A matrix wind farm level active power control system based on wind speed measurement

Abstract

The invention relates to a field-level active power control system for a matrix wind farm based on wind speed measurement. By setting a field-level operation decision library that is continuously updated and iterated, a precise wind measuring device is used to identify accurate incoming wind conditions, and the yaw target angle of each wind turbine in the field-level operation decision library is found. The yaw target angle of the wind turbine is issued and executed through a field-level control system and a wind turbine main control system. The system adopts a dynamically updated field-level operation decision library. After simulating and calculating the initial value of the field-level operation decision library, each operation T period is dynamically updated using the operation result in the previous operation T period, thereby reducing the simulation calculation deviation of the field-level operation decision library. At the same time, the system adopts field-level active power optimal control of the matrix wind farm, and based on the full-field maximum power tracking of the matrix wind farm, the yaw angle of the wind turbine is controlled, thereby reducing the influence and interference between wind turbines and reducing the wake loss of the matrix wind farm by changing the yaw angle of the front-row wind turbine.

Description

Technical Field

[0001] The invention relates to the technical field of wind farms, and in particular to a matrix wind farm level active power control system based on wind speed measurement. Background Art

[0002] As we all know, wind power generation, as a clean energy, has developed very rapidly and plays an important role in the country's "carbon peak and carbon neutrality" action. How to maximize the utilization efficiency of wind energy in wind power generation is an important direction for improving the rate of return of wind farms and improving the quality and efficiency of wind power generation. Traditional research on increasing power generation only considers the optimization of the operating status of a single unit. Wind turbine control generally adopts the maximum power tracking control method, and the maximum power tracking control of a single unit is not the maximum power control of the entire wind farm, because there is mutual influence and mutual interference between the wind turbines in the wind farm, such as the wake effect, among which the wake effect of the matrix wind farm is more prominent, and the traditional control does not consider the influence between units, which causes mutual interference between units due to the wake effect, causing the loss of power generation in the entire field. Therefore, the wind farm-level active power control system can effectively improve the power generation of the entire wind farm. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the purpose of the present invention is to provide a matrix wind farm-level active power control system based on wind speed measurement. By controlling the operation mode of the wind turbines, the problem of mutual influence between wind turbines in the wind farm can be effectively solved, thereby increasing the total active power at the wind farm level, and then increasing the power generation of the wind farm.

[0004] The technical solution provided by the present invention is:

[0005] A matrix wind farm field-level active power control system based on wind speed measurement, the system includes a precise wind measurement system, a field-level operation decision library, a field-level control system and a wind turbine master control system, wherein:

[0006] Precision wind measurement system: used to measure the real-time wind conditions of the front row wind turbines, including the wind speed and wind direction, and search the field-level operation decision database according to the real-time wind direction to match the yaw target angle of each wind turbine in the wind farm corresponding to the real-time wind conditions;

[0007] Field-level operation decision library: includes the target yaw angle of wind turbines when the field-level active power of wind farms is optimal under various incoming wind conditions, and is continuously updated and iterated;

[0008] Field-level control system: calls the target yaw angle of the wind turbine in this state, and sends the target yaw angle of the wind turbine to the main control system of each wind turbine;

[0009] Wind turbine main control system: adjusts the wind turbine's yaw system to the target angle according to the yaw target angle and maintains this operating state.

[0010] Compared with the prior art, the present invention effectively solves the problem of mutual influence between wind turbines in a wind farm by controlling the operation mode of the wind turbine, thereby increasing the total active power of the wind farm level and improving the power generation of the wind farm. The present invention sets a continuously updated and iterated field-level operation decision library, uses a precise wind measuring device to identify accurate incoming wind conditions, searches for the yaw target angle of each wind turbine in the field-level operation decision library, and issues and executes the yaw target angle of the wind turbine through the field-level control system and the wind turbine main control system. On the one hand, the system adopts a dynamically updated field-level operation decision library. After simulating and calculating the initial value of the field-level operation decision library, each operation T period is dynamically updated using the operation results in the previous operation T period, thereby reducing the simulation calculation deviation of the field-level operation decision library; on the other hand, the system adopts the field-level active power optimal control of the matrix wind farm, based on the full-field maximum power tracking of the matrix wind farm, to realize the control of the yaw angle of the wind turbine, thereby reducing the influence and interference between the wind turbines by changing the yaw angle of the front-row wind turbines, and reducing the wake loss of the matrix wind farm. It is easy to use, has good effects, and has good social and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 The present invention is a flow chart of a control method of a matrix wind farm-level active power control system based on wind speed measurement.

