A configuration method for energy storage capacity based on unbalanced wind power

Abstract

The present invention provides a method for configuring energy storage capacity based on unbalanced wind power, belonging to the technical field of power systems. The method comprises: preprocessing historical wind power data according to historical meteorological data of a target area, dividing the preprocessed data according to time scales, obtaining a dual analysis algorithm at each time scale from a scale-algorithm comparison table, decomposing and analyzing the corresponding divided data to obtain an unbalanced power array at the corresponding time scale, determining the unbalance characteristics at the corresponding time scale based on the unbalanced power array, optimizing the energy storage capacity configuration model, and determining a capacity configuration plan based on the current energy storage demand and actual energy storage situation of the target area to meet the actual needs of the power system.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a method for configuring energy storage capacity based on unbalanced wind power. Background Art

[0002] As global demand for clean energy continues to grow, wind power, as a clean, renewable energy source, is increasingly contributing to the growth of power systems. However, due to the randomness and volatility of wind resources, wind power output has significant uncertainty. This unbalanced power poses numerous challenges to the stability, reliability, and power quality of power systems.

[0003] To address the impact of unbalanced wind power, configuring an appropriate energy storage system is an effective solution. Currently, there are some issues with energy storage capacity configuration. For example, traditional configuration methods are mostly based on empirical data or simple statistical analysis, without fully considering the dynamic changes in wind power and the characteristics of unbalanced power at different time scales. As a result, the energy storage capacity configuration is either too large, resulting in wasted resources and increased costs, or too small, unable to effectively balance the unbalanced wind power and making it difficult to meet the actual needs of the power system.

[0004] Therefore, the present invention proposes a method for configuring energy storage capacity based on unbalanced wind power. Summary of the Invention

[0005] The present invention provides a method for configuring energy storage capacity based on unbalanced wind power to solve the above-mentioned technical problems.

[0006] The present invention provides a method for configuring energy storage capacity based on unbalanced wind power, comprising:

[0007] Step 1: Collect historical wind power data of a target area and pre-process the historical wind power data according to historical meteorological data of the target area, wherein the historical wind power data includes wind power output power in different seasons and under different weather conditions;

[0008] Step 2: Divide the preprocessed data according to the time scale, and obtain the dual analysis algorithm at each time scale from the scale-algorithm comparison table to decompose and analyze the corresponding divided data to obtain an unbalanced power array of the corresponding time scale, wherein the unbalanced power array includes N1 unbalanced power components obtained based on the first analysis algorithm and N2 unbalanced power components obtained based on the second analysis algorithm;

[0009] Step 3: Establish an energy storage capacity configuration model, taking the power system's requirement to smooth wind power fluctuations as a constraint and maximizing the energy storage system's full lifecycle operating efficiency as the objective function.

[0010] Step 4: Determine the imbalance characteristics at the corresponding time scale based on the unbalanced power array, optimize the energy storage capacity configuration model, and determine the capacity configuration plan based on the current energy storage demand and the current actual energy storage situation of the target area.

[0011] Preferably, preprocessing the historical wind power data according to the historical meteorological data of the target area includes:

[0012] Sort the historical wind power at each historical collection moment in sequence to determine whether there is missing power;

[0013] If it exists, obtain the historical meteorological data at the missing time and analyze it to obtain the initial filling power;

[0014] Acquire a first reference power and first meteorological data at a first non-missing time closest to the missing time, and at the same time, acquire a second reference power and second meteorological data at a second non-missing time second closest to the missing time;

[0015] Obtaining a first predicted power based on the first meteorological data and an operating mode of the wind power system at a first non-missing moment, and simultaneously obtaining a second predicted power based on the second meteorological data and an operating mode of the wind power system at a second non-missing moment;

[0016] Correcting the initial filling power according to the first reference power, the second reference power, the first predicted power, and the second predicted power;

[0017]

[0018]

[0019]

[0020] in, represents the fine-tuning function; Represents the historical meteorological data Q at the missing moment and the historical meteorological data at each remaining historical collection moment The number of similar moments with a similarity greater than or equal to 0.9; represents the historical wind power at the jth similar moment; Represents positive and negative functions; 、 represent the first reference power and the first predicted power respectively; represent the second reference power and the second predicted power respectively; Indicates the power after a correction; represents the initial filling power; Represents the historical meteorological data Q at the missing time and the first meteorological data at the first non-missing time Similarity function of Represents the historical meteorological data Q at the missing time and the second meteorological data at the second non-missing time The similarity function of .

