations, as well as the short-term random and intermittent fluctuations of new energy power generation, necessitate dynamic optimization of the power system across time scales such as days, months, quarters, and even years. The volatility of new energy sources at each time scale gives rise to different demand scenarios, and single-dimensional regulation methods are insufficient to meet the flexibility requirements of multiple scenarios. Therefore, facing the problem of coupled demands across multiple scenarios, constructing a scenario identification mechanism and formulating differentiated configuration strategies are particularly important.

[0003] Current research on the optimization of flexible resource allocation largely focuses on single-scale or single-demand scenarios, lacking differentiated allocation methods for different scenarios. In multi-scenario integration, how to coordinate the flexible adjustment resources of the power generation, grid, load, and storage sides across different time scales to achieve more comprehensive collaborative optimization still requires in-depth research. Furthermore, in complex power grid environments where multiple adjustable resources participate, there are complex coupling relationships between these resources. Most current research focuses on the independent optimization of single resources, lacking collaborative coupling optimization mechanisms between resources, particularly in the overall optimization control under large-scale distributed adjustable resources. Summary of the Invention

[0004] Based on the above analysis, the embodiments of the present invention aim to provide a flexible resource adaptive configuration method based on multi-scenario fusion, in order to solve the technical problems that existing methods mostly focus on the optimization of a single time scale or scenario, lack research on differentiated configuration methods under multi-scenario fusion, and emphasize the independent optimization of a single resource without fully considering the synergistic coupling between multiple adjustable resources, making it difficult to meet the flexibility adjustment requirements of the power system under a high proportion of new energy access.

[0005] The objective of this invention is mainly achieved through the following technical solutions:

[0006] This invention provides a flexible resource adaptive configuration method based on multi-scenario fusion, comprising the following steps:

[0007] Obtain forecasts of wind power, solar power, and load power in the regional power grid;

[0008] A scenario identification objective function is constructed with the goal of minimizing the weighted sum of flexibility characteristic indices for power balance, peak shaving, and frequency regulation scenarios. A flexibility characteristic index vector is obtained by solving the scenario identification objective function based on scenario identification constraints. The corresponding scenario identification result is obtained by judging based on the flexibility characteristic index vector.

[0009] If the scenario identification result is one or more of the power balance scenario, peak shaving scenario, and frequency regulation scenario, a two-stage adaptive configuration model is constructed with the goal of minimizing the capacity configuration of flexible resources on each side of the source, grid, load, and storage, and minimizing the control loss of flexible resources on each side; the scale of flexible resource configuration is obtained by solving based on the scenario identification result and the first stage constraints.

[0010] Based on the predicted values ​​of wind power, photovoltaic power, and load power, the scale of newly added flexible resource allocation, and the deterministic and uncertain constraints of the second stage, the flexible resource operation plan corresponding to the scale of flexible resource allocation under the scenario is obtained.

[0011] Furthermore, it also includes constructing a source-load model; and performing data fitting and solving on the source-load model based on historical power data of wind power, photovoltaic power, and various types of users in the regional power grid to obtain predicted values ​​of wind power, photovoltaic power, and load power, including:

[0012] Obtain load sample data of wind power, photovoltaic and various types of users in the historical period of the regional power grid, and preprocess it to obtain preprocessed load sample data of wind power, photovoltaic and various types of users;

[0013] Using SVM and cross-validation, the source-load model is fitted and solved based on the load sample data of wind power, photovoltaics, and various types of users to obtain the predicted values ​​of wind power, photovoltaics, and load power.

[0014] Furthermore, the source-load model includes wind power, photovoltaic power, and load power models;

[0015] The wind power model is fitted and solved using the wind power sample data to obtain the predicted wind power value.

[0016] The photovoltaic power model is fitted and solved using the photovoltaic sample data to obtain the predicted photovoltaic power value.

[0017] The load power model is fitted and solved using load sample data of each type of user to obtain the load power prediction value corresponding to each type of user; the load power prediction value of each type of user is integrated with the historical load average to obtain the load power prediction value.

[0018] Furthermore, the scene recognition objective function is expressed as follows:

[0019] ;

[0020] in, , , These are the flexibility characteristics indicators for power balance, peak shaving, and frequency regulation scenarios, respectively. , , They are respectively , , The weights; , These represent the maximum and minimum output power of the flexibility resources required for the scenario; , These are the rated values ​​for the maximum and minimum output power, respectively. , These represent the upward and downward ramp rates of the flexibility resources required by the scenario, respectively. , These represent the maximum upward and downward ramp rates of the flexibility resources, respectively. , These are the indicators for the standard frequency modulation capability and the fast frequency modulation capability of the flexible resources required by the scenario. These refer to the upper and lower backup adjustment capabilities of the conventional frequency modulation capability, respectively.

[0021] The scenario identification constraints include conventional power balance constraints, conventional non-adjustable unit operation constraints, wind power generation constraints, transmission line constraints, flexible adjustment resource operation constraints, full load constraints, spinning reserve capacity constraints, and fast frequency regulation reserve constraints.

[0022] Furthermore, based on the current power grid parameters, the scene recognition objective function is solved using the Gurobi solver based on the scene recognition constraints, resulting in... ;

[0023] Will The flexibility feature index vector constitutes the above. ,as follows:

[0024] ;

[0025] The corresponding scenario is determined based on the aforementioned flexibility feature index vector, including:

[0026] like or If the value is non-zero, a power balance scenario is determined to have occurred;

[0027] like or If the value is non-zero, then a peak-shaving scenario is determined to have occurred;

[0028] like or If the value is non-zero, then a frequency modulation scenario is determined to have occurred.

[0029] Furthermore, the two-stage adaptive configuration model includes a two-stage objective function that minimizes the capacity configuration of flexible resources on each side of the source, grid, load, and storage, as well as the control loss of flexible resources on each side, along with first-stage constraints and second-stage constraints.

[0030] The constraints in the first stage include configuration scale constraints, power demand constraints, ramp-up demand constraints, and frequency adjustment demand constraints.

[0031] The second stage constraints include second stage deterministic constraints and second stage uncertainty constraints;

[0032] The second phase of deterministic constraints includes source-side, network, load, and side flexibility resource constraints.

[0033] Furthermore, the two-stage objective function is as follows:

[0034] ;

[0035] in, , , , These are the unit capacity configuration parameters for source, grid, load, and storage resources, respectively. Scale allocation for various resources including power sources, grids, loads, and storage; , These are timescales for the entire year and typical days, respectively. , , , These are the control losses of resources on the source, grid, load, and storage sides, respectively. For various flexible resource categories including source, grid, load, and storage; Let be the set of load uncertainties.

[0036] Furthermore, the second stage uncertainty constraints ,as follows:

[0037] ;

[0038] in, Represents an uncertain variable; , , These represent the output power of photovoltaic and wind power, and the electrical load power, respectively, after considering uncertainties. , , These are the predicted values ​​for wind power, solar power, and load power, respectively. , , These are the maximum allowable fluctuation coefficients for wind power, photovoltaic power output, and electrical load power, respectively.

[0039] Furthermore, the scale of the flexible resource allocation This includes the scale of new flexible resources on the source, grid, load, and storage sides. , , and ;

[0040] The flexibility resource allocation scale corresponds to the flexibility resource operation plan, which includes the power adjustment of newly added source-side flexibility resources. Added grid-side flexibility resource adjustment power Newly added load-side resources can transfer load out. New load-side resources can transfer loads and interruptible load power. New power generation capacity from storage-side resources , Increased charging power of energy storage resources And newly added load-side resources can transfer loads. .

