A method and system for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction

Abstract

The present invention discloses a method and system for smoothing power fluctuations of supercapacitor energy storage based on wind speed prediction, including: collecting direction data and size data of wind speed, dividing the direction data into a calculation group and a proofreading group, calculating and obtaining a direction prediction coefficient, and performing direction prediction of wind speed; dividing the size data into a calculation group and an adjustment group, using a GRU network to predict the size of wind speed, combining the direction prediction value with the size prediction value to obtain wind speed prediction data; dynamically adjusting the charge and discharge rate of the supercapacitor energy storage system according to the wind speed prediction data and the load demand of the power grid to achieve supercapacitor energy storage power fluctuation smoothing. The method of the present invention significantly improves the overall accuracy of wind speed prediction by independently collecting and processing the direction and size data of wind speed, and then combining the high-precision prediction of wind speed size by the GRU network. Combined with the prediction data of direction and size, it is possible to achieve dynamic adjustment of the supercapacitor energy storage system, smooth power fluctuations, and ensure the stability of the power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation, and in particular to a method and system for smoothing power fluctuations of an over-capacity energy storage system based on wind speed prediction. Background Art

[0002] In recent years, the proportion of new energy access to the grid has increased year by year, and wind power has been greatly developed due to its cleanliness and sustainability. However, wind power is greatly affected by wind speed, has random fluctuations and intermittent characteristics, and direct grid connection will lead to grid voltage fluctuations, unstable power system frequency, and other problems. The introduction of supercapacitors provides a solution to the above problems. Supercapacitors can quickly adjust the balance of power supply and demand, and configuring energy storage on the power generation side can smooth out wind power fluctuations and improve grid connection stability.

[0003] Supercapacitors have the characteristics of high power density and fast charging and discharging speed. They have good application prospects in quickly compensating for fluctuations in wind farm output and regulating voltage quality. The existing method is to derive the wind power value based on the predicted results of wind speed. The supercapacitor system smoothes the fluctuations based on the calculated wind power value, so that the output value after smoothing meets the grid connection requirements. However, wind speed is a vector, which has the dual attributes of size and direction. Predicting only the size of the speed will lead to inaccurate estimated wind power values, which will in turn lead to the misalignment of the supercapacitor's smoothing, so that the output value after smoothing does not meet the grid connection requirements.

[0004] In order to solve this problem, the present invention proposes a bidirectional prediction model for the magnitude and direction of wind speed, which independently predicts the magnitude and direction of wind speed, and finally converts the prediction of direction into the magnitude of wind speed, making the prediction of wind speed more accurate. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is that the existing wind speed prediction method only predicts the magnitude of the speed, resulting in inaccurate estimated wind power value, which in turn leads to over-capacity leveling dislocation, so that the output value after leveling does not meet the requirements of grid connection.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: a method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction, comprising: collecting direction data and magnitude data of wind speed, dividing the direction data into a calculation group and a proofreading group, calculating and obtaining a direction prediction coefficient, and performing direction prediction of wind speed; dividing the magnitude data into a calculation group and an adjustment group, using a GRU network to predict the magnitude of wind speed, combining the direction prediction value with the magnitude prediction value to obtain wind speed prediction data; dynamically adjusting the charge and discharge rate of the super-capacity energy storage system according to the wind speed prediction data and the grid load demand to achieve super-capacity energy storage power fluctuation smoothing.

[0008] As a preferred solution of the method for smoothing power fluctuations of ultra-capacity energy storage based on wind speed prediction described in the present invention, the direction data and magnitude data of the wind speed include: assuming that the direction perpendicular to the wind turbine blade is 0°, the direction of the wind speed is defined as the angle θ that deviates from the perpendicular to the wind turbine blade, and the collected wind speed magnitude sequence is [x1, x2, x3, ..., x n ], the corresponding angles are [θ1, θ2, θ3, …, θ n ].

[0009] As a preferred solution of the method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction described in the present invention, wherein: the directional prediction coefficient includes: [θ1, θ2, θ3, ..., θ n ] is divided into calculation groups [θ1, θ2, θ3, …, θ 0.95n ] and proofreading group [θ 0.95n+1 ,θ 0.95n+2 ,θ 0.95n+3 ,…,θ n ], use the calculation group to calculate the value of the proofreading group, expressed as:

[0010]

[0011]

[0012]

[0013] Among them, θ′ 0.95n+1 represents the θ calculated using the calculation group 0.95n+1 The value of; α represents the activation regression coefficient; j represents the number of terms in the weighted sum; s represents the number of terms used for factorial weighted sum calculation; represents the value of the i-th calculation group; ELU represents the activation function; α is the activation regression coefficient, which is expressed as:

[0014]

