Frequency control method for new energy power systems considering power stochasticity

Abstract

This invention discloses a frequency control method for a new energy power system considering power stochasticity, comprising: establishing a stochastic optimal frequency control model for the power system considering wind power participation based on the PDF shape control method; the stochastic optimal frequency control model uses the change in the active power reference value of the wind farm as the control variable, and sets a feedback control strategy for the change in the active power reference value of the wind farm, so that when the system frequency deviates, the active power injected into the AC grid by the wind farm can be changed by adjusting the change in the active power reference value of the wind farm, thereby realizing the active participation of the wind farm in the power system frequency support; the PDF is a probability density function; and the optimal feedback control strategy for wind farm output is obtained by solving the stochastic optimal frequency control model of the power system. This invention can improve the frequency stability of the power system under stochastic power disturbances and ensure that the steady-state frequency deviation under power disturbances does not exceed the safety limit.

Description

Technical Field

[0001] This invention relates to the field of power technology, and more specifically to a frequency control method for a new energy power system that takes into account the randomness of power output. Background Technology

[0002] Existing power system frequency control technologies mainly employ primary frequency regulation based on synchronous generator speed governors and secondary frequency regulation based on synchronous generator speed governors combined with automatic generation control systems. For new energy sources such as wind power, droop control or virtual synchronous machine control are mostly used. Existing methods for solving the dynamic model of frequency control are mainly for deterministic dynamic models. They mainly solve the optimal frequency control problem through rolling optimization and feedback correction of Model Predictive Control (MPC), which can optimize multiple control variables simultaneously.

[0003] Although existing technologies have incorporated the output of new energy sources such as wind power as part of the system frequency support, most of them adopt conventional control strategies, which have poor flexibility and cannot provide effective frequency support for complex scenarios such as random power fluctuations. Although the MPC method can solve the frequency control problem, it will significantly reduce the accuracy of the model solution when the system experiences random load fluctuations. Furthermore, MPC requires solving dynamic optimization problems online in each control cycle, resulting in a large amount of computation. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a frequency control method for new energy power systems that considers the randomness of power, which can improve the frequency stability of the power system under random power disturbances and ensure that the steady-state frequency deviation under power disturbances does not exceed the safety limit range.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] A frequency control method for a new energy power system considering power stochasticity includes:

[0007] A stochastic optimal frequency control model for the power system considering wind power participation is established based on the PDF shape control method. The stochastic optimal frequency control model uses the change in the active power reference value of the wind farm as the control variable, and sets a feedback control strategy for the change in the active power reference value of the wind farm. When the system frequency deviates, the active power injected into the AC grid by the wind farm can be changed by adjusting the change in the active power reference value of the wind farm, so as to realize the active participation of the wind farm in the power system frequency support. The PDF is a probability density function.

[0008] The optimal feedback control strategy for wind farm output is obtained by solving the stochastic optimal frequency control model of the power system.

[0009] Compared with the prior art, the advantages of this invention are as follows:

[0010] This invention sets a feedback control strategy for changes in the active power reference value of wind farms. When power system frequency deviations occur, the active power injected into the AC grid from the wind farm can be altered by adjusting the change in the wind farm's active power reference value. This approach is no longer limited to conventional control strategies such as virtual synchronous machines. Leveraging the high controllability of the wind farm's grid-connected converter, a control law for wind farm active power changes with frequency can be obtained, improving the system's frequency stability under random power disturbances. This enables wind farms to actively participate in power system frequency support. Simultaneously, this invention establishes a stochastic optimal frequency control model for the power system considering wind power participation based on the PDF shape control method. Using the change in the wind farm's active power reference value as the control variable, and by setting a reasonable target PDF and effectively tracking the actual PDF of the steady-state frequency deviation, the steady-state frequency deviation under power disturbances is guaranteed to remain within safe limits. Attached Figure Description

[0011] Figure 1 A flowchart of a frequency control method for a new energy power system that considers power randomness, provided in an embodiment of this application;