[0021] Preferably, after correcting the initial filling power once, the method further includes:

[0022] Lock the continuous time segment according to the time distribution of all similar moments, draw a power curve based on the continuous time segment, and intercept the power curve with the power after the first correction as a straight line to obtain the excess increment of the upper intercepted part and the proportion of the upper intercepted time, the excess attenuation of the lower intercepted part and the proportion of the lower intercepted time, and perform a secondary correction and completion on the power after the first correction;

[0023]

[0024]

[0025]

[0026] in, represents the power after secondary correction; 、 Respectively represent the existence of the upper cut part and The maximum absolute difference between the lower intercepted part and The maximum absolute difference; 、 Respectively represent the enclosed area of ​​the upper cut-off part and the lower cut-off part; Indicates the total number of historical collection moments; Represents positive and negative functions; The constant value is 1; 、 Respectively represent the proportion of upper interception time and the proportion of lower interception time; They represent excess attenuation and excess gain respectively.

[0027] Preferably, the unbalanced power component of each analysis algorithm is set in advance.

[0028] Preferably, determining the imbalance characteristics at a corresponding time scale based on the imbalance power array includes:

[0029] Randomly shuffling and splitting the order of unbalanced powers in each unbalanced power array to obtain a plurality of new arrays, and performing time domain feature extraction on the new arrays;

[0030] Performing a fast Fourier transform on the new array, analyzing the main frequency components and assigning values ​​to obtain the dominant frequency, quantifying the energy distribution of the dominant frequency components, and extracting frequency characteristic parameters;

[0031] Decomposing the new array into sub-signals of different frequency bands based on wavelet basis functions and the number of decomposition layers, and obtaining the variation characteristics of the unbalanced power in time and frequency by analyzing the time-frequency graph;

[0032] The time domain features, frequency domain features and time-frequency domain features are integrated and dimensionality reduced to obtain a feature vector;

[0033] Clustering algorithm is used to perform cluster analysis on cluster vectors to divide the imbalanced power arrays at different time scales into different categories. The imbalance characteristics of each category are determined by analyzing the feature mean and distribution of each category.

[0034] Preferably, optimizing the energy storage capacity configuration model includes:

[0035] Integrate the imbalance features at different time scales to construct a multi-dimensional feature matrix;

[0036] The grey correlation analysis algorithm is used to calculate the correlation between the characteristics of each time scale, explore the potential connection between the imbalance characteristics of different time scales, and introduce the influencing factors reflecting the imbalance characteristics of each time scale;

[0037] Based on the influencing factors, the constraints in the energy storage capacity configuration model are refined to achieve optimization.

[0038] Preferably, the capacity configuration plan is determined based on the current energy storage demand and the current actual energy storage situation of the target area, including:

[0039] Analyze the current energy storage demand to obtain the power difference between the output power and the power load, and at the same time, predict the demand trend after the current energy storage demand based on several consecutive demands before the current energy storage demand;

[0040] Determine the energy storage system status and charging and discharging capacity based on the actual current energy storage situation;

[0041] Taking the power difference, demand trend, energy storage system status, and charging and discharging capabilities as input, the energy storage capacity and power values ​​are continuously adjusted to find the optimal solution that meets the constraints and minimizes the objective function, thereby obtaining the current optimal capacity configuration plan.

[0042] Preferably, the time scale includes: an ultra-short-term time scale, a short-term time scale, and a medium- to long-term time scale.

[0043] Compared with the prior art, the present invention has the following advantages:

[0044] After preprocessing, accurate, complete, and standardized historical wind power data is obtained, which can truly reflect the wind power output of the target area in different seasons and weather conditions. By processing data at different time scales through a dual analysis algorithm, the imbalance characteristics of wind power over different timeframes can be deeply explored from multiple perspectives, resulting in a comprehensive and detailed imbalance power array. The established energy storage capacity configuration model is based on actual power system demand and energy storage system performance. By optimizing the objective function and complying with constraints, it can calculate the configuration plan that maximizes the full lifecycle operating efficiency of the energy storage system while meeting wind power balancing requirements. By determining the imbalance characteristics based on the imbalance power array and optimizing the model, the energy storage capacity configuration model can more accurately reflect the imbalance characteristics of wind power, improving the model's adaptability and accuracy. The capacity configuration plan is then determined based on current energy storage demand and actual conditions, achieving dynamic and accurate energy storage capacity configuration, ensuring that the energy storage system can effectively balance wind power fluctuations and meet power system needs at all times. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] Figure 1 This is a flow chart of a method for configuring energy storage capacity based on unbalanced wind power provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0047] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0048] The present invention provides a method for configuring energy storage capacity based on unbalanced wind power, such as Figure 1 Shown, including:

[0049] Step 1: Collect historical wind power data of a target area and pre-process the historical wind power data according to historical meteorological data of the target area, wherein the historical wind power data includes wind power output power in different seasons and under different weather conditions;

[0050] Step 2: Divide the preprocessed data according to the time scale, and obtain the dual analysis algorithm at each time scale from the scale-algorithm comparison table to decompose and analyze the corresponding divided data to obtain an unbalanced power array of the corresponding time scale, wherein the unbalanced power array includes N1 unbalanced power components obtained based on the first analysis algorithm and N2 unbalanced power components obtained based on the second analysis algorithm;

[0051] Step 3: Establish an energy storage capacity configuration model, taking the power system's requirement to smooth wind power fluctuations as a constraint and maximizing the energy storage system's full lifecycle operating efficiency as the objective function.

