Hybrid energy storage system capacity configuration method based on ISODATA and random simulation scene generation

Abstract

The invention discloses a hybrid energy storage system capacity configuration method generated based on an ISODATA and a random simulation scene, and the method comprises the steps: obtaining the year-round historical data of wind speed, illumination intensity and temperature with hours as a time scale, forming a year-round historical data set, and carrying out the preprocessing of the data set; clustering the preprocessed data set by using an ISODATA clustering algorithm to obtain a data set taking hours as a time scale in a typical day, taking the data set as an average value of minute-level wind speed, illumination intensity and temperature data in each hour, and randomly simulating to form wind speed, illumination intensity and temperature data taking minutes as a time scale; the method comprises the following steps: establishing a hybrid energy storage capacity configuration model for a wind-light-load-storage micro-grid system; and inputting data taking minutes as a time scale into the hybrid energy storage capacity configuration model, and obtaining an optimal capacity configuration scheme of hybrid energy storage by using a particle swarm algorithm. According to the invention, the advantages of the storage battery and the super capacitor can be brought into full play during hybrid energy storage capacity configuration.

Description

Technical Field

[0001] The present invention relates to the technical field of microgrid capacity planning, and in particular to a hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation. Background Art

[0002] With the shortage of fossil energy and the aggravation of environmental pollution in recent years, renewable energy power generation has developed rapidly. Wind power generation and photovoltaic power generation have become the main force of clean energy power generation due to their low cost and abundant resources. However, in the process of wind and solar power generation, the strong volatility and randomness of wind and light resources have a serious impact on the stability and reliability of the power system. Therefore, new power systems with a high proportion of renewable energy face huge challenges.

[0003] In order to meet this challenge, the power system must improve its flexibility and adaptability. An effective way is to use energy storage systems, which can help adjust the imbalance between power generation and load. When the power generation is greater than the load demand, the energy storage system is charged, and when the power generation is less than the load demand, the energy storage system is discharged, thereby maintaining the balance between supply and demand and alleviating the impact on the power grid. However, the new power system has higher requirements on the response speed, charging and discharging performance, and economy of the energy storage system. Therefore, a hybrid energy storage system composed of batteries and supercapacitors is used to solve this problem. At the same time, the size of the hybrid energy storage configuration capacity is also a key factor. Too large a capacity configuration leads to higher costs and economic waste. Too small a capacity configuration does not have a significant effect on suppressing wind and solar fluctuations.

[0004] Optimizing the capacity of hybrid energy storage systems is a necessary step to achieve system economy and load power supply reliability and stability. Due to the different characteristics of batteries and supercapacitors in hybrid energy storage systems, batteries have high energy density and low power density; supercapacitors have low energy density, high power density, and faster charging / discharging speed than batteries. Therefore, batteries can be used to store energy for a long time (such as 1 hour, 15 minutes, etc.), but cannot play its advantages when the time scale is very short (such as a few seconds); on the contrary, supercapacitors charge / discharge very quickly, and the charging / discharging process can be completed in only a few seconds, so supercapacitors cannot play their advantages when the time scale is long. Therefore, a scene generation method is urgently needed to simultaneously characterize the wind and light seasons, day and night regular changes, and instantaneous random fluctuations, so that the advantages of batteries and hybrid energy storage can be fully utilized. At present, the research on wind and light scene generation for hybrid energy storage capacity configuration often uses the empirical mode decomposition method to divide the fluctuating renewable energy power generation into high-frequency and low-frequency parts, so that the two energy storage devices can bear them separately, but this method is still limited by the original sampling time scale of wind and light data, and most of them still cannot fully utilize the advantages of supercapacitors. Summary of the invention

[0005] The purpose of the present invention is to provide a hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation, and to apply this wind-solar scenario generation method to the capacity configuration of a hybrid energy storage system of a wind-solar-load-storage microgrid. This method can not only describe the seasonal and day-night changes of wind-solar data, but also characterize its random fluctuations within each hour. It can also improve the economy of the wind-solar-load-storage microgrid and better reduce the impact of random fluctuations of wind and solar on the power grid.

[0006] The present invention adopts the following technical solution:

[0007] The hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation includes the following steps:

[0008] S1. Obtain the annual historical data of wind speed, light intensity and temperature on an hourly time scale to form an annual historical data set, and preprocess the data set.

[0009] S2. Use the ISODATA clustering algorithm to cluster the preprocessed data set to obtain a data set with an hourly time scale under typical days.

[0010] S3. Use the data obtained in step S2 as the average value of the wind speed, light intensity, and temperature data at the hourly and minute levels, set the standard deviation of random noise, and randomly simulate to form wind speed, light intensity, and temperature data with minutes as the time scale.

[0011] S4. Establish a hybrid energy storage capacity configuration model for the wind-solar-load-storage microgrid system.

[0012] S5. Input the data obtained in step S3 into the hybrid energy storage capacity configuration model, and use the particle swarm algorithm to obtain the optimal capacity configuration plan of the hybrid energy storage.

[0013] Furthermore, in step S1, the preprocessing includes the following:

[0014] The historical data is normalized using the linear function normalization method. The specific formula is:

[0015]

[0016] Among them, X norm represents the processed data, X represents the original data, and X max Indicates the maximum value in the original data, X min Indicates the minimum value in the original data.

[0017] Furthermore, in step S2, the data obtained on a typical day with an hourly time scale includes the following:

[0018] S201, according to the ISODATA clustering algorithm, select K in the preprocessed data set 0 The data are used as the initial cluster center, expressed as Indicates K 0 Initial cluster centers.