[0012] Figure 2 Schematic diagram of the wind turbine layout of a matrix wind farm and the first reference front row of wind turbines.

[0013] Figure 3 Flowchart for solving the particle swarm optimization algorithm based on Monte Carlo simulation.

[0014] Figure 4 This is a comparison chart of the total active power of wind turbines in wind farms with three control methods under different wind directions.

[0015] Figure 5 The following is a comparison chart of the average active power of wind turbines in wind farms with three control methods at different wind speeds. DETAILED DESCRIPTION

[0016] The specific implementation modes of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0017] The present invention provides a matrix-type wind farm field-level active power control system based on wind speed measurement, comprising a precise wind measurement system, a field-level operation decision library, a field-level control system and a wind turbine master control system, wherein:

[0018] Precision wind measurement system: used to measure the real-time wind conditions of the front row wind turbines, including the wind speed and wind direction, and search the field-level operation decision database according to the real-time wind direction to match the yaw target angle of each wind turbine in the wind farm corresponding to the real-time wind conditions;

[0019] Field-level operation decision library: includes the target yaw angle of wind turbines when the field-level active power of wind farms is optimal under various incoming wind conditions, and is continuously updated and iterated;

[0020] Field-level control system: calls the target yaw angle of the wind turbine in this state, and sends the target yaw angle of the wind turbine to the main control system of each wind turbine;

[0021] Wind turbine main control system: adjusts the wind turbine's yaw system to the target angle according to the yaw target angle and maintains this operating state.

[0022] To ensure the use effect, the precise wind measurement system includes a precise wind measurement device arranged on the sentinel fan, which can obtain the real-time wind conditions of the wind speed and wind direction of the wind in front of the fan;

[0023] The sentinel wind turbines are wind turbines installed at the four corners of the matrix wind farm. Figure 2 As shown, fan a 11 、a m1 、a 1n 、a mn There are four sentinel wind turbines, each of which is equipped with a precise wind measuring device; the remaining wind turbines are still traditional mechanical or ultrasonic anemometers for wind speed and direction. The precise wind measuring device is a prior art, which can obtain the real-time wind conditions of the accurate incoming wind speed and wind direction in front of the wind turbine, such as the MolasNL Doppler wind measurement lidar wind measuring device produced and sold by Nanjing Mulei Laser Technology Co., Ltd.

[0024] According to the incoming wind direction, 0-360° is equally divided into four intervals, corresponding to the four sentinel wind turbines of the matrix wind farm. The first sentinel wind turbine is determined according to the interval where the incoming wind direction is located, and the accurate incoming wind conditions are obtained based on the precise wind measurement system of the first sentinel wind turbine.

[0025] The initial value calculation process of the field-level operation decision library is as follows:

[0026] According to the incoming wind speed, based on the single wind turbine wake model, the single wind turbine yaw wake model and the multi-wind turbine yaw wake superposition model, the calculated wind speed of other wind turbines in the wind farm is obtained:

[0027] Based on the single wind turbine wake model, the wind turbine wake velocity distribution is calculated and the wind turbine wake loss Δv is obtained:

[0028]

[0029] Where: Δv is the wind turbine wake loss, V ∞ is the incoming wind speed, C T is the thrust coefficient of the wind turbine, σ is the characteristic width of the wake, r0 is the radius of the wind wheel, and d0 is the impeller diameter.

[0030] Based on the single wind turbine yaw wake model, the distribution of the wind turbine wake in the yaw state is calculated:

[0031]

[0032] Where: y d is the center offset of the wind turbine wake, θ is the yaw angle, k is the wake expansion coefficient, and x is the length of the wake core area.

[0033] Based on the superposition model of multiple wind turbine yaw wakes, the wind turbine wake loss ΔV under the influence of multiple wind turbines is calculated:

[0034]

[0035] Where: ΔV is the wind turbine wake loss under the influence of multiple wind turbines, is the characteristic propagation velocity of the wake when the ith wind turbine is operating in isolation, V C is the wake characteristic propagation speed after the operation of multiple wind turbines, u i is the wake velocity of the ith wind turbine when it operates in isolation.