[0052] Step 4: Determine the imbalance characteristics at the corresponding time scale based on the unbalanced power array, optimize the energy storage capacity configuration model, and determine the capacity configuration plan based on the current energy storage demand and the current actual energy storage situation of the target area.

[0053] Preferably, the unbalanced power component of each analysis algorithm is pre-set. Preferably, the time scale includes: an ultra-short-term time scale, a short-term time scale, and a medium- to long-term time scale.

[0054] In this embodiment, the target area is a specific wind farm and its surrounding power coverage area, for example, a large wind farm and the regional power grid to which it is connected.

[0055] In this embodiment, historical wind power data refers to the wind power output values ​​of the wind farm at different times over the past period (e.g., the past three years), such as 1000 kW wind power output at 00:00 on January 1, 2022, and 3500 kW wind power output at 12:00 on July 15, 2022. This data covers power output in spring, summer, autumn, and winter, as well as under various weather conditions, including sunny, rainy, windy, and thunderstorm days. The historical meteorological data and historical wind power data correspond to meteorological observation data for the target area, including wind speed (e.g., 8 m / s wind speed at 15:00 on March 20, 2022), wind direction (e.g., northwest wind at 09:00 on October 5, 2022), temperature (e.g., -20°C temperature at 00:00 on January 10, 2023), and humidity (e.g., 60% humidity at 14:00 on June 25, 2023).

[0056] Preprocessing is to fill in the missing data to make it meet the requirements of subsequent analysis.

[0057] In this embodiment, there are ultra-short term time scales (0-4 hours), short term time scales (4-24 hours), and medium- to long-term time scales (1-several days).

[0058] In this embodiment, the scale-algorithm comparison table is a pre-established table that records the applicable analysis algorithms for each time scale. For example, for the ultra-short time scale, the strain modal decomposition (VMD) algorithm and the improved empirical mode decomposition (EEMD) algorithm are used; for the short time scale, the Fourier transform-based seasonal decomposition method (STL) and the gated recurrent unit-autoregressive integrated moving average (GRU-ARIMA) hybrid algorithm are used; and for the medium- and long-term time scales, the long short-term memory network-grey prediction model (LSTM-GM) fusion algorithm and the wavelet transform-based time series analysis method are used. Two algorithms are used to analyze data at each time scale. The array containing multiple unbalanced power components obtained through algorithmic analysis is a quantitative representation of the unbalanced characteristics of wind power.

[0059] In this embodiment, the wind power data is decomposed to obtain power variation parts that can reflect different frequencies or characteristics. Each component represents a certain characteristic of wind power imbalance. The number of components (N1, N2) is determined according to the algorithm and data characteristics and is set in advance.

[0060] In this embodiment, the power system's requirements for smoothing wind power fluctuations are to ensure stable operation of the power system, and limit standards are set for the wind power fluctuation range, fluctuation frequency, etc. For example, it is required that the wind power fluctuation rate does not exceed 10% within a certain time period, or that the power change rate cannot exceed a certain value (such as 50kW / min).

[0061] In this embodiment, the full life cycle operating efficiency of the energy storage system is an indicator that measures the comprehensive performance of the energy storage system throughout its entire service life, including energy storage and release efficiency, operational stability, and maintenance cost. Maximum operating efficiency means that, under the premise of meeting the requirements of the power system, the energy storage system achieves the highest energy utilization rate, the lowest cost, and the best benefits throughout its entire life cycle.

[0062] In this embodiment, the energy storage capacity configuration model is a mathematical model established based on the objective function and constraint conditions, and is used to calculate and determine the optimal capacity and power configuration scheme of the energy storage system.

[0063] In this embodiment, imbalance characteristics are extracted from the imbalance power array and are parameters or patterns that can reflect the essential characteristics of wind power imbalance at different time scales, such as the intensity of ultra-short-term high-frequency fluctuations, short-term cyclical changes, and medium- to long-term trends. The structure, parameters, and objective function of the energy storage capacity configuration model are adjusted and improved to better meet actual needs and enhance the model's accuracy and practicality.

[0064] In this embodiment, the capacity configuration plan is a plan for determining specific configuration parameters of the energy storage system, such as the capacity size (e.g., 1000 kWh) and power parameters (e.g., 500 kW), and is used to guide the construction and operation of the energy storage system.