[0015] The proofreading group obtained by calculating the calculation group is [θ′ 0.95n+1 ,θ′ 0.95n+2 ,θ′ 0.95n+3 ,…,θ′ n ], and the proofreading group [θ 0.95n+1 ,θ 0.95n+2 ,θ 0.95n+3 ,…,θ n ]Comparison calculation direction prediction coefficient, expressed as:

[0016]

[0017]

[0018] Among them, sigmoid represents the sigmoid activation function; tanh represents the tanh activation function; represents the value of the i-th proofreading group; θ′i represents the value of the i-th proofreading group calculated using the calculation group; c i Represents the directional prediction difference coefficient, [c1, c2, c3, …, c 0.05n ]; b i represents the directional prediction ratio coefficient, [b1, b2, b3, …, b 0.05n ].

[0019] As a preferred solution of the method for smoothing power fluctuation of super-capacity energy storage based on wind speed prediction described in the present invention, wherein: the wind speed direction prediction includes, assuming that the prediction group is [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ], the calculation of the prediction group is expressed as:

[0020]

[0021]

[0022] Among them, e represents the base of the natural logarithm of the mathematical constant; represents the adjustment parameter; q represents the fitting parameter, [q1, q2, q3, …, q 0.05n ] is calculated as:

[0023] Initially randomly generate a set [q1, q2, q3, ..., q 0.05n ] (1) , the range of each value is min(b i , c i )≤q i ≤max(b i , c i ), get the initial value

[0024] According to [q1, q2, q3, …, q 0.05n ] (1) The maximum value q in max and the minimum value q min , and randomly generate a set of [q1, q2, q3, …, q 0.05n ] (2) , the range of each value is min(b i / q min , c i / q min )≤q i ≤max(q max ×b i ,qmax ×c i ),get

[0025] Repeat the process of generating new sequences of q until satisfy:

[0026]

[0027] at this time is θ n+1 The final prediction result is calculated to θ n+0.05n , the obtained prediction group [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ] is the wind speed direction prediction result.

[0028] As a preferred solution of the method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction described in the present invention, the wind speed prediction includes: n ] is divided into calculation groups [x1, x2, x3, …, x 0.95n ] and adjustment group [x 0.95n+1 , x 0.95n+2 , x 0.95n+3 , …, x n ], using the calculation group decoupling adjustment group value, x′ 0.95n+1 The calculation method is expressed as:

[0029] x′ 0.95n+1 =α1x1+α2x2+α3x3+…+α0. 95n x 0.95n

[0030] Among them, x′ 0.95n+1 represents the x coupled by the calculation group 0.95n+1 α represents the fitting parameter; 20 groups [α1, α2, α3, …, α 0.95n ], calculate 20 x′ 0.95n+1 , expressed as Find 20 x' 0.95n+1 Actual value of the middle distance x 0.95n+1 The most recent value is counted as The corresponding α value is

[0031] Calculate the advance coefficient:

[0032]

[0033]

[0034] Among them, ω represents the forward coefficient, m represents the fitting parameter; avg represents the average value. According to the forward coefficient ω, 20 groups [α1, α2, α3, …, α 0.95n ] Move forward randomly again, and the degree of random advancement is expressed as:

[0035] α i =α i +rand(-ω,ω)

[0036] Among them, α i represents the i-th fitting parameter α; calculate x′ again 0.95n+1 , repeatedly calculate x′ 0.95n+1 And move forward 30 times, find the closest to the actual value X among the 30 times 0.95n+1 x' 0.95n+1 As the final output, the adjusted group value obtained by calculating the group coupling is [x′ 0.95n+1 , x′ 0.95n+2 , x′ 0.95n+3 , …, x′ n ];

[0037] Use GRU network to predict wind speed sequence [x n+1 , x n+2 , x n+3 , …, x n+0.95n ], the weights of the GRU network update gate are optimized, expressed as:

[0038]

[0039] Among them, W z represents the optimized update gate weight; W represents the initial update gate weight; n represents the total number of wind speed sequences.

[0040] As a preferred solution of the method for smoothing power fluctuation of super-capacity energy storage based on wind speed prediction described in the present invention, wherein: the wind speed prediction data includes: the wind speed direction prediction result [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ] projected onto the wind speed prediction result [x n+1 , x n+2 , x n+3 , …, x n+0.95n ], the final wind speed forecast data is expressed as [x n+1 cos(θ n+1 ), x n+2 cos(θ n+2 ), …, x n+0.95n cos(θ n+0.95n )].