[0041] Furthermore, the power demand constraint is as follows:

[0042] ;

[0043] in, , , , These are the quantities of conventional non-adjustable generating units, external transmission lines, flexible regulating generating units, and loads. For the first Taiwan's conventional non-adjustable units at time contribution; For photovoltaic units in Efforts made at all times; For wind turbine units Efforts made at all times; For the first External transmission lines in Efforts made at all times; For flexible adjustment of units exist Efforts made at all times; Adjusting power for newly added flexibility resources; For load nodes exist The load demand at any given moment; for The load reduction power at any given time.

[0044] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0045] 1. Multi-scenario integration and optimization of differentiated configuration capabilities: This invention constructs a multi-scenario identification model (power balance scenario, peak shaving scenario, and frequency regulation scenario), comprehensively considers the flexibility characteristics of power balance scenario, peak shaving scenario, and frequency regulation scenario, solves the limitations of single-scenario optimization in the prior art, realizes differentiated resource adaptation under multi-scenario coupling, and significantly improves the dynamic response capability of the power system to new energy fluctuations (such as wind and solar intermittency and load mutation).

[0046] 2. Global Coordinated Optimization of Multiple Resource Types (Source, Grid, Load, and Storage): This invention designs a two-stage adaptive configuration model (capacity configuration + operation regulation), integrating the dynamic coupling relationship of resources on the source side (thermal power, energy storage), grid side (transmission lines), load side (interruptible / transferable loads), and storage side (energy storage). It breaks through the bottleneck of independent optimization of a single resource in traditional research, realizes full-cycle coordinated optimization across time scales and resource types, improves overall regulation efficiency, realizes coordinated optimization configuration of flexible resources on the source, grid, load, and storage sides, improves resource utilization efficiency, and overcomes the shortcomings of existing technologies that focus on independent optimization of a single resource.

[0047] 3. Dynamic Adaptation and Uncertainty Management: This invention introduces a two-stage adaptive configuration model. The first stage focuses on resource capacity configuration, while the second stage considers the dynamics and uncertainties of operation adjustment, making the resource configuration scheme more adaptable and robust, effectively addressing the uncertainties brought about by the volatility of new energy sources; ensuring the stable operation of the system under complex scenarios such as extreme weather and forecast deviations, and reducing the risk of wind and solar curtailment and load reduction.

[0048] 4. Multi-scenario identification and differentiated configuration: Based on the flexibility feature index vector (such as ramp rate and frequency regulation capability), the demand scenario is dynamically determined, and resources are configured differently according to different scenarios (one or more) to avoid the difficulty of a single configuration method to adapt to the ever-changing and complex power grid operation conditions.

[0049] 5. Improved accuracy of data-driven prediction and configuration: This invention uses SVM (Support Vector Machine) and cross-validation to fit wind power, photovoltaic and load power prediction models, and combines historical data and meteorological response regression analysis to significantly improve prediction accuracy, provide highly reliable input for multi-scenario identification and resource allocation, and support the refined operation of the power system.

[0050] In this invention, the above-described technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of this invention will be set forth in the following description, and some advantages may become apparent from the description or be learned by practicing the invention. The objects and other advantages of this invention can be realized and obtained from what is particularly pointed out in the description and drawings. Attached Figure Description

[0051] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts.

[0052] Figure 1 This is a flowchart of a flexible resource adaptive configuration method based on multi-scenario fusion in an embodiment of the present invention. Detailed Implementation

[0053] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0054] This invention proposes a flexible resource adaptive allocation method based on multi-scenario fusion, starting from various power interaction demand scenarios. First, it introduces multi-scenario characteristic indicators to quantitatively describe multi-dimensional demands such as power balance and renewable energy consumption, constructing a scenario demand identification model and establishing the coupling relationship between multiple scenario demands. Second, it aggregates the scenario identification results and filters out the most severe demand scenario. Finally, based on the flexibility requirements of the most severe scenario and considering the uncertainties at both the source and load ends, it constructs a two-stage adaptive allocation model for flexible resources. This method can solve the uncertainty problem caused by the volatility of renewable energy and, by dynamically adapting to derived scenario demands, achieve more efficient collaborative and differentiated allocation of flexible resources across multiple scenarios to cope with complex power interaction scenario demands.

[0055] A specific embodiment of the present invention discloses a flexible resource adaptive configuration method based on multi-scenario fusion, such as... Figure 1 As shown, it includes the following steps:

[0056] Step S1: Obtain the predicted values ​​of wind power, photovoltaic power and load power of the regional power grid;

[0057] Step S2: Construct a scenario identification objective function with the goal of minimizing the weighted sum of flexibility characteristic indices for power balance scenario, peak shaving scenario, and frequency modulation scenario; solve the scenario identification objective function based on scenario identification constraints to obtain a flexibility characteristic index vector; and make a judgment based on the flexibility characteristic index vector to obtain the corresponding scenario identification result.

[0058] Step S3: If the scenario identification result is one or more of the power balance scenario, peak shaving scenario, and frequency regulation scenario, construct a two-stage adaptive configuration model with the goal of minimizing the capacity configuration of flexible resources on each side of the source, grid, load, and storage and minimizing the control loss of flexible resources on each side; solve for the scale of flexible resource configuration based on the scenario identification result and the first stage constraints.

[0059] Step S4: Based on the predicted values ​​of wind power, photovoltaic power and load power, the scale of newly added flexible resource allocation, and the deterministic and uncertain constraints of the second stage, solve to obtain the flexible resource operation plan corresponding to the scale of flexible resource allocation in the scenario.

[0060] Step S1 includes steps S11-S13.

[0061] Step S11: Obtain historical data on wind power, photovoltaic power, and load power within the regional power grid.

[0062] Because wind and solar power output is highly dependent on meteorological conditions (such as wind speed, solar irradiance, and temperature), it exhibits significant nonlinearity, volatility, and uncertainty. Meanwhile, the load side displays clear daily and weekly cycles over time and is also affected by various external factors such as temperature, humidity, and holidays. Therefore, building a sensing module that considers wind and solar resources and load characteristics can enable advance mapping of available power sources and electricity demand in future periods, providing data support for resource allocation, scheduling strategy generation, and reserve capacity arrangement.

[0063] Based on historical wind and solar power output data and meteorological data, correlation coefficients between wind and solar power output and meteorological factors were calculated. The data showed that wind power output is related to wind speed, wind direction, and temperature; photovoltaic power output is related to total solar irradiance and temperature; and load power consumption also showed a positive correlation with changes in temperature, humidity, and precipitation. Based on these correlations, wind, solar, and load sample sets were constructed to obtain digital characteristics of the source and load.

[0064] By utilizing historical extreme weather load data and corresponding meteorological data within the region, sample data of wind power, photovoltaic power, and various types of user loads are obtained.

[0065] The wind power sample data is constructed as follows:

[0066] Formula (1)

[0067] in, This represents wind power sample data; , , They are respectively Wind speed, wind direction, and temperature at any given time; for Historical wind power output at any given time; This represents the total number of observation periods.

