System frequency-fan rotating speed collaborative recovery and control method considering wind speed fluctuation

Abstract

The invention discloses a system frequency-fan rotating speed collaborative recovery and control method considering wind speed fluctuation, relates to the technical field of wind power wind speed control, and realizes fan rotating speed recovery control by considering the influence of the wind speed fluctuation on the system frequency and the fan rotating speed. Comprising the following steps: S1, constructing a fan multi-period exit recovery strategy; s2, constructing a frequency-rotating speed collaborative recovery model based on multi-period exit; s3, constructing a system frequency response analytical model; s4, constructing a fan rotating speed analysis model; s5, constructing a fan rotating speed smooth recovery criterion; s6, constructing a frequency lowest point analytical model; and S7, carrying out self-adaptive setting of multi-period exit parameters. On the basis of ensuring smooth recovery of the rotating speed of the fan, the frequency secondary drop is effectively relieved, and compared with an existing quit strategy, the quit strategy has great advantages.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power speed control, and specifically to a system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations. Background Art

[0002] The process of fan participating in frequency regulation can be divided into two stages: the frequency support stage and the withdrawal stage. In the frequency support stage, the fan can release the kinetic energy inside the rotor and adjust the pitch angle to provide active power support through an additional power control link. The secondary frequency drop of the system starts at the transition moment between the frequency support and withdrawal stages, and the sudden decrease in the output electromagnetic power caused by the change of the control strategy is the root cause of the secondary drop phenomenon. Currently, the time for fans to participate in frequency regulation has reached the minute - scale time scale. The limit of the 1 - minute active power change of the wind farm caused by the decrease in wind speed or the wind speed exceeding the cut - out wind speed is 10% of the rated power. At the same time, referring to the actual wind farm data, the active power change caused by the wind speed change can also reach 5% or even 15% of the rated power within one minute. Such a large power fluctuation may lead to the failure of the fan speed recovery. Currently, few studies on fan frequency regulation consider the problem of wind speed fluctuations, and lack the analysis of the influence of wind speed fluctuations on the system frequency and fan speed. Summary of the Invention

[0003] The purpose of the present invention is to provide a system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations, and to realize the fan speed recovery control by considering the influence of wind speed fluctuations on the system frequency and fan speed.

[0004] The technical solution adopted by the present invention to solve its technical problems is as follows: A system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations, comprising the following steps:

[0005] S1. Construct a multi - period withdrawal and recovery strategy for the fan;

[0006] S2. Construct a frequency - speed collaborative recovery model based on multi - period withdrawal;

[0007] S3. Construct a system frequency response analysis model;

[0008] S4. Construct a fan speed analysis model;

[0009] S5. Construct a criterion for the successful recovery of the fan speed;

[0010] S6. Construct a frequency lowest point analysis model;

[0011] S7. Perform adaptive tuning of multi - period withdrawal parameters;

[0012] S7.1. For the droop coefficient of multi - segment withdrawal and the segment time tdi Carry out setting;

[0013] S7.2. Determine the number of segments for the fan to withdraw.

[0014] Furthermore, in step S1, the multi-period withdrawal and restoration strategy divides the frequency modulation process of the fan into four stages:

[0015] The first stage: t ∈ [0, T off , the fan participates in frequency modulation in a certain control mode, and the power increment during frequency modulation where Δf is the frequency deviation, k p is the droop control coefficient, k d is the virtual inertia control coefficient; T off is the end time of frequency support;

[0016] The second stage: t ∈ (T off , T cov , control the fan to gradually withdraw from frequency modulation, and T cov is the start time of constant power;

[0017] The third stage: t ∈ (T cov , T mpp , the fan keeps the electromagnetic power at time T cov unchanged, and the speed recovers. T mpp is the time when the fan enters the following MPPT curve;

[0018] The fourth stage: t > T mpp , the fan operates following the MPPT curve.

[0019] Furthermore, the droop coefficient is expressed as where represents the droop coefficient in the first stage, represents the droop coefficient in the i-th period of the second stage, i = (2, 3,..., n); t di represents the segmentation point between the i-th segment and the (i + 1)-th segment, i = (1, 2,..., n), and t dn = T cov .

[0020] Furthermore, before the fan operates following the MPPT curve, the electromagnetic power output by the fan through the additional power link is where ΔP eW,cov represents the additional electromagnetic power increment output by the fan at time T cov , Δf(T cov ) represents the system frequency at time T cov , and k p (t) is the expression of the droop coefficient.

[0021] Further, in step S2, a dynamic frequency modulation model considering the fan speed is established by adding a power control link of the wind turbine to the traditional system frequency response SFR model, where ΔP L (t) is the power disturbance, s is the complex variable, F H is the work ratio of the high-pressure cylinder of the steam turbine, T R is the reheater time constant, R is the synchronous machine regulation coefficient, M is the synchronous machine inertia time constant, D is the synchronous machine damping coefficient, ΔP G is the change in the mechanical power of the synchronous machine, ΔP eW is the power change during the frequency modulation of the wind turbine, X(s) is the relationship between the frequency modulation power of the fan and the frequency, Δf is the difference between the actual frequency and the rated frequency f n , α is the wind power penetration rate, ω0 is the initial speed, and Δω is the speed change;

[0022] In the fourth stage, the fan follows the MPPT curve to recover, and the electromagnetic power output by the fan is: The increment of the electromagnetic power input by the fan to the system after the disturbance P ref takes the value of P opt .

[0023] Further, in step S3, a frequency domain analytical model is established, and the initial conditions of the segmentation points are solved by the method of integration by parts to obtain the frequency response analytical model of each time period, and finally the frequency response expression for t ∈ (0, T cov is obtained: where Δf i (t) is the frequency response of the i-th time period, i = (1, 2,..., n); draw the simulation structure diagram corresponding to the system part, list and simplify the state equations, and establish a variable coefficient second-order non-homogeneous linear differential equation for the system frequency: where a1, a2, a3, a4 are coefficients, is the first derivative of the frequency, is the second derivative of the frequency; is the first derivative of the power disturbance. By solving in segments, Equation (8) is simplified into multiple constant coefficient second-order non-homogeneous linear differential equations, and the general solution is obtained by further solving the initial conditions.