[0012] Figure 2 A schematic diagram illustrating the effect of implementing probability density function shape control;

[0013] Figure 3 Flowchart for solving the stochastic optimal frequency control problem;

[0014] Figure 4 A schematic diagram of a model for connecting a wind farm to the IEEE 39-node system;

[0015] Figure 5 A comparison is made between the solution results of the example and the target probability density function;

[0016] Figure 6 Δf before and after frequency control s The probability density function; Detailed Implementation

[0017] Example:

[0018] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0019] See Figure 1 As shown, the frequency control method for new energy power systems that considers power randomness provided in this embodiment includes:

[0020] 110. A stochastic optimal frequency control model for the power system considering wind power participation is established based on the probability density function (PDF) shape control method. The stochastic optimal frequency control model for the power system uses the change in the active power reference value of the wind farm as the control variable, and sets a feedback control strategy for the change in the active power reference value of the wind farm. When the power system frequency deviates, the active power injected into the AC grid by the wind farm can be changed by adjusting the change in the active power reference value of the wind farm, so as to realize the active participation of the wind farm in the power system frequency support.

[0021] Considering the randomness of power output, the system frequency changes from a fixed value to a random variable, making it difficult for traditional deterministic control methods to meet the requirements of stochastic optimal frequency control. Meanwhile, according to relevant regulations, the rated frequency of my country's power system is f... N =50Hz, with an allowable frequency deviation range of ±0.2 to ±0.5Hz. In other words, for a power system, as long as the frequency deviation does not exceed the allowable range, the system's frequency safety can be guaranteed. The PDF (Power Frequency Deviation) can contain all the random information in a random process, and its shape can intuitively reflect the probability distribution of the random process. It is superior for state variable control problems that only need to be limited to the allowable range. Therefore, in this step, the PDF shape control method is used to describe the frequency control problem considering power randomness. The PDF of the system's steady-state frequency deviation is used as the control objective. By requiring the target PDF of the steady-state frequency deviation to satisfy a chance constraint at a certain confidence level, the actual PDF of the steady-state frequency deviation is controlled within the allowable safe operating range, thereby ensuring the system's frequency safety under random power disturbances. The steady-state frequency deviation Δf... s The effect of PDF shape control is illustrated in the following diagram. Figure 2 As shown in the figure, it can be seen that when the system experiences random power disturbances, the steady-state frequency deviation Δf using PDF shape control... s The probability of exceeding ±0.2Hz is significantly lower than the probability without control, thus reducing the risk of system frequency instability. Therefore, a PDF shape control method is adopted, by setting a reasonable target PDF and adjusting Δf... s Effective tracking of actual PDFs can make Δf s The frequency of the system is kept within safe limits to ensure frequency safety under random power fluctuations.

[0022] 120. Solve the stochastic optimal frequency control model of the power system to obtain the optimal feedback control strategy for the wind farm output.

[0023] Therefore, this method sets a feedback control strategy for the change in the active power reference value of the wind farm. When the power system frequency deviates, it can change the active power injected into the AC grid by adjusting the change in the active power reference value of the wind farm. It is no longer limited to conventional control strategies such as virtual synchronous machines. It can obtain the control law of wind farm active power changing with frequency based on the high controllability of the wind farm grid-connected interface converter, which can improve the frequency stability of the system under random power disturbances. It realizes the active participation of wind farm in power system frequency support. At the same time, this method establishes a stochastic optimal frequency control model of the power system considering wind power participation based on the PDF shape control method. The change in the active power reference value of the wind farm is used as the control variable. By setting a reasonable target PDF and effectively tracking the actual PDF of steady-state frequency deviation, the steady-state frequency deviation under power disturbances is guaranteed not to exceed the safety limit.