[0036] Corresponding to fan a ij When the wind speed is v ij :

[0037] v ij =V ∞ -ΔV ij

[0038] Where: v ij For fan a ij Calculated wind speed, ΔV ij For fan a ij of wake loss.

[0039] The calculated active power of each wind turbine is derived based on the calculated wind speed, the maximum value of the sum of the active power of the entire field under the current wind conditions is obtained, and the target yaw angle of all wind turbines is obtained, which is specifically:

[0040] The angle between the yaw angle of the wind turbine and the incoming wind direction is the yaw angle θ. When the yaw angle of the wind turbine changes, the calculated wind speed of the wind turbine will also change. When the yaw angle of all wind turbines in the matrix wind farm is When , the corresponding calculated wind speed of each fan is Calculating active power and

[0041] Where: θ ij For fan a ij The yaw angle, v ij (θ ij ) is fan a ij Calculated wind speed, P ij (θ ij ) is fan a ij The calculation of active power, ρ air density, A is the wind wheel swept area, C P is the wind energy utilization coefficient.

[0042] The sum of the active power of the matrix wind farm is P, and the objective function of the control optimization is:

[0043]

[0044] Where: P is the sum of the active power of the matrix wind farm, P ij (θ ij ) is fan a ij The calculated active power is m, the number of rows of wind turbines in the matrix wind farm, and n is the number of columns of wind turbines in the matrix wind farm.

[0045] The particle swarm optimization algorithm of Monte Carlo simulation is used to optimize the objective function and obtain the target yaw angle of each wind turbine in the field under the current wind conditions.

[0046] The specific method of optimizing the objective function using the particle swarm optimization algorithm of Monte Carlo simulation includes the following steps:

[0047] Step 1: Initialize a particle population: randomly generate m*n particles with position and velocity within the set range, calculate the fitness of each particle as the local optimal fitness, compare the local optimal fitness of m*n particles, select the best one and record it in the global optimal fitness, and record this particle as the global optimal vector;

[0048] Step 2: Update the weight factor w and learning factors c1, c2:

[0049]

[0050] Where: w max 、w min is the maximum and minimum value of the inertia weight factor; f is the current fitness value, f avg and f min They are the average fitness value and minimum fitness value of all particles at present;

[0051]

[0052]

[0053] Where: c 1end and c 1start is the final and initial value of c1, c 2end and c 2start is the final value and initial value of c2, t is the current iteration number, t max is the maximum number of iterations;

[0054] Step 4: Calculate the objective function to obtain the fitness of the current particle;

[0055] Step 5: Update the local optimal fitness and the local optimal vector of the body;

[0056] Step 6: Update the global optimal fitness and the global optimal vector;

[0057] Step 7: Update the position and velocity of each particle;

[0058]

[0059]

[0060] In the formula, 1≤i≤m*n, m*n is the size of the population, represents the d-dimensional velocity of the i-th particle in the k-th iteration; represents the d-dimensional position of the i-th particle in the k-th iteration; ω represents the inertia weight; c1 and c2 represent the learning factors of the particles, represents the d-dimensional monomer extreme value of particle i in the k-th iteration; It represents the global extreme value of the d-dimensional particle swarm in the k-th iteration; is a random number distributed in the (0,1) interval;

[0061] Step 8: If the number of iterations reaches the maximum value, stop searching and output the results, otherwise return to step 2 to continue iterative calculation;

[0062] The objective function is solved by the particle swarm optimization algorithm of Monte Carlo simulation, and the optimal solution of the objective function θ is obtained. ij , thus obtaining the target yaw angle of all wind turbines in the current wind condition;

[0063] By solving the objective function under different wind conditions, the target yaw angle of all wind turbines in each wind condition can be obtained.

[0064] According to the above steps, simulate and calculate the initial value of the field-level operation decision library.