[0068] The photovoltaic sample data was constructed as follows:

[0069] Formula (2)

[0070] in, This represents photovoltaic sample data; , They are respectively Total irradiance and temperature at any given time; for Historical photovoltaic power output at any given time.

[0071] The load sample data is constructed as follows:

[0072] Formula (3)

[0073] in, This represents the load sample data; , , They are respectively Time relative to Changes in temperature, humidity, and precipitation over time; for Time relative to Historical load changes at any given time; User type.

[0074] Step S12 also includes constructing a source-load model; and performing data fitting and solving on the source-load model based on historical power data of wind power, photovoltaic power and various types of users in the regional power grid to obtain predicted values ​​of wind power, photovoltaic power and load power.

[0075] Based on the constructed sample data of regional power grid wind power, photovoltaic and load data and key meteorological variables, we will explore the response patterns and correlations of the sample data in the time series dimension, and establish the mapping relationship between wind power, photovoltaic and load power output and environmental factors.

[0076] Wind power output fluctuates primarily with changes in wind speed and direction; photovoltaic power output is affected by both irradiance and temperature; and load power output exhibits a superimposed response to changes in temperature, humidity, and precipitation.

[0077] By fitting these source load sample data, accurate predictions of wind and solar availability and load demand trends can be achieved, providing fundamental support for subsequent operation scenario identification, regulation demand characterization, and flexible resource allocation optimization.

[0078] A source-load model is constructed, and the model is solved by data fitting based on historical power data of wind power, photovoltaic power, and various types of users in the regional power grid, to obtain predicted values ​​of wind power, photovoltaic power, and load power, including:

[0079] Obtain load sample data of wind power, photovoltaic and various types of users in the historical period of the regional power grid, and preprocess it to obtain preprocessed load sample data of wind power, photovoltaic and various types of users;

[0080] Using SVM and cross-validation, the source-load model is fitted and solved based on the load sample data of wind power, photovoltaics, and various types of users to obtain the predicted values ​​of wind power, photovoltaics, and load power.

[0081] The source-load model includes wind power, photovoltaic power, and load power models;

[0082] The wind power model is fitted and solved using the wind power sample data to obtain the predicted wind power value.

[0083] The photovoltaic power model is fitted and solved using the photovoltaic sample data to obtain the predicted photovoltaic power value.

[0084] The load power model is fitted and solved using load sample data of each type of user to obtain the load power prediction value corresponding to each type of user; the load power prediction value of each type of user is integrated with the historical load average to obtain the load power prediction value.

[0085] Wind power model ,as follows:

[0086] Formula (4)

[0087] in, , , They are respectively The proportional coefficient of wind power increment caused by changes in meteorological factors such as wind speed, wind direction, and temperature at any given time; This represents the error term in the wind power fitting.

[0088] For example, The value ranges from 0.4 to 0.8; The value ranges from -0.1 to 0.1; The value is 0.01~0.05; The value can be between 0.1 and 0.3; it can be changed according to specific needs.

[0089] Photovoltaic power model ,as follows:

[0090] Formula (5)

[0091] in, , They represent The proportionality coefficient of photovoltaic power increment caused by changes in meteorological factors such as total irradiance and temperature at any given time; This represents the photovoltaic power fitting error term.

[0092] For example, The value ranges from 0.6 to 0.9; The value ranges from -0.002 to -0.005; The value can be between 0.05 and 0.2; it can be changed according to specific needs.

[0093] Load power model ,as follows:

[0094] Formula (6)

[0095] in, For the first time under extreme weather conditions User Total load at any given time; Assuming no extreme weather events occur, the first User Load forecast values ​​at any given time; The first is caused by meteorological factors such as temperature, humidity and precipitation. User The load increment at any given time is related to various meteorological factors such as temperature, humidity, and precipitation. , and They are respectively The actual changes in temperature, humidity, and precipitation under extreme weather conditions at any given time; , and These represent the changes in meteorological factors such as temperature, humidity, and precipitation, respectively. User The proportional coefficient of the increase in electricity consumption at any given time; ε is the error term, representing the consumption differences that cannot be reflected by changes in meteorological factors such as temperature, humidity, and precipitation.

[0096] For example,

[0097] (1) Temperature variation coefficient This indicates the effect of a unit temperature change (e.g., 1°C) on the load increment.

[0098] Residential users: 0.05~0.15 (for every 1°C increase in temperature, the air conditioning load increases by 5%~15%).

[0099] Commercial users: 0.03~0.10 (temperature sensitivity is slightly lower than that of residential users);

[0100] Industrial users: 0.01~0.05 (production equipment is not sensitive to temperature changes);

[0101] For special scenarios:

[0102] At extreme high temperatures (>35°C), It may increase to 0.2~0.3;

[0103] Heating demand is significant in cold regions (<0°C). It may be a negative value (e.g., -0.1 to -0.2).

[0104] (2) Humidity variation coefficient This indicates the effect of a unit change in humidity (e.g., 1%) on the load increment.

[0105] General range: 0.005~0.02 (for every 1% increase in humidity, the load increases slightly by 0.5%~2%).

[0106] High humidity areas: 0.01~0.03 (If dehumidification equipment is needed during the rainy season, the load will increase significantly).

[0107] (3) Precipitation variation coefficient This indicates the impact of a unit change in precipitation (e.g., 1 mm) on the load increment.

[0108] General range: 0.001~0.01 (for every 1mm increase in precipitation, the load increases slightly by 0.1%~1%).

[0109] Heavy rain scenario: The value may increase to 0.02~0.05 (e.g., due to a surge in demand for drainage pumps and emergency power).

[0110] Arid regions: May be negative (reduced rainfall leads to a decrease in agricultural irrigation load).

[0111] In this invention, extreme weather is exemplarily defined as follows: extreme weather occurs when any of the following events occur:

[0112] (1) Temperature: Temperatures exceeding the statistical range of local historical temperature data, such as high temperatures above 35°C or low temperatures below -10°C;

[0113] (2) Humidity: Relative humidity is above 80% or below 20% for an extended period of time;

[0114] (3) Precipitation: Rainfall exceeding 50 mm within 24 hours (e.g., heavy rain), or extreme precipitation events such as blizzards.

[0115] (4) Wind speed: When the wind speed reaches the storm range of 17–25 m / s, the wind turbine will frequently trigger overspeed protection or derating operation.

[0116] (5) Irradiance: Under the influence of rainstorms, thunderstorms, sandstorms, etc., the irradiance can drop sharply from 800–1000W / m² on sunny days to 100–300W / m², or even below 100W / m², accompanied by minute-level fluctuations of more than 500W / m², resulting in a "sawtooth" rapid drop in photovoltaic output.

[0117] For example, user load types include industrial, commercial, residential, agricultural, and public service users.

[0118] Due to differences in user load types (industrial, commercial, and residential), geographical location, climate, economy, industry, population, and other factors, the values ​​of coefficients and parameters in the formula vary.

[0119] Source and load digital feature acquisition, constructing wind and load sample data , , Establish a mapping relationship between power output and environmental factors.

[0120] Step S13: Train the source-load model and perform wind power, photovoltaic, and various types of load forecasting.

[0121] After constructing the source-load model, it is trained using an SVM (Support Vector Machine). During training, cross-validation is employed to select the optimal parameters. These trained parameters are then used to predict the source-load, ultimately yielding the predicted wind power output. Photovoltaic power forecast Forecasted load power .