[0019] S202, calculate the distance between each data and all initial cluster centers, classify each data into the category of the cluster center closest to it, and obtain K categories, that is, there are K cluster centers.

[0020] S203, when the number of data in a certain category is less than θ N , the category is discarded and the data in the category is assigned to the category that is closest to it.

[0021] S204, update the cluster center of each category, the specific formula is:

[0022]

[0023] Among them, u j represents the cluster center of the jth category, C j Represents the data set of the jth category, and x represents the data.

[0024] Calculate the average distance from each category to its cluster center The specific formula is:

[0025]

[0026] S205, calculate the average distance from all samples to their cluster center The specific formula is:

[0027]

[0028] S206. If some categories meet the conditions for splitting and merging operations, splitting and merging operations are performed until the categories, cluster centers and number of categories obtained no longer change or the maximum number of iterations is reached; if not, return to step S202 and perform the operation again.

[0029] S207, performing a linear inverse transformation on all cluster centers obtained in step S206 to form data of typical daytime wind speed, light intensity, and temperature on an hourly time scale; the specific formula for the linear inverse transformation is:

[0030] X hour =X norm (X max -X min )+X min

[0031] Among them, X hour Represents data with hourly time scale.

[0032] Furthermore, in step S206, if the number of iterations of this iteration operation is an even number, or the current number of categories K>2K 0 , the merge operation is started, and the specific contents are as follows:

[0033] Calculate the distances between all cluster centers and arrange them from small to large; if two cluster centers u g and u j The distance between them is less than the minimum distance θ c , then the two categories are merged to form a new category, and the cluster center of the new category is u new , the specific formula is:

[0034]

[0035] Among them, C g represents the set of data of the gth category, u g Represents the cluster center of the g-th category.

[0036] Calculate the variance of the data in each category. If the variance in a category is greater than the maximum variance Then split the category; or if the current number of clusters When , the category with the largest variance is split.

[0037] The formula for calculating variance is:

[0038]

[0039] in, represents the variance of the jth category; x jm Represents the mth data of the jth category, m∈[1,2,...,number j ]; x avej represents the average value of all data in the jth category, number j Represents the total number of data in the jth category.

[0040] The category that undergoes the split operation can obtain two new category cluster centers u' a and u' b , the specific expression is:

[0041] u′ a =u j +γ

[0042] u′ b =u j -γ

[0043] Where γ represents the splitting matrix, γ=α[0,......,τ max ,......,0] T , α represents the splitting parameter, 0<α≤1, τ max Represents the largest component in the matrix.

[0044] Furthermore, in step S3, the random simulation includes the following contents:

[0045] S301, the formula for generating the time vector is:

[0046] time=0:T s :(T-rate)

[0047] Where time represents the time vector, T s represents the sampling time interval, T represents the total time length, and rate represents the sampling frequency.

[0048] S302, generating random noise, the formula is:

[0049] noise(m,q)=σ×rand(size(time))

[0050] Among them, σ represents the standard deviation of noise, rand(...) represents the function of generating random numbers, size(...) represents the length, and noise(m,q) represents the random noise of the qth minute of the mth hour.

[0051] S303, set the hourly average values ​​of wind speed, light intensity, and temperature. The specific formula is:

[0052] mean_wind(m)=H_wind(m)

[0053] mean_pv(m)=H_pv(m)

[0054] mean_Temp(m)=H_Temp(m)

[0055] Among them, mean_wind(m) represents the average wind speed in the mth hour, mean_pv(m) represents the average light intensity in the mth hour, mean_Temp(m) represents the average temperature in the mth hour, H_wind(m) represents the wind speed in the mth hour with hours as the time scale, H_pv(m) represents the light intensity in the mth hour with hours as the time scale, and H_Temp(m) represents the temperature in the mth hour with hours as the time scale.

[0056] S304: The formula for generating wind speed, light intensity, and temperature data with a time scale of minutes is:

[0057] wind_speed(m,q)=mean_wind(m)+noise(m,q)

[0058] pv(m,q)=mean_pv(m)+noise(m,q)

[0059] Temp(m,q)=mean_Temp(m)+noise(m,q)

[0060] Among them, wind_speed(m,q) represents the wind speed at the mth hour and qth minute, pv(m,q) represents the light intensity at the mth hour and qth minute, and Temp(m,q) represents the temperature at the mth hour and qth minute.

[0061] Furthermore, in step S4, establishing a hybrid energy storage capacity configuration model includes the following contents:

[0062] S401. The wind-solar-load-storage microgrid system includes a wind-solar power generation system and a hybrid energy storage system. The wind-solar power generation system includes a wind turbine and a photovoltaic generator. The hybrid energy storage system includes batteries and supercapacitors. The load of the wind-solar-load-storage microgrid system is a constant power load.

[0063] When the wind and solar power generation power is greater than the load demand power, the hybrid energy storage system is charged; when the wind and solar power generation power is less than the load demand power, the hybrid energy storage system is discharged; the excess power generation power that cannot be used is regarded as wind and solar power abandonment, and the power that cannot meet the load demand is regarded as load shedding.

[0064] The calculation formula for wind power generation is:

[0065]

[0066] Among them, P wt (t) represents the real-time output power of the wind turbine at time t, P r represents the rated power of the wind turbine, v(t) represents the real-time wind speed at the hub height at time t, and v in Indicates the fan cut-in wind speed, v out Indicates the cut-out wind speed, v r Indicates the rated wind speed of the fan.

