Wind-solar-hydro storage coordinated operation optimization method based on weather classification and volatility characteristics

Abstract

The present application relates to the technical field of power system operation optimization, and discloses a wind-solar-hydro-storage collaborative operation optimization method based on weather typing and fluctuation characteristics, comprising: collecting historical output data and meteorological data of wind power, photovoltaic, hydroelectric and energy storage systems, and constructing meteorological feature vectors and fluctuation feature vectors; inputting the comprehensive feature vectors fusing the meteorological feature vectors and the fluctuation feature vectors into a self-organizing mapping network for clustering; for each weather scenario, calculating the linear correlation, nonlinear mutual information and timing similarity among wind, solar, hydro and storage, constructing a collaborative relationship matrix, and extracting the average synergy degree, fluctuation offset rate and energy dominant coefficient as the scenario features; establishing a multi-energy capacity optimization model, and solving the optimal capacity configuration by using a particle swarm optimization algorithm. The present application has the advantages of significantly improving the multi-energy complementary utilization efficiency in a high-proportion new energy system, reducing the energy storage configuration cost, and enhancing the stability and reliability of system operation.

Description

Technical Field

[0001] This invention relates to the field of power system operation optimization technology, and in particular to a method for optimizing the coordinated operation of wind, solar, hydro, and storage systems based on weather patterns and volatility characteristics. Background Technology

[0002] With the large-scale integration of renewable energy sources such as wind and solar power, the randomness, volatility, and weather dependence of new energy output are gradually becoming key factors affecting the safe and stable operation of the power system. Wind power output is significantly affected by wind speed changes, exhibiting large-amplitude, high-frequency short-term fluctuations; solar power is easily affected by irradiance, temperature, and cloud cover, showing significant nonlinear fluctuations between sunny and cloudy days, and between morning / evening and midday; hydropower output, while possessing regulation capabilities, is also constrained by inflow, rainfall, and seasonal hydrological characteristics, exhibiting medium- to long-term variation characteristics. Energy storage systems can compensate for wind and solar fluctuations, but their capacity and power configuration must be matched with weather and output characteristics to be effective. Therefore, in the context of multi-energy joint operation of "wind-solar-hydro-storage," revealing the output fluctuation patterns and multi-energy coupling relationships under different meteorological conditions is fundamental to achieving optimal system configuration and coordinated control.

[0003] Currently, there is some research on weather scene recognition, mainly focusing on weather classification for wind and solar power generation. For example, methods such as K-means, SOM, self-organizing mapping networks, and fuzzy clustering are used to classify meteorological data into types such as sunny, cloudy, weak wind, and high wind, and then analyze the wind and solar power output characteristics under different weather conditions. However, such studies mostly focus on single or dual energy output modes, lacking unified modeling of wind, solar, hydro, and energy storage systems, and failing to incorporate hydrological conditions, wind-solar complementarity, and energy storage response mechanisms into the weather classification system. At the same time, existing studies mostly focus on power output levels (mean, maximum) or curve shape classification, and rarely systematically characterize the stability and regulation requirements of multi-energy output from the perspectives of power fluctuation, ramp rate, and spectral energy distribution.

[0004] Furthermore, some studies on multi-energy complementarity have proposed methods such as power correlation coefficient, coefficient of variation, energy complementarity index, and high-dimensional complementarity index to assess the complementary relationship between wind and solar or wind, solar, and hydropower. However, these methods are usually based on full-time statistics or single-timescale calculations, without combining weather scenarios for classification analysis, and lack coupling descriptions of volatility indicators with weather types. More importantly, existing results mostly remain at the "evaluation of complementarity" stage, without establishing a complete technical chain of "weather type—power volatility—multi-energy coordinated regulation—capacity configuration," making it difficult to directly apply to practical engineering applications such as energy storage configuration optimization, hydropower peak-shaving strategy formulation, or wind and solar power output smoothing control. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide an optimization method for the coordinated operation of wind, solar, hydro, and storage based on weather patterns and volatility characteristics.