[0024] Further, the calculation method of the frequency response analytical model in each stage is as follows:

[0025] (1) The frequency response analytical model in the first stage

[0026] First, solve for t ∈ (0, t d1 (T off)] For the frequency response, the fan participates in frequency regulation in the comprehensive inertia control mode, and the expression is: Among them, is a constant coefficient, is the characteristic root corresponding to the homogeneous equation of Equation (7) in (0, t d1 )], a5 is a coefficient, and X 1 is a variable; Integrate both sides of Equation (9) in the interval [0-, 0+] respectively, and simplify to get:

[0027] And further simplify to obtain the initial conditions: Among them, is the rate of change of the frequency deviation at time 0 + , Δf 1 (0 + ) is the response value of the frequency deviation at time 0 + , and Δf 1 (0 + ) = 0;

[0028] Substitute Δf 1 (0 + ) and into Equation (10), and simplify to obtain the expressions of the two coefficients: a6 is the coefficient after combining system parameters. Substitute it into Equation (9) to obtain the frequency response expression in the first stage:

[0029] (2) Analytical model of frequency response in the second stage

[0030] The analytical solution of the frequency response in the i-th time period (t d(i-1) , t di ) is: Among them, is a constant coefficient, is the characteristic root of the corresponding homogeneous equation of Equation (8) in (t d(i-1) , t di ); X i is a variable; Write the time-domain analytical model at time t = t d1 according to Equation (6), and integrate it in the interval [t d1 -, t d1+ , and simplify to get:

[0031]

[0032] Use the integration by parts method to expand the product of k p (t) and to get: Substitute Equation (17) into Equation (16), and simplify to obtain the initial condition equation:

[0033] Write the time-domain analytical model at \(t = t di \) according to Equation (6), and integrate it in the interval \([t di -, t di+ \) to simplify and obtain: Use the integration by parts method for simplification to obtain the initial conditions for the \(i\)-th time period:

[0034] \(i = 3, \ldots, n\);

[0035] Substitute the initial condition Equation (18) and the initial condition Equation (20) into Equation (15) to obtain the constant coefficients where \(B1 i \), \(B2 i \) are simplification coefficients. Substitute Equation (21) into Equation (15) to obtain the frequency response expression for the \(i\)-th time period:

[0036] Furthermore, in step S4, the rotational speed function is piecewise linearized. In the time period \([T a-1 , T a \), the rotational speed linearly changes and is expressed as where \(\omega a-1 \) represents the rotational speed at \(t = T a-1 \), \(\omega a \) represents the rotational speed at \(t = T a \), \(T0 = 0\), and \(A\) is the total number of segments; in the time period \([T a-1 , T a \), the energy equation of the wind speed fluctuation is satisfied: The electromagnetic power integral term is expressed as: Substitute Equations (2), (14), and (22) into (25) and further simplify to obtain Equation [T a-1 , t d(i+1) interval \(k p \) value takes [t d(i+1) , T a interval \(k p \) value takes Substitute them into Equation (26) respectively, where \(Y1\) is a coefficient; (28), \(A1, A2, A3, A4, A5\) are the simplified coefficients; to obtain the mechanical power integral term: where \(Y2\) is a transcendental function of \(\omega\), \(A6\) is the simplified coefficient, and the exponential integral function is defined as Substitute Equations (25) and (29) into Equation (24) and simplify to obtain: Calculate the corresponding rotational speed ω according to Equation (32) a , and simplify Equation (32) to: H(ω a ) = B (33), where H(ω a ) represents a function of the rotational speed ω a , and B is a function independent of the rotational speed ω a .

[0037] Furthermore, in step S5, when the rotational speed of the fan can be successfully restored, there exists a time T mpp such that ω = ω E , where ω E is the rotational speed on the MPPT curve, that is, in the time interval [T cov , T mpp , the energy equation has a real solution. According to Equation (33), iteratively obtain the rotational speed ω cov at time T Tcov ; according to the maximum power P MPPT formula on the MPPT curve: Obtain: where k opt is the MPPT control coefficient, ω and the power value P MPPT are in one-to-one correspondence, λ opt is the optimal tip speed ratio of the fan in the MPPT region, and the wind power fluctuation function is ΔP b = b·t, where b represents the wind power change rate. Substitute Equations (22) and (29) into Equation (34) to obtain When Equation (37) has a real solution, the discriminant must be greater than or equal to zero.

[0038] Furthermore, in step S6, the lowest frequencies in the four stages during the frequency modulation process are expressed as: where, represents the lowest frequency in the first stage, represents the second lowest frequency drop in the i-th time period of the second stage; according to the frequency response analytical expressions of Equations (14) and (22), obtain its derivative and set it equal to zero to get the time of the lowest frequency as: where, represents the time of the lowest frequency in the first stage, represents the time of the lowest frequency in the i-th time period, i = 2, 3,..., n; a7 is the simplified coefficient after combining system parameters, and Z i is a coefficient;

[0039] Substituting Equation (40) into Equations (14) and (22) gives the expression for the lowest frequency point in the first stage: Substituting Equation (41) into Equations (14) and (22) gives the expression for the lowest frequency point in the \(i\)-th time period of the second stage:

[0040] Furthermore, in step S7.1, given the number of segments \(n\), the lowest point of the second - order frequency drop in the multi - segment exit stage is expressed as: where represents the maximum value of the lowest point of the second - order frequency drop when the number of segments is \(n\), and \(\Delta f\) min n is the lowest point of the second - order frequency drop in the \(n\)-th time period of the second stage, i.e., the multi - segment exit stage; a multi - segment exit parameter optimization model is established, and the objective function includes the second - order frequency drop and the lowest rotational speed, expressed as: where \(\omega\) m represents the lowest rotational speed during the fan frequency modulation process, and \(\mu_1\) and \(\mu_2\) are constants; during the fan frequency modulation process, the lowest rotational speed constraint equation is satisfied: \(\omega>\omega\) min (46), where \(\omega\) min represents the lowest limit value of the fan operating rotational speed. The total frequency modulation time of the fan is defined as the time until the rotational speed returns to the MPPT curve, expressed as \(T\) mpp \(< 80s\) (47). At the same time, satisfying Equation (38), when the second - order frequency drop is the smallest, the lowest - point constraint is satisfied: Adding the lowest - frequency - point constraint equation: \(\varepsilon_1\) is the minimum value of the optimization objective; using the particle swarm intelligence algorithm, with Equation (45) as the objective function and Equations (38), (46), (47), (48), (49) as the constraints of the optimization model, the droop coefficient and the segment time \(t\) di .

[0041] Furthermore, in step S7.2, if the difference between the objective - function value \(n\_F\) for the number of segments \(n\) and the objective - function value \((n - 1)\_F\) for the number of segments \(n - 1\) is less than the threshold \(\varepsilon_2\), then the number of segments selected for the multi - segment exit strategy is \(n - 1\).