[0065] The above technical solution has the following beneficial effects: after preprocessing, accurate, complete, and standardized historical wind power data is obtained, which can truly reflect the wind power output of the target area under different seasons and weather conditions. By processing data at different time scales through a dual analysis algorithm, the imbalance characteristics of wind power over different time frames can be deeply explored from multiple perspectives, resulting in a comprehensive and detailed imbalance power array. The established energy storage capacity configuration model is based on actual power system demand and energy storage system performance. By optimizing the objective function and complying with constraints, it can calculate the configuration plan that maximizes the full lifecycle operating efficiency of the energy storage system while meeting wind power balancing requirements. By determining the imbalance characteristics based on the imbalance power array and optimizing the model, the energy storage capacity configuration model can more accurately reflect the imbalance characteristics of wind power, improving the model's adaptability and accuracy. The capacity configuration plan is then determined based on current energy storage demand and actual conditions, achieving dynamic and accurate configuration of energy storage capacity, ensuring that the energy storage system can effectively balance wind power fluctuations and meet power system needs at all times.

[0066] The present invention provides a method for configuring energy storage capacity based on unbalanced wind power, which pre-processes historical wind power data according to historical meteorological data of the target area, including:

[0067] Sort the historical wind power at each historical collection moment in sequence to determine whether there is missing power;

[0068] If it exists, obtain the historical meteorological data at the missing time and analyze it to obtain the initial filling power;

[0069] Acquire a first reference power and first meteorological data at a first non-missing time closest to the missing time, and at the same time, acquire a second reference power and second meteorological data at a second non-missing time second closest to the missing time;

[0070] Obtaining a first predicted power based on the first meteorological data and an operating mode of the wind power system at a first non-missing moment, and simultaneously obtaining a second predicted power based on the second meteorological data and an operating mode of the wind power system at a second non-missing moment;

[0071] Correcting the initial filling power according to the first reference power, the second reference power, the first predicted power, and the second predicted power;

[0072]

[0073]

[0074]

[0075] in, represents the fine-tuning function; Represents the historical meteorological data Q at the missing moment and the historical meteorological data at each remaining historical collection moment The number of similar moments with a similarity greater than or equal to 0.9; represents the historical wind power at the jth similar moment; Represents positive and negative functions; 、 represent the first reference power and the first predicted power respectively; represent the second reference power and the second predicted power respectively; Indicates the power after a correction; represents the initial filling power; Represents the historical meteorological data Q at the missing time and the first meteorological data at the first non-missing time Similarity function of Represents the historical meteorological data Q at the missing time and the second meteorological data at the second non-missing time The similarity function of .

[0076] In this embodiment, multiple variables related to similar historical meteorological data are introduced to comprehensively consider the influence of power reference values ​​and predicted power values ​​at different times. It is based on the principle that wind power under similar meteorological conditions is similar. The power under similar conditions is statistically analyzed to provide a basis for correction. The method is to find the average power of these similar moments to make the result more representative. Add C (initial filling power) and then divide by 2. This is to comprehensively consider the power conditions at similar moments based on the initial filling power, and balance the relationship between the initial estimate and the statistics of similar situations.

[0077] The average value of the absolute value of the difference between the two reference powers and the corresponding predicted powers is calculated, which reflects the degree of deviation between the reference power and the predicted power. The degree of deviation is weighted using the minimum similarity between the meteorological data at the missing time and the two adjacent non-missing times. Lower similarity indicates greater differences in meteorological conditions. Even if the reference power deviates significantly from the predicted power, this deviation should not be overly relied upon for correction. This weighting approach makes fine-tuning more reasonable.

[0078] The beneficial effects of this technical solution are: filling in and correcting missing wind power data effectively improves the integrity of historical wind power data. By rationally correcting missing power data, the subsequent determination of imbalance power arrays and imbalance characteristics based on this data becomes more accurate. This, in turn, allows for more accurate reflection of actual wind power fluctuations when optimizing the energy storage capacity configuration model, making the configuration model's constraints and objective functions more realistic.

[0079] The present invention provides a method for configuring energy storage capacity based on unbalanced wind power, which, after correcting the initial filling power, further includes:

[0080] Lock the continuous time segment according to the time distribution of all similar moments, draw a power curve based on the continuous time segment, and intercept the power curve with the power after the first correction as a straight line to obtain the excess increment of the upper intercepted part and the proportion of the upper intercepted time, the excess attenuation of the lower intercepted part and the proportion of the lower intercepted time, and perform a secondary correction and completion on the power after the first correction;

[0081]

[0082]

[0083]

[0084] in, represents the power after secondary correction; 、 Respectively represent the existence of the upper cut part and The maximum absolute difference between the lower intercepted part and The maximum absolute difference; 、 Respectively represent the enclosed area of ​​the upper cut-off part and the lower cut-off part; Indicates the total number of historical collection moments; Represents positive and negative functions; The constant value is 1; 、 Respectively represent the proportion of upper interception time and the proportion of lower interception time; They represent excess attenuation and excess gain respectively.