[0006] The objective of this invention is achieved through the following technical solution: an optimization method for the coordinated operation of wind, solar, hydro, and storage systems based on weather typology and volatility characteristics, the method comprising:

[0007] S1. Multi-source data construction and normalization: Collect historical output data and meteorological data of wind power, photovoltaic, hydropower and energy storage systems, construct meteorological feature vectors and fluctuation feature vectors, and perform normalization processing.

[0008] S2. Weather scenario classification based on self-organizing map network: The comprehensive feature vector that integrates meteorological feature vector and fluctuation feature vector is input into the self-organizing map network for clustering, and several typical weather-fluctuation scenarios are divided.

[0009] S3. Modeling the collaborative output law of wind, solar, hydro and storage under various weather types. For each weather scenario, calculate the linear correlation, nonlinear mutual information and time series similarity between wind, solar, hydro and storage, construct a collaborative relationship matrix, and extract the average collaborative degree, fluctuation offset rate and energy dominance coefficient as scenario features.

[0010] S4. Using the installed capacity of wind, solar, hydro, and storage as decision variables, and with the goals of system fluctuation suppression, economic efficiency, and coordinated absorption, a multi-energy capacity optimization model is established, and the optimal capacity configuration is solved using the particle swarm optimization algorithm.

[0011] Specifically, the meteorological feature vector is: ;

[0012] The fluctuation feature vector includes power gradient, ramp rate, sliding window local variance, and window mean.

[0013] The power gradient is:

[0014] ;

[0015] In the formula, For energy At any moment The power gradient; For the first Energy-like energy at all times The actual power; For time step;

[0016] The slope distance is:

[0017] ;

[0018] In the formula, For energy At any moment The amount of uphill climb;

[0019] The local variance of the sliding window is:

[0020] ;

[0021] In the formula, For energy In length Local power variance within the time window; The average power within the same window;

[0022] The mean value of the window is:

[0023] .

[0024] Specifically, the specific steps of S2 include:

[0025] The integrated feature vectors are combined into a system-wide input matrix along the time dimension:

[0026] ;

[0027] In the formula, For a moment The overall system feature vector; For energy The comprehensive feature vector, Number of energy types;

[0028] The optimal matching unit is determined by calculating the Euclidean distance to all nodes using a self-organizing map network.

[0029] ;

[0030] In the formula, For a moment The best matching node index; The input feature vector; For nodes The weight vector; It is the Euclidean norm;

[0031] Perform adaptive updates on the weight vectors of the best-matching node and its neighbors:

[0032] ;

[0033] In the formula, The learning rate; It is a neighborhood function;

[0034] ;

[0035] In the formula, For nodes Topological coordinates of the best node in the competing layer of the self-organizing mapping network; This is the neighborhood width parameter;

[0036] Assign each time-instance sample to the nearest class node:

[0037] ;

[0038] In the formula, For a moment The weather type category to which it belongs; This is the node weight vector after training.

[0039] Specifically, in S3, the linear correlation between wind, solar, hydro, and storage is defined as:

[0040] ;

[0041] In the formula, For weather type Lower energy and The linear correlation coefficient; For covariance; This represents the standard deviation of the corresponding energy in this type of sample; In terms of weather type Lower energy , A set of power time series samples;

[0042] The nonlinear mutual information between wind, solar, hydro, and storage in S3 is:

[0043] ;

[0044] In the formula, Weather type Lower energy and Mutual information; For energy Output in discrete amplitude range Marginal distribution probability; It is a joint probability distribution;

[0045] The temporal similarity between wind, solar, hydro, and storage in S3 is as follows:

[0046] ;

[0047] In the formula, For energy Weather type A measure of temporal similarity; This represents the set of all allowed time-matching paths; p and q are time-matching index pairs.

[0048] Specifically, the collaborative relationship matrix is ​​as follows:

[0049] ;

[0050] In the formula, Weather type Lower energy and The overall synergy coefficient; For the weighting coefficients, satisfying ;

[0051] The average degree of coordination is:

[0052] ;

[0053] In the formula, Weather type The average synergy coefficient of the system under the following conditions; Indicates the number of energy types; It is a weather scenario. Lower energy and The overall synergy coefficient;

[0054] The volatility offset rate is:

[0055] ;

[0056] In the formula, Weather type Lower system fluctuation offset rate; The standard deviation of the combined power series of wind power and photovoltaic power; The standard deviation of the net power of the wind and solar power system after being smoothed by hydropower and energy storage;

[0057] The energy dominance coefficient is:

[0058] ;

[0059] In the formula, Indicates weather type Lower energy The dominant coefficient; Weather type Lower energy With energy The overall synergy coefficient.