[0042] The beneficial effects of the present invention are as follows: Through the analysis of the measured operation data, it is found that during the process of the fan participating in the system frequency modulation, if the wind speed continuously decreases, it may lead to the failure of the fan speed to smoothly return to the MPPT curve, providing a practical basis for the influence mechanism of wind speed fluctuation on the fan frequency modulation performance. A multi-period withdrawal strategy for the fan is proposed to cope with the influence of wind speed fluctuation. A coupled mathematical model among the system frequency, fan speed, and wind speed under fluctuating wind speed conditions is established. On this basis, a criterion for the smooth recovery of the fan speed is constructed to ensure the smooth recovery of the fan speed after frequency modulation. With the criterion for the smooth recovery of the fan speed as a constraint and the suppression of the secondary frequency drop as the goal, a parameter adaptive tuning method for the multi-period withdrawal strategy is proposed. On the basis of ensuring the smooth recovery of the fan speed, the secondary frequency drop is effectively alleviated, showing great advantages compared with the current withdrawal strategy. Description of the Drawings

[0043] Figure 1 It is a fan power-rotor speed curve diagram based on the rotor kinetic energy to support the system frequency;

[0044] Figure 2 It is a curve diagram of power and speed changing with time for the first type of recovery strategy;

[0045] Figure 3 It is a curve diagram of power and speed changing with time for the second type of recovery strategy;

[0046] Figure 4 It is a droop coefficient diagram for the multi-period withdrawal strategy;

[0047] Figure 5 It is a dynamic frequency modulation model diagram considering the fan speed. Detailed Implementation Manner

[0048] The present invention first discusses the influence of wind speed fluctuation on the fan frequency modulation process, and then elaborates in detail on the system frequency-fan speed collaborative recovery and control method considering wind speed fluctuation.

[0049] I. Influence of Wind Speed Fluctuation on Fan Frequency Modulation Process

[0050] The fan captures mechanical power from the air, and the mechanical power Among them, P mW is the converted mechanical power, ρ is the air density, R is the fan rotation radius, v is the ambient wind speed, C p is the wind energy utilization coefficient, λ is the tip speed ratio, λ k is the simplified coefficient, θ is the pitch angle, ω t is the fan speed, ω is the generator rotor speed, and K G is the gearbox gear ratio.

[0051] The wind turbine participates in frequency regulation by adding an incremental signal to the active power reference value. Since the wind turbine is connected to the grid through a converter and its electromagnetic power is controllable, the wind speed fluctuation affects the operation of the wind turbine by influencing the mechanical power captured by the wind turbine. The power-rotor speed curve of the wind turbine based on the frequency of the rotor kinetic energy support system is as Figure 1 shown. Among them, point A is the initial operating point of the wind turbine. The curve AC represents the curve of mechanical power varying with the speed, and the curve D′A represents the maximum power point tracking MPPT curve of the wind turbine. The maximum power P MPPT on the MPPT curve is: where k opt is the MPPT control coefficient; ω corresponds one-to-one with the power value P MPPT , and λ opt is the optimal tip speed ratio of the wind turbine in the MPPT region.

[0052] When the frequency drops, the wind turbine participates in frequency regulation through an additional power control link. Its electromagnetic power increases along the curve ABC, providing power support for the system. At this time, the electromagnetic power is greater than the mechanical power, and the wind turbine speed drops. After point C, the wind turbine exits the frequency support stage and enters the speed recovery stage. The electromagnetic power drops to less than the mechanical power, and the rotor speed begins to recover, and finally returns to the initial operating state. In the traditional wind turbine exit strategy, at the exit moment, the electromagnetic power directly drops to the corresponding MPPT operating point, that is, directly drops from point C to point D′. This process will generate a large power deficit, resulting in a serious secondary frequency drop. The improved wind turbine exit strategy no longer directly drops the operating point from point C to the MPPT operating curve, but alleviates the secondary drop by reducing the power deficit when the wind turbine exits. The improved wind turbine exit strategy is divided into two types: the first type directly drops from point C to point D and runs to point E as the speed recovers; the second is to slowly reduce the electromagnetic power from point C until it runs to point E. Finally, both return to the initial operating state along the MPPT operating curve. The wind speed fluctuation types are summarized into three fluctuation scenarios: the wind speed remains unchanged, the wind speed continuously decreases, and the wind speed continuously increases. The following analyzes the influence of the three wind speed fluctuation scenarios on the wind turbine speed.

[0053] By setting the same frequency support parameter electromagnetic power P eW , the curves of power and speed varying with time of the two recovery strategies under the three wind speed scenarios are obtained, as Figure 2 and Figure 3 shown. The operation of the wind turbine satisfies the rotor motion equation: where ΔP b is the fluctuating power, and J is the inertia time constant of the wind turbine. When the wind speed increases, ΔP b takes a positive value, and when the wind speed decreases, ΔP b takes a negative value. The wind turbine provides frequency support by releasing the rotor kinetic energy, expressed as: ∫(P mW +ΔPb ) - P eW dt = J∫ωdω (5). The following analyzes the first type of recovery strategy.

[0054] (1) Wind speed remains constant scenario

[0055] As Figure 2 shown in Figure a, when the wind speed remains constant, during the process of the wind turbine running to point C, the electromagnetic power P eW is greater than the mechanical power and the rotational speed decreases from ω0 to The area of ABC represents the rotor kinetic energy released by the wind turbine, expressed as: After that, the electromagnetic power decreases from point C to point D and remains P E unchanged. At this time, the mechanical power is greater than the electromagnetic power, the rotational speed gradually recovers, and the mechanical power further increases until ω recovers to the rotational speed ω E corresponding to the MPPT curve, that is, running to point E. The area of CDEE' is equal to the rotor kinetic energy absorbed by the wind turbine, expressed as: E Subsequently, the wind turbine gradually recovers to the rated operating state following the MPPT curve. During this process, the wind turbine goes from point A - B - C - D - E - A, and the rotational speed of the wind turbine recovers smoothly.

[0056] (2) Wind speed continuously decreases scenario

[0057] Figure 2 If the wind speed continuously decreases, as Figure 2 shown in Figure b, the mechanical power considering only the influence of the rotational speed is expressed as and the fluctuating power caused by the wind speed is expressed as ΔP b , and the actual mechanical power is expressed as The wind turbine goes from A - B - C, and the rotational speed decreases from ω0 to The energy change is expressed as: After the wind turbine drops to point D, at first the mechanical power is greater than the electromagnetic power and the rotational speed is in the recovery stage. However, due to the continuous decrease of the wind speed, the mechanical power continues to decrease and gradually decreases to be equal to the electromagnetic power, that is Figure 2 point F in Figure b, and then continues to decrease to be below the electromagnetic power. The area of CDF is the rotor kinetic energy absorbed by the wind turbine, expressed as: where ω F is the rotational speed at point F. During this process, the wind turbine goes from A - B - C - D - F, and the rotational speed recovery of the wind turbine fails.