[0067] The calculation formula for photovoltaic power generation is:

[0068]

[0069] Among them, P pv (t) represents the real-time output power of the photovoltaic generator at the tth moment, P sta Indicates the rated power of the photovoltaic panel under standard parameters, fpv Indicates the power reduction coefficient, G sta , G(t) represent the standard light intensity and the light intensity at the tth moment, α T Indicates the power temperature coefficient, T a,sta , T a (t) represent the standard ambient temperature and the photovoltaic panel temperature at the tth moment respectively.

[0070] The calculation formula of wind and solar power generation is:

[0071] P pw (t) = P pv (t)+P wt (t)

[0072] Among them, P pw (t) represents the wind and solar power generation power at the tth moment.

[0073] S402. Set the decision variables of the hybrid energy storage capacity configuration model to be the battery capacity and the supercapacitor capacity, and the constraints to be the system operation power balance and the maximum power limit of each device in the system.

[0074] S403. The objective function of the hybrid energy storage capacity configuration model is:

[0075] min f=min C b +min C wave

[0076] Among them, C b represents the construction cost, α 1 , β 1 , μ 1 , 1 They represent the maintenance cost ratio coefficients of photovoltaic generators, wind turbines, batteries, and supercapacitors respectively; Respectively represent the unit capacity investment cost of photovoltaic generators, wind turbines, batteries, and supercapacitors; E pv 、E wt 、E b 、E c Respectively represent the configuration capacity of photovoltaic generator, wind turbine, battery and supercapacitor; ξ represents the fluctuation penalty coefficient; C wave represents the load power fluctuation penalty cost, P L (t) represents the load working power at time t; P L,mean represents the average value of load power; f represents the objective function.

[0077] Furthermore, in step S5, obtaining the optimal capacity configuration scheme of hybrid energy storage includes the following contents:

[0078] S501, initializing the parameters of the particle swarm algorithm and the positions and velocities of the particles, setting the number of particles and the maximum number of iterations.

[0079] S502, iterate each particle, calculate the fitness value of each particle at the current position, and compare it with the minimum fitness value pbest among the positions it has passed i Compare, if the fitness value of the current position is smaller, then the current fitness value is taken as the fitness value of the optimal position;

[0080] pbest i =min{fitness 1 , fitness 2 ,..., fitness i ,..., fitness N}

[0081] fitness i and fitness N They represent the fitness values ​​of the particle at the i-th and N-th iteration positions, respectively, 1≤i≤N, and N represents the maximum number of iterations.

[0082] S503, compare the fitness value of each particle with the minimum fitness value gbest among all the positions where the particles pass e Compare, if the fitness value of the current position is smaller, then the current fitness value is taken as the fitness value of the optimal position;

[0083] gbest e =min{pbest 1 , pbest 2 , ..., pbest e , ..., pbest Z}

[0084] pbest e and pbest Z They represent the fitness values ​​of the e-th and Z-th particles respectively, 1≤e≤Z, and Z represents the total number of particles.

[0085] S504, update the speed and position of the particle, the specific formula is:

[0086] v i =v i-1 +c 1 ×rand()×(Pbest i -x i )+c 2 ×rand()×(gbest i -x i )

[0087] x i =x i-1 +v i

[0088] Among them, v i 、v i-1 Respectively represent the particle speed of the i-th and i-1-th iterations, x i 、x i-1 Respectively represent the particle positions of the i-th and i-1-th iterations, c 1 、c 2 They represent individual learning factor and social learning factor respectively.

[0089] S505, return to step S502 and repeat the operation until the maximum number of iterations is reached or the results of the two iterations are the same, then stop and obtain the optimal capacity configuration plan of the hybrid energy storage.

[0090] Furthermore, the present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the hybrid energy storage system capacity configuration method generated based on ISODATA and random simulation scenarios are implemented.

[0091] Furthermore, the present invention also proposes a computer-readable storage medium, which stores a computer program. When the computer program is run by a processor, it executes the hybrid energy storage system capacity configuration method generated based on ISODATA and random simulation scenarios.

[0092] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0093] 1. The method proposed in the present invention can not only describe the seasonal and day-night changes of wind and solar data, but also characterize its random fluctuations within each hour. It can comprehensively consider the time characteristics of batteries and supercapacitors, so that the advantages of both can be fully utilized in the process of capacity configuration.

[0094] 2. The method proposed in the present invention shows that the wind-solar scene generation method has adaptability and superiority in hybrid energy storage capacity configuration. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 It is an overall implementation flow chart of the present invention.

[0096] Figure 2 This is a historical data chart of the scenery throughout the year according to an embodiment of the present invention.

[0097] Figure 3This is a graph of the clustering results of the scenery data according to an embodiment of the present invention.

[0098] Figure 4 This is a typical scene scenery data diagram of an embodiment of the present invention.

[0099] Figure 5 This is a landscape data diagram after random simulation according to an embodiment of the present invention.

[0100] Figure 6 It is a structural diagram of a wind-solar-load-storage microgrid system according to an embodiment of the present invention.

[0101] Figure 7 It is the charging and discharging power of the battery in the embodiment of the present invention.

[0102] Figure 8 It is the charging and discharging power of the supercapacitor of the embodiment of the present invention.