[0024] In one specific embodiment, step 120 above includes:

[0025] The stochastic optimal frequency control model of the power system is characterized by the Iton process stochastic differential equation. Based on the FPK (Fokker-Planck-Kolmogorov) equation, the differential equations of the mixed moments of the power system state variables are derived, transforming the stochastic differential equations into deterministic partial differential equations. Then, a stochastic optimal control solution method based on statistical moment approximation is proposed, which transforms the partial differential equations into deterministic algebraic equations, thereby solving the stochastic optimal frequency control model of the power system and obtaining the optimal feedback control strategy for wind farm output.

[0026] Thus, by using the Fokker-Planck-Kolmogorov (FPK) equations to derive the differential equations of the mixed moments of the system state variables, the stochastic differential equations are transformed into deterministic partial differential equations. Then, a stochastic optimal control solution method based on statistical moment approximation is proposed to transform the partial differential equations into deterministic algebraic equations, thereby solving the optimal control model to obtain the control law of wind farm active power variation with frequency that meets the steady-state frequency deviation requirements.

[0027] In one specific embodiment, the change in the reference value of active power of the wind farm is ΔP. w,ref ΔP w,ref The feedback control strategy is expressed as a polynomial of the frequency deviation Δf as follows:

[0028]

[0029] Where g(Δf) represents ΔP w,ref Regarding the functional relationship of Δf, i is the power exponent of Δf in the feedback control strategy polynomial, b iThese are the undetermined coefficients for the feedback control strategy.

[0030] The stochastic optimal frequency control model of the power system uses the steady-state frequency deviation Δf s The objective is to minimize the deviation between the current PDF and the target PDF, as follows:

[0031]

[0032] Where p(Δf) s The steady-state frequency deviation Δf is determined by the system's stochastic dynamic model. s PDF, p d (Δf s ) is a given Δf s The target PDF.

[0033] In one specific embodiment, the constraints of the power system stochastic optimal frequency control model include a system frequency dynamic model containing wind farms:

[0034] The AC power grid, excluding wind farms, can be equivalently represented as a single-unit system frequency response model through parameter equivalence. Therefore, considering the active participation of wind farms in frequency control, the system's frequency dynamic model is as follows:

[0035]

[0036] Where Δf is the system frequency deviation, ΔP t It is the change in the equivalent steam turbine output power, ΔP g It is the position change of the equivalent speed governor, P w P is the active power injected into the system of the wind farm. w,ref Here, ω is the reference value for the active power of the wind farm, H is the equivalent inertia time constant of the power grid, D is the load regulation coefficient, and T is the reference value for the active power of the wind farm. t It is the time constant of the equivalent steam turbine, T g R is the time constant of the equivalent governor, and R is the droop coefficient of the equivalent steam turbine.

[0037] The constraints of the stochastic optimal frequency control model for power systems also include the control dynamic constraints of wind farm converter stations:

[0038] The power from the wind farm is collected and transmitted to the rectifier station of the flexible DC transmission line. After being converted to AC via DC lines and an inverter station, it is then connected to the AC system via AC lines. The converter station employs dq decoupling control based on the phase-locked loop (PLL) principle. Active and reactive power can be controlled by adjusting the d-axis and q-axis currents. Generally, converter stations use a dual-loop control system consisting of an outer power loop and an inner current loop. However, because the inner current loop responds very quickly while the system frequency control has a long timescale, the effect of the inner current loop control is ignored in this model. Therefore, the dynamic model of the converter station control is as follows:

[0039]

[0040] Among them, P w and Q w These are the active and reactive power injected into the AC system by the wind farm, respectively; P w,ref and Q w,ref These are reference values ​​for the active and reactive power of the wind farm, respectively; I gd and I gq These are the current phasors I injected into the AC grid from the wind farm's grid connection point. g The d-axis and q-axis components; x1 and x2 are introduced intermediate variables; K pgd and K igd These are the proportional and integral coefficients for the d-axis outer loop control, respectively; K pgq and K igq These are the proportional and integral coefficients for the q-axis outer loop control, respectively.