[0060] Specifically, the objective function of the multi-energy capacity optimization model is:

[0061] ;

[0062] In the formula, The system synthesis objective function; Total number of weather types; Weather type The probability of occurrence; This is the net power fluctuation index of the system under this weather type; For this weather type The following is the operating or capacity cost function; This is a volatility offset rate indicator; These are the target weighting coefficients, satisfying... ; For energy The installed capacity;

[0063] Specifically, the constraints of the multi-energy capacity optimization model include:

[0064] Power balance constraints under weather scenarios:

[0065] ;

[0066] In the formula, It is a weather scenario. The load demand below; energy In the context Average utilization rate; It is a standardized output curve per unit capacity; It refers to energy storage capacity; Number of energy types; For energy The installed capacity;

[0067] Energy storage energy constraints:

[0068] ;

[0069] In the formula, For energy storage time The energy state; These are charging power and discharging power, respectively. For charging and discharging efficiency; It is the time step.

[0070] Specifically, the particle swarm optimization algorithm is used to solve for the optimal capacity configuration, including:

[0071] Encode the position vector of each particle into a capacity solution:

[0072] ;

[0073] In the formula, Indicates the particle position; For wind power installed capacity; This refers to the installed capacity of photovoltaic power. The installed capacity of hydropower; The configured capacity of the energy storage system.

[0074] The fitness function of the particle swarm optimization algorithm is the objective function plus a penalty term:

[0075] ;

[0076] In the formula, Fitness for particle swarm optimization algorithm; For the overall goal; This serves as a measure of violations of constraints such as power balance and energy storage boundaries. It is a punishment factor;

[0077] Particle velocity and position updates:

[0078] ;

[0079] ;

[0080] In the formula, For the first The position and velocity of the particles in the generation; For the individual's historical best; This is the best historical record for the group. Inertial weights; It is the cognitive / social coefficient; It is a random number;

[0081] Velocity and position boundary handling:

[0082] ;

[0083] ;

[0084] In the formula, This is a boundary truncation function used to limit the amplitude of the updated variable, ensuring that it always remains within the given upper and lower bounds. For the first The position and velocity of the particles in the generation; This represents the maximum speed. These represent the lower / upper limits of each energy capacity; For update assignment operations;

[0085] Inertia weight decreases algebraically:

[0086] ;

[0087] In the formula, Initial / final inertia weights; It is the maximum iteration algebra;

[0088] PSO terminates when the maximum number of iterations is reached or the improvement in the population's optimal fitness falls below a threshold, and outputs the globally optimal capacity solution:

[0089] ;

[0090] In the formula, , , , These represent the optimal configuration capacities of wind power, photovoltaic, hydropower, and energy storage systems, respectively, obtained through particle swarm optimization algorithms.

[0091] The present invention has the following advantages:

[0092] This invention constructs a meteorological-fluctuation fusion feature for refined weather classification and independently quantifies the synergistic patterns among wind, solar, hydro, and storage under different weather scenarios. It then establishes a multi-scenario adaptive capacity optimization model, which significantly improves the multi-energy complementary utilization efficiency in high-proportion new energy systems, reduces energy storage configuration costs, and enhances the stability and reliability of system operation. Attached Figure Description

[0093] Figure 1 This is a schematic diagram of the operation optimization method of the present invention. Detailed Implementation

[0094] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention; that is, the described embodiments are merely some embodiments of the invention, and not all embodiments. The components of the embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0095] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0096] The present invention will be further described below with reference to the accompanying drawings, but the scope of protection of the present invention is not limited to the following description.