[0065] The process of updating and iterating the field-level operation decision library is as follows:

[0066] In the next running T period, the running results in the previous running T period are used for dynamic update. The specific updating steps are as follows:

[0067] Define the other three sentinel fans except the first sentinel fan as correction fans, among which the sentinel fan diagonally opposite to the first sentinel fan is the main correction fan, and the other two sentinel fans are slave correction fans. 11 When it is the first sentinel fan, fan a mn Main correction fan, fan a m1 and a 1n In order to calibrate the fan, the accurate wind measurement system of the calibrated fan is used to calibrate the calculated wind speed to obtain the effective wind speed of each fan. m1 Complete (a 11 ,a m1 ) The calculated wind speed correction of this column of fans is performed from the correction fan a 1n Complete (a 11 ,a 1n ) Calculated wind speed correction for this row of fans, main correction fan a mn , from the correction fan a m1 , from the correction fan a 1n Complete the calculated wind speed correction for all other wind turbines in the matrix wind farm;

[0068] With a 11 Taking the first sentinel fan as an example, fan a mn For main correction fan, fan a m1 and a 1n For the correction fan.

[0069] From the correction fan a m1 Complete (a 11 ,a m1 ) The calculated wind speed correction steps for this column of fans:

[0070] Read the value of the correction fan a during the T period m1 The actual measured wind speed and calculate the wind speed v m1 (t), then at time t, the correction fan a m1 Correction factor k m1 (t) is:

[0071]

[0072] Calculate all the correction coefficients in the T period and obtain the correction fan a in the T period m1 The average correction factor k m1 (T) is:

[0073]

[0074] Then in the period (T, 2T), (a 11 ,a m1 ) Fan a in the fan series i1 The actual correction factor k i1 (T) is:

[0075]

[0076] From the correction fan a 1n Complete (a 11 ,a 1n ) The calculated wind speed correction steps for this row of fans:

[0077] Read the value of the correction fan a during the T period 1n The actual measured wind speed and calculate the wind speed v 1n (t), then at time t, the correction fan a 1n Correction factor k 1n (t) is:

[0078]

[0079] Calculate all the correction coefficients in the T period and obtain the correction fan a in the T period 1n The average correction factor k 1n (T) is:

[0080]

[0081] Then in the period (T, 2T), (a 11 ,a 1n ) Air blower in the row 1j The actual correction factor k 1j (T) is:

[0082]

[0083] Main correction fan a mn , from the correction fan a m1 , from the correction fan a 1n Finish Correction steps for calculated wind speed of fan:

[0084] Read the main correction fan a during the T period mn The actual measured wind speed and calculate the wind speed v mn (t), then at time t the main correction fan a mn Correction factor k mn (t) is:

[0085]

[0086] Calculate all correction systems in the T period and obtain the main correction fan a in the T period mn The average correction system k mn (T) is:

[0087]

[0088] According to the previous derivation, during the T period, the correction fan a m1 The average correction factor k m1 (T) is:

[0089]

[0090] According to the previous derivation, during the T period, the correction fan a 1n The average correction factor k 1n (T) is:

[0091]

[0092] Then in the (T, 2T) period, Fan in fan ij The actual correction factor k ij (T) is:

[0093] k ij (T) = λ 11 +λ m1 k m1 (T)+λ 1n k 1n (T)+λ mn k mn (T)

[0094] λ 11 +λ m1 +λ 1n +λ mn =1

[0095]

[0096] In summary, fan a ij The actual correction factor k ij (T) is:

[0097]

[0098] According to the above formula, the correction coefficient k of the fan calculated wind speed in each time period is obtained. ij , and calculate the effective wind speed V of each fan ij :

[0099] V ij =k ij *vij

[0100] The effective active power of each wind turbine is derived according to the effective wind speed, and the maximum value of the sum of the active power of the entire field under the current wind conditions is obtained, and the target yaw angle of all wind turbines can be obtained.

[0101] The specific method is:

[0102] When the yaw angle of all wind turbines in the matrix wind farm is When the effective wind speed of each fan is Effective active power and

[0103] Where: θ ij For fan a ij The yaw angle, V ij For fan a ij The effective wind speed, For fan a ij The effective active power, ρ air density, A is the wind wheel swept area, C P is the wind energy utilization coefficient.

[0104] The sum of the active power of the matrix wind farm is P 1 , which is the objective function of control optimization:

[0105]

[0106] Where: P 1 is the sum of the active power of the matrix wind farm, For fan a ij The calculated active power is m, the number of rows of wind turbines in the matrix wind farm, and n is the number of columns of wind turbines in the matrix wind farm.

[0107] The particle swarm optimization algorithm of Monte Carlo simulation is used to optimize the field-level active power objective function of the matrix wind farm and obtain the target yaw angle of each wind turbine in the field under the current wind conditions.