[0122] (1) Preprocessing of wind power, photovoltaic and load sample data, including missing value filling and data normalization;

[0123] Missing value imputation: Missing values ​​are imputed using linear interpolation or the mean of adjacent time periods;

[0124] Data normalization:

[0125] Normalize the characteristics of different dimensions such as wind speed, irradiance, and temperature;

[0126] Load data is normalized according to user type to avoid differences in magnitude affecting model performance.

[0127] (2) Train the source load model.

[0128] Configure the SVM, setting the task type to regression; select the radial basis function (RBF) for the kernel function, which is suitable for nonlinearity.

[0129] Hyperparameters, including:

[0130] (a) Penalty coefficient The value range is [0.1, 100], which controls the model's tolerance to error;

[0131] (b) Nuclear parameters The width of the RBF kernel is controlled, with a value range of [0.001, 0.1].

[0132] (c) Insensitive loss function Set to 5% of the target dimension (e.g., in load forecasting). =0.05 × maximum load value).

[0133] k-fold cross-validation is used, where k equals 10, for example.

[0134] The sample dataset is divided into one subset as the validation set and the remaining nine subsets as the training set.

[0135] For each set of hyperparameter combinations ( , , ), calculate the average validation error (e.g., mean squared error, MSE); select the hyperparameter combination with the smallest validation error as the final model parameters.

[0136] The training process includes:

[0137] Input sample data as follows:

[0138] Wind power sample data: wind speed, wind direction, temperature , ,

[0139] Photovoltaic sample data: irradiance, temperature ,

[0140] Load sample data: changes in temperature, humidity, and precipitation , , Historical load variation and corresponding user types .

[0141] SVMs for wind power, solar power, and various types of users are trained separately:

[0142] (a) Wind power and photovoltaic power are trained with separate independent SVM models;

[0143] (b) The load is trained into an independent SVM model according to the user type, and then the total load prediction is output through weighted ensemble.

[0144] (3) Call the trained SVM model to obtain the predicted values ​​of wind power output, photovoltaic power output, and load increment of each type of user; integrate the predicted values ​​of load increment of each type of user into the load prediction value, as follows:

[0145] Load forecast = Baseline load + Incremental load forecast for each type of user (Formula 7)

[0146] The baseline load is the average of historical load data.

[0147] Support vector machines (SVMs) are used to train the system on various influencing factors. During training, cross-validation is employed to select the optimal parameters. These trained parameters are then used to predict source loads, ultimately yielding the predicted wind power output. Photovoltaic power forecast and load power forecast .

[0148] The purpose of step S1 is to acquire historical data of the regional power grid and construct a source-load model, use methods such as SVM to fit and train the data, and finally obtain the predicted values ​​of wind power, photovoltaic power and load power, so as to provide data support for subsequent scene identification and flexible resource allocation.

[0149] Step S2 includes steps S21-S22.

[0150] Step S21: Construct a scene recognition model.

[0151] This invention proposes a multi-scenario power grid configuration identification mechanism. In order to quantify the needs of each scenario, multiple scenario feature indicators are introduced as relaxation variables to identify scenarios such as power balance, peak shaving, and frequency regulation that exist in power grid operation.

[0152] With the large-scale integration of new energy sources, the fluctuating and intermittent nature of wind and solar power has created demands for power balancing, peak shaving, and frequency regulation. Different resources offer varying capabilities in these areas, requiring more flexible resources to balance the needs of different scenarios given the current resource allocation.

[0153] To identify the most prominent demand scenarios in the current power grid, the scenario identification model takes the minimum weighted sum of the new resource flexibility characteristic indicators as the objective function. It introduces six new resource flexibility characteristic indicators: maximum output power, minimum output power, ramp rate, ramp rate, conventional frequency regulation capability, and fast frequency regulation capability. A multi-scenario demand identification model is constructed, and then the worst demand scenarios are identified under the condition of minimizing the input of new resources.

[0154] The objective function for scene recognition is to minimize the weighted average of the newly added resource flexibility indicators, that is, to minimize the increase in the resource flexibility adjustment parameters required to balance scene demands.

[0155] Construct a scene recognition model, which includes a scene recognition objective function and scene recognition constraints.

[0156] The scene recognition objective function is expressed as follows:

[0157] Formula (8)

[0158] in, , , These are the flexibility characteristics indicators for power balance, peak shaving, and frequency regulation scenarios, respectively. , , They are respectively , , The weights; , These represent the maximum and minimum output power of the flexibility resources required for the scenario; , These are the rated values ​​for the maximum and minimum output power, respectively. , These represent the upward and downward ramp rates of the flexibility resources required by the scenario, respectively. , These represent the maximum upward and downward ramp rates of the flexibility resources, respectively. , These are the indicators for the standard frequency modulation capability and the fast frequency modulation capability of the flexible resources required by the scenario. These refer to the upper and lower backup adjustment capabilities of the conventional frequency modulation capability, respectively.

[0159] The scenario identification constraints include conventional power balance constraints, conventional non-adjustable unit operation constraints, wind power generation constraints, transmission line constraints, flexible adjustment resource operation constraints, full load constraints, spinning reserve capacity constraints, and fast frequency regulation reserve constraints.

[0160] , , weight For example, , , The values ​​are as follows, and can be revised according to specific needs.

[0161] Power balance scenario Dominant

[0162] Peak shaving pressure scenario Key Scenarios:

[0163] FM scene priority: .

[0164] Scene recognition constraints are as follows:

[0165] (1) Power balance constraints of conventional systems

[0166] Formula (9)

[0167] in, For the first Taiwan's conventional non-adjustable units at time contribution; For photovoltaic units in Efforts made at all times; For wind turbine units Efforts made at all times; For the first External transmission lines in Efforts made at all times; For flexible adjustment of units exist Efforts made at all times; , , , These are the types of conventional non-adjustable generating units, external transmission lines, flexible regulating generating units, and the number of loads. For load nodes exist The load demand at any given moment; for The load reduction power at any given time.

[0168] Formula (9) is an indirect constraint condition of the objective function formula (8). Power balance affects the output of the unit, and the output of each unit corresponds to each flexibility adjustment index. The minimum weighted sum of the flexibility adjustment index is the objective function for scenario identification.

[0169] Conventional non-adjustable generator sets refer to generator sets with fixed output power that are difficult to adjust quickly according to the real-time needs of the power system. These units operate according to a pre-set plan and rarely participate in the real-time power regulation of the system; examples include traditional thermal power generator sets, nuclear power generator sets, and hydroelectric generator sets.

[0170] Flexible generating units refer to generator sets that can quickly adjust their output power according to grid demand. They are designed to help the grid better cope with load fluctuations and the uncertainty of new energy generation, and improve the stability and flexibility of the power system. They have rapid adjustment capabilities, flexible start and stop, and can quickly respond to changes in grid load, including short-term adjustments from minutes to hours.

[0171] Source-side flexibility resources include: gas turbine generator sets and pumped storage units;

[0172] Grid-side flexibility adjustment resources refer to power transmission between different regions;

[0173] Load-side flexibility adjustment resources refer to electric vehicles, air conditioning loads, big data centers, etc.

[0174] Storage-side flexibility adjustment resources refer to new energy storage devices such as electrochemical energy storage and flywheel energy storage.

[0175] Load shedding power at any time It refers to The load power that is reduced due to the needs of power system operation or emergency situations is used to reduce part of the load through human or automatic control when the power system cannot meet the full load demand in order to maintain the stable operation of the power system.