[0097] like Figure 1 As shown, an optimization method for the coordinated operation of wind, solar, hydro, and storage based on weather typology and volatility characteristics is presented. This method includes:

[0098] S1. Multi-source data construction and normalization: Collect historical output data and meteorological data of wind power, photovoltaic, hydropower and energy storage systems, construct meteorological feature vectors and fluctuation feature vectors, and perform normalization processing.

[0099] To describe the operational characteristics of multi-energy systems including wind power, photovoltaics, hydropower, and energy storage under different weather conditions, a multi-energy output time series and a meteorological driving characteristic series are constructed. Let the first... Energy-like ( (representing wind power, solar power, hydropower, and energy storage respectively) at time... The actual power is Simultaneously, meteorological variables contributing to its output, such as wind speed, irradiance, temperature, cloud cover, water flow, and humidity, are collected to form a meteorological feature vector. To characterize the fluctuation characteristics of renewable energy, this invention introduces fluctuation indicators such as output gradient, ramp rate, and short-time variance to construct a fluctuation feature vector. To avoid modeling bias caused by differences in dimensions and scales, the power is normalized as follows:

[0100] ;

[0101] In the formula, For energy At any moment Normalized power; For energy Minimum / maximum power values ​​in historical samples;

[0102] The fluctuation feature vector includes power gradient, ramp rate, sliding window local variance, and window mean.

[0103] The short-time fluctuation intensity is described by the power gradient, which is:

[0104] ;

[0105] In the formula, For energy At any moment The power gradient; For the first Energy-like energy at all times The actual power; The time step (e.g., 15 min or 1 h);

[0106] The gradient is used to characterize the change in output between adjacent time points, and the gradient is:

[0107] ;

[0108] In the formula, For energy At any moment The larger the slope, the more drastic the short-term fluctuations.

[0109] The continuous volatility over a period of time is quantified using the local variance of a sliding window. The local variance of the sliding window is:

[0110] ;

[0111] In the formula, For energy In length Local power variance within the time window; The average power within the same window; It is the length of the sliding window (e.g., 4, 6, or 24 time points);

[0112] The mean value of the window is:

[0113] .

[0114] Ultimately, building energy The comprehensive feature vector integrates normalized power, volatility, and meteorological data;

[0115] ;

[0116] In the formula, It is energy At any moment The comprehensive feature vector, including normalized power ,gradient Climbing distance Local variance ,as well as Weather characteristics .

[0117] S2. Weather scenario classification based on self-organizing map network: A comprehensive feature vector integrating meteorological and fluctuation feature vectors is input into a self-organizing map network for clustering, dividing the system into several typical weather-fluctuation scenarios. To extract typical power output patterns and fluctuation laws of wind-solar-hydro-storage systems under different meteorological conditions, a self-organizing map (SOM) is used for weather scenario classification based on the constructed multi-source features. This method can map high-dimensional inputs to a two-dimensional topological space while preserving nonlinear feature relationships, thereby achieving coupled clustering of meteorological and power output fluctuation information.

[0118] The specific steps of S2 include:

[0119] The comprehensive feature vectors of all energy sources are integrated into a total system input matrix along the time dimension:

[0120] ;

[0121] In the formula, For a moment The overall system feature vector; For energy The comprehensive feature vector, Number of energy types (wind, solar, hydro, storage);

[0122] Self-organizing map networks consist of an input layer and a competition layer. Assume the competition layer contains... Each node (potential weather type) Having weight vectors with the same dimensions as the input features In each iteration, for the input sample The optimal matching unit is determined by calculating the Euclidean distance to all nodes using a self-organizing map network.

[0123] ;

[0124] In the formula, For a moment The best matching node index; The input feature vector; For nodes The weight vector; It is the Euclidean norm;

[0125] Subsequently, an adaptive update is performed on the weight vectors of the best matching unit node and its neighbors:

[0126] ;

[0127] In the formula, The learning rate decreases over time; For neighborhood functions, Gaussian form is often used:

[0128] ;

[0129] In the formula, For nodes Topological coordinates of the best node in the competing layer of the self-organizing mapping network; The neighborhood width parameter gradually converges with training iterations;

[0130] After training, the self-organizing map network forms a stable set of cluster centers, with each node representing a typical weather-fluctuation scenario, assigning each time-series sample to the nearest category node:

[0131] ;

[0132] In the formula, For a moment The weather type category to which it belongs; This is the node weight vector after training.