[0058] As Figure 2 shown in Figure a, the area of ABC minus the area of CDEE' is equal to the rotor kinetic energy released by the rotor when it drops from ω0 to the set rotational speed ω E , expressed as: At Figure 2In Figure b, the area of ABC minus the area of CDF is expressed as: Figure 2 In Figure b, ω F <ω E , more rotational speed is released during the support phase, and insufficient energy is absorbed during the rotational speed recovery phase, resulting in the rotational speed not recovering to P E The rotational speed ω corresponding to the MPPT curve E , the mechanical power then drops below the electromagnetic power, leading to the power P E not intersecting the MPPT curve and the failure of rotational speed recovery.

[0059] (3) Scenario of continuously increasing wind speed

[0060] As can be seen from the foregoing analysis, as long as the rotational speed recovers to P E The corresponding rotational speed ω E , the wind turbine can smoothly recover following the MPPT curve. Figure 2 In Figure c, in the scenario of continuously increasing wind speed, the mechanical power continuously increases. After , the mechanical power will not drop below the electromagnetic power, and the rotor absorbs sufficient energy to recover the rotational speed to ω E , and there is no risk of rotational speed recovery failure with the increase in wind speed.

[0061]

[0062] Similar to the first type of recovery strategy, in the scenario of continuously decreasing wind speed for the second type of recovery strategy, before the rotational speed of the wind turbine recovers to the rotational speed ω E corresponding to the MPPT curve of P E , the mechanical power has already dropped below the electromagnetic power, and the power P E does not intersect the MPPT curve, resulting in the failure of rotational speed recovery. While there is no risk of rotational speed recovery failure in the scenario of continuously increasing wind speed.

[0063] In summary, for the wind turbine rotational speed recovery strategy that improves the secondary frequency dip by reducing the power deficit, in the scenario of continuously decreasing wind speed during the wind turbine frequency modulation period, the mechanical power captured by the wind turbine continuously decreases. If the rotational speed of the wind turbine does not recover to the rotational speed ω E corresponding to the electromagnetic power P on the MPPT curve E before the mechanical power drops below the electromagnetic power, resulting in the power P E not intersecting the MPPT curve, it will further lead to the failure of wind turbine rotational speed recovery.

[0064] II. System frequency - wind turbine rotational speed coordinated recovery and control method considering wind speed fluctuations

[0065] Figure 1In it, when the electromagnetic power of the CD segment drops significantly, it will cause a large secondary frequency drop; however, if the electromagnetic power of the CD segment drops slightly, in the scenario where the wind speed continues to decrease, in the later stage of speed recovery, due to the continuous decrease of mechanical power, it may lead to the non-intersection of the electromagnetic power and the MPPT curve, resulting in the failure of speed recovery and the risk of the wind turbine being cut off. Therefore, in order to slow down the secondary drop and avoid the risk of speed recovery failure caused by wind speed fluctuations, the present invention proposes a multi-stage exit wind turbine recovery strategy, by gradually reducing the droop coefficient of the wind turbine in multiple stages to control the wind turbine to gradually exit frequency modulation.

[0066] S1. Construct a multi-stage exit recovery strategy for the wind turbine

[0067] Divide the frequency modulation process of the wind turbine into the following four stages:

[0068] The first stage: Frequency support stage, t ∈ [0, T off . The wind turbine participates in frequency modulation in a certain control mode to provide frequency support to the system, and ends at the end moment T off of frequency support. Taking the most commonly used synthetic inertia as an example for analysis, the power increment during frequency modulation where f is the frequency, Δf is the frequency deviation, k p is the droop control coefficient, and k d is the virtual inertia control coefficient.

[0069] The second stage: Segmented exit stage, t ∈ (T off , T cov . The wind turbine exits the virtual inertia control, and at the same time, the droop coefficient gradually decreases in segments to control the wind turbine to gradually exit frequency modulation. T cov is the start moment of constant power.

[0070] The third stage: Constant power stage, t ∈ (T cov , T mpp . The wind turbine keeps the electromagnetic power at the moment of T cov unchanged, and the speed recovers until it intersects with the MPPT curve at the moment of T mpp . T mpp is the moment when the wind turbine enters the stage of following the MPPT curve.

[0071] The fourth stage: Following the MPPT curve recovery stage, t > T mpp . The wind turbine operates following the MPPT curve.

[0072] In order to more intuitively represent the change of the droop coefficient during the frequency modulation process of the wind turbine, the droop coefficient is expressed as Figure 4 shown and formula where, represents the droop coefficient in the first stage, It represents the droop coefficient in the i-th time period of the second stage, where i = (2, 3, …, n); t di It represents the segmentation point between the i-th segment and the (i + 1)-th segment, where i = (1, 2, …, n), t dn = T cov . It can be seen from this that the droop coefficient of the fan is divided into n time periods during the entire frequency modulation process, and in the segmented exit stage, the droop coefficient is gradually reduced in n - 1 time periods. In addition, the virtual inertia coefficient of the multi-time period exit strategy is only assigned during the frequency support stage.

[0073] Based on this, before the fan follows the MPPT curve, the electromagnetic power output by the fan through the additional power link is expressed as where ΔP eW,cov represents the increment of the additional electromagnetic power output by the fan at time T cov , Δf(T cov ) represents the system frequency at time T cov , and k p (t) is the expression of the droop coefficient.

[0074] Through the multi-time period exit recovery strategy of the fan, the large one-time drop of electromagnetic power is avoided, and the secondary frequency drop is alleviated; at the same time, with the gradual decrease of the droop coefficient in the later stage, the electromagnetic power also gradually decreases. By reasonably setting parameters, the problem that the mechanical power drops below the electromagnetic power due to the continuous decrease of the wind speed, further leading to the failure of the rotational speed recovery, is solved.

[0075] S2. Construct a frequency - rotational speed collaborative recovery model based on multi-time period exit

[0076] In order to reasonably set the parameters of the multi-segment exit strategy and solve the problem of wind speed fluctuation while slowing down the secondary drop, it is necessary to establish the relationship between the fluctuating power, the fan rotational speed, the system frequency and the multi-segment exit parameters. For this purpose, a fan rotational speed - frequency collaborative recovery model considering wind speed fluctuation is established.

[0077] By adding the power control link of the wind turbine generator set to the traditional system frequency response SFR model, a dynamic frequency modulation model considering the fan rotational speed as shown in Figure 5 is established. Where: ΔP L (t) is the power disturbance; s is the complex variable; F H is the work ratio of the high-pressure cylinder of the steam turbine; T R is the reheater time constant; R is the synchronous machine regulation coefficient; M is the synchronous machine inertia time constant; D is the synchronous machine damping coefficient; ΔP G is the change in the mechanical power of the synchronous machine; ΔP eW is the change in the power during the frequency modulation of the wind turbine generator set; X(s) is the relationship between the frequency modulation power of the fan and the frequency; Δf is the difference between the actual frequency and the rated frequency fn The difference, α is the wind power penetration rate, ω0 is the initial speed, and Δω is the speed change.