[0085] In this embodiment, the redundant increment refers to the average of the absolute values ​​of all differences between the power value at each moment in the upper cut portion and the power after the first correction.

[0086] The excess attenuation refers to the average of the absolute values ​​of all differences between the power value at each moment in the lower cut portion and the power after one correction.

[0087] In this embodiment, the time ratio is the number of time involved in the corresponding intercepted part / the total number of time.

[0088] In this embodiment, Taking into account the time proportion of the upper and lower intercepted parts, as well as the excess increment and excess attenuation, It reflects the difference in the proportion of the upper and lower interception times, multiplied by The power amplitude that needs further adjustment is measured by combining the difference between the upper and lower power margins.

[0089] according to The positive or negative value is used to determine the value, and the adjustment direction is accurately judged to ensure that the secondary correction can reasonably determine whether to adjust the power value in the positive or negative direction based on the area difference between the upper and lower parts of the power curve.

[0090] The beneficial effect of the above technical solution is: through secondary correction, combined with information such as the time proportion of the upper and lower intercepted parts of the power curve, redundant increments and attenuation, the power value can be adjusted more finely, so that the filled power data is more in line with the actual power change trend, thereby improving the accuracy of the power data.

[0091] The present invention provides a method for configuring energy storage capacity based on unbalanced wind power, which determines the unbalance characteristics at a corresponding time scale based on the unbalanced power array, including:

[0092] Randomly shuffling and splitting the order of unbalanced powers in each unbalanced power array to obtain a plurality of new arrays, and performing time domain feature extraction on the new arrays;

[0093] Performing a fast Fourier transform on the new array, analyzing the main frequency components and assigning values ​​to obtain the dominant frequency, quantifying the energy distribution of the dominant frequency components, and extracting frequency characteristic parameters;

[0094] Decomposing the new array into sub-signals of different frequency bands based on wavelet basis functions and the number of decomposition layers, and obtaining the variation characteristics of the unbalanced power in time and frequency by analyzing the time-frequency graph;

[0095] The time domain features, frequency domain features and time-frequency domain features are integrated and dimensionality reduced to obtain a feature vector;

[0096] Clustering algorithm is used to perform cluster analysis on cluster vectors to divide the imbalanced power arrays at different time scales into different categories. The imbalance characteristics of each category are determined by analyzing the feature mean and distribution of each category.

[0097] In this embodiment, an unbalanced array, such as [10kW, 15kW, 8kW, 12kW, 9kW], is randomly reordered to obtain [12kW, 9kW, 15kW, 8kW, 10kW], which can be split into [12kW, 9kW], [15kW, 8kW], and [10kW] to obtain several new arrays. A random function in a computer programming language, such as the random.shuffle() function in Python, is used to randomly reorder the elements in the unbalanced power array.

[0098] In this embodiment, the numpy.fft.fft() function in Python performs a fast Fourier transform on the new array to obtain frequency domain data, analyzes the amplitude of each frequency component in the frequency domain data, finds the main frequency components with larger amplitudes, determines the dominant frequency with the largest amplitude, and calculates the frequency characteristic parameters based on the energy values ​​of the dominant frequency and other frequency components. For example, the main frequency energy ratio is obtained by dividing the dominant frequency energy by the total energy.

[0099] In this embodiment, the signals of different frequency bands are obtained after wavelet transform decomposition. For example, the new array [12kW, 9kW, 15kW, 8kW, 10kW] is decomposed based on the db4 wavelet basis function and the decomposition level is 3 to obtain the low-frequency sub-signals [10kW, 11kW], the medium-frequency sub-signals [3kW, 2kW], and the high-frequency sub-signals [1kW, 0.5kW].

[0100] In this embodiment, the time-frequency diagram uses time as the horizontal axis and frequency as the vertical axis to display the energy distribution of the signal at different times and frequencies. The time-frequency diagram can be used to see which frequency components are dominant at a certain moment.

[0101] In this embodiment, it is assumed that the time domain features have a mean of 10kW and a variance of 2kW2. Frequency domain features: such as the main frequency energy accounting for 40%, the high frequency energy accounting for 10%, etc. Time-frequency domain features: such as the energy density value of a specific area in the time-frequency graph.

[0102] In this embodiment, the mean and variance of time-domain features, the dominant frequency energy ratio and high frequency energy ratio of frequency-domain features, and the energy density of time-frequency domain features are combined and then reduced in dimension using an algorithm. For example, principal component analysis (PCA) is used to convert high-dimensional features into low-dimensional feature vectors. For example, a two-dimensional vector [0.5, 0.3] combines key information from the time domain, frequency domain, and time-frequency domain.