[0108] In this way, the target yaw angles of all wind turbines in the wind farm under different wind speeds and wind conditions are obtained, and the results are stored in the field-level operation decision library. As the wind farm operates, the field-level operation decision library is dynamically updated every T time interval according to the previous T time interval. The dynamic update of the field-level operation decision library is completed through the above steps.

[0109] The field-level control system calls the target yaw angle of the wind turbine in this state, and sends the target yaw angle of the wind turbine to the main control system of each wind turbine; each wind turbine main control system adjusts the yaw system of the wind turbine to the target angle according to the target yaw angle, and maintains operation in this state.

[0110] The present invention has achieved the same or similar results through practical application:

[0111] Taking a 300MW wind farm as an example, a total of 150 2MW wind turbines are installed.

[0112] The three control methods are compared: traditional control, offline field-level control (without considering the dynamic update of the field-level operation decision library), and online field-level control (the control method of the present invention considers the dynamic update of the operation decision library). The wind speed and wind direction statistics are based on the incoming wind speed and wind direction statistics of the sentinel fan.

[0113] The average active power of wind turbines in wind farms at different wind speeds is as follows: Figure 5 It can be seen that under different wind speeds, the offline field-level control system has improved the average active power of wind turbines in wind farms compared with the traditional control system, and the online field-level control system has significantly improved the total active power of wind farms compared with the offline field-level control system and the traditional control system.

[0114] The comparison results of the average active power of wind turbines in wind farms at different wind speeds are as follows: Figure 4 As shown in the figure, traditional control is the worst, offline field-level control is the second, and online field-level control is the best. The increase in the average active power of wind turbines is equivalent to the increase in the total active power of wind turbines in the wind farm.

[0115] The comparison results of the total active power of wind turbines in wind farms under different wind directions are as follows: Figure 5 As shown, traditional control is the worst, offline field-level control is second, and online field-level control is the best.