[0103] Fig. 9 is an unbalanced power diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0104] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0105] To achieve the above objectives, the present invention proposes a hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation, such as Figure 1 As shown, the specific steps are as follows:

[0106] S1. Obtain the annual historical data of wind speed, light intensity and temperature on an hourly time scale to form an annual historical data set, and preprocess the data set; the specific contents are as follows:

[0107] When using the ISODATA clustering algorithm to cluster wind and solar data, the order of magnitude of the sample data for each day may vary greatly, which will have a serious adverse impact on the establishment of the clustering model and thus affect the clustering effect. Therefore, before formally clustering the data, the data must be normalized and preprocessed, which will help speed up the clustering speed and improve the clustering effect.

[0108] The historical data is normalized using the linear function normalization method, and the data is proportionally reduced to the range of [0,1] through a certain linear transformation. The specific formula is:

[0109]

[0110] Among them, X norm represents the processed data, X represents the original data, and X maxIndicates the maximum value in the original data, X min Indicates the minimum value in the original data.

[0111] In this embodiment, the annual historical data of wind speed, light intensity, and temperature in Xinjiang Uygur Autonomous Region in 2019 with a time scale of 1 hour are obtained, such as Figure 2 As shown, Figure 2 (a) is the historical data of wind speed throughout the year, Figure 2 (b) is the historical data of annual light intensity. Figure 2 (c) is the historical temperature data for the whole year.

[0112] S2. Use ISODATA clustering algorithm to cluster the preprocessed data set to obtain a data set with hourly time scale under typical days. A total of 6 typical days, i.e. 144 hours of wind speed, light intensity and temperature data are obtained. The specific contents are as follows:

[0113] S201, according to the ISODATA clustering algorithm, select K in the preprocessed data set 0 The data are used as the initial cluster center, expressed as Indicates K 0 Initial cluster centers.

[0114] S202, K represents the total number of cluster centers, The initial value is K 0 After splitting, merging, and discarding operations, the value of K will change. The distance from each data to all initial cluster centers is calculated separately, and each data is classified into the category with the nearest cluster center, and K categories are obtained.

[0115] S203, when the number of data in a certain category is less than θ N , the category is discarded and the data in the category is assigned to the category that is closest to it.

[0116] S204, update the cluster center of each category, the specific formula is:

[0117]

[0118] Among them, u j represents the cluster center of the jth category, C j Represents the data set of the jth category, and x represents the data.

[0119] Calculate the average distance from each category to its cluster center The specific formula is:

[0120]

[0121] S205, calculate the average distance from all samples to their cluster center The specific formula is:

[0122]

[0123] S206. If the conditions for splitting and merging operations are met, the splitting and merging operations are performed until the result does not change, that is, the category to which each sample belongs, the cluster center and the number of categories remain unchanged, or the maximum number of iterations of 100 is reached; if not met, return to step S202 and perform the operation again.

[0124] If the number of iterations of this iteration is an even number, or the current number of categories K>2K 0 , the merge operation is started, and the specific contents are as follows:

[0125] Calculate the distances between all cluster centers and arrange them from small to large; if two cluster centers u g and u j The distance between them is less than the minimum distance θ c , then these two categories are merged to form a new category, and the cluster center of the new category is u new :

[0126]

[0127] Among them, C g represents the set of data of the gth category, u g Represents the cluster center of the g-th category.

[0128] Calculate the variance of the data in each category. If the variance in a category is greater than the maximum variance Then split the category; or, if the current number of clusters When , the category with the largest variance is split.

[0129] The formula for calculating variance is:

[0130]

[0131] in, represents the variance of the jth category; x jm Represents the mth data of the jth category, m∈[1,2,...,number j ]; x avej represents the average value of all data in the jth category, number j Represents the total number of data in the jth category.

[0132] The category that undergoes the split operation can obtain two new category cluster centers u'a and u' b , the specific expression is:

[0133] u′ a =u j +γ

[0134] u′ b =u j -γ

[0135] Where γ represents the splitting matrix, γ=α[0,......,τ max ,......,0] T , α represents the randomly set splitting parameter, 0<α≤1, τ max Represents the largest component in the matrix.

[0136] In this embodiment, 6 typical daily normalized data are obtained as follows: Figure 3 As shown, Figure 3 (a)- Figure 3 The upper figure in (f) represents the normalized cluster centers of categories 1-6. The cluster centers of wind speed, light intensity, and temperature data of these six typical days are representative and can represent the wind and light conditions throughout the year after inverse transformation; the lower figure represents the clustering results of the data of categories 1-6 throughout the year. The 365 colored lines in the lower figure represent 365 groups of data throughout the year. The horizontal axis sampling point 72 represents the dimension of the cluster. The 1st to 24th sampling points represent the wind speed of a certain day, the 25th to 48th sampling points represent the light intensity of a certain day, and the 49th to 72nd sampling points represent the temperature of a certain day. As can be seen from the figure, the final clustering results have 6 categories, that is, 6 typical days, corresponding to different wind and light weather conditions.

[0137] S207, perform a linear inverse transformation on all cluster centers of the clustering results obtained in step S206 to restore the data from [0,1] to the original order of magnitude, forming data of typical daytime wind speed, light intensity, and temperature with an hourly time scale. The specific formula for the linear inverse transformation is:

[0138] X hour =X norm (X max -X min )+X min

[0139] Among them, X hour Represents data with hourly time scale.