[0041] As a preferred solution of the method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction described in the present invention, the super-capacity energy storage power fluctuation smoothing includes predicting the power generation power P according to the wind speed prediction data, and if P exceeds the expected load P f The wind speed is judged to be high if P is lower than the expected load P f The wind speed is judged to be low wind speed;

[0042] Based on historical data, calculate the average load within the time window H like The expected load is high if Then the expected load is low load;

[0043] When the wind speed forecast shows that the power generation will exceed the grid demand, the power surplus response strategy is triggered. Under high wind speed and high load conditions, if the energy storage is greater than or equal to 50%, it is charged at the maximum charging rate while meeting the grid load demand, expressed as:

[0044] P 充 =PP f

[0045] P 放 =P f

[0046] If the energy storage is less than 50%, it is charged at 50% of the charging rate, and the remaining energy is used for grid output, expressed as:

[0047] P 充 =0.5×(PP f )

[0048] P 放 =0.5×P

[0049] At high wind speed and low load, if the energy storage is greater than or equal to 50%, it is charged at the maximum charging rate and the excess energy is exported to the grid, expressed as:

[0050] P 充 =P

[0051] P 放 =P f

[0052] If the energy storage is less than 50%, it is charged at 80% of the charging rate, and 20% of the energy is used for grid output, expressed as:

[0053] P 充 =0.8P

[0054] P 放 =0.2P

[0055] When the wind speed forecast shows that the generated power will be lower than the grid demand, the power shortage response strategy is triggered. When the wind speed is low and the load is high, if the energy storage is greater than or equal to 50%, it is discharged at the maximum discharge rate to meet the grid load demand, which is expressed as:

[0056] P 放 =P f

[0057] If the energy storage is less than 50%, it is discharged at a discharge rate of 50%, and the rest is supplemented by wind energy, expressed as:

[0058] P 充 =0.5×P f

[0059] P = 0.5 × P f

[0060] At low wind speed and low load, if the energy storage is greater than or equal to 50%, it is discharged at the minimum discharge rate, and wind energy is preferentially used to meet the load, which is expressed as:

[0061] P 放 =P f

[0062] If the energy storage is less than 50%, it is discharged at a discharge rate of 60%, and 40% of the energy is supplemented by wind energy, expressed as:

[0063] P 充 =0.6×P f

[0064] P = 0.4 × P f

[0065] in, P 充 Indicates the charging power of the energy storage system; P 放 Represents the discharge power of the energy storage system.

[0066] In a second aspect, the present invention also provides an overcapacity energy storage power fluctuation smoothing system based on wind speed prediction, including: a wind direction prediction module, which defines the angle of wind speed, divides the angle sequence into a calculation group and a proofreading group, calculates the direction prediction coefficient, and predicts the future wind direction; a wind speed prediction module, which divides the wind speed magnitude sequence into a calculation group and an adjustment group, uses a GRU network to predict the wind speed magnitude sequence, combines the direction prediction value with the magnitude prediction value, and obtains wind speed prediction data; an energy storage adjustment module, which adjusts the charging and discharging power of the energy storage system according to the wind speed prediction data and the grid load demand, to achieve overcapacity energy storage power fluctuation smoothing.

[0067] In a third aspect, the present invention further provides a computing device, comprising: a memory and a processor;

[0068] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction are implemented.

[0069] In a fourth aspect, the present invention further provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction.

[0070] Beneficial effects of the present invention: The method of the present invention significantly improves the overall accuracy of wind speed prediction by independently collecting and processing wind speed direction and size data, and then combining the GRU network with high-precision prediction of wind speed size. Combined with the prediction data of direction and size, the dynamic adjustment of the supercapacitor energy storage system can be achieved, so that it can better respond to wind speed changes, smooth power fluctuations, and ensure grid stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0072] Figure 1 An overall flow chart of a method for smoothing power fluctuations of an ultra-capacity energy storage system based on wind speed prediction provided by an embodiment of the present invention;

[0073] Figure 2 A schematic diagram of wind speed direction of a method for smoothing power fluctuations of an ultra-capacity energy storage system based on wind speed prediction provided in accordance with a second embodiment of the present invention. DETAILED DESCRIPTION

[0074] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0075] Example 1

[0076] Reference Figure 1 and Figure 2 , which is an embodiment of the present invention, provides a method for smoothing power fluctuations of an ultra-capacity energy storage system based on wind speed prediction, comprising:

[0077] S1: Collect the direction data and magnitude data of the wind speed, divide the direction data into a calculation group and a proofreading group, calculate the direction prediction coefficient, and make a direction prediction of the wind speed.

[0078] Supercapacitor energy storage power fluctuation smoothing refers to the use of supercapacitors (supercapacitors) to buffer and stabilize power output fluctuations caused by wind speed changes. Due to the instability of wind power generation, wind speed changes will cause fluctuations in power generation. By connecting the supercapacitor energy storage system to the power grid, energy can be quickly stored and released when power fluctuations are caused by wind speed changes, thereby smoothing power output and improving the stability and reliability of the power grid.