Claims

1. A matrix wind farm level active power control system based on wind speed measurement, characterized in that: The system includes a precise wind measurement system, a field-level operation decision library, a field-level control system and a wind turbine master control system, including: Precision wind measurement system: used to measure the real-time wind conditions of the front row wind turbines, including the wind speed and wind direction, and search the field-level operation decision database according to the real-time wind direction to match the yaw target angle of each wind turbine in the wind farm corresponding to the real-time wind conditions; Field-level operation decision library: includes the target yaw angle of wind turbines when the field-level active power of wind farms is optimal under various incoming wind conditions, and is continuously updated and iterated; Field-level control system: calls the target yaw angle of the wind turbine under the current incoming wind conditions, and sends the target yaw angle of the wind turbine to the main control system of each wind turbine; Wind turbine main control system: adjusts the wind turbine's yaw system to the target angle according to the yaw target angle, and maintains this state of operation; The precise wind measurement system includes a precise wind measurement device arranged on the sentinel fan, which can obtain the real-time wind conditions of the wind speed and wind direction of the wind in front of the fan; The sentinel wind turbines are wind turbines installed at the four corners of the matrix wind farm, i.e. wind turbine a 11 、a m1 、a 1n 、a mn There are four sentinel fans, each of which is equipped with a precise wind measuring device; According to the incoming wind direction, 0-360° is equally divided into four intervals, corresponding to the four sentinel wind turbines of the matrix wind farm. The first sentinel wind turbine is determined according to the interval where the incoming wind direction is located, and the accurate incoming wind conditions are obtained based on the precise wind measurement system of the first sentinel wind turbine; The initial value calculation process of the field-level operation decision library is as follows: According to the incoming wind speed, based on the single wind turbine wake model, the single wind turbine yaw wake model and the multi-wind turbine yaw wake superposition model, the calculated wind speed of other wind turbines in the wind farm is obtained: Based on the single wind turbine wake model, the wind turbine wake velocity distribution is calculated and the wind turbine wake loss Δv is obtained: Where: Δv is the wind turbine wake loss, V ∞ is the incoming wind speed, C T is the thrust coefficient of the wind turbine, σ is the characteristic width of the wake, r0 is the radius of the wind wheel, and d0 is the diameter of the impeller; Based on the single wind turbine yaw wake model, the distribution of the wind turbine wake in the yaw state is calculated: Where: y d is the center offset of the wind turbine wake, θ is the yaw angle, k is the wake expansion coefficient, and x is the length of the wake core area; Based on the superposition model of multiple wind turbine yaw wakes, the wind turbine wake loss ΔV under the influence of multiple wind turbines is calculated: Where: ΔV is the wind turbine wake loss under the influence of multiple wind turbines, is the characteristic propagation velocity of the wake when the ith wind turbine is operating in isolation, V C is the wake characteristic propagation speed after the operation of multiple wind turbines, u i is the wake velocity of the ith wind turbine when it operates in isolation; Corresponding to fan a ij When the wind speed is v ij : v ij =V ∞ -ΔV ij Where: v ij For fan a ij Calculated wind speed, ΔV ij For fan a ij wake loss; The calculated active power of each wind turbine is derived based on the calculated wind speed, the maximum value of the sum of the active power of the entire field under the current wind conditions is obtained, and the target yaw angle of all wind turbines is obtained, which is specifically: The angle between the yaw angle of the wind turbine and the incoming wind direction is the yaw angle θ. When the yaw angle of the wind turbine changes, the calculated wind speed of the wind turbine will also change. When the yaw angle of all wind turbines in the matrix wind farm is When , the corresponding calculated wind speed of each fan is Calculating active power and Where: θ ij For fan a ij The yaw angle, vij(θ ij ) is fan a ij Calculated wind speed, P ij (θ ij ) is fan a ij The calculation of active power, ρ air density, A is the wind wheel swept area, C P is the wind energy utilization coefficient; The sum of the active power of the matrix wind farm is P, and the objective function of the control optimization is: Where: P is the sum of the active power of the matrix wind farm, P ij (θ ij ) is fan a ij The calculated active power is m, which is the number of rows of wind turbines in the matrix wind farm, and n is the number of columns of wind turbines in the matrix wind farm; The particle swarm optimization algorithm of Monte Carlo simulation is used to optimize the objective function and obtain the target yaw angle of each wind turbine in the field under the current wind conditions. The process of updating and iterating the field-level operation decision library is as follows: In the next running T period, the running results in the previous running T period are used for dynamic update. The specific updating steps are as follows: Define the other three sentinel fans except the first sentinel fan as correction fans, among which the sentinel fan diagonally opposite to the first sentinel fan is the main correction fan, and the other two sentinel fans are slave correction fans. 11 When it is the first sentinel fan, fan a mn Main correction fan, fan a m1 and a 1n In order to calibrate the fan, the accurate wind measurement system of the calibrated fan is used to calibrate the calculated wind speed to obtain the effective wind speed of each fan. m1 Complete (a 11 ,a m1 ) The calculated wind speed correction of this column of fans is performed from the correction fan a 1n Complete (a 11 ,a 1n ) Calculated wind speed correction for this row of fans, main correction fan a mn , from the correction fan a m1 , from the correction fan a 1n Complete the calculated wind speed correction for all other wind turbines in the matrix wind farm; With a 11 Taking the first sentinel fan as an example, fan a mn For main correction fan, fan a m1 and a 1n To calibrate the fan; From the correction fan a m1 Complete (a 11 ,a m1 ) The calculated wind speed correction steps for this column of fans: Read the value of the correction fan a from the calibration fan during the T period m1 The actual measured wind speed and calculate the wind speed v m1 (t), then at time t, the correction fan a m1 Correction factor k m1 (t) is: Calculate all the correction coefficients in the T period and obtain the correction fan a in the T period m1 The average correction factor k m1 (T) is: Then in the period (T, 2T), (a 11 ,a m1 ) Fan a in the fan series i1 The actual correction factor k i1 (T) is: From the correction fan a 1n Complete (a 11 ,a 1n ) The calculated wind speed correction steps for this row of fans: Read the value of the correction fan a from the calibration fan during the T period 1n The actual measured wind speed and calculate the wind speed v 1n (t), then at time t, the correction fan a 1n Correction factor k 1n (t) is: Calculate all the correction coefficients in the T period and obtain the correction fan a in the T period 1n The average correction factor k 1n (T) is: Then in the period (T, 2T), (a 11 ,a 1n ) Air blower in the row 1j The actual correction factor k 1j (T) is: Main correction fan a mn , from the correction fan a m1 , from the correction fan a 1n Finish Correction steps for calculated wind speed of fan: Read the main correction fan a during the T period mn The actual measured wind speed and calculate the wind speed v mn (t), then at time t the main correction fan a mn Correction factor k mn (t) is: Calculate all correction systems in the T period and obtain the main correction fan a in the T period mn The average correction system k mn (T) is: According to the previous derivation, during the T period, the correction fan a m1 The average correction factor k m1 (T) is: According to the previous derivation, during the T period, the correction fan a 1n The average correction factor k 1n (T) is: Then in the (T, 2T) period, Fan in fan ij The actual correction factor k ij (T) is: k ij (T)=λ 11 +λ m1 k m1 (T)+λ 1n k 1n (T)+λ mn k mn (T) l 11 +λ m1 +λ 1n +λ mn =1 In summary, fan a ij The actual correction factor k ij (T) is: According to the above formula, the correction coefficient k of the fan calculated wind speed in each time period is obtained. ij , and calculate the effective wind speed V of each fan ij : V ij =k ij *v ij The effective active power of each wind turbine is calculated according to the effective wind speed, and the maximum value of the sum of the active power of the entire field under the current wind conditions is obtained, and the target yaw angle of all wind turbines can be obtained; The specific method is: When the yaw angle of all wind turbines in the matrix wind farm is When the effective wind speed of each fan is Effective active power and Where: θ ij For fan a ij The yaw angle, V ij For fan a ij The effective wind speed, For fan a ij The effective active power, ρ air density, A is the wind wheel swept area, C P is the wind energy utilization coefficient; The sum of the active power of the matrix wind farm is P 1 , which is the objective function of control optimization: Where: P 1 is the sum of the active power of the matrix wind farm, For fan a ij The calculated active power is m, which is the number of rows of wind turbines in the matrix wind farm, and n is the number of columns of wind turbines in the matrix wind farm; The particle swarm optimization algorithm of Monte Carlo simulation is used to optimize the field-level active power objective function of the matrix wind farm and obtain the target yaw angle of each wind turbine in the field under the current wind conditions. In this way, the target yaw angles of all wind turbines in the wind farm under different wind speeds and wind conditions are obtained, and the results are stored in the field-level operation decision library. As the wind farm operates, the field-level operation decision library is dynamically updated every T time interval according to the previous T time interval. The dynamic update of the field-level operation decision library is completed through the above steps.