[0103] In this embodiment, the clustering algorithm adopts the K-means clustering algorithm, and the different groups divided by the clustering algorithm are assumed to be divided into three categories, namely category A, category B, and category C, using the K-means clustering algorithm.

[0104] In this embodiment, for each category's feature vector, the mean value of each dimension is calculated. For example, if the values ​​of the feature vectors in category A in the first dimension are 0.4, 0.5, and 0.6, then the feature mean value in this dimension is (0.4 + 0.5 + 0.6) ÷ 3 = 0.5. The distribution describes the distribution of the feature vectors in the category in the feature space, such as whether they are concentrated and the degree of dispersion. Category A may represent an unbalanced feature with high-frequency fluctuations and large fluctuations.

[0105] The beneficial effects of the above technical solution are as follows: random shuffling and splitting can break the sequential correlation of the original data, increase the randomness and diversity of the data, and help to more comprehensively explore data features. Through fast Fourier transform and correlation analysis, the characteristics of unbalanced power can be revealed from the frequency dimension, and the frequency components that play a major role in power fluctuations can be found. Wavelet transform can analyze unbalanced power in both time and frequency dimensions simultaneously, and compared with simple time domain or frequency domain analysis, it can more comprehensively capture the local characteristics of the signal. Feature fusion can comprehensively utilize feature information from different dimensions to fully describe the characteristics of unbalanced power. Cluster analysis can classify unbalanced power arrays with similar unbalanced characteristics into the same category, realizing the classification management and analysis of unbalanced power. The imbalance characteristics determined by analyzing the feature mean and distribution of each category help to more clearly understand the different types of wind power imbalance.

[0106] The present invention provides a method for configuring energy storage capacity based on unbalanced wind power, which optimizes the energy storage capacity configuration model, including:

[0107] Integrate the imbalance features at different time scales to construct a multi-dimensional feature matrix;

[0108] The grey correlation analysis algorithm is used to calculate the correlation between the characteristics of each time scale, explore the potential connection between the imbalance characteristics of different time scales, and introduce the influencing factors reflecting the imbalance characteristics of each time scale;

[0109] Based on the influencing factors, the constraints in the energy storage capacity configuration model are refined to achieve optimization.

[0110] In this embodiment, the numerical value measuring the closeness of the correlation between two or more features generally ranges from 0 to 1. For example, if the correlation between the ultra-short-term high-frequency fluctuation intensity and the short-term cyclical fluctuation amplitude is calculated to be 0.6, it indicates that there is a certain correlation between the two.

[0111] Hidden and difficult-to-observe relationships between imbalance characteristics at different time scales. For example, ultra-short-term high-frequency fluctuations may affect short-term power fluctuation cycles to a certain extent. This influence relationship is a potential connection.

[0112] A coefficient, determined based on the correlation and potential connections between timescale features, is used to measure the impact of a particular timescale feature on the energy storage capacity configuration model. For example, the impact factor of an ultra-short-term feature might be 0.3, indicating its relative importance in the model.

[0113] In this embodiment, the formula of the correlation coefficient is as follows:

[0114]

[0115] Among them, x0(k) is the value of the reference sequence at time k, xi(k) is the value of the comparison sequence at time k, ρ is the resolution coefficient (generally taken as 0.5), and the correlation coefficient is averaged to obtain the correlation degree.

[0116] Based on the calculated correlations, combined with professional knowledge and practical experience, we analyze the potential connections between features at different timescales. For example, we found a high correlation between the intensity of ultra-short-term high-frequency fluctuations and the amplitude of short-term cyclical fluctuations, which may indicate that ultra-short-term high-frequency fluctuations have an impact on the short-term cyclical fluctuations. Then, based on the magnitude of the correlation and the importance of the potential connection, we determine the impact factor of each timescale feature. Features with high correlations and a significant impact on energy storage configuration have relatively large impact factors.

[0117] In this embodiment, the ultra-short-term characteristic impact factor is 0.3, the short-term characteristic impact factor is 0.2, and the medium- to long-term characteristic impact factor is 0.5.