[0176] (2) Operating constraints of conventional non-adjustable units:

[0177] Formula (10)

[0178] in, , The first The minimum and maximum generating power of a conventional non-adjustable generator unit.

[0179] (3) Wind power generation constraints:

[0180] Formula (11)

[0181] in, , These represent the minimum and maximum generating capacities of the wind turbine, respectively.

[0182] (4) Constraints on photovoltaic power generation:

[0183] Formula (12)

[0184] in, , These represent the minimum and maximum power generation of the photovoltaic unit, respectively.

[0185] (5) Transmission line constraints:

[0186] Formula (13)

[0187] in, , The first Minimum and maximum transmission power of each line; Indicates the first Safety margin coefficient of each line.

[0188] For example, safety margin coefficient The value is between 0 and 1.

[0189] This indicates the maximum transmission power limit; at any given time, the actual transmission power of a transmission line cannot exceed its maximum transmission power. The line cannot operate at full load; a certain margin must be maintained.

[0190] This indicates the minimum transmission power limit; at any given time, the actual transmission power of a power transmission line cannot be lower than its maximum transmission power. This means that the transmission power of the line cannot exceed this limit in reverse.

[0191] (6) Flexibly adjust resource operation constraints:

[0192] Formula (14)

[0193] in, For the first Flexible resource adjustment The start / stop flag at any time, with values ​​of 1 and 0 indicating that the unit is in the start-up and shutdown state, respectively; For the first Minimum output of flexible adjustable units; For the first The maximum output of the flexible-adjustable unit; , These are the minimum and maximum output power of the new flexible control unit, respectively.

[0194] (7) Full load constraint:

[0195] Formula (15)

[0196] in, This represents the maximum load percentage. For example, The value ranges from 0.8 to 0.95, with the specific value determined based on the actual power grid operation requirements and load characteristics. This ratio is used to limit the upper limit of the load, prevent system overload, and at the same time leave a certain margin to cope with emergencies and ensure the safe and stable operation of the system.

[0197] (8) Climbing constraint:

[0198] Formula (16)

[0199] in, For flexible adjustment of units exist Efforts made at all times; , These are the maximum uphill and maximum downhill speeds, respectively. , They are respectively time, Time of the first Start / stop status of flexible resources; , These refer to the ability to adjust for the required resources to climb uphill and downhill, respectively. , These represent the duration of the uphill and downhill climbs, respectively.

[0200] (9) Spinning reserve capacity constraint

[0201] Formula (17)

[0202] in, The percentage of spinning reserve capacity to total load; The proportion of spinning reserve capacity to installed renewable energy capacity; , These refer to the upper and lower standby rotary adjustment capabilities of the newly added flexible adjustment units; The quantity of renewable energy; For renewable energy in Efforts are made at all times.

[0203] For example, The value ranges from 0.05 to 0.1; The value is between 0.1 and 0.2; it is adjusted according to the specific grid operation requirements and the scale of renewable energy access.

[0204] (10) Fast frequency modulation standby constraint:

[0205] Formula (18)

[0206] in, For the first Flexible resource adjustment Efforts made at all times; For the first Adjusting resource ramp-up rate with flexibility; Indicates the duration of frequency modulation; To provide the rapid rotational standby capability for the newly added flexible adjustment units; , These represent the frequency regulation capacity ratios of load and renewable energy, respectively.

[0207] For example, , The value should be between 0.05 and 0.15; the specific value needs to be adjusted according to the frequency regulation requirements of the power grid, load characteristics and the scale of renewable energy access, so as to ensure the stability and reliability of the system during the frequency regulation process.

[0208] Step S22: Solve the scene recognition objective function based on scene recognition constraints to obtain a flexibility feature index vector; make a judgment based on the flexibility feature index vector to obtain the corresponding scene.

[0209] Based on the scene recognition objective function and constraints defined in the scene recognition model, the defined flexibility characteristic index is solved. The maximum output power is then obtained. Minimum output power Uphill climbing rate Downhill climbing rate Conventional frequency modulation capability Fast frequency modulation capability .

[0210] To further optimize the allocation of flexibility resources for the most prominent scenarios in different regions of the power grid, a flexibility characteristic index vector is constructed based on the solution of each flexibility characteristic index, as shown in formula (18). If the flexibility characteristic index vector is a non-zero vector, it indicates the existence of one or more scenarios. Further, it is determined which flexibility characteristic index is non-zero, which represents the existence of the scenario corresponding to that index.

[0211] Based on the current power grid parameters, the scene recognition objective function is solved using the Gurobi solver based on the scene recognition constraints, yielding the following results: ;

[0212] Will The flexibility feature index vector constitutes the above. ,as follows:

[0213] Formula (19)

[0214] The corresponding scenario is determined based on the aforementioned flexibility feature index vector, including:

[0215] like or If the value is non-zero, a power balance scenario is determined to have occurred;

[0216] like or If the value is non-zero, then a peak-shaving scenario is determined to have occurred;

[0217] like or If the value is non-zero, then a frequency modulation scenario is determined to have occurred.

[0218] The current power grid parameters include the installed capacity data of wind turbines, photovoltaic units, non-adjustable units, and adjustable resources, as well as load data; these data are input into the scene recognition model and solved using the gurobi solver in MATLAB.

[0219] If the flexibility characteristic index vector If all variables are zero, then steps S3-S4 do not need to be executed.

[0220] Step S2 aims to identify the most prominent demand scenarios (power balance scenario, peak shaving scenario, and frequency regulation scenario) of the power grid by constructing a scenario recognition model and solving the flexibility characteristic index vector, thus providing a basis for subsequent differentiated resource collaborative allocation.

[0221] Step S3 includes steps S31-S32.

[0222] Step S31: Construct a two-stage adaptive configuration model.

[0223] To achieve systematic and scientific allocation of flexible resources based on scenario identification, a two-stage adaptive configuration model integrating capacity allocation and operational regulation was constructed. This two-stage adaptive configuration model aims to assess the capacity deployment pressure and dynamic control costs of various types of power source-grid-load-storage flexible resources throughout their entire lifecycle, ensuring that the configuration scheme balances physical feasibility and operational regulation performance.

[0224] The two-stage adaptive configuration model includes a two-stage objective function that minimizes the capacity configuration of flexible resources on each side of the source, grid, load, and storage, as well as the control loss of flexible resources on each side, along with first-stage constraints and second-stage constraints.

[0225] The constraints in the first stage include configuration scale constraints, power demand constraints, ramp-up demand constraints, and frequency adjustment demand constraints.

[0226] The second stage constraints include second stage deterministic constraints and second stage uncertainty constraints;

[0227] The second phase of deterministic constraints includes source-side, network, load, and side flexibility resource constraints.

[0228] In the two-stage adaptive configuration model, the first stage measures the capacity deployment pressure and construction cost of various flexible resources under multiple nodes and multiple types through capacity configuration parameters.

[0229] The second stage, based on the time scale of typical periods, introduces dynamic adjustment of load and operational losses for various resources, quantifying the real-time losses and adjustment intensity of resources during operation. The two-stage objective function, through comprehensive optimization of capacity allocation and operational adjustment, achieves efficient allocation of flexible resources and ensures operational adaptability.