[0079] The regulation of excess capacity energy storage is also based on the predicted wind speed to deduce the wind power value. However, in most existing technologies, only the size of the wind speed is predicted without considering the influence of wind direction and the angle of the wind turbine blades, resulting in inaccurate estimated wind power values, which in turn leads to the misalignment of excess capacity leveling, making the output value after leveling not meet the requirements of grid connection. Therefore, this method predicts the size and direction of the wind speed separately, and then integrates the prediction results to make the prediction of wind speed more accurate.

[0080] like Figure 2 As shown in the figure, assuming that the angle perpendicular to the fan blade is 0°, the direction of the wind speed is defined as the angle θ that deviates from the angle perpendicular to the fan blade. The collected wind speed sequence is [x1, x2, x3, …, x n ], the corresponding angles are [θ1, θ2, θ3, …, θ n ].

[0081] Furthermore, [θ1, θ2, θ3, …, θ n ] is divided into calculation groups [θ1, θ2, θ3, …, θ 0.95n ] and proofreading group [θ 0.95n+1 ,θ 0.95n+2 ,θ 0.95n+3 ,…,θ n ], first use the calculation group to calculate the value of the known proofreading group, generate a coefficient based on the difference between the calculated value and the actual proofreading group value, and then use the generated coefficient to calculate the unknown prediction group value.

[0082] Use the calculation group to calculate the value of the proofreading group, expressed as:

[0083]

[0084]

[0085]

[0086] Among them, θ′ 0.95n+1 represents the θ calculated using the calculation group0.95n+1 The value of; α represents the activation regression coefficient; j represents the number of terms in the weighted sum; s represents the number of terms used for factorial weighted sum calculation; represents the value of the i-th calculation group; ELU represents the activation function.

[0087] It should be noted that in θ' 0.95n+2 And in the subsequent calculations, the ELU function is added, and correction is made at each step according to the error between the actual value and the predicted value. ELU is used to smoothly adjust small negative errors to avoid the impact of excessive negative values, and linear adjustments are made to positive errors to maintain the flexibility and accuracy of the prediction, making subsequent predictions more accurate.

[0088] α is the activation regression coefficient, expressed as:

[0089]

[0090] Confirmation method of j and s: Randomly give a set of j and s to calculate θ' 0.95n+1 , then calculate whether If it does not meet the requirement, continue to randomly assign a set of values ​​until j and s meet the requirement. until.

[0091] The proofreading group obtained by calculating the calculation group is [θ′ 0.95n+1 ,θ′ 0.95n+2 ,θ′ 0.95n+3 ,…,θ′ n ], and the proofreading group [θ 0.95n+1 ,θ 0.95n+2 ,θ 0.95n+3 ,…,θ n ]Comparison calculation direction prediction coefficient, expressed as:

[0092]

[0093]

[0094] Among them, sigmoid represents the sigmoid activation function; tanh represents the tanh activation function; represents the value of the i-th proofreading group; θ′i represents the value of the i-th proofreading group calculated using the calculation group; c i Represents the directional prediction difference coefficient, [c1, c2, c3, …, c 0.05n ]; b i represents the directional prediction ratio coefficient, [b1, b2, b3, …, b 0.05n ].

[0095] It should be noted that when calculating θ′ through the calculation group, the impact of outliers and noise on the prediction results is reduced through scaling and exponential smoothing, making the prediction more stable and reliable. i The / i! form introduces an exponential decay factor, which gives lower weight to older data points and higher weight to newer data points, emphasizing recent data and making the wind direction forecast closer to the most recent situation.

[0096] Furthermore, let the prediction group be [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ], the calculation of the prediction group is expressed as:

[0097]

[0098]

[0099] Among them, e represents the base of the natural logarithm of the mathematical constant; represents the adjustment parameter; q represents the fitting parameter, and the calculation method of [q1, q2, q3, …, q0.05n] is expressed as:

[0100] Initially randomly generate a set [q1, q2, q3, ..., q 0.05n ] (1) , the range of each value is min(b i , c i )≤q i ≤max(b i , c i ), get the initial value

[0101] According to [q1, q2, q3, …, q 0.05n ] (1) The maximum value q in max and the minimum value q min , and randomly generate a set of [q1, q2, q3, …, q 0.05n ] (2) , the range of each value is min(b i / q min , c i / q min )≤q i ≤max(q max ×b i ,q max ×c i ),get

[0102] Repeat the process of generating new sequences of q until satisfy:

[0103]

[0104] at this time is θ n+1 The final prediction result is calculated to θ n+0.05n , the obtained prediction group [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ] is the wind speed direction prediction result.

[0105] S2: Divide the size data into a calculation group and an adjustment group, use the GRU network to predict the size of the wind speed, combine the direction prediction value with the size prediction value, and obtain the wind speed prediction data.