[0041] The power output from the converter station will be injected into the AC power grid through AC lines. Ignoring the resistance of the AC lines, the following voltage equation applies:

[0042]

[0043] in, U is the phasor of the AC bus voltage of the converter station. s X is the voltage phasor at the node where the AC line connects to the AC power grid. L It is the reactance of the AC line.

[0044] The voltage equation (5) is transformed by dq decomposition, and the resulting equation is as follows:

[0045]

[0046] Among them, e w and f w These represent the real and imaginary parts of the voltage at the grid connection point of the converter station on the grid side of the wind farm.

[0047] The constraints of the stochastic optimal frequency control model for power systems also include power balance constraints at each node of the AC power grid:

[0048] The power balance constraints for each node are as follows:

[0049]

[0050] Among them, P si and Q si P represents the active and reactive power output of the generator at node i; wi and Q wi P represents the active and reactive power output of wind power at node i; Li and Q Li G represents the active and reactive loads of node i; ij and B ij Let e ​​be the mutual conductance and mutual susceptance between nodes i and j; i and e j f represents the real part of the voltage at nodes i and j, respectively; i and f j Let n be the imaginary parts of the voltages at nodes i and j, respectively; b It represents the total number of nodes in an AC power grid.

[0051] Equations (2)-(4) and (6)-(7) above constitute the stochastic optimal frequency control model of the power system considering wind power participation.

[0052] Since the above-mentioned stochastic optimal frequency control model of the power system contains stochastic differential equation constraints (3) and nonlinear equation constraints in equations (6)-(7), it is difficult to solve directly. Therefore, this embodiment uses the FPK equation to obtain the mixed moment differential equation of the system state variables, transforming the stochastic differential equation into a deterministic partial differential equation, and further proposes a solution method for the stochastic optimal control model based on statistical moment approximation, transforming the partial differential equation constraints into algebraic equation constraints, thereby realizing the effective solution of the stochastic optimal control model.

[0053] Mixed moment differential equations based on the FPK equation:

[0054] Linearizing equations (6)-(7) in the stochastic optimal frequency control model and then representing the algebraic variables with state variables, and substituting them into the differential equations (3)-(4), yields a pure differential equation model. Since the processed model is a pure differential equation model containing stochasticity, it can be uniformly represented using Iton's stochastic differential equations, as shown in equation (8). Among them, the random disturbance ω of the load power is a random variable in the model.

[0055]

[0056] Where X is the state variable vector in the stochastic optimal frequency control model, expressed as in equation (9). f(·) and G(·) are the function vector and function matrix of the state variables.

[0057] X=[Δf,ΔP t ,ΔP g [x1,x2] T (9)

[0058] The FPK equation is one of the classic state variable probability density function evolution equations. It transforms the analysis problem of dynamic systems containing stochasticity into the analysis problem of deterministic partial differential equations (PDEs), and is an important tool for analyzing stochastic differential equations. Let p(X,t) be the joint probability density function of the state variable vector X, then it satisfies the following FPK equation:

[0059]

[0060] Among them, f i (X,t) is the i-th component of the function vector f(X,t), and S is the power spectral density matrix of the Gaussian white noise vector ω. T ) ij It refers to the matrix GSG T The element in the i-th row and j-th column is n, where n is the number of system state variables.

[0061] The stochastic optimal frequency control model established in this embodiment contains multiple stochastic state variables, making the commonly used concept of statistical moments inapplicable. Therefore, the concept of mixed moments is introduced. Mixed moments, also known as joint moments, are used to represent the joint numerical characteristics of two or more random variables. If η and γ are two random variables, then η... k γ l The mathematical expectation E[η k γ l This is called the k+l order mixed moment of η and γ.

[0062] Suppose that the mixing moments of the state variable X can be obtained from its joint probability density function, as follows:

[0063]

[0064] in, Let r = r1 + r2 + r3 + r4 + r5, then It is the r-th order mixture moment of the state variable X. To obtain the mixture moments of each order of the state variable X in the stochastic frequency control model, we can multiply both sides of the equation (10) by h and then integrate over the state variable, that is:

[0065]

[0066] By processing the left side of the equal sign in equation (12), we can obtain:

[0067]

[0068] in, It is the derivative of the r-th order mixed moment of the state variable X.