[0230] The two-stage objective function is as follows:

[0231] Formula (20)

[0232] in, , , , These are the unit capacity configuration parameters for source, grid, load, and storage resources, respectively. Scale allocation for various resources including power sources, grids, loads, and storage; , These are timescales for the entire year and typical days, respectively. , , , These are the control losses of resources on the source, grid, load, and storage sides, respectively. For various flexible resource categories including source, grid, load, and storage; Let be the set of load uncertainties.

[0233] , The time scales are the whole year and the typical day, respectively. In this invention, the selection of the typical day is based on the historical data of annual load and wind power and photovoltaic power output. The daily curves are classified by K-means clustering analysis, and the day with the largest peak-to-valley difference is selected as the typical day for study.

[0234] (1) Source-side flexibility, resource control, and loss ,as follows:

[0235] Formula (21)

[0236] in, , , These are the constant, primary, and secondary control loss coefficients for power supply-side flexibility resources (coal-fired power and gas-fired power). For the first Power-side flexibility resources Power generation during a given time period; , Indicates fuel consumption and emissions The cost; , They represent the first Power-side flexibility, resource fuel consumption rate and Emission rate; This indicates the total number of flexibility resources.

[0237] Taking coal-fired power as an example, exemplarily, , , The possible values ​​are as follows:

[0238] constant term : Typically in the range of 100~500 yuan / MW, used to characterize the fixed costs of flexible regulation of coal-fired power plants, such as equipment maintenance, personnel on-duty expenses, etc.

[0239] First item Generally, it is in the range of 5 to 15 yuan / MW², reflecting the loss cost that is proportional to the power generation during coal-fired power regulation, and is mainly related to factors such as fuel consumption;

[0240] Quadratic terms The cost is approximately RMB 0.1 to 0.5 per MW³, reflecting the losses related to the square of the power generation during the regulation process of coal-fired power plants due to reduced equipment operating efficiency and decreased fuel utilization efficiency.

[0241] (2) Grid-side flexibility resource control losses ,as follows:

[0242] Formula (22)

[0243] in, Indicates the first Inter-network lines Transmission power during a given time period; , , These represent voltage deviation cost, active power cost, and reactive power cost of line operation and maintenance, respectively. , , These represent the voltage, active power boundaries, and reactive power boundaries of each node and line, respectively. This indicates the total number of interconnected lines. Indicates the reference voltage level; , These represent the starting and ending nodes connecting the two ends of the line, respectively.

[0244] (3) Load-side flexibility resource control loss ,as follows:

[0245] Formula (23)

[0246] in, , The first The unit compensation cost of load-side flexibility resources with two load response modes: time-shiftable and interruptible. , The first The load-side flexibility resources can respond to load loads in two ways: time-shifting and interruptible. This represents the total number of load-side flexibility resources.

[0247] (4) Energy storage side flexibility resource control loss ,as follows:

[0248] Formula (24)

[0249] in, , , The first Daily maintenance costs of energy storage resources, unit costs during charging, and unit costs during discharging; , The first Energy storage resources Charging and discharging power during different time periods; To allow for flexible adjustment of the total number of resources on the energy storage side.

[0250] For example, Energy storage resources include lithium battery energy storage, flow battery energy storage, sodium-sulfur battery energy storage, lead-acid battery energy storage, supercapacitor energy storage, flywheel energy storage, and compressed air energy storage.

[0251] The first phase of constraints includes configuration scale constraints, power demand constraints, ramp-up demand constraints, and frequency adjustment demand constraints, as follows:

[0252] (1) Configuration size constraints are as follows:

[0253] Formula (25)

[0254] in, The total limit for configurable flexibility resources; For the first The scale of flexible resource allocation has been increased. This represents the total scale of the newly added flexible resource allocation; Configure parameters for storage-side flexibility resource units; Configure parameters for source-side flexibility resource units; Configure parameters for load-side flexibility resource units; Configure parameters for network-side flexibility resource units.

[0255] (2) Power demand constraints:

[0256] The power demand constraints are as follows:

[0257] Formula (26)

[0258] in, , , , These are the quantities of conventional non-adjustable generating units, external transmission lines, flexible regulating generating units, and loads. For the first Taiwan's conventional non-adjustable units at time contribution; For photovoltaic units in Efforts made at all times; For wind turbine units Efforts made at all times; For the first External transmission lines in Efforts made at all times; For flexible adjustment of units exist Efforts made at all times; Adjusting power for newly added flexibility resources; For load nodes exist The load demand at any given moment; for The load reduction power at any given time.

[0259] , These are the newly added source side and network side. Flexible resource adjustment power at any time; , The newly added storage-side resources are respectively Power generation and charging capacity at any given time; , , The newly added load-side resources are respectively Power transferred into the load, power transferred out of the load, and power of the interruptible load at any given time; , , , The figures represent the number of new flexible resources added to each of the power source, grid, load, and storage sides.

[0260] (3) The ramp-up requirement constraint is as follows:

[0261] Formula (27)

[0262] in, , , , , , , , ,as follows:

[0263] Formula (28)

[0264] in, , , , These are the coefficients of the upward ramping capacity that can be provided by resources on each side of the unit source-grid-load-storage system; , , , These are the downward ramping capacity coefficients that can be provided by resources on each side of the unit source-grid-load-storage system; For the new source side in Flexible resource adjustment power at any time; For the newly added network side Flexible resource adjustment power at any time; To add new load-side resources Power can be transferred out at any time; To add new load-side resources The transferable load power at any given time; For new storage resources in Power generation at any given moment; For the addition of storage-side resources The charging power at any given time.

[0265] (4) Frequency regulation requirement constraints are as follows:

[0266] Formula (29)

[0267] in, , , , , , ,as follows:

[0268] Formula (30)

[0269] in, , , , These refer to the resource allocation scale for each side of the source, grid, load, and storage; , , These are the coefficients of conventional frequency regulation capability that can be provided by resources on each side of the unit source-grid-load-storage system; , , These are the rapid frequency regulation capability coefficients that can be provided by resources on each side of the source-grid-load-storage unit; This indicates the activation status of network-side flexibility resources; 1 indicates enabled, and 0 indicates disabled. This refers to the maximum power generation capacity of various generating units on the source side; The maximum transmittable power for network-side flexibility resources; The maximum rechargeable power for energy-side flexibility resources; , , These are the upper limits of the ramp rate for resource flexibility adjustment on the source, grid, and storage sides, respectively. , , This provides a continuous adjustable frequency time for the source, grid, and storage side to flexibly adjust resources.

[0270] The second stage constraints include second stage deterministic constraints and second stage uncertainty constraints; the second stage deterministic constraints include source-side, network, load, and side flexibility resource constraints.

[0271] Second-stage deterministic constraints:

[0272] (1) Source-side flexibility adjustment of resource constraints, as follows:

[0273] Formula (31)

[0274] Formula (32)

[0275] in, , respectively source-side flexibility resources , Power generation at any given moment; , These represent the minimum and maximum generating capacities of various types of generator units on the source side, respectively. , These represent the maximum uphill and downhill ramping capabilities of various types of generator units on the source side.

[0276] (2) Network-side flexibility resource constraints are as follows:

[0277] Formula (33)

[0278] in, , These are the minimum and maximum values ​​for power transmission in external power transmission channels, respectively.