[0133] After the above SOM clustering training, all samples were divided into Typical weather-fluctuation types (e.g., low fluctuation type for sunny days, high fluctuation type for cloudy days, steady-state type during wet season, high fluctuation type during dry season, etc.). Each type Corresponding feature center This can serve as a representative feature of the typical power output and fluctuation pattern of the system under this type of weather conditions. Subsequent steps will be based on this to model the collaborative power output law and optimize the capacity.

[0134] S3. Modeling the synergistic output patterns of wind, solar, hydro, and storage under various weather types: For each weather scenario, the linear correlation, nonlinear mutual information, and time series similarity among wind, solar, hydro, and storage are calculated to construct a synergistic relationship matrix. The average synergy degree, fluctuation offset rate, and energy dominance coefficient are extracted as scenario features. After completing the SOM clustering in S2, the samples are divided into several typical weather-fluctuation types. This invention addresses each type of weather scenario. Power sequences of wind power, photovoltaic power, hydropower, and energy storage were extracted from the grouped samples. This method calculates the synergistic output patterns of various energy sources. Unlike traditional global analysis based on continuous time series, this method independently calculates synergy indicators within each weather category, revealing the complementary patterns, regulation characteristics, and fluctuation response patterns of multiple energy sources under different weather conditions.

[0135] Intraclass linear co-operation analysis, for each weather type ,energy and The linear coherence in this type of sample is defined by the Pearson correlation coefficient:

[0136] ;

[0137] In the formula, For weather type Lower energy and The linear correlation coefficient; For covariance; This represents the standard deviation of the corresponding energy in this type of sample; In terms of weather type Lower energy , The power time series sample set; when This indicates a significant complementary relationship between energy sources.

[0138] Intra-class nonlinear cooperative relationship analysis: To characterize nonlinear wave coupling, this invention calculates the mutual information index:

[0139] ;

[0140] In the formula, Weather type Lower energy and Mutual information; For energy Output in discrete amplitude range Marginal distribution probability; It represents a joint probability distribution; the greater the mutual information, the stronger the nonlinear dependence, characterizing the intensity of the common response of different energy sources to weather fluctuations.

[0141] Intra-class temporal similarity analysis, considering the lag in the response of different energy sources to weather changes, this invention uses Dynamic Time Warping (DTW) to measure the similarity of intra-class output sequences:

[0142] ;

[0143] In the formula, For energy Weather type A measure of temporal similarity; This represents the set of all allowed time-matching paths; p and q are time-matching index pairs. The smaller the value, the more synchronized the output modes of energy sources i and j are, and the smaller the response lag.

[0144] The construction of a collaborative relationship matrix under weather scenarios is aimed at integrating linear, nonlinear, and temporal collaborative features to construct weather types. The following multi-energy synergy matrix:

[0145] ;

[0146] In the formula, Weather type Lower energy and The overall synergy coefficient; For the weighting coefficients, satisfying ;matrix It is a symmetric array, describing the coupling modes among multiple energy sources under this weather scenario.

[0147] Synergistic pattern extraction and feature output yield each weather type Multi-energy synergy matrix Subsequently, the present invention further extracts a set of parameters reflecting the overall complementarity and regulatory dominance of the system from the matrix features to form a feature output that can be used for capacity optimization and operation strategy design;

[0148] First, define the average degree of coordination index. This is used to measure the overall energy coordination level of the system under this type of weather scenario:

[0149] ;

[0150] In the formula, Weather type The average synergy coefficient of the system under the following conditions; Indicates the number of energy types; It is a weather scenario. Lower energy and The overall synergy coefficient;

[0151] Secondly, to quantitatively assess the offsetting effect of wind and solar power fluctuations on hydropower and energy storage, this invention proposes a fluctuation offsetting rate index. :

[0152] ;

[0153] In the formula, Weather type Lower system fluctuation offset rate; The standard deviation of the combined power series of wind power and photovoltaic power; The standard deviation of the net power of the wind and solar power system after being smoothed by hydropower and energy storage; when When the value approaches 1, it indicates that the fluctuations in scenery are almost completely offset.