[0140] After the inverse transformation, the changes of 6 typical daytime wind speeds, light intensity, and temperature on an hourly time scale are as follows: Figure 4 As shown, Figure 4 (a) Figure 4 (b) Figure 4 (c) are wind speed, light intensity, and temperature data. As shown in the figure, after the inverse transformation, the wind speed, light intensity, and temperature data have returned to their original order of magnitude, and the wind speed, light intensity, and temperature fluctuation trends and sizes on the six typical days are different, indicating that the clustering results in step 2 are widely representative and can represent the situation throughout the year.

[0141] S3, taking the 144 data obtained in step S2 as the average value of the hourly and minute-level wind speed, light intensity, and temperature data, setting the random noise standard deviation, and performing simulation in the software matlab, randomly simulating to form wind speed, light intensity, and temperature data with minutes as the time scale; in this process, the wind speed, light intensity, and temperature with hours as the time scale obtained in step 2 need to be taken as the average value, and then a certain standard deviation is set. In this embodiment, the standard deviation is set to 1; the specific contents are:

[0142] S301, the formula for generating the time vector is:

[0143] time=0:T s :(T-rate)

[0144] Where time represents the time vector, T s represents the sampling time interval, T represents the total time length, and rate represents the sampling frequency.

[0145] S302, generating random noise, the formula is:

[0146] noise(m,q)=σ×rand(size(time))

[0147] Among them, σ represents the standard deviation of noise, rand(...) represents the function of generating random numbers, size(...) represents the length, noise(m,q) represents the random noise of the qth minute of the mth hour, and each hour has 60 minutes, for a total of 144 hours, that is, 1≤q≤60,1≤m≤144.

[0148] S303, set the hourly average values ​​of wind speed, light intensity, and temperature. The specific formula is:

[0149] mean_wind(m)=H_wind(m)

[0150] mean_pv(m)=H_pv(m)

[0151] mean_Temp(m)=H_Temp(m)

[0152] Among them, mean_wind(m) represents the average wind speed in the mth hour, mean_pv(m) represents the average light intensity in the mth hour, mean_Temp(m) represents the average temperature in the mth hour, H_wind(m) represents the wind speed in the mth hour with hours as the time scale, H_pv(m) represents the light intensity in the mth hour with hours as the time scale, and H_Temp(m) represents the temperature in the mth hour with hours as the time scale.

[0153] S304: The formula for generating wind speed, light intensity, and temperature data with a time scale of minutes is:

[0154] wind_speed(m,q)=mean_wind(m)+noise(m,q)

[0155] pv(m,q)=mean_pv(m)+noise(m,q)

[0156] Temp(m,q)=mean_Temp(m)+noise(m,q)

[0157] Among them, wind_speed(m,q) represents the wind speed at the mth hour and qth minute, pv(m,q) represents the light intensity at the mth hour and qth minute, and Temp(m,q) represents the temperature at the mth hour and qth minute.

[0158] The wind speed, light intensity, and temperature data of six typical days with a time scale of 1 minute obtained after random simulation are as follows: Figure 5 As shown, Figure 5 (a) Figure 5 (b) Figure 5 (c) are the data graphs of wind speed, light intensity, and temperature. It can be seen from the figure that the wind speed varies between 0 and 25 m / s. When the wind speed is large, it can represent windy weather in autumn and winter, and when the wind speed is small, it can represent sunny weather. The light intensity follows the trend of large in the middle and small on both sides during the day, which shows that the data is in line with the actual situation and is correct. The temperature varies between -40 and 40℃, which can represent the temperature of the four seasons, indicating the correctness of the random simulation method in step 3.

[0159] S4. For the wind-solar-load-storage microgrid system, a hybrid energy storage capacity configuration model is established; the specific contents are as follows:

[0160] S401, such as Figure 6 As shown, the wind-solar-load-storage microgrid system includes a wind-solar power generation system and a hybrid energy storage system. The wind-solar power generation system includes a wind turbine and a photovoltaic generator. The hybrid energy storage system includes batteries and supercapacitors. The load of the wind-solar-load-storage microgrid system is a constant power load.

[0161] The coordinated operation strategy of the system is: the wind and solar power generation system supplies power to the load. When the wind and solar power generation power is greater than the load demand power, the hybrid energy storage system charges; when the wind and solar power generation power is less than the load demand power, the hybrid energy storage system discharges; the excess power generation power that cannot be used is regarded as abandoned wind and solar power, and the power that cannot meet the load demand is regarded as load shedding.

[0162] The calculation formula for wind power generation is:

[0163]

[0164] Among them, P wt (t) represents the real-time output power of the wind turbine at time t, P r represents the rated power of the wind turbine, v(t) represents the real-time wind speed at the hub height at time t, and v in Indicates the fan cut-in wind speed, v out Indicates the cut-out wind speed, v r Indicates the rated wind speed of the fan.

[0165] The calculation formula for photovoltaic power generation is:

[0166]

[0167] Among them, P pv (t) represents the real-time output power of the photovoltaic generator at the tth moment, P sta Indicates the rated power of the photovoltaic panel under standard parameters, f pv Indicates the power reduction coefficient, G sta , G(t) represent the standard light intensity and the light intensity at the tth moment, α T Indicates the power temperature coefficient, T a,sta , T a (t) are the standard ambient temperature and the photovoltaic panel temperature at the tth moment respectively.

[0168] The calculation formula of wind and solar power generation is:

[0169] P pw (t) = P pv (t)+P wt (t)

[0170] Among them, P pw (t) represents the wind and solar power generation power at the tth moment.