[0069] Given boundary conditions If true, by processing the right side of the equation (12), we can obtain:

[0070]

[0071] From equations (12), (13), and (14), we can obtain the differential equation expressions for the mixed moments of the state variables of each order, as follows:

[0072]

[0073] By introducing the concept of mixed moments, the mixed moment differential equation (15) of the dynamic model (8) can be obtained using the FPK equation, thereby transforming the random differential equation into a deterministic partial differential equation.

[0074] Methods for solving stochastic optimal control problems:

[0075] For the power system stochastic optimal frequency control model proposed in this embodiment, its stochastic optimal control problem refers to finding the optimal feedback control strategy for wind farm output as frequency changes, so that the steady-state frequency deviation of the system is controlled within the range that meets the requirements for safe operation under the condition of considering the randomness of load power.

[0076] From equation (8), we can see that f(·) is the undetermined coefficient b of the state variable X and the feedback control strategy. i A polynomial function vector (i = 1, 2, 3) can be represented by the following formula (taking the first component as an example):

[0077] f1(X,t)=a0+(a1+b1)Δf+(a2+b2)Δf 2 +(a3+b3)Δf 3 +a4ΔP t +a5ΔP g +a6x1+a7x2 (16)

[0078] Where a0, a1, ..., a7 are the coefficients of the polynomial function, which are definite constants.

[0079] The expected calculation term in the first term on the right side of equation (15) can be expressed as (taking one of the terms as an example):

[0080]

[0081] The expected calculation term in the second term on the right side of equation (15) can be expressed as (taking one of the terms as an example):

[0082]

[0083] Substituting equations (17) and (18) into equation (15) yields the following system of mixed moment differential equations:

[0084]

[0085] in, It means about The functional relationship is shown in equation (19), which indicates that the differentials of the mixing moments of each order of the system state variables are functions of the mixing moments of each order.

[0086] Since there exists a form of Δf in f(·) 3 The non-first-order terms may lead to higher-order mixed moments above r in the solved mixed moment differential equation system, making equation (19) a non-closed mixed moment differential equation system. To effectively solve the stochastic optimal control problem, higher-order mixed moments of the objective probability density function can be used to replace the higher-order mixed moments in equation (19). When the stochastic dynamic system reaches steady state, the derivatives of the mixed moments of each order of the state variables are zero, and equation (19) can be rewritten as the following algebraic equation system:

[0087]

[0088] In probability theory, the frequency deviation (PDF) of a random variable is equivalent to its moments of all orders. Therefore, when the mixed moments of the state variables can approximate the mixed moments of the target, then the PDF of the state variables can also approximate the PDF of the target. However, in the stochastic optimal frequency control model of this invention, the optimal control problem only concerns the steady-state frequency deviation Δf. s The PDF does not concern itself with the joint distribution function of all state variables. Therefore, solving the optimal control problem only requires statistical moments m. 10000 ,m 20000 ,m 30000 ,...,m r0000 Therefore, the optimization control objective (2) of the model can be rewritten as the steady-state frequency deviation Δf. s The deviations of each order of statistical moments from the objective value are minimized as follows:

[0089]

[0090] Where, m dk It is the steady-state frequency deviation Δf s The k-th order origin moment of the target PDF. By solving the optimization model composed of equations (20)-(21), the undetermined coefficients b of the optimal feedback control strategy for wind farm output can be obtained. i This enables wind farms to actively participate in system frequency support control.

[0091] In summary, the mixed moment differential equations of state variables of various orders can be derived from the FPK equations. Then, the multidimensional PDF shape control method with statistical moment approximation can be used to transform the originally difficult-to-solve optimization models (2)-(4), (6)-(7) containing stochastic differential equations into deterministic algebraic optimization models (20)-(21), thereby realizing the effective solution of the stochastic optimal frequency control model of the system. The solution flowchart is as follows. Figure 3 As shown.