[0118] In this embodiment, the instantaneous power change rate limit in the power constraint is adjusted based on the ultra-short-term characteristic impact factor; the long-term capacity adjustment range in the capacity constraint is modified based on the medium- and long-term characteristic impact factor, which is a refinement process. If the ultra-short-term characteristic impact factor is high, it means that the ultra-short-term imbalance characteristic has high requirements for the power response speed of the energy storage system, so it is necessary to focus on adjusting the instantaneous power change rate limit in the power constraint; if the medium- and long-term characteristic impact factor is large, it means that the medium- and long-term imbalance characteristic has a great impact on the long-term planning of the energy storage capacity, and attention should be paid to the adjustment of the capacity constraint. If the ultra-short-term characteristic impact factor is large, the instantaneous power change rate limit value is reduced, such as from the original 50kW / min to 30kW / min, to better adapt to the power change requirements of ultra-short-term high-frequency fluctuations; if the medium- and long-term characteristic impact factor is large, the capacity adjustment range in the capacity constraint is appropriately expanded according to the medium- and long-term trend changes, such as increasing the capacity upper limit from 80% of the original rated capacity to 90%.

[0119] The beneficial effects of this technical solution are: by integrating and constructing a multidimensional feature matrix, it is possible to systematically organize and present imbalance characteristics at different time scales, making this feature information clearer and more organized. Using a gray correlation analysis algorithm, it is possible to quantitatively analyze the degree of correlation between imbalance characteristics at different time scales, uncovering potential connections. By refining the constraints of the energy storage capacity configuration model based on influencing factors, the model's constraints can be made more consistent with the actual imbalance characteristics at different time scales.

[0120] The present invention provides a method for configuring energy storage capacity based on unbalanced wind power, which determines a capacity configuration scheme based on the current energy storage demand and the current actual energy storage situation of the target area, including:

[0121] Analyze the current energy storage demand to obtain the power difference between the output power and the power load, and at the same time, predict the demand trend after the current energy storage demand based on several consecutive demands before the current energy storage demand;

[0122] Determine the energy storage system status and charging and discharging capacity based on the actual current energy storage situation;

[0123] Taking the power difference, demand trend, energy storage system status, and charging and discharging capabilities as input, the energy storage capacity and power values ​​are continuously adjusted to find the optimal solution that meets the constraints and minimizes the objective function, thereby obtaining the current optimal capacity configuration plan.

[0124] In this embodiment, demand trend forecasting utilizes time series forecasting algorithms, such as exponential smoothing and long short-term memory (LSTM) networks. Taking LSTM as an example, several consecutive demand data points preceding the current energy storage demand are normalized and then used as input to the LSTM network. The trained LSTM model then predicts demand for the short term in the future. During model training, historical data is used, and model parameters are adjusted to minimize prediction error.

[0125] In this embodiment, the energy storage system status primarily refers to indicators such as the battery's SOC and SOH, which reflect the health and availability of the energy storage system. The charge and discharge capacity refers to the maximum charging power, maximum discharge power, and continuous charge and discharge time that the energy storage system can safely and efficiently achieve in its current state. For example, in its current state, the energy storage system has a maximum charging power of 150kW, meaning that, under other conditions, it can charge a maximum of 150kWh per hour. A maximum discharge power of 200kW and a continuous discharge time of 2 hours means that it can discharge at 200kW for a maximum of 2 hours.

[0126] In this embodiment, the optimal solution is the combination of energy storage capacity and power that satisfies all constraints and minimizes the objective function. For example, when the energy storage capacity is 1200kWh and the power is 180kW, the objective function value is minimized, and this combination is the optimal solution.

[0127] The beneficial effect of this technical solution is that, by using multiple key factors as inputs for optimization calculations, an optimal capacity configuration can be obtained that comprehensively considers current demand, system status, and future trends. This solution not only meets the power system's requirements for smoothing wind power fluctuations, but also minimizes the lifecycle cost of the energy storage system, improving its economic performance and operational efficiency.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for configuring energy storage capacity based on unbalanced wind power, characterized in that: include: Step 1: Collect historical wind power data of a target area and pre-process the historical wind power data according to historical meteorological data of the target area, wherein the historical wind power data includes wind power output power in different seasons and under different weather conditions; Step 2: Divide the preprocessed data according to the time scale, and obtain the dual analysis algorithm at each time scale from the scale-algorithm comparison table to decompose and analyze the corresponding divided data to obtain an unbalanced power array of the corresponding time scale, wherein the unbalanced power array includes N1 unbalanced power components obtained based on the first analysis algorithm and N2 unbalanced power components obtained based on the second analysis algorithm; Step 3: Establish an energy storage capacity configuration model, taking the power system's requirement to smooth wind power fluctuations as a constraint and maximizing the energy storage system's full lifecycle operating efficiency as the objective function. Step 4: Determine the imbalance characteristics at the corresponding time scale based on the unbalanced power array, optimize the energy storage capacity configuration model, and determine the capacity configuration plan based on the current energy storage demand and current actual energy storage situation of the target area; The method for determining the imbalance characteristics at the corresponding time scale based on the imbalance power array includes: randomly shuffling and splitting the order of the imbalance power in each imbalance power array to obtain several new arrays, and extracting time domain features from the new arrays; performing fast Fourier transform on the new array, analyzing the main frequency components and amplitudes to obtain the dominant frequency, quantifying the energy distribution of the main frequency components, and extracting frequency feature parameters; decomposing the new array into sub-signals of different frequency segments based on wavelet basis functions and the number of decomposition layers, and obtaining the variation characteristics of the imbalance power in time and frequency by analyzing the time-frequency diagram; fusing the time domain features, frequency domain features, and time-frequency domain features, and performing dimensionality reduction processing to obtain feature vectors; using a clustering algorithm to perform cluster analysis on the cluster vectors to divide the imbalance power arrays at different time scales into different categories, and determining the imbalance characteristics of each category by analyzing the feature mean and distribution of each category.