[0078] During normal operation, the fan operates at the MPPT point, and the output electromagnetic power is P opt After the disturbance occurs, the electromagnetic power output by the fan in the first three stages is the power reference value P ref plus the increment ΔP eW , and in the fourth stage, the fan follows the MPPT curve to recover. The electromagnetic power output by the fan is: Compared with before the disturbance, the increment of the electromagnetic power input by the fan to the system after the disturbance occurs From Figure 5 It can be seen that by controlling the fan power control link ΔP eW , on the one hand, it affects ΔP add and further affects the frequency change; on the other hand, it affects the electromagnetic power P of the fan eW and further affects the speed change. The fluctuating power ΔP b affects the fan speed change by affecting the mechanical power. Similarly, the value of P ref will not only affect the frequency change but also affect the speed change. The value of P ref is generally taken as P MPPT , P mW or P opt . From equations (1) and (2), it can be seen that both P MPPT and P mW have a non-linear relationship with the speed. If the value of P ref is taken as P MPPT or P mW , it will cause non-linear coupling of ΔP add with the speed, further resulting in non-linear coupling of the frequency and the fan speed, making the frequency response unable to be solved analytically. Therefore, in order to decouple the frequency and the fan speed and ensure that the SFR model can be analyzed, the value of P ref is taken as P opt . Next, the analytical expressions between ΔP eW and the frequency, as well as between ΔP eW , ΔP b and the speed are solved respectively.

[0079] S3. Construct an analytical model of the system frequency response

[0080] First, according to equations (14), (15) and (17), analyze the frequency response under multi-period withdrawal control. In the first and second stages, when t ∈ (0, T cov , since the droop coefficient k p step-decreases at the segmentation point t di , it causes ΔP add to also step-decrease at the segmentation point. ΔPadd Multiple step - downs will result in multiple frequency minimum points in the frequency. In the third stage, when t ∈ (T cov , T mpp , the electromagnetic power at the moment of the second - stage T cov is kept unchanged. At t = T cov , ΔP add will not have a step - change, and no frequency extreme points will be generated in the third stage. The transition from the third stage to the fourth stage is achieved by finding the intersection point of the electromagnetic power and the MPPT curve. So at t = T cmpp , ΔP add will not have a step - change, and no frequency minimum points will be generated either. Meanwhile, in the fourth stage, the wind turbine operates following the MPPT curve, and the electromagnetic power ΔP add gradually increases. Therefore, no frequency minimum points will be generated in the fourth stage either. Thus, the frequency minimum points only exist during the first and second stages. Establish a frequency - domain analytical model and solve the initial conditions at the segmented points by the method of integration by parts to obtain the frequency - response analytical model for each time period, and finally obtain the frequency - response expression for t ∈ (0, T cov . According to Equation (14), the frequency response for t ∈ (0, T cov is (18). Where Δf i (t) is the frequency response of the i - th time period, i=(1, 2, …, n). According to Figure 5 in the system part, draw the corresponding simulation structure diagram, then list and simplify the state - equation group to establish a variable - coefficient second - order non - homogeneous linear differential equation for the system frequency: Where, a1, a2, a3, a4 are coefficients, is the first - order derivative of the frequency, is the second - order derivative of the frequency; is the first - order derivative of the power perturbation. By solving in segments, Equation (20) is simplified into multiple constant - coefficient second - order non - homogeneous linear differential equations, and the general solution is obtained by further solving the initial conditions. Assume that the system has a perturbation ΔP L (t)=u(t)ΔP L , where u(t) represents the unit - step function.

[0081] (1) Frequency - response analytical model in the first stage

[0082] First, solve the frequency response for the frequency - support stage t ∈ (0, t d1 (T off ). The wind turbine participates in frequency regulation in the comprehensive inertia - control mode, and the expression is: Where, is a constant coefficient, is the characteristic root corresponding to the homogeneous equation of Equation (19) in (0, t d1 , a5 is a coefficient, and X 1 is a variable.

[0083] Since the system undergoes a step change at t = 0, the coefficients and in the general solution (21) are required to obtain + two initial conditions Δf 1 (0 + ) and Δf 1 (0 + ) are the response values of the frequency deviation at 0 + moment, is the change rate of the frequency deviation at 0 + moment. Since the frequency is continuous, Δf 1 (0 + ) = Δf 1 (0-) = 0. To solve , integrate both sides of Equation (21) in the interval [0-, 0+], and simplify to get:

[0084] And further simplify to obtain the initial conditions:

[0085] Substitute the two initial conditions into Equation (22) and simplify to obtain the expressions of the two coefficients: a6 is the coefficient after combining system parameters. Substitute it into Equation (21) to obtain the frequency response expression in the first stage:

[0086] (2) Analytical model of frequency response in the second stage

[0087] Solve the frequency response in the second stage, that is, the general solution of the frequency response in the i-th time period. The analytical solution of the frequency response in the i-th time period (t d(i-1) , t di ) is: Among them, is a constant coefficient, is the characteristic root of the corresponding homogeneous equation of Equation (20) in (t d(i-1) , t di ); X i is a variable. To solve , it is necessary to obtain Δf i (t d(i-1)+ ) and Among them,

[0088] Since \(t = t\) d1 At this moment, the virtual inertia coefficient becomes 0, and both the droop coefficient and the virtual inertia coefficient have step changes. While for \(t = t\) di (\(i = 2, 3, \cdots, n - 1\)) only the droop coefficient changes. Therefore, it is necessary to separately solve another initial condition for the frequency response in the second time period and the \(i\)th (\(i = 3, \cdots, n\)) time period.