2. The matrix wind farm level active power control system according to claim 1, characterized in that: The specific method of optimizing the objective function using the particle swarm optimization algorithm of Monte Carlo simulation includes the following steps: Step 1: Initialize a particle population: randomly generate m*n particles with position and velocity within the set range, calculate the fitness of each particle as the local optimal fitness, compare the local optimal fitness of m*n particles, select the best one and record it in the global optimal fitness, and record this particle as the global optimal vector; Step 2: Update the weight factor w and learning factors c1, c2: Where: w max 、w min is the maximum and minimum value of the inertia weight factor; f is the current fitness value, f avg and f min They are the average fitness value and minimum fitness value of all particles at present; Where: c 1end and c 1start is the final and initial value of c1, c 2end and c 2start is the final value and initial value of c2, t is the current iteration number, t max is the maximum number of iterations; Step 4: Calculate the objective function to obtain the fitness of the current particle; Step 5: Update the local optimal fitness and the local optimal vector of the body; Step 6: Update the global optimal fitness and the global optimal vector; Step 7: Update the position and velocity of each particle; In the formula, 1≤i≤m*n, m*n is the size of the population, represents the d-dimensional velocity of the i-th particle in the k-th iteration; represents the d-dimensional position of the i-th particle in the k-th iteration; ω represents the inertia weight; c1 and c2 represent the learning factors of the particles, represents the d-dimensional monomer extreme value of particle i in the k-th iteration; It represents the global extreme value of the d-dimensional particle swarm in the k-th iteration; is a random number distributed in the (0,1) interval; Step 8: If the number of iterations reaches the maximum value, stop searching and output the results, otherwise return to step 2 to continue iterative calculation; The objective function is solved by the particle swarm optimization algorithm of Monte Carlo simulation, and the optimal solution of the objective function θ is obtained. ij , thus obtaining the target yaw angle of all wind turbines in the current wind condition; By solving the objective function under different wind conditions, the target yaw angle of the wind turbines in the whole field under various wind conditions can be obtained; According to the above steps, simulate and calculate the initial value of the field-level operation decision library.

Images