[0279] (3) Load-side flexibility resource constraints, including interruptible load constraints and transferable load constraints, as follows:

[0280] a) Interruptible load constraints, as follows:

[0281] Formula (34)

[0282] Formula (35)

[0283] b) Transferable load constraints, as follows:

[0284] Formula (36)

[0285] Formula (37)

[0286] in, , , They are respectively Time of the first The new load-side flexibility resources include load transfer in, load transfer out, and interruptible load power. for Time of the first The newly added load-side flexibility resources can transfer load outflow power; for Time of the first The new load-side flexibility resources are based on the original load. , , These are the maximum interrupted load ratio, the maximum continuous time interrupted load ratio, and the maximum transferable load ratio, respectively. , These are 0-1 state variables for load transfer out and load transfer in, respectively. A variable of 1 indicates that the state is enabled, and a variable of 0 indicates that the state is disabled.

[0287] (4) Resource constraints for storage-side flexibility adjustment are as follows:

[0288] Formula (38)

[0289] Formula (39)

[0290] in, , These are the minimum and maximum discharge power of the storage-side flexibility resources, respectively. , These are the minimum and maximum charging power for the energy storage flexibility resources, respectively. The 0-1 flag indicates the activation of storage-side flexibility resource discharge, with 1 indicating activation of storage-side flexibility resource discharge mode and 0 indicating deactivation of storage-side flexibility resource discharge mode. The 0-1 activation flag for charging storage-side flexibility resources indicates that the charging mode for storage-side flexibility resources is enabled, and that the charging mode for storage-side flexibility resources is disabled. , Time periods , Storage-side flexibility resources and energy; , The efficiency is calculated under charging and discharging conditions, respectively. , Time periods The charging and discharging power; The scheduling time step; Indicates the rated capacity of storage-side flexibility resources; For time period Storage-side flexibility resources SOC (State of Charge). For storage-side flexibility, available SOC resources are required. This indicates the minimum available State of Charge (SOC) for storage-side flexibility resources; The ratio of the maximum energy difference between the start and end of a day for storage-side flexibility resources to the rated capacity; , These represent the initial and final energy levels of the storage-side flexibility resources, respectively.

[0291] Second stage uncertainty constraints ,as follows:

[0292] Formula (40)

[0293] in, Represents an uncertain variable; , , These represent the output power of photovoltaic and wind power, and the electrical load power, respectively, after considering uncertainties. , , These are the predicted values ​​for wind power, solar power, and load power, respectively. , , These are the maximum allowable fluctuation coefficients for wind power, photovoltaic power output, and electrical load power, respectively.

[0294] in, To represent an uncertain variable, an uncertain set can be taken. Any value in; , and These represent the output power of photovoltaic and wind power, and the electrical load power, respectively, after considering uncertainties. , , These represent the predicted wind power, solar power output, and electrical load power, respectively. , , These represent the maximum allowable fluctuation coefficients for wind power, photovoltaic power output, and electrical load power, respectively.

[0295] Step S32: Based on the scene recognition results and the first-stage constraints, solve to obtain the flexible resource allocation scale.

[0296] The two-stage adaptive configuration model transformation is as follows:

[0297] Formula (41)

[0298] in, This indicates the configuration parameters for each flexibility resource; This represents the set of decision variables for increasing capacity of various flexible resources. -Scale of various resources), number of newly added flexible resources on the source, grid, load and storage sides. , , , ; This represents a set of uncertain variables, including the wind and solar power output and load at each node. , , ), This represents the set of decision variables, including the power generation of each wind and solar power unit, the power generation of conventional non-adjustable units, the power generation of the original flexible adjustable units, the power transmission between the grids, and the power of newly added flexible resources on each side of the source, grid, load, and storage. This represents the first phase of constraints; This indicates the second phase of deterministic constraints; This indicates the second stage of uncertainty constraints.

[0299] Input various grid parameters for the current scenario, including historical meteorological data, historical wind and solar power generation data, historical load data, inter-grid transmission power data, the proportion of controllable load, conventional non-adjustable units, the installed capacity of various flexible adjustment units, and the technical parameters of flexible resources on each side of the source, grid, load, and storage system (including the constant term, primary term, and secondary term coefficients of the control loss function on each side; coal consumption rate; carbon dioxide emission rate, etc.).

[0300] The scale of flexible resource allocation This includes the scale of new flexible resources on the source, grid, load, and storage sides. , , and ;

[0301] Step S3 aims to construct a two-stage adaptive configuration model integrating capacity configuration and operational adjustment based on the identified scenarios, and to solve for the flexibility resource allocation scale of each side (source, grid, load, and storage) based on the constraints of the first stage. , , , .

[0302] Step S4, specifically.

[0303] Based on the allocation scale of various resources of source, grid, load and storage , , , Based on various resource control loss parameters, wind and solar power output, and load power uncertainty data (including predicted values ​​based on wind power, solar power, and load power), and using the second-stage deterministic and uncertainty constraints, an optimization solution is performed to obtain the operation plan for various newly added flexible resources of source, grid, load, and storage. , , , , , , .

[0304] The flexibility resource allocation scale corresponds to the flexibility resource operation plan, which includes the power adjustment of newly added source-side flexibility resources. Added grid-side flexibility resource adjustment power Newly added load-side resources can transfer load out. New load-side resources can transfer loads and interruptible load power. New power generation capacity from storage-side resources , Increased charging power of energy storage resources And newly added load-side resources can transfer loads. .

[0305] Step S4 is based on the determined scale of flexible resource allocation, various resource control loss parameters, and uncertain data on wind and solar power output and load power. Combined with the deterministic and uncertain constraints of the second stage, it optimizes the solution to formulate specific operation plans for the newly added flexible resources on each side of the power source, grid, load, and storage.

[0306] In summary, the flexible resource adaptive configuration method based on multi-scenario fusion of this invention has the following beneficial effects:

[0307] 1. Multi-scenario integration and optimization of differentiated configuration capabilities: This invention constructs a multi-scenario identification model (power balance scenario, peak shaving scenario, and frequency regulation scenario), comprehensively considers the flexibility characteristics of power balance scenario, peak shaving scenario, and frequency regulation scenario, solves the limitations of single-scenario optimization in the prior art, realizes differentiated resource adaptation under multi-scenario coupling, and significantly improves the dynamic response capability of the power system to new energy fluctuations (such as wind and solar intermittency and load mutation).

[0308] 2. Global Coordinated Optimization of Multiple Resource Types (Source, Grid, Load, and Storage): This invention designs a two-stage adaptive configuration model (capacity configuration + operation regulation), integrating the dynamic coupling relationship of resources on the source side (thermal power, energy storage), grid side (transmission lines), load side (interruptible / transferable loads), and storage side (energy storage). It breaks through the bottleneck of independent optimization of a single resource in traditional research, realizes full-cycle coordinated optimization across time scales and resource types, improves overall regulation efficiency, realizes coordinated optimization configuration of flexible resources on the source, grid, load, and storage sides, improves resource utilization efficiency, and overcomes the shortcomings of existing technologies that focus on independent optimization of a single resource.

[0309] 3. Dynamic Adaptation and Uncertainty Management: This invention introduces a two-stage adaptive configuration model. The first stage focuses on resource capacity configuration, while the second stage considers the dynamics and uncertainties of operation adjustment, making the resource configuration scheme more adaptable and robust, effectively addressing the uncertainties brought about by the volatility of new energy sources; ensuring the stable operation of the system under complex scenarios such as extreme weather and forecast deviations, and reducing the risk of wind and solar curtailment and load reduction.