[0092] The following application scenario example will be used to further verify and illustrate this method:

[0093] Using a large wind farm connected to the IEEE 39-bus system via a flexible DC transmission line as an example, the effectiveness of the proposed stochastic optimal frequency control model and solution method for the power system is verified. The hardware environment of the test system for the example is an Intel(R) Xeon(R) E3-1270 CPU@3.60GHz, 32GB of RAM, and a Win10 64-bit operating system, programmed in GAMS win6424.5.6 software.

[0094] After the power from the wind farm is collected, it is injected into the IEEE 39-bus system via flexible DC transmission lines and AC lines. A schematic diagram of the example model is shown below. Figure 4As shown. The power reference value of the model is 100MVA, and the frequency reference value is 50Hz. Subsequent electrical parameters are expressed in per-unit values. The IEEE 39-bus system can be equivalently represented as a single-unit system frequency response model consisting of a governor and a non-reheat turbine through parameter equivalence. The equivalent inertia time constant of the power grid is 9.9385s, the equivalent load regulation coefficient is 10, the equivalent governor time constant is 0.1s, the equivalent turbine time constant is 0.5s, and the equivalent unit droop coefficient is 0.04. After the wind farm power is collected, it is injected into the AC grid through a flexible DC transmission line and an AC line, and then through 3 nodes. The node at the converter station grid connection is taken as the 40th node of the model. The AC line reactance is 0.024, the active power reference value of the wind farm is 5, and the reactive power reference value is 0. The feedback control strategy for the active power output of the wind farm adopts a third-order polynomial form. It is assumed that the maximum disturbance value of the random load disturbance in the model is ±0.2. The example uses ±0.004 as the upper and lower limits of the safe range for the steady-state frequency deviation of the system, and requires a confidence level of 0.99. Therefore, the steady-state frequency deviation Δf of the system is set. s The target probability density function follows a Gaussian distribution with a mean of 0 and a standard deviation of 0.001553.

[0095] Solving the stochastic optimal control model (20)-(21) of the example, we obtain ΔP. w,ref The results of the feedback control strategy are shown in Table 1.

[0096] Table 1 shows the coefficient values ​​of the feedback control strategy obtained from the example solution.

[0097]

[0098] The Monte Carlo method was used to sample and statistically analyze the solved model. 20,000 samples were randomly selected from the normal distribution of the load disturbance. The steady-state frequency deviation of each sample was obtained through time-domain simulation calculations of models (3)-(4) and (6)-(7), thus yielding the statistical results of the steady-state frequency deviation. Steady-state frequency deviation Δf s The comparison between the sampling inspection results and the target probability density function is as follows: Figure 5 As shown in Table 2, the statistical characteristics of the solution results of the example model are presented. Figure 5 As can be seen from Table 2, after implementing the feedback control strategy obtained from the solution for the wind farm, the steady-state frequency deviation Δf s The sampling statistics are very close to the target PDF, and the statistical characteristics of the solution are basically consistent with the target value, with the relative error of the standard deviation being only about 0.71%. Therefore, it can be seen that after implementing the stochastic optimal frequency control proposed in this invention, the steady-state frequency deviation Δf of the system is significantly reduced. s The actual PDF can track the target PDF well, effectively achieving PDF shape control.

[0099] To verify the effectiveness of the method proposed in this embodiment, Monte Carlo sampling statistics were performed on the normal distribution of load random disturbances without implementing frequency control. The statistical results of the steady-state frequency deviation of all samples were then fitted with a normal distribution to obtain the corresponding PDF. This allows for a comparison of the PDF before and after implementing frequency control. Figure 6 As shown in the figure, before implementing frequency control, the maximum steady-state frequency deviation of the system could reach ±0.007, far exceeding the maximum value of the system's safe allowable range. After implementing the proposed stochastic optimal frequency control, the steady-state frequency deviation of the system was effectively controlled within the safe allowable range of [-0.004, 0.004], and the statistical confidence probability of falling within this range reached 99.06%, which is greater than the given confidence level of 0.99. Therefore, the proposed stochastic optimal frequency control method can effectively maintain the safety of the system's steady-state frequency under random load disturbances.