[0089] Solving the initial condition for the second time period Analyze \(t = t\) d1 at this moment separately. Write the time-domain analytical model at \(t = t\) d1 according to Equation (18). Integrate it in the interval \([t\) d1 -, \(t\) d1+ and simplify to get:

[0090]

[0091] Use the integration by parts method to expand the product of \(k\) p (\(t\)) and to get: Comparing Equation (28) with (29), the integral term of the impulse function contained in Equation (28) can exactly cancel out the integral term obtained after expanding Equation (29). Substitute Equation (29) into Equation (28) and simplify to get the initial condition equation:

[0092] Solving the initial condition for the \(i\)th (\(i = 3, \cdots, n\)) time period Analyze \(t = t\) di (\(i = 2, 3, \cdots, n - 1\)) at this moment separately. Write the time-domain analytical model at \(t = t\) di according to Equation (18), and integrate it in the interval \([t\) di -, \(t\) di+ and simplify to get: Similarly, use the integration by parts method to simplify and get the initial condition for the \(i\)th (\(i = 3, \cdots, n\)) time period:

[0093]

[0094] Substitute the initial condition equation (30) and the initial condition equation (32) into Equation (27) to get the constant coefficients \(B1\) i 、\(B2\) i as simplified coefficients. Substitute Equation (33) into Equation (27) to get the frequency response expression for the \(i\)th time period:

[0095] S4. Construct a fan speed analysis model

[0096] To solve the analytical expression of the speed, the speed function is piecewise linearized, that is, in the time period [T a-1 , T a , the speed linearly changes and is expressed as where ω a-1 represents the speed at t = T a-1 , ω a represents the speed at t = T a , T0 = 0, and A is the total number of segments. In the time period [T a-1 , T a , the energy equation of the wind speed fluctuation is satisfied as follows:

[0097] The speed analysis model is used to judge whether the speed is lower than the minimum limit and whether it can be smoothly restored. Therefore, only the integral in the interval (0, T mpp ) needs to be solved. According to Equation (16), the electromagnetic power integral term is expressed as:

[0098] Substitute Equations (14), (26), and (34) into (37) and further simplify to obtain Equation If [T a-1 , T a is included in one time period of Equation (14), for example, t di ≤ T a-1 < t d(i+1) and T a ≤ t d(i+1) , then substitute the corresponding k p value into Equation (38) to obtain the integral value of the electromagnetic power in the interval [T a-1 , T a ; if [T a-1 , T a straddles two time periods shown in (14), for example, t di ≤ T a-1 < t d(i+1) and t d(i+1) < T a , then divide the interval [T a-1 , T a into two parts: [T a-1 , t d(i+1) and [t d(i+1) , T a . The k a-1 value in the interval [T d(i+1) , t p takes [t d(i+1) , Ta Interval k p Value taken Substitute into Equation (38) for calculation respectively. Among them, Y1 is a coefficient. Simplify Equation (1) to A1, A2, A3, A4, A5 are the coefficients after simplifying Equation (1). Further obtain the mechanical power integral term: Among them, Y2 is a transcendental function of ω, and A6 is the simplified coefficient. The Ei() function is an exponential integral function, defined as: Substitute Equations (37) and (41) into Equation (36). Since Y2 is a transcendental function of ω, ω cannot be directly obtained a The display function of, so it is simplified into an implicit function about ω a Implicit function of: At the known starting point (T a-1 , ω a-1 ), and given the cut-off time T a , calculate the corresponding rotational speed ω according to Equation (44) a . The implicit function formula (44) about ω a is simplified to: H(ω a ) = B(45). Among them, H(ω a ) represents a function about the rotational speed ω a , and B is a function independent of the rotational speed ω a .

[0099] In order to alleviate the secondary drop to the greatest extent and ensure the smooth recovery of the rotational speed under wind speed fluctuations, based on the frequency response analysis model and the fan rotational speed analysis model, construct a criterion for the smooth recovery of the fan rotational speed, and with the goal of minimizing the secondary drop degree, propose the control parameter tuning steps of the multi-segment exit strategy. S5. Construct a criterion for the smooth recovery of the fan rotational speed

[0100] Due to the continuous decrease of the mechanical power, the fan rotational speed cannot recover to the rotational speed on the MPPT curve corresponding to the electromagnetic power, and finally the fan rotational speed recovery fails. The rotational speed on the MPPT curve corresponding to the electromagnetic power is expressed as ω E , and the rotational speed recovery failure means that there is no T mpp moment satisfying ω = ω E . On the contrary, if the rotational speed can be smoothly recovered, that is, there is a T mpp moment satisfying ω = ω E , that is, in the time interval [T cov , T mpp , the energy equation has a real solution. According to Equation (45), iteratively obtain T covThe rotational speed ω at a moment Tcov ; Obtained according to Equation (3): The wind power fluctuation function is ΔP b = b·t, where b represents the wind power change speed. Substitute Equations (34) and (41) into Equation (46) to obtain Regard Equation (48) as a quadratic equation of one variable about T mpp When the equation (48) has real solutions, it is necessary to satisfy the discriminant Greater than or equal to zero.

[0101] To sum up, the condition for the smooth recovery of the fan is the formula: Δ≥0.

[0102] S6. Construct an analytical model for the lowest frequency point

[0103] Under the criterion of smooth recovery of the rotational speed, in order to relieve the secondary frequency drop as much as possible, it is necessary to first obtain the analytical expression of the lowest frequency point. From Figure 5 It can be seen that at the segmented moment, the droop coefficient will step down, and the output electromagnetic power of the fan will also have a step drop, which will cause the system frequency to have a downward process in each time period, that is, there is a lowest frequency point in each time period. Based on this, the lowest frequency points in the four stages of the frequency modulation process are expressed as: Among them, Represents the lowest frequency point in the first stage, Represents the lowest point of the secondary frequency drop in the i-th time period of the second stage.

[0104] The frequency response is a non-zero initial response, and there must be an extreme point. Therefore, the lowest frequency point is obtained by derivative calculation. According to the frequency response analytical expressions of Equations (26) and (34), its derivative is obtained and equal to zero, and the time of the lowest frequency point is obtained as: Among them, Represents the time of the lowest frequency point in the first stage, (i = 2, 3,..., n) represents the time of the lowest frequency point in the i-th time period. a7 is the simplified coefficient after the system parameter combination, and Z i Is the coefficient.

[0105] Substitute Equation (51) into Equations (26) and (34) to obtain the lowest frequency point expression in the first stage: Substitute Equation (52) into Equations (26) and (34) to obtain the lowest frequency point expression in the i-th time period of the second stage: S7. Adaptive tuning of multi-time period exit parameters

[0106] Given the parameters of the comprehensive inertia frequency modulation system, the goal of parameter tuning is to minimize the secondary frequency drop under the constraint of a smooth recovery of the rotational speed. The controllable parameters of the multi-period exit strategy include: the exit time T off (t d1 ) during the frequency support stage, the number of segments n - 1 (n ≥ 2) during the exit stage, and the droop coefficients (i = 2, 3, …, n) and the segment times t di (i = 2, 3, …, n) during the multi-segment exit.

[0107] S7.1. The droop coefficients and the segment times t di are tuned

[0108] Given the number of segments n and the scenario of frequency modulation parameter determination, the lowest point of the secondary frequency drop during the multi-segment exit stage is expressed as: where represents the maximum value of the lowest point of the secondary frequency drop when the number of segments is n. Δf min n is the lowest point of the secondary frequency drop at the nth time period of the second stage, i.e., the segment exit stage.

[0109] An optimization model for multi-segment exit parameters is established, and the objective function includes the secondary frequency drop and the lowest rotational speed, expressed as: where ω m represents the lowest rotational speed during the fan frequency modulation process, and μ1 and μ2 are constants.