[0171] S402. The decision variables of the hybrid energy storage capacity configuration model are set to be the battery capacity and the supercapacitor capacity, and the constraints of the actual operating conditions are the system operating power balance and the maximum power limit of each device in the system.

[0172] S403. The objective function of the hybrid energy storage capacity configuration model is:

[0173] min f=min C b +min C wave

[0174] Among them, C b represents the construction cost, α 1 , β 1 , μ 1 , 1 They represent the maintenance cost ratio coefficients of photovoltaic generators, wind turbines, batteries, and supercapacitors, respectively. They represent the unit capacity investment costs of photovoltaic generators, wind turbines, batteries, and supercapacitors, respectively. pv 、E wt 、E b 、E c Respectively represent the configuration capacity of photovoltaic generator, wind turbine, battery and supercapacitor; C wave represents the load power fluctuation penalty cost, ξ represents the volatility penalty coefficient, P L (t) represents the load working power at time t, P L,mean represents the average value of load power; f represents the objective function.

[0175] S5. Input the data obtained in step S3 into the hybrid energy storage capacity configuration model, and use the particle swarm algorithm to obtain the optimal capacity configuration plan for the hybrid energy storage; the specific content is:

[0176] S501, initialize the parameters of the particle swarm algorithm, set the number of particles to 100, the maximum number of iterations to 50, and initialize the position and speed of the particles.

[0177] S502, iterate each particle, calculate the fitness value of each particle at the current position, and compare it with the minimum fitness value (local optimal fitness value) pbest among the positions it has passed i Compare, if the fitness value of the current position is smaller, then the current fitness value is taken as the fitness value of the optimal position;

[0178] pbest i =min{fitness 1 , fitness 2 ,..., fitness i ,..., fitness N}

[0179] fitness i and fitness NThey represent the fitness values ​​of the particle at the i-th and N-th iteration positions, respectively, 1≤i≤N, and N represents the maximum number of iterations.

[0180] S503, compare the fitness value of each particle with the minimum fitness value gbest among all the positions where the particles pass e Compare, if the fitness value of the current position is smaller, then the current fitness value is taken as the fitness value of the optimal position;

[0181] gbest e =min{pbest 1 , pbest 2 , ..., pbest e , ..., pbest Z}

[0182] pbest e and pbest Z They represent the fitness values ​​of the e-th and Z-th particles respectively, 1≤e≤Z, and Z represents the total number of particles.

[0183] S504, adjusting the particle speed and position according to the following formulas, the particle speed and position update formulas are:

[0184] v i =v i-1 +c 1 ×rand()×(Pbest i -x i )+c 2 ×rand()×(gbest i -x i )

[0185] x i =x i-1 +v i

[0186] Among them, v i 、v i-1 Respectively represent the particle speed of the i-th and i-1-th iterations, x i 、x i-1 Respectively represent the particle positions of the i-th and i-1-th iterations, c 1 、c 2 They represent individual learning factor and social learning factor respectively.

[0187] S505, return to step S502 and repeat the operation until the maximum number of iterations is 50, or the results of the two iterations are the same, then stop, and obtain the optimal capacity configuration scheme of the hybrid energy storage, that is, the value of the decision variable corresponding to the global optimal fitness value when the iteration stops, that is, the optimal configuration scheme of the battery and supercapacitor capacity.

[0188] The basic parameters of the wind-solar-load-storage microgrid system in this embodiment are shown in Table 1.

[0189] Table 1 Basic parameters in wind-solar-load-storage microgrid system

[0190] parameter Value Unit investment cost of photovoltaic equipment / yuan / kW 5000 Annual operation and maintenance cost of photovoltaic equipment 0.02 Photovoltaic installed capacity / MW 100 Unit investment cost of wind power equipment / yuan / kW 4000 Annual operation and maintenance cost coefficient of wind power equipment 0.02 Wind turbine installed capacity / MW 100 Constant power load size / MW 100 Rated capacity of a single battery / kWh / unit 1.2 Rated power of a single battery / kW / unit 0.12 Single investment cost of battery / yuan / unit 400 Annual battery operation and maintenance cost coefficient 0.2 Supercapacitor single capacity / kWh / unit <![CDATA[3.54×10 -3 ]]> Maximum power of a single supercapacitor / kW / unit 4.05 Supercapacitor single investment cost / yuan / unit 300 Annual operation and maintenance cost coefficient of supercapacitor 0.09

[0191] The optimal configuration of hybrid energy storage capacity is shown in Table 2.

[0192] Table 2 Hybrid energy storage capacity optimization configuration scheme

[0193] index Numeric Battery capacity / MWh 380.12 Supercapacitor capacity / MWh 6.69 Load power relative standard deviation 0.42 Annual hybrid energy storage cost / 10,000 yuan 17728.24 Power shortage rate / % 20.3 Wind and solar power abandonment rate / % 7.2

[0194] As can be seen from the table, the battery capacity in the hybrid energy storage system is much larger than the supercapacitor capacity, which means that the hybrid energy storage system mainly works with batteries and supplemented by supercapacitors. This means that the use of the wind and solar scene generation method based on ISODATA clustering algorithm and random simulation can enable the hybrid energy storage system to work according to the prescribed charging and discharging plan, and can reflect the charging and discharging characteristics of batteries and supercapacitors.