[0100] The calculation results show that the wind power participation in the stochastic optimal frequency control method of the power system proposed in this embodiment can effectively achieve PDF shape control of the steady-state frequency deviation of the system and achieve good control effect, thus verifying the feasibility and effectiveness of the proposed method.

[0101] In summary, the frequency control method for new energy power systems that considers power stochasticity proposed in this invention has the following advantages compared with existing technologies:

[0102] (1) The proposed flexible wind power feedback control strategy is no longer limited to conventional control strategies such as virtual synchronous machines. It can obtain the control law of wind farm active power changing with frequency based on the high controllability of the wind farm grid-connected interface converter, and can improve the frequency stability of the system under random power disturbances.

[0103] (2) An optimal frequency control model for a new energy power system considering power randomness was established based on the probability density function (PDF) shape control method. The PDF of the steady-state frequency deviation of the system was used as the control target. By setting a reasonable target PDF and making the actual PDF of the steady-state frequency deviation effectively tracked, the steady-state frequency deviation under power disturbance is guaranteed not to exceed the safety limit.

[0104] (3) The stochastic optimal frequency control model is characterized by the Iton process stochastic differential equation. Based on the FPK equation, the differential equations of the mixed moments of the system state variables are derived, and the stochastic differential equation is transformed into a deterministic partial differential equation. Then, a stochastic optimal control solution method based on statistical moment approximation is proposed to transform the partial differential equation into a deterministic algebraic equation, thereby realizing the solution of the optimal frequency control model.

[0105] The above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made based on the essence of the content of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A frequency control method for a new energy power system considering power stochasticity, characterized in that, include: A stochastic optimal frequency control model for the power system considering wind power participation is established based on the PDF shape control method. The stochastic optimal frequency control model uses the change in the active power reference value of the wind farm as the control variable, and sets a feedback control strategy for the change in the active power reference value of the wind farm. When the system frequency deviates, the active power injected into the AC grid by the wind farm can be changed by adjusting the change in the active power reference value of the wind farm, so as to realize the active participation of the wind farm in the power system frequency support. The PDF is a probability density function. Solving the stochastic optimal frequency control model of the power system yields the optimal feedback control strategy for wind farm output, including: The stochastic optimal frequency control model of the power system is characterized by the Iton process stochastic differential equation. Based on the FPK equation, the differential equations of the mixed moments of the power system state variables are derived, transforming the stochastic differential equations into deterministic partial differential equations. Then, based on the stochastic optimal control solution method of statistical moment approximation, the partial differential equations are transformed into deterministic algebraic equations, thereby solving the stochastic optimal frequency control model of the power system and obtaining the optimal feedback control strategy for wind farm output. The change in the reference value of active power of the wind farm is ， The feedback control strategy is expressed as the following frequency deviation polynomials: (1) in, represent about The functional relationship, where i is the polynomial of the feedback control strategy. The power exponent, b i These are the undetermined coefficients for the feedback control strategy.

2. The frequency control method for a new energy power system considering power stochasticity as described in claim 1, characterized in that, The power system stochastic optimal frequency control model minimizes the deviation of the PDF of the steady-state frequency deviation Δf s from the target PDF as follows: (2) in, The steady-state frequency deviation is determined by the stochastic optimal frequency control model of the power system. PDF, It is given. The target PDF.

3. The frequency control method for a new energy power system considering power stochasticity as described in claim 2, characterized in that, The constraints of the stochastic optimal frequency control model for the power system include the dynamic frequency model of the system containing wind farms, as follows: (3) in, It is the system frequency deviation. It is the change in the equivalent steam turbine output power. It is the position change of the equivalent speed governor, P w P is the active power injected into the system of the wind farm. w,ref Here, ω is the reference value for the active power of the wind farm, H is the equivalent inertia time constant of the power grid, D is the load regulation coefficient, and T is the reference value for the active power of the wind farm. t Tg is the time constant of the equivalent steam turbine, R is the time constant of the equivalent governor, and R is the droop coefficient of the equivalent steam turbine.