[0110] To ensure the safe operation of the fan, during the fan frequency modulation process, the lowest rotational speed constraint is satisfied: ω > ω min (57). where ω min represents the lowest limit value of the fan operating rotational speed. To avoid an overly long fan rotational speed recovery time, a constraint on the total fan frequency modulation time is imposed. The total frequency modulation time is defined as the time until the rotational speed recovers to the MPPT curve, expressed as: T mpp < 80 s (58). At the same time, the successful recovery constraint of the fan rotational speed (49) is satisfied. In addition, analyzing equation (55), when the secondary frequency drop is minimized, the lowest point constraint is satisfied: Considering the large number of tuning parameters, to increase the convergence speed, a constraint equation for the lowest frequency point is added: ε1 is the minimum value of the optimization objective.

[0111] The particle swarm intelligence algorithm is used to solve the optimization model with equation (56) as the objective function and equations (49), (57), (58), (59), and (60) as the constraints. Finally, the droop coefficients and the segment times t di。

[0112] S7.2. Determine the number of segments for the turbine to withdraw.

[0113] The more segments for the turbine to withdraw, the less the electromagnetic power drops each time, the smaller the frequency drop caused by the segments, and the smaller the secondary frequency drop. However, too many segments will lead to excessive control difficulty and too high control accuracy requirements. Therefore, it is also very important to determine the appropriate number of segments. Considering both the frequency modulation effect and control convenience, if the difference between the objective function value n_F for the number of segments n and the objective function value n-1_F for the number of segments n-1 is less than the threshold ε2, then the number of segments selected for the multi-segment withdrawal strategy is n-1. It should be noted that the parameters obtained by tuning are the parameters after the single-machine equivalence at the substation level. In actual application, the wind farm distributes the parameters or power reference values to the wind turbines according to the tuned parameters and the operating conditions of each unit in the wind farm to control the turbines to participate in frequency modulation.

[0114] Through the analysis of the measured operation data, the present invention finds that during the process of the turbine participating in the system frequency modulation, if the wind speed continuously decreases, it may cause the turbine speed to not be able to recover smoothly to the MPPT curve, providing a practical basis for the influence mechanism of wind speed fluctuations on the turbine frequency modulation performance. A multi-period withdrawal strategy for the turbine is proposed to cope with the influence of wind speed fluctuations. A coupled mathematical model among the system frequency, turbine speed and wind speed under fluctuating wind speed conditions is established. On this basis, a criterion for the smooth recovery of the turbine speed is constructed to ensure the smooth recovery of the turbine speed after frequency modulation. With the criterion for the smooth recovery of the turbine speed as the constraint and the suppression of the secondary frequency drop as the goal, a parameter adaptive tuning method for the multi-period withdrawal strategy is proposed. On the basis of ensuring the smooth recovery of the turbine speed, the secondary frequency drop is effectively alleviated, which has great advantages compared with the current withdrawal strategy.

Claims

1. A system frequency - fan speed coordinated restoration and control method considering wind speed fluctuations, characterized in that, It includes the following steps: S1. Construct the multi-period withdrawal and restoration strategy of the fan; S2. Construct the frequency-rotational speed collaborative restoration model based on multi-period withdrawal; S3. Construct the analytical model of the system frequency response; S4. Construct the analytical model of the fan rotational speed; S5. Construct the criterion for the smooth restoration of the fan rotational speed; S6. Construct the analytical model of the lowest frequency point; S7. Perform the adaptive tuning of the multi-period withdrawal parameters; S7.

1. Set the droop coefficient for multi-segment exit and the segment time t di for calibration; S7.

2. Determine the number of segments for the fan withdrawal.

2. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 1, characterized in that, In step S1, the multi-period withdrawal and restoration strategy divides the frequency regulation process of the fan into four stages: The first stage: t ∈ [0, T off , the fan participates in frequency regulation in a certain control mode, and the power increment during frequency regulation where Δf is the frequency deviation, k p is the droop control coefficient, k d is the virtual inertia control coefficient; T off is the end time of frequency support; The second stage: t ∈ (T off , T cov , control the fan to gradually exit the frequency modulation, and T cov is the starting moment when the power is constant; The third stage: t ∈ (T cov , T mpp , the electromagnetic power of the fan remains unchanged at the moment of T cov , the rotational speed recovers, and T mpp is the moment when the fan enters the moment of following the MPPT curve; Fourth stage: t > T mpp , the fan operates following the MPPT curve.

3. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 2, characterized in that Express the droop coefficient as where represents the droop coefficient in the first stage represents the droop coefficient in the i-th period of the second stage, i = (2, 3, …, n); t di represents the segmentation point between the i-th segment and the (i + 1)-th segment, i = (1, 2, …, n), t dn = T cov .

4. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 3, wherein Before the fan follows the MPPT curve, the electromagnetic power output by the fan through the additional power link is where ΔP eW,cov represents the increment of additional electromagnetic power output by the fan at time T cov , Δf(T cov ) represents the system frequency at time T cov , and k p (t) is the expression of the droop coefficient.

5. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 4, characterized in that, In step S2, a dynamic frequency modulation model considering the fan speed is established by adding the power control link of the wind turbine to the traditional system frequency response (SFR) model. Among them, ΔP L (t) is the power disturbance, s is the complex variable, F H is the work ratio of the high-pressure cylinder of the steam turbine, T R is the reheater time constant, R is the synchronous machine regulation coefficient, M is the synchronous machine inertia time constant, D is the synchronous machine damping coefficient, ΔP G is the change in the mechanical power of the synchronous machine, ΔP eW is the power change during the frequency modulation of the wind turbine, X(s) is the relationship between the frequency modulation power of the fan and the frequency, Δf is the difference between the actual frequency and the rated frequency f n , α is the wind power penetration rate, ω0 is the initial speed, and Δω is the speed change; In the fourth stage, the fan follows the MPPT curve to recover, and the electromagnetic power output by the fan is: The increment of electromagnetic power input by the fan to the system after the disturbance P ref Take the value of P opt .

6. The system frequency - fan speed coordinated recovery and control method considering wind speed fluctuations according to claim 5, characterized in that In step S3, a frequency-time domain analysis model is established, and the initial conditions of the segmentation points are solved by the method of integration by parts to obtain the frequency response analysis model for each time period, and finally the frequency response expression for t ∈ (0, T cov : where, Δf i (t) is the frequency response in the i-th period, i = (1, 2, …, n); draw the analog structure diagram corresponding to the system part, list and simplify the state equations, and establish a variable-coefficient second-order non-homogeneous linear differential equation for the system frequency: Among them, a1, a2, a3, a4 are coefficients, is the first derivative of frequency, is the second derivative of frequency; is the first derivative of power perturbation. By solving piecewise, Equation (8) is simplified into multiple second-order non-homogeneous linear differential equations with constant coefficients, and the general solution is further obtained by solving the initial conditions.

7. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 6, characterized in that The calculation methods of the frequency response analytical models for each stage are as follows: (1) The frequency response analytical model for the first stage First, solve for the frequency response of \(t\in(0,t d1 (T off )]. The fan participates in frequency regulation in the comprehensive inertia control mode, and the expression is: Among them, are constant coefficients, is the characteristic root corresponding to the homogeneous equation of Equation (7) in (0, t d1 , a5 is a coefficient, and X 1 is a variable; integrating both sides of Equation (9) in the interval [0-, 0+] respectively and simplifying, we get: And further simplify to obtain the initial conditions: wherein, is the rate of change of the frequency deviation at time 0 + , Δf 1 (0 + ) is the response value of the frequency deviation at time 0 + , and Δf 1 (0 + ) = 0; Substitute Δf 1 (0 + ) and into Equation (10), and simplify to obtain the expressions for the two coefficients: a6 is the coefficient after the combination of system parameters. Substituting it into Equation (9), the frequency response expression for the first stage is obtained: (2) The frequency response analytical model for the second stage The analytical solution of the frequency response during the i-th period (t d(i-1) , t di ) is as follows: Among them, is a constant coefficient, is the characteristic root of the corresponding homogeneous equation of Equation (8) in (t d(i-1) , t di ; X i is a variable; according to Equation (6), write the time-domain analytical model at time t = t d1 , and integrate it in the interval [t d1 -, t d1+ , and simplify to obtain: Integrate by parts for k p (t) and The product expansion is as follows: Substitute Equation (17) into Equation (16) and simplify to obtain the initial condition equation: Write the time-domain analytical model at time t = t according to Equation (6), and integrate it in the interval [t di -, t di to simplify and obtain: di+ Use the integration by parts method to simplify and obtain the initial conditions for the i-th time period: i = 3, …, n; Substitute the initial condition formula (18) and the initial condition formula (20) into formula (15) to obtain the constant coefficient Among them, B1 i and B2 i are reduction coefficients. Substituting Equation (21) into Equation (15) gives the frequency response expression for the i-th time period:

8. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 7, characterized in that, In step S4, the rotational speed function is piecewise linearized. In the time period [T a-1 , T a , the linear change of the rotational speed is expressed as where ω a-1 represents the rotational speed at t = T a-1 , ω a represents the rotational speed at t = T a , T0 = 0, and A is the total number of segments; in the time period [T a-1 , T a , the energy equation of the wind speed fluctuation is satisfied as: The electromagnetic power integral term is expressed as: Substituting equations (2), (14), and (22) into (25) and further simplifying, we get equation [T a-1 , t d(i+1) interval k p value takes [t d(i+1) , T a interval k p value takes and substituting them into equation (26) respectively, where Y1 is a coefficient; A1, A2, A3, A4, and A5 are the simplified coefficients; the mechanical power integral term is obtained as: where Y2 is a transcendental function of ω, A6 is the simplified coefficient, and the exponential integral function is defined as Substituting equations (25) and (29) into equation (24) and simplifying to: Calculating the corresponding rotational speed ω a according to equation (32), and simplifying equation (32) to: H(ω a ) = B(33), where H(ω a ) represents a function of the rotational speed ω a , and B is a function independent of the rotational speed ω a .

9. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 8, characterized in that In step S5, there exists a time T when the fan speed can be successfully restored mpp such that ω = ω E , where ω E is the rotational speed on the MPPT curve, that is, in the time interval [T cov , T mpp , the energy equation has a real solution. According to Equation (33), the rotational speed ω cov at time T Tcov is obtained by iteration; according to the maximum power P MPPT formula on the MPPT curve: it is obtained that: where k opt is the MPPT control coefficient, ω corresponds one-to-one with the power value P MPPT , λ opt is the optimal tip speed ratio of the fan in the MPPT region, and the wind power fluctuation function is ΔP b = b·t, where b represents the wind power change rate. Substituting Equations (22) and (29) into Equation (34), we get When Equation (37) has a real solution, the discriminant should be greater than or equal to zero.

10. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 9, characterized in that, In step S6, the lowest frequency points in the four stages during the frequency regulation process are expressed as: Among them, represents the lowest frequency point in the first stage, represents the lowest point of the second - stage frequency secondary drop at the i - th time period; according to the frequency - response analytical expressions of equations (14) and (22), its derivative is obtained and set equal to zero, and the time of the lowest frequency point is obtained as: Among them, represents the time of the lowest frequency point in the first stage, represents the time of the lowest frequency point in the i-th period, i = 2, 3, …, n; a7 is the simplified coefficient after combining system parameters, and Z i is a coefficient; substituting Equation (40) into Equations (14) and (22) gives the expression of the lowest frequency point in the first stage: Substitute Equation (41) into Equations (14) and (22) to obtain the expression for the lowest frequency point in the i-th period of the second stage:

11. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 10, characterized in that, In step S7.1, given the number of segments n, the lowest point of the second frequency drop in the multi-segment exit stage is expressed as: where Δf″ min represents the maximum value of the lowest point of the second frequency drop when the number of segments is n, and Δf min n is the lowest point of the second frequency drop at the nth time period in the second stage, i.e., the multi-segment exit stage; a multi-segment exit parameter optimization model is established, and the objective function includes the second frequency drop and the lowest speed, which is expressed as: where ω m represents the lowest speed during the fan frequency modulation process, and μ1 and μ2 are constants; during the fan frequency modulation process, the lowest speed constraint equation is satisfied: ω > ω min (46), where ω min represents the minimum limit value of the fan operating speed; the total fan frequency modulation time is defined as the time until the speed returns to the MPPT curve, which is expressed as T mpp <80s(47). At the same time, when Equation (38) is satisfied and the second frequency drop is the smallest, the lowest point constraint is satisfied: Add the lowest point constraint equation of the frequency: ε1 is the minimum value of the optimization objective; the particle swarm intelligence algorithm is used, with Equation (45) as the objective function and Equations (38), (46), (47), (48), and (49) as the constraints of the optimization model. Finally, the droop coefficient with the smallest second frequency drop under the condition of releasing less rotor kinetic energy and the segment time t di .

12. The system frequency - fan speed collaborative recovery and control method considering wind speed fluctuations according to claim 11, characterized in that In step S7.2, if the difference between the objective function value n_F for the number of segments n and the objective function value n-1_F for the number of segments n-1 is less than the threshold ε2, then the number of segments selected for the multi-segment withdrawal strategy is n-1.

Images