[0106] Furthermore, the wind speed sequence is [x1, x2, x3, …, x n ] is divided into calculation groups [x1, x2, x3, …, x 0.95n ] and adjustment group [x 0.95n+1 , x 0.95n+2 , x 0.95n+3 , …, x n ], using the calculation group decoupling adjustment group value, x′ 0.95n+1 The calculation method is expressed as:

[0107] x′ 0.95n+1 =α1x1+α2x2+α3x3+…+α 0.95n x 0.95n

[0108] Among them, x′ 0.95n+1 Denotes X using the computational group coupling 0.95n+1 The value of ; α represents the fitting parameter.

[0109] Randomly generate 20 groups [α1, α2, α3, …, α 0.95n ], calculate 20 x′ 0.95n+1 , expressed as Find 20 x' 0.95n+1 Actual value of the middle distance x 0.95n+1 The most recent value is counted as The corresponding α value is Calculate the advance coefficient:

[0110]

[0111]

[0112] Among them, ω represents the forward coefficient, m represents the fitting parameter; avg represents the average value. According to the forward coefficient ω, 20 groups [α1, α2, α3, …, α 0.95n ] Move forward randomly again, and the degree of random advancement is expressed as:

[0113] α i =α i +rand(-ω,ω)

[0114] Among them, α i represents the i-th fitting parameter α; that is, each α in the 20 groups i Both advance by the amount of rand(-ω, ω).

[0115] Calculate x′ again 0.95n+1 , repeatedly calculate x′ 0.95n+1 And move forward 30 times, find the closest to the actual value X among the 30 times 0.95n+1 x' 0.95n+1 As the final output, the adjusted group value obtained by calculating the group coupling is [x′ 0.95n+1 , x′ 0.95n+2 , x′ 0.95n+3 , …, x′ n ].

[0116] Use GRU network to predict wind speed sequence [x n+1 , x n+2 , x n+3 , …, x n+0.95n ], the weights of the GRU network update gate are optimized, expressed as:

[0117]

[0118] Among them, W z represents the optimized update gate weight; W represents the initial update gate weight; n represents the total number of wind speed sequences.

[0119] Furthermore, the wind speed direction prediction result [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ] projected onto the wind speed prediction result [x n+1 , x n+2 , x n+3 , …, x n+0.95n ], the final wind speed forecast data is expressed as [x n+1 cos(θ n+1 ), x n+2 cos(θ n+2 ), …, x n+0.95n cos(θ n+0.95n )].

[0120] S3: Dynamically adjust the charge and discharge rate of the over-capacity energy storage system according to the wind speed prediction data and grid load demand to achieve the suppression of power fluctuations of the over-capacity energy storage.

[0121] Furthermore, converting the wind speed x into the power generation P is usually achieved through the power curve of the wind turbine. The power curve of the wind turbine usually includes the following key regions: the minimum wind speed x1 at which the wind turbine starts generating electricity, the wind speed x2 at which the wind turbine can output the maximum rated power, and the maximum wind speed x3 at which the wind turbine stops operating for safety reasons.

[0122] When the predicted wind speed x < x1, the power generation P = 0;

[0123] When x1 ≤ x < x2, the power generation is proportional to the wind speed, expressed as:

[0124]

[0125] When x2 ≤ x < x3, the power generation P = P 额定 ;

[0126] When x ≥ x3, the power generation P = 0.

[0127] Furthermore, predict the power generation P according to the wind speed prediction data. If P exceeds the expected load P f then determine that the wind speed is a high wind speed. If P is lower than the expected load P f then determine that the wind speed is a low wind speed.

[0128] Calculate the average load within the time window H according to historical data If then the expected load is a high load. If then the expected load is a low load.

[0129] When the wind speed prediction shows that the power generation will exceed the grid demand, trigger the power surplus response strategy. At high wind speed - high load, if the energy storage is greater than or equal to 50%, then charge at the maximum charge rate while meeting the grid load demand, expressed as:

[0130] P 充 = P - P f

[0131] P 放 = P f

[0132] If the energy storage is less than 50%, then charge at a 50% charge rate while using the remaining electric energy for grid output, expressed as:

[0133] P 充=0.5×(PP f )

[0134] P 放 =0.5×P

[0135] At high wind speed and low load, if the energy storage is greater than or equal to 50%, it is charged at the maximum charging rate and the excess energy is exported to the grid, expressed as:

[0136] P 充 =P

[0137] P 放 =P f

[0138] If the energy storage is less than 50%, it is charged at 80% of the charging rate, and 20% of the energy is used for grid output, expressed as:

[0139] P 充 =0.8P

[0140] P 放 =0.2P

[0141] When the wind speed forecast shows that the generated power will be lower than the grid demand, the power shortage response strategy is triggered. When the wind speed is low and the load is high, if the energy storage is greater than or equal to 50%, it is discharged at the maximum discharge rate to meet the grid load demand, which is expressed as:

[0142] P 放 =P f

[0143] If the energy storage is less than 50%, it is discharged at a discharge rate of 50%, and the rest is supplemented by wind energy, expressed as:

[0144] P 充 =0.5×P f

[0145] P = 0.5 × P f

[0146] At low wind speed and low load, if the energy storage is greater than or equal to 50%, it is discharged at the minimum discharge rate, and wind energy is preferentially used to meet the load, which is expressed as:

[0147] P 放 =P f

[0148] If the energy storage is less than 50%, it is discharged at a discharge rate of 60%, and 40% of the energy is supplemented by wind energy, expressed as:

[0149] P 充 =0.6×P f

[0150] P = 0.4 × Pf

[0151] Among them, P 充 Represents the charging power of the energy storage system; P 放 Represents the discharge power of the energy storage system.

[0152] It should be noted that the adjustment strategy assumes by default that the grid demand can be met. Considering that in actual situations the grid demand may not be met according to the fixed adjustment strategy, it is necessary to dynamically adjust the charging and discharging power of the energy storage.

[0153] Taking the power surplus response strategy as an example, when the wind speed is high and the load is low, and the energy storage is less than 50%, P 放 =0.2P Under normal circumstances, only 20% of wind energy is converted into usable power, which may lead to failure to meet grid demand and cause power shortage. Therefore, dynamic adjustment is required in practical applications.

[0154] When P f >0.2P, then increase the discharge power P 放 To meet the grid load demand, but not exceed the discharge capacity of the supercapacitor; when P f When ≤0.2P, the discharge power is set to P 放 =P f and use the remaining wind energy for charging or storage.

[0155] Through dynamic adjustment, it can ensure that the grid load is met, avoid power shortages, make full use of wind energy when there is sufficient wind energy, and avoid energy waste.

[0156] This embodiment also provides an overcapacity energy storage power fluctuation smoothing system based on wind speed prediction, including: a wind direction prediction module, which defines the angle of wind speed, divides the angle sequence into a calculation group and a proofreading group, calculates the direction prediction coefficient, and predicts the future wind direction; a wind speed prediction module, which divides the wind speed magnitude sequence into a calculation group and an adjustment group, uses a GRU network to predict the wind speed magnitude sequence, combines the direction prediction value with the magnitude prediction value, and obtains wind speed prediction data; an energy storage adjustment module, which adjusts the charging and discharging power of the energy storage system according to the wind speed prediction data and the grid load demand, and realizes overcapacity energy storage power fluctuation smoothing.

[0157] This embodiment also provides a computing device, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the method for smoothing power fluctuations of ultra-capacity energy storage based on wind speed prediction as proposed in the above embodiment.

[0158] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction proposed in the above embodiment is implemented.

[0159] The storage medium proposed in this embodiment and the method for smoothing power fluctuations of ultra-capacity energy storage based on wind speed prediction proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0160] Through the above description of the implementation method, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccessMemory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0161] Example 2

[0162] The following is an embodiment of the present invention, which provides a method for smoothing power fluctuations of an ultra-capacity energy storage system based on wind speed prediction. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0163] Use MATLAB / Simulink software to conduct simulation experiments, create a Simulink model, use the RandomNumber module to simulate the random changes in wind speed, and use the SineWave module and Product module to simulate the changes in wind speed direction.

[0164] The required wind speed direction and size data can be directly generated through simulation, and the degree of improvement in prediction accuracy when considering wind direction can be better compared. In the Simulink model, the prediction of wind direction is implemented by a custom algorithm through the MATLABFunction module, and the wind speed size prediction is implemented by adding a GRU network model on this basis. Six groups of different wind speed and wind direction conditions are simulated, and this method is used for prediction and energy storage adjustment. The experimental data are shown in the following table.

[0165] Table 1 Experimental data table

[0166]

[0167]

[0168] From the table data, it is obvious that the bidirectional prediction model of wind speed direction and magnitude improves the accuracy of wind speed prediction. For example, in test object 1, the actual wind speed magnitude is 10.5m / s, and the predicted wind speed magnitude is 10.3m / s, with an error of only 0.2m / s. The predicted wind speed direction is also very close to the actual direction. This shows that the prediction accuracy is improved by independently processing the data of wind speed direction and magnitude.