4. The frequency control method for a new energy power system considering power stochasticity as described in claim 3, characterized in that, The constraints of the power system stochastic optimal frequency control model also include dynamic control constraints for wind farm converter stations: (4) Among them, P w and Q w These are the active and reactive power injected into the AC system by the wind farm, respectively; P w,ref and Q w,ref These are reference values ​​for the active and reactive power of the wind farm, respectively; I gd and I gq These are the current phasors I injected into the AC grid from the wind farm's grid connection point. g The d-axis and q-axis components; x1 and x2 are introduced intermediate variables; K pgd and K igd These are the proportional and integral coefficients for the d-axis outer loop control, respectively; K pgq and K igq These are the proportional and integral coefficients of the q-axis outer loop control, respectively. The power output from the converter station will be injected into the AC power grid through AC lines. Ignoring the resistance of the AC lines, the following voltage equation applies: (5) in, U is the phasor of the AC bus voltage of the converter station. s X is the voltage phasor at the node where the AC line connects to the AC power grid. L It is the reactance of the AC line; The voltage equation (5) is transformed by dq decomposition, and the resulting equation is as follows: (6) where e w and f w are the real and imaginary parts of the grid-side converter station grid point voltage of the wind farm, respectively.

5. The frequency control method for a new energy power system considering power stochasticity as described in claim 4, characterized in that, The constraints of the stochastic optimal frequency control model for the power system also include power balance constraints at each node of the AC power grid: (7) Among them, P si and Q si P represents the active and reactive power output of the generator at node i; wi and Q wi P represents the active and reactive power output of wind power at node i; Li and Q Li G represents the active and reactive loads of node i; ij and B ij Let e ​​be the mutual conductance and mutual susceptance between nodes i and j; i and e j f represents the real part of the voltage at nodes i and j, respectively; i and f j Let n be the imaginary parts of the voltages at nodes i and j, respectively; b It represents the total number of nodes in an AC power grid.

6. The frequency control method for a new energy power system considering power stochasticity as described in claim 5, characterized in that, Linearize equations (6)-(7), then represent the algebraic variables as state variables, and substitute them into the differential equations (3)-(4) to obtain a pure differential equation model; use Iton's stochastic differential equation for unified characterization, as shown in equation (8), where the random disturbance ω of the load power is a random variable in the model: (8) Where X is the state variable vector in the stochastic optimal frequency control model, and its expression is as shown in equation (9); and These are function vectors and function matrices relating to state variables; (9) Let p(X, t) be the joint probability density function of the state variable vector X, then it satisfies the following FPK equation: (10) in, It is a function vector The i-th component, S is the power spectral density matrix of the Gaussian white noise vector ω. Refers to matrix The element in the i-th row and j-th column is n, where n is the number of system state variables.

7. The frequency control method for a new energy power system considering power stochasticity as described in claim 6, characterized in that, Suppose that the mixing moments of the state variable X are obtained from its joint probability density function, as follows: (11) in, ;set up ,but It is a state variable X of r First-order mixing moment; Multiplying both sides of equation (10) by h and integrating over the state variable, we get: (12) Processing the left side of the equals sign in equation (12) yields: (13) in, It is the derivative of the r-th order mixed moment of the state variable X; Given boundary conditions If true, processing the right side of the equation (12) yields: (14) The differential equations for the mixed moments of the state variables of each order are obtained from equations (12), (13), and (14) as follows: (15) By introducing the concept of mixed moments, the mixed moment differential equation (15) of the dynamic model (8) can be obtained using the FPK equation, which can transform the differential equation containing stochasticity into a deterministic partial differential equation.

Images