[0154] Furthermore, to identify energy dominance under different weather scenarios, this invention defines an energy dominance coefficient:

[0155] ;

[0156] In the formula, Weather type Lower energy With energy The overall synergy coefficient; Indicates weather type Lower energy The dominant coefficient; the molecule is the energy source. The total strength of synergistic relationships with other energy sources, with the denominator being the total synergistic strength of the system; this indicator is used to identify which energy source plays a dominant regulatory role in this type of weather, such as solar power dominating on sunny days, hydropower dominating on cloudy days, and wind power dominating on windy days, etc.

[0157] Ultimately, this invention will , and The weather scenario feature set is composed of indicators such as:

[0158]

[0159] In the formula, Weather type The collaborative feature vector includes three categories of features: overall system synergy, fluctuation offsetting capability, and energy dominance. This feature set serves as the input parameter set for subsequent capacity optimization models, reflecting the interaction characteristics and regulation potential among multiple energy sources in the system under different weather scenarios, thereby achieving scenario-driven capacity optimization modeling.

[0160] S4. Using the installed capacity of wind, solar, hydro, and storage as decision variables, and aiming at system fluctuation suppression, economic efficiency, and coordinated absorption, a multi-energy capacity optimization model is established, and the optimal capacity configuration is solved using the particle swarm optimization algorithm. This is done under various weather scenarios. Collaborative feature set Subsequently, this invention aims to improve the system's clean energy absorption rate and power output stability by constructing a multi-energy capacity optimization model. This model uses the installed capacity of wind power, photovoltaic power, hydropower, and energy storage as decision variables, comprehensively considering weather scenario probabilities, power smoothness, energy storage utilization, and economic constraints to achieve "weather-based adaptive" capacity matching.

[0161] The objective function of the multi-energy capacity optimization model is:

[0162] ;

[0163] In the formula, The system synthesis objective function; Total number of weather types; Weather type The probability of occurrence; This is the net power fluctuation index of the system under this weather type; For this weather type The following is the operating or capacity cost function; This is a volatility offset rate indicator; These are the target weighting coefficients, satisfying... ; For energy The installed capacity; the objective function achieves a multi-objective balance with minimal fluctuations, lowest cost, and greatest synergy.

[0164] The constraints of the multi-energy capacity optimization model include:

[0165] Power balance constraints under weather scenarios: To ensure a balance between system output and load demand, under each weather type... Establish the power balance equation internally:

[0166] ;

[0167] In the formula, It is a weather scenario. The load demand below; energy In the context Average utilization rate; It is a standardized output curve per unit capacity; This refers to energy storage capacity (positive for discharging, negative for charging). Number of energy types; For energy The installed capacity;

[0168] Energy storage constraints: Energy storage systems must meet energy conservation and charge / discharge power limits.

[0169] ;

[0170] In the formula, For energy storage time The energy state; These are charging power and discharging power, respectively. For charging and discharging efficiency; It is the time step.

[0171] Simultaneously meet the upper and lower limits of capacity and power:

[0172] ;

[0173] Solving for the optimal capacity configuration using the particle swarm optimization algorithm includes:

[0174] Encode the position vector of each particle into a capacity solution:

[0175] ;

[0176] In the formula, The particle position; For wind power installed capacity; This refers to the installed capacity of photovoltaic power. The installed capacity of hydropower; The configured capacity of the energy storage system.

[0177] The fitness function of the particle swarm optimization algorithm is the objective function plus a penalty term (constraint violation penalty):

[0178] ;

[0179] In the formula, Fitness for particle swarm optimization algorithm; For the overall goal; This serves as a measure of violations of constraints such as power balance and energy storage boundaries (e.g., a weighted sum of over-limits). It is a punishment factor;

[0180] PSO core update formula and boundary feasibility, particle velocity and position update:

[0181] ;

[0182] ;

[0183] In the formula, For the first The position and velocity of the particles in the generation; For the individual's historical best; This is the best historical record for the group. Inertial weight; It is the cognitive / social coefficient; It is a random number;

[0184] Velocity and position boundary handling:

[0185] ;

[0186] ;

[0187] In the formula, This is a boundary truncation function; For the first The position and velocity of the particles in the generation; This represents the maximum speed. These are the lower / upper limits of each energy capacity; if there are still violations of these constraints, then... Increase the penalty or project to the nearest feasible solution; For update assignment operations.