[0169] Combined with the dynamic adjustment strategy of the supercapacitor energy storage system, when the energy storage state is higher than 50%, it is charged at a higher charging rate, and when the energy storage state is lower than 50%, it is charged at a moderate charging rate, and part of the electric energy is used for grid output. For example, in test object 2, the energy storage state is 48%, and 0.8 times the wind power is used for charging (i.e. 6.5kW), and the remaining electric energy (i.e. 4.1kW) is used for grid output, which not only ensures the stability of the grid, but also avoids overcharging of the energy storage system.

[0170] In summary, the method of the present invention has significant advantages in wind speed prediction and energy storage system regulation. Compared with the prior art that only considers the size of wind speed, the present invention improves the accuracy of wind speed prediction by considering the direction and size of wind speed, and further optimizes the charge and discharge regulation strategy of the supercapacitor energy storage system. This method effectively solves the problems of inaccurate wind speed prediction and unreasonable energy storage regulation in the prior art, and significantly improves the stability and reliability of wind power grid connection.

[0171] 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 preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction, characterized in that: include: Collect the direction data and magnitude data of wind speed, divide the direction data into calculation group and proofreading group, calculate the direction prediction coefficient, and make the direction prediction of wind speed; The wind speed direction data and magnitude data include: assuming that the angle perpendicular to the fan blade is 0°, the direction of the wind speed is defined as the angle θ that deviates from the angle perpendicular to the fan blade, and the collected wind speed magnitude sequence is [x1, x2, x3, ..., x n ], the corresponding angles are [θ1, θ2, θ3, …, θ n ]; The directional prediction coefficients include: [θ1, θ2, θ3, ..., θ n ] is divided into calculation groups [θ1, θ2, θ3, …, θ 0.95n ] and proofreading group [θ 0.95n+1 ,θ 0.95n+2 ,θ 0.95n+3 ,…,θ n ], use the calculation group to calculate the value of the proofreading group, expressed as: Among them, θ′ 0.95n+1 represents the θ calculated using the calculation group 0.95n+1 The value of; α represents the activation regression coefficient; j represents the number of terms in the weighted sum; s represents the number of terms used for factorial weighted sum calculation; represents the value of the i-th calculation group; ELU represents the activation function; α is the activation regression coefficient, expressed as: The proofreading group obtained by calculating the calculation group is [θ′ 0.95n+1 ,θ′ 0.95n+2 ,θ′ 0.95n+3 ,…,θ′ n ], and the proofreading group [θ 0.95n+1 ,θ 0.95n+2 ,θ 0.95n+3 ,…,θ n ]Comparison calculation direction prediction coefficient, expressed as: Among them, sigmoid represents the sigmoid activation function; tanh represents the tanh activation function; represents the value of the i-th proofreading group; θ′ i represents the value of the proofreading group obtained by using the calculation group; c i Represents the directional prediction difference coefficient, [c1, c2, c3, …, c 0.05n ]; b i represents the directional prediction ratio coefficient, [b1, b2, b3, …, b 0.05n ]; The wind speed data are divided into a calculation group and an adjustment group. The GRU network is used to predict the wind speed. The direction prediction value is combined with the size prediction value to obtain the wind speed prediction data. The wind speed prediction data includes the wind speed direction prediction result [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ] projected onto the wind speed prediction result [x n+1 , x n+2 , x n+3 , …, x n+0.95n ], the final wind speed forecast data is expressed as [x n+1 cos(θ n+1 ),x n+2 cos(θ n+2 ), …, x n+0.95n cos(θ n+0.95n )]; According to the wind speed forecast data and grid load demand, the charging and discharging rates of the super-capacity energy storage system are dynamically adjusted to achieve super-capacity energy storage power fluctuation smoothing.

2. The method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction according to claim 1 is characterized in that: The wind speed direction prediction includes: assuming the prediction group is [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ], the calculation of the prediction group is expressed as: Among them, e represents the base of the natural logarithm of the mathematical constant; represents the adjustment parameter; q represents the fitting parameter, [q1, q2, q3, …, q 0.05n The calculation method of ] is as follows: initially randomly generate a set of [q1, q2, q3, …, q 0.05n ] (1) , the range of each value is min(b i , c i )≤q i ≤max(b i , c i ), get the initial value According to [q1, q2, q3, …, q 0.05n ] (1) The maximum value q in max and the minimum value q min , and randomly generate a set of [q1, q2, q3, …, q 0.05n ] (2) , the range of each value is min(b i / q min , c i / q min )≤q i ≤max(q max ×b i ,q max ×c i ),get Repeat the process of generating new sequences of q until satisfy: at this time is θ n+1 The final prediction result is calculated to θ n+0.05n , the obtained prediction group [θ n+1 ,θ n+2 ,θ n+3 ,…,θ n+0.05n ] is the wind speed direction prediction result.