[0195] The partial diagram of the charging and discharging power of the battery and supercapacitor of the hybrid energy storage system is as follows Figure 7 , Figure 8 As shown, the system unbalanced power changes as Fig. 9 As shown in the figure, the system unbalanced power is equal to the wind and solar power generation power minus the load demand power. As can be seen from the figure, the battery only charges and discharges at the rated power, and the charging and discharging speed is relatively slow, which is used to smooth the system unbalanced power for a long time. The supercapacitor charges and discharges very quickly, and the charging and discharging power is relatively large, which can be used to smooth the system's relatively large unbalanced power in a short time. The simulation results in matlab software can illustrate the applicability and effectiveness of the capacity configuration method for the hybrid energy storage system based on ISODATA and random simulation scenarios proposed in this paper.

[0196] The embodiment of the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. It should be noted that when the processor executes the computer program, the specific steps of the method provided in the embodiment of the present invention are corresponding to the specific steps of the method provided in the embodiment of the present invention, and the processor has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the method provided in the embodiment of the present invention.

[0197] The embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. It should be noted that when the computer program is executed by the processor, it corresponds to the specific steps of the method provided in the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the method provided in the embodiment of the present invention.

[0198] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation, characterized in that: include: S1. Obtain the annual historical data of wind speed, light intensity and temperature on an hourly time scale to form an annual historical data set, and preprocess the data set; S2, clustering the preprocessed data set using the ISODATA clustering algorithm to obtain a data set with an hourly time scale under a typical day; S3, taking the data obtained in step S2 as the average value of the wind speed, light intensity, and temperature data at the hourly and minute levels, setting the standard deviation of random noise, and randomly simulating to form wind speed, light intensity, and temperature data with a time scale of minutes; S4. Establish a hybrid energy storage capacity configuration model for the wind-solar-load-storage microgrid system; S5. Input the data obtained in step S3 into the hybrid energy storage capacity configuration model, and use the particle swarm algorithm to obtain the optimal capacity configuration plan of the hybrid energy storage.

2. The hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation according to claim 1 is characterized in that: In step S1, the preprocessing includes the following: The historical data is normalized using the linear function normalization method. The specific formula is: Among them, X norm represents the processed data, X represents the original data, and X max Indicates the maximum value in the original data, X min Indicates the minimum value in the original data.

3. The hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation according to claim 1 is characterized in that: In step S2, the data obtained on a typical day with an hourly time scale includes the following: S201. According to the ISODATA clustering algorithm, K0 data are selected as the initial clustering centers in the preprocessed data set, which is expressed as represents the K0th initial cluster center; S202, calculate the distance between each data and all initial cluster centers, classify each data into the category of the cluster center closest to it, and obtain K categories, that is, there are K cluster centers. S203, when the number of data in a certain category is less than θ N When , the category is discarded and the data in the category is assigned to the category with the closest distance; S204, update the cluster center of each category, the specific formula is: Among them, u j represents the cluster center of the jth category, C j represents the data set of the jth category, and x represents the data; Calculate the average distance from each category to its cluster center The specific formula is: S205, calculate the average distance from all samples to their cluster center , the specific formula is: S206, if some categories meet the conditions for splitting and merging operations, splitting and merging operations are performed until the obtained categories, cluster centers and number of categories do not change or the maximum number of iterations is reached; if not, return to step S202 and perform the operation again; S207, performing a linear inverse transformation on all cluster centers obtained in step S206 to form data of typical daytime wind speed, light intensity, and temperature on an hourly time scale; the specific formula for the linear inverse transformation is: X hour =X norm (X max -X min )+X min Among them, X hour Represents data with hourly time scale.

4. The hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation according to claim 3 is characterized in that: In step S206, if the number of iterations of this iterative operation is an even number, or the current number of categories K>2K0, then enter the merging operation, the specific content is: Calculate the distances between all cluster centers and arrange them from small to large; if two cluster centers u g and u j The distance between them is less than the minimum distance θ c , then the two categories are merged to form a new category, and the cluster center of the new category is u new , the specific formula is: Among them, C g represents the set of data of the gth category, u g represents the cluster center of the g-th category; Calculate the variance of the data in each category. If the variance in a category is greater than the maximum variance Then split the category; or if the current number of clusters When , the category with the largest variance is split; The formula for calculating variance is: in, represents the variance of the jth category; x jm Represents the mth data of the jth category, m∈[1,2,...,number j ]; x avej represents the average value of all data in the jth category, number j represents the total number of data in the jth category; The category that undergoes the split operation can obtain two new category cluster centers u' a and u' b , the specific expression is: in' a =in j +γ in' b =in j -γ Where γ represents the splitting matrix, γ=α[0,......,τ max ,......,0] T , α represents the splitting parameter, 0<α≤1, τ max Represents the largest component in the matrix.

5. The hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation according to claim 1 is characterized in that: In step S3, the random simulation includes the following contents: S301, the formula for generating the time vector is: time=0:T s :(T-rate) Where time represents the time vector, T s represents the sampling time interval, T represents the total time length, and rate represents the sampling frequency; S302, generating random noise, the formula is: noise(m,q)=σ×rand(size(time)) Among them, σ represents the standard deviation of noise, rand(...) represents the function of generating random numbers, size(...) represents the length, and noise(m,q) represents the random noise of the qth minute of the mth hour; S303, set the hourly average values ​​of wind speed, light intensity, and temperature. The specific formula is: mean_wind(m)=H_wind(m) mean_pv(m)=H_pv(m) mean_Temp(m)=H_Temp(m) Where, mean_wind(m) represents the average wind speed in the mth hour, mean_pv(m) represents the average light intensity in the mth hour, mean_Temp(m) represents the average temperature in the mth hour, H_wind(m) represents the wind speed in the mth hour with hours as the time scale, H_pv(m) represents the light intensity in the mth hour with hours as the time scale, and H_Temp(m) represents the temperature in the mth hour with hours as the time scale; S304: The formula for generating wind speed, light intensity, and temperature data with a time scale of minutes is: wind_speed(m,q)=mean_wind(m)+noise(m,q) pv(m,q)=mean_pv(m)+noise(m,q) Temp(m,q)=mean_Temp(m)+noise(m,q) Among them, wind_speed(m,q) represents the wind speed at the mth hour and qth minute, pv(m,q) represents the light intensity at the mth hour and qth minute, and Temp(m,q) represents the temperature at the mth hour and qth minute.