[0188] Inertia weight decreases algebraically:

[0189] ;

[0190] In the formula, Initial / final inertia weights; It is the maximum iteration algebra;

[0191] PSO terminates when the maximum number of iterations is reached or the improvement in the population's optimal fitness falls below a threshold, and outputs the globally optimal capacity solution:

[0192] ;

[0193] In the formula, , , , These represent the optimal configuration capacities of wind power, photovoltaic, hydropower, and energy storage systems, respectively, obtained through particle swarm optimization algorithms.

[0194] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the present invention. Any person skilled in the art can make many possible variations and modifications to the technical solution of the present invention, or modify it into equivalent embodiments, without departing from the scope of the present invention. Therefore, any modifications, equivalent changes, and alterations made to the above embodiments based on the technology of the present invention without departing from the scope of the present invention are within the protection scope of the present invention.

Claims

1. An optimization method for the coordinated operation of wind, solar, hydro, and storage based on weather patterns and volatility characteristics, characterized by: The method includes, S1. Multi-source data construction and normalization: Collect historical output data and meteorological data of wind power, photovoltaic, hydropower and energy storage systems, construct meteorological feature vectors and fluctuation feature vectors, and perform normalization processing. S2. Weather scenario classification based on self-organizing map network: The comprehensive feature vector that integrates meteorological feature vector and fluctuation feature vector is input into the self-organizing map network for clustering, and several typical weather-fluctuation scenarios are divided. S3. Modeling the collaborative output law of wind, solar, hydro and storage under various weather types. For each weather scenario, calculate the linear correlation, nonlinear mutual information and time series similarity between wind, solar, hydro and storage, construct a collaborative relationship matrix, and extract the average collaborative degree, fluctuation offset rate and energy dominance coefficient as scenario features. The collaborative relationship matrix is ​​as follows: ； In the formula, Weather type Lower energy and The overall synergy coefficient; For the weighting coefficients, satisfying ; For weather type Lower energy and The linear correlation coefficient; Weather type Lower energy and Mutual information; For energy Weather type A measure of temporal similarity; The average degree of coordination is: ； In the formula, Weather type The average synergy coefficient of the system under the following conditions; Indicates the number of energy types; It is a weather scenario. Lower energy and The overall synergy coefficient; The volatility offset rate is: ； In the formula, Weather type Lower system fluctuation offset rate; The standard deviation of the combined power series of wind power and photovoltaic power; The standard deviation of the net power of the wind and solar power system after being smoothed by hydropower and energy storage; The energy dominance coefficient is: ； In the formula, Indicates weather type Lower energy The dominant coefficient; Weather type Lower energy With energy The comprehensive synergy coefficient; the average synergy degree, fluctuation offset rate and energy dominance coefficient constitute the weather scenario feature set, and participate in capacity optimization as the input parameter set of the multi-energy capacity optimization model; S4. Using the installed capacity of wind, solar, hydro, and storage as decision variables, and with the goals of system fluctuation suppression, economy, and coordinated absorption, a multi-energy capacity optimization model is established, and the optimal capacity configuration is solved using the particle swarm optimization algorithm. The objective function of the multi-energy capacity optimization model is: ； In the formula, The system synthesis objective function; Total number of weather types; Weather type The probability of occurrence; This is the net power fluctuation index of the system under this weather type; For this weather type The following is the operating or capacity cost function; This is a volatility offset rate indicator; These are the target weighting coefficients, satisfying... ; For energy The installed capacity.