3. The method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction according to claim 2 is characterized in that: The wind speed magnitude prediction includes converting the wind speed magnitude sequence into [x1, x2, x3, ..., x n ] is divided into calculation groups [x1, x2, x3, …, x 0.95n ] and adjustment group [x 0.95n+1 , x 0.95n+2 , x 0.95n+3 , …, x n ], using the calculation group decoupling adjustment group value, x' 0.95n+1 The calculation method is expressed as: x' 0.95n+1 =α1x1+α2x2+α3x3+...+α 0.95n x 0.95n Among them, x′ 0.95n+1 Denotes X using the computational group coupling 0.95n+1 α represents the fitting parameter; 20 groups [α1, α2, α3, …, α 0.95n ], calculate 20 x′ 0.95n+1 , expressed as Find 20 x' 0.95n+1 Actual value of mid-range X 0.95n+1 The most recent value is counted as The corresponding α value is Calculate the advance coefficient: Among them, ω represents the forward coefficient, m represents the fitting parameter; avg represents the average value. According to the forward coefficient ω, 20 groups [α1, α2, α3, …, α 0.95n ] Move forward randomly again, and the degree of random advancement is expressed as: a i =a i +rand(-ω,ω) Among them, α i represents the i-th fitting parameter α; calculate x′ again 0.95n+1 , repeatedly calculate x′ 0.95n+1 And move forward 30 times, find the closest to the actual value X among the 30 times 0.95n+1 x' 0.95n+1 As the final output, the adjusted group value obtained by calculating the group coupling is [x′ 0.95n+1 , x' 0.95n+2 , X′ 0.95n+3 ,…,x′ n ]; Use GRU network to predict wind speed sequence [x n+1 , x n+2 , x n+3 , …, x n+0.95n ], the weights of the GRU network update gate are optimized, expressed as: Among them, W z represents the optimized update gate weight; W represents the initial update gate weight; n represents the total number of wind speed sequences.

4. The method for smoothing power fluctuations of super-capacity energy storage based on wind speed prediction according to claim 3 is characterized in that: The super-capacity energy storage power fluctuation smoothing includes predicting the power generation power P according to the wind speed forecast data, and if P exceeds the expected load P f The wind speed is judged to be high if P is lower than the expected load P f The wind speed is judged to be low wind speed; Based on historical data, calculate the average load within the time window H like The expected load is high if Then the expected load is low load; When the wind speed forecast shows that the power generation will exceed the grid demand, the power surplus response strategy is triggered. Under high wind speed and high load conditions, if the energy storage is greater than or equal to 50%, it is charged at the maximum charging rate while meeting the grid load demand, expressed as: P 充 =P-P f P 放 =P f If the energy storage is less than 50%, it is charged at 50% of the charging rate, and the remaining energy is used for grid output, expressed as: P 充 =0.5×(P-P f ) P 放 =0.5×P At high wind speed and low load, if the energy storage is greater than or equal to 50%, it is charged at the maximum charging rate and the excess energy is exported to the grid, expressed as: P 充 =P P 放 =P f If the energy storage is less than 50%, it is charged at 80% of the charging rate, and 20% of the energy is used for grid output, expressed as: P 充 =0.8P P 放 =0.2P When the wind speed forecast shows that the generated power will be lower than the grid demand, the power shortage response strategy is triggered. When the wind speed is low and the load is high, if the energy storage is greater than or equal to 50%, it is discharged at the maximum discharge rate to meet the grid load demand, which is expressed as: P 放 =P f If the energy storage is less than 50%, it is discharged at a discharge rate of 50%, and the rest is supplemented by wind energy, expressed as: P 充 =0.5×P f P=0.5×P f At low wind speed and low load, if the energy storage is greater than or equal to 50%, it is discharged at the minimum discharge rate, and wind energy is preferentially used to meet the load, which is expressed as: P 放 =P f If the energy storage is less than 50%, it is discharged at a discharge rate of 60%, and 40% of the energy is supplemented by wind energy, expressed as: P 充 =0.6×P f P=0.4×P f Among them, P 充 Represents the charging power of the energy storage system; P 放 Represents the discharge power of the energy storage system.

5. A super-capacity energy storage power fluctuation smoothing system based on wind speed prediction using the method as described in any one of claims 1 to 4, characterized in that: include, The wind direction prediction module defines the angle of wind speed, divides the angle sequence into calculation group and proofreading group, calculates the direction prediction coefficient, and predicts the future wind direction; The wind speed prediction module divides the wind speed sequence into a calculation group and an adjustment group, uses the GRU network to predict the wind speed sequence, combines the direction prediction value with the size prediction value, and obtains the wind speed prediction data; The energy storage adjustment module adjusts the charging and discharging power of the energy storage system according to wind speed forecast data and grid load demand, thereby smoothing out fluctuations in over-capacity energy storage power.

6. A computing device comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the method described in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the method according to any one of claims 1 to 4.

Images