[0310] 4. Multi-scenario identification and differentiated configuration: Based on the flexibility feature index vector (such as ramp rate and frequency regulation capability), the demand scenario is dynamically determined, and resources are configured differently according to different scenarios (one or more) to avoid the difficulty of a single configuration method to adapt to the ever-changing and complex power grid operation conditions.

[0311] 5. Improved accuracy of data-driven prediction and configuration: This invention uses SVM (Support Vector Machine) and cross-validation to fit wind power, photovoltaic and load power prediction models, and combines historical data and meteorological response regression analysis to significantly improve prediction accuracy, provide highly reliable input for multi-scenario identification and resource allocation, and support the refined operation of the power system.

[0312] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0313] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A flexible resource adaptive allocation method based on multi-scenario fusion, characterized in that, include: Obtain forecasts of wind power, solar power, and load power in the regional power grid; A scenario identification objective function is constructed with the goal of minimizing the weighted sum of flexibility characteristic indices for power balance, peak shaving, and frequency regulation scenarios. A flexibility characteristic index vector is obtained by solving the scenario identification objective function based on scenario identification constraints. The corresponding scenario identification result is obtained by judging based on the flexibility characteristic index vector. If the scenario identification result is one or more of the power balance scenario, peak shaving scenario, and frequency regulation scenario, a two-stage adaptive configuration model is constructed with the goal of minimizing the capacity configuration of flexible resources on each side of the source, grid, load, and storage, and minimizing the control loss of flexible resources on each side; the scale of flexible resource configuration is obtained by solving based on the scenario identification result and the first stage constraints. Based on the predicted values ​​of wind power, photovoltaic power and load power, the scale of newly added flexible resource allocation, and the deterministic and uncertain constraints of the second stage, the flexible resource operation plan corresponding to the scale of flexible resource allocation under the scenario is obtained by solving the problem. The two-stage adaptive configuration model includes a two-stage objective function that minimizes the capacity configuration of flexible resources on each side of the source, grid, load, and storage, as well as the control loss of flexible resources on each side, along with first-stage constraints and second-stage constraints. The constraints in the first stage include configuration scale constraints, power demand constraints, ramp-up demand constraints, and frequency adjustment demand constraints. The second-stage constraints include second-stage deterministic constraints and second-stage uncertainty constraints; The second phase of deterministic constraints includes source-side, network, load, and side flexibility resource constraints; The two-stage objective function is as follows: ； in, , , , These are the unit capacity configuration parameters for source, grid, load, and storage resources, respectively. Scale allocation for various resources including power sources, grids, loads, and storage; , These are timescales for the entire year and typical days, respectively. , , , These are the control losses of resources on the source, grid, load, and storage sides, respectively; For various flexible resource categories including source, grid, load, and storage; For the set of load uncertainties; Second stage uncertainty constraints ,as follows: ； in, Represents an uncertain variable; , , These represent the output power of photovoltaic and wind power, and the electrical load power, respectively, after considering uncertainties. , , These are the predicted values ​​for wind power, solar power, and load power, respectively. , , These are the maximum allowable fluctuation coefficients for wind power, photovoltaic power output, and electrical load power, respectively.

2. The flexible resource adaptive allocation method based on multi-scenario fusion according to claim 1, characterized in that, It also includes constructing a source-load model; and performing data fitting and solving on the source-load model based on historical power data of wind power, photovoltaic power, and various types of users in the regional power grid to obtain predicted values ​​of wind power, photovoltaic power, and load power, including: Obtain load sample data of wind power, photovoltaic and various types of users in the historical period of the regional power grid, and preprocess the data to obtain preprocessed load sample data of wind power, photovoltaic and various types of users; Using SVM and cross-validation, the source-load model is fitted and solved based on the load sample data of wind power, photovoltaics, and various types of users to obtain the predicted values ​​of wind power, photovoltaics, and load power.

3. The flexible resource adaptive allocation method based on multi-scenario fusion according to claim 2, characterized in that, The source-load model includes a wind power model, a photovoltaic power model, and a load power model; The wind power model is fitted and solved using wind power sample data to obtain the predicted wind power value. The photovoltaic power model is fitted and solved using photovoltaic sample data to obtain the predicted photovoltaic power value. The load power model is fitted and solved using load sample data of each type of user to obtain the load power prediction value corresponding to each type of user; the load power prediction value of each type of user is integrated with the historical load average to obtain the load power prediction value.

4. The flexible resource adaptive allocation method based on multi-scenario fusion according to claim 1, characterized in that, The scene recognition objective function is expressed as follows: ； in, , , These are the flexibility characteristics indicators for power balance, peak shaving, and frequency regulation scenarios, respectively. , , They are respectively , , The weights; , These represent the maximum and minimum output power of the flexibility resources required by the scenario; , These are the rated values ​​for the maximum and minimum output power, respectively. , These represent the upward and downward ramp rates of the flexibility resources required by the scenario, respectively. , These represent the maximum upward and downward ramp rates of the flexibility resources, respectively. , These are the indicators for the standard frequency modulation capability and the fast frequency modulation capability of the flexible resources required by the scenario. These refer to the upper and lower backup adjustment capabilities of the conventional frequency modulation capability, respectively. The scenario identification constraints include conventional power balance constraints, conventional non-adjustable unit operation constraints, wind power generation constraints, transmission line constraints, flexible adjustment resource operation constraints, full load constraints, spinning reserve capacity constraints, and fast frequency regulation reserve constraints.

5. The flexible resource adaptive allocation method based on multi-scenario fusion according to claim 4, characterized in that, Based on the current power grid parameters, the scene recognition objective function is solved using the Gurobi solver based on the scene recognition constraints, yielding the following results: ; Will The flexibility feature index vector constitutes the above. ,as follows: ； The corresponding scenario is determined based on the aforementioned flexibility feature index vector, including: like or If the value is non-zero, a power balance scenario is determined to have occurred; like or If the value is non-zero, then a peak-shaving scenario is determined to have occurred; like or If the value is non-zero, then a frequency modulation scenario is determined to have occurred.

6. The flexible resource adaptive allocation method based on multi-scenario fusion according to claim 1, characterized in that, The scale of flexible resource allocation This includes the scale of new flexible resources on the source, grid, load, and storage sides. , , and ; The flexibility resource allocation scale corresponds to the flexibility resource operation plan, which includes the power adjustment of newly added source-side flexibility resources. Added grid-side flexibility resource adjustment power Newly added load-side resources can transfer load out. New load-side resources can transfer loads and interruptible load power. New power generation capacity from storage-side resources , Increased charging power of energy storage resources And newly added load-side resources can transfer load into .

7. The flexible resource adaptive allocation method based on multi-scenario fusion according to claim 1 or 6, characterized in that, The power demand constraints are as follows: ； in, , , , These are the quantities of conventional non-adjustable generating units, external transmission lines, flexible regulating generating units, and loads. For the first Taiwan's conventional non-adjustable units at time contribution; For photovoltaic units in Efforts made at all times; For wind turbine units Efforts made at all times; For the first External transmission lines in Efforts made at all times; For flexible adjustment of units exist Efforts made at all times; Adjusting power for newly added flexibility resources; For load nodes exist The load demand at any given moment; for The load reduction power at any given time.

Images