6. The hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation according to claim 1 is characterized in that: In step S4, establishing a hybrid energy storage capacity configuration model includes the following contents: S401. The wind-solar-load-storage microgrid system includes a wind-solar power generation system and a hybrid energy storage system. The wind-solar power generation system includes a wind turbine and a photovoltaic generator. The hybrid energy storage system includes a battery and a supercapacitor. The load of the wind-solar-load-storage microgrid system is a constant power load. When the wind and solar power generation power is greater than the load demand power, the hybrid energy storage system is charged; when the wind and solar power generation power is less than the load demand power, the hybrid energy storage system is discharged; the excess power that cannot be used is regarded as abandoned wind and solar power, and the power that cannot meet the load demand is regarded as load shedding; The calculation formula for wind power generation is: Among them, P wt (t) represents the real-time output power of the wind turbine at time t, P r represents the rated power of the wind turbine, v(t) represents the real-time wind speed at the hub height at time t, and v in Indicates the fan cut-in wind speed, v out Indicates the cut-out wind speed, v r Indicates the rated wind speed of the fan; The calculation formula for photovoltaic power generation is: Among them, P pv (t) represents the real-time output power of the photovoltaic generator at the tth moment, P sta Indicates the rated power of the photovoltaic panel under standard parameters, f pv Indicates the power reduction coefficient, G sta , G(t) represent the standard light intensity and the light intensity at the tth moment, α T Represents the power temperature coefficient, T a,sta 、T a (t) represent the standard ambient temperature and the photovoltaic panel temperature at the tth moment respectively; The calculation formula of wind and solar power generation is: P pw (t)=P pv (t)+P wt (t) Among them, P pw (t) represents the wind and solar power generation power at the tth moment; S402, setting the decision variables of the hybrid energy storage capacity configuration model to be the battery capacity and the supercapacitor capacity, and the constraints to be the system operation power balance and the maximum power limit of each device in the system; S403. The objective function of the hybrid energy storage capacity configuration model is: minf=minC b +minC wave Among them, C b represents the construction cost, α1, β1, μ1, and ρ1 represent the maintenance cost proportional coefficients of photovoltaic generators, wind turbines, batteries, and supercapacitors, respectively; Respectively represent the unit capacity investment cost of photovoltaic generators, wind turbines, batteries, and supercapacitors; E pv 、E wt 、E b 、E c Respectively represent the configuration capacity of photovoltaic generator, wind turbine, battery and supercapacitor; ξ represents the fluctuation penalty coefficient; T represents the total time length; C wave represents the load power fluctuation penalty cost, P L (t) represents the load working power at time t; P L,mean represents the average value of load power; f represents the objective function.

7. The hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation according to claim 1 is characterized in that: In step S5, the optimal capacity configuration scheme of hybrid energy storage is obtained, including the following contents: S501, initializing the parameters of the particle swarm algorithm and the positions and velocities of the particles, setting the number of particles and the maximum number of iterations; S502, iterate each particle, calculate the fitness value of each particle at the current position, and compare it with the minimum fitness value pbest among the positions it has passed i Compare and if the fitness value of the current position is smaller, then the current fitness value is taken as the fitness value of the optimal position; in p best i =min{fitness1,fitness2,...,fitness i ,...,fitness N } fitness i and fitness N Respectively represent the fitness values ​​of the particle at the i-th and N-th iteration positions, 1≤i≤N, N represents the maximum number of iterations; S503, compare the fitness value of each particle with the minimum fitness value gbest among all the positions where the particles pass e Compare and if the fitness value of the current position is smaller, then the current fitness value is taken as the fitness value of the optimal position; in best e =min{pbest1,pbest2,...,pbest e ,...,pbest Z } pbest e and pbest Z They represent the fitness values ​​of the e-th and Z-th particles, respectively, 1≤e≤Z, and Z represents the total number of particles; S504, update the speed and position of the particle, the specific formula is: v i =v i-1 +c1×rand()×(Pbest i -x i )+c2×rand()×(gbest i -x i ) x i =x i-1 +v i Among them, v i 、v i-1 Respectively represent the particle speed of the i-th and i-1-th iterations, x i 、x i-1 Respectively represent the particle positions of the i-th and i-1-th iterations, c1 and c2 represent the individual learning factor and social learning factor, respectively, and rand() represents the random number generation function; S505, return to step S502 and repeat the operation until the maximum number of iterations is reached or the results of the two iterations are the same, then stop and obtain the optimal capacity configuration plan of the hybrid energy storage.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the hybrid energy storage system capacity configuration method generated based on ISODATA and random simulation scenarios as described in any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the hybrid energy storage system capacity configuration method based on ISODATA and random simulation scenario generation described in any one of claims 1 to 7 is executed.

Images