2. The wind-solar-hydro-storage coordinated operation optimization method based on weather morphology and volatility characteristics as described in claim 1, characterized in that: The meteorological feature vector is: ; The fluctuation feature vector includes power gradient, ramp rate, sliding window local variance, and window mean. The power gradient is: ； In the formula, For energy At any moment The power gradient; For the first Energy-like energy at all times The actual power; For time step; The gradient is: ； In the formula, For energy At any moment The amount of uphill climb; The local variance of the sliding window is: ； In the formula, For energy In length Local power variance within the time window; The average power within the same window; The mean value of the window is: 。 3. The wind-solar-hydro-storage coordinated operation optimization method based on weather morphology and volatility characteristics as described in claim 1, characterized in that: The specific steps of S2 include: The integrated feature vectors are combined into a system-wide input matrix along the time dimension: ； In the formula, For a moment The overall system feature vector; For energy The comprehensive feature vector, Number of energy types; The optimal matching unit is determined by calculating the Euclidean distance to all nodes using a self-organizing map network. ； In the formula, For a moment The best matching node index; The input feature vector; For nodes The weight vector; It is the Euclidean norm; Perform adaptive updates on the weight vectors of the best-matching node and its neighbors: ； In the formula, The learning rate; It is a neighborhood function; ； In the formula, For nodes Topological coordinates of the best node in the competing layer of the self-organizing mapping network; This is the neighborhood width parameter; Assign each time-instance sample to the nearest class node: ； In the formula, For a moment The weather type category to which it belongs; This is the node weight vector after training.

4. The wind-solar-hydro-storage coordinated operation optimization method based on weather typology and volatility characteristics as described in claim 1, characterized in that: In S3, the linear correlation between wind, solar, hydro, and storage is defined as: ； In the formula, For weather type Lower energy and The linear correlation coefficient; For covariance; This represents the standard deviation of the corresponding energy in this type of sample; In terms of weather type Lower energy , A set of power time series samples; The nonlinear mutual information between wind, solar, hydro, and storage in S3 is: ； In the formula, Weather type Lower energy and Mutual information; For energy Output in discrete amplitude range Marginal distribution probability; It is a joint probability distribution; The temporal similarity between wind, solar, hydro, and storage in S3 is as follows: ； In the formula, For energy Weather type A measure of temporal similarity; This represents the set of all allowed time-matching paths; p and q are time-matching index pairs.

5. The wind-solar-hydro-storage coordinated operation optimization method based on weather typology and volatility characteristics as described in claim 4, characterized in that: The constraints of the multi-energy capacity optimization model include: Power balance constraints under weather scenarios: ； In the formula, It is a weather scenario. The load demand below; energy In the context Average utilization rate; It is a standardized output curve per unit capacity; It refers to energy storage capacity; Number of energy types; For energy The installed capacity; Energy storage energy constraints: ； In the formula, For energy storage time The energy state; These are charging power and discharging power, respectively. For charging and discharging efficiency; It is the time step.

6. The wind-solar-hydro-storage coordinated operation optimization method based on weather typology and volatility characteristics as described in claim 1, characterized in that: Solving for the optimal capacity configuration using the particle swarm optimization algorithm includes: Encode the position vector of each particle into a capacity solution: ； In the formula, The particle position; For wind power installed capacity; This refers to the installed capacity of photovoltaic power. The installed capacity of hydropower; The configured capacity of the energy storage system; The fitness function of the particle swarm optimization algorithm is the objective function plus a penalty term: ； In the formula, Fitness for particle swarm optimization algorithm; For the overall goal; This serves as a measure of violations of power balance and energy storage boundary constraints. It is a punishment factor; Particle velocity and position updates: ； ； In the formula, For the first The position and velocity of the particles in the generation; For the individual's historical best; This is the best historical record for the group. Inertial weights; It is the cognitive / social coefficient; It is a random number; Velocity and position boundary handling: ； ； In the formula, This is a boundary truncation function; For the first The position and velocity of the particles in the generation, This represents the maximum speed. These are the lower / upper limits of each energy capacity; For update assignment operations; Inertia weight decreases algebraically: ； In the formula, Initial / final inertia weights; It is the maximum iteration algebra; PSO terminates when the maximum number of iterations is reached or the improvement in the population's optimal fitness falls below a threshold, and outputs the globally optimal capacity solution: ； In the formula, , , , These represent the optimal configuration capacities of wind power, photovoltaic, hydropower, and energy storage systems, respectively, obtained through particle swarm optimization algorithms.

Images