2. The method for configuring energy storage capacity based on unbalanced wind power according to claim 1, characterized in that: Preprocessing the historical wind power data according to the historical meteorological data of the target area includes: Sort the historical wind power at each historical collection moment in sequence to determine whether there is missing power; If it exists, obtain the historical meteorological data at the missing time and analyze it to obtain the initial filling power; Acquire a first reference power and first meteorological data at a first non-missing time closest to the missing time, and at the same time, acquire a second reference power and second meteorological data at a second non-missing time second closest to the missing time; Obtaining a first predicted power based on the first meteorological data and an operating mode of the wind power system at a first non-missing moment, and simultaneously obtaining a second predicted power based on the second meteorological data and an operating mode of the wind power system at a second non-missing moment; Correcting the initial filling power according to the first reference power, the second reference power, the first predicted power, and the second predicted power; in, represents the fine-tuning function; Represents the historical meteorological data Q at the missing moment and the historical meteorological data at each remaining historical collection moment The number of similar moments with a similarity greater than or equal to 0.9; represents the historical wind power at the jth similar moment; Represents positive and negative functions; 、 represent the first reference power and the first predicted power respectively; represent the second reference power and the second predicted power respectively; Indicates the power after a correction; represents the initial filling power; Represents the historical meteorological data Q at the missing time and the first meteorological data at the first non-missing time Similarity function of Represents the historical meteorological data Q at the missing time and the second meteorological data at the second non-missing time The similarity function of .

3. The method for configuring energy storage capacity based on unbalanced wind power according to claim 2, characterized in that: After correcting the initial filling power, the method further includes: Lock the continuous time segment according to the time distribution of all similar moments, draw a power curve based on the continuous time segment, and intercept the power curve with the power after the first correction as a straight line to obtain the excess increment of the upper intercepted part and the proportion of the upper intercepted time, the excess attenuation of the lower intercepted part and the proportion of the lower intercepted time, and perform a secondary correction and completion on the power after the first correction; in, represents the power after secondary correction; 、 Respectively represent the existence of the upper cut part and The maximum absolute difference between the lower intercepted part and The maximum absolute difference; 、 Respectively represent the enclosed area of ​​the upper cut-off part and the lower cut-off part; Indicates the total number of historical collection moments; Represents positive and negative functions; The constant value is 1; 、 Respectively represent the proportion of upper interception time and the proportion of lower interception time; They represent excess attenuation and excess gain respectively.

4. The method for configuring energy storage capacity based on unbalanced wind power according to claim 1, characterized in that: The unbalanced power components of each analysis algorithm are pre-set.

5. The method for configuring energy storage capacity based on unbalanced wind power according to claim 1, characterized in that: Optimizing the energy storage capacity configuration model includes: Integrate the imbalance features at different time scales to construct a multi-dimensional feature matrix; The grey correlation analysis algorithm is used to calculate the correlation between the characteristics of each time scale, explore the potential connection between the imbalance characteristics of different time scales, and introduce the influencing factors reflecting the imbalance characteristics of each time scale; Based on the influencing factors, the constraints in the energy storage capacity configuration model are refined to achieve optimization.

6. The method for configuring energy storage capacity based on unbalanced wind power according to claim 1, characterized in that: Determine the capacity allocation plan based on the current energy storage demand and actual energy storage situation in the target area, including: Analyze the current energy storage demand to obtain the power difference between the output power and the power load, and at the same time, predict the demand trend after the current energy storage demand based on several consecutive demands before the current energy storage demand; Determine the energy storage system status and charging and discharging capacity based on the actual current energy storage situation; Taking the power difference, demand trend, energy storage system status, and charging and discharging capabilities as input, the energy storage capacity and power values ​​are continuously adjusted to find the optimal solution that meets the constraints and minimizes the objective function, thereby obtaining the current optimal capacity configuration plan.

7. The method for configuring energy storage capacity based on unbalanced wind power according to claim 1, characterized in that: The time scale includes: an ultra-short-term time scale, a short-term time scale, and a medium- to long-term time scale.

Images