Wind farm virtual inertia optimization allocation method based on improved particle swarm algorithm

Abstract

The application discloses a wind farm virtual inertia optimization distribution method based on an improved particle swarm algorithm, and steps are as follows: firstly, the critical inertia of a power grid for maintaining frequency dynamic stability is solved according to a power grid frequency safety constraint index, and a power grid inertia compensation target is obtained in combination with an actual inertia of the power grid. Then, the minimum system maximum frequency deviation is taken as an optimization target, the wind farm virtual inertia compensation target is taken as an optimization object, the power grid inertia level constraint, the wind turbine virtual inertia support capability constraint and the system frequency stability are taken as constraint conditions, and an optimization distribution model is established. Then, the model is solved by using the improved particle swarm algorithm, and the virtual inertia optimization distribution scheme of each wind farm is obtained. Finally, the correctness and effectiveness of the method are verified in an IEEE-39 node example system. Under the premise of ensuring that the power grid inertia is at a safe level and the wind turbine runs stably, the optimal distribution scheme of the virtual inertia of each wind farm can be obtained, and the algorithm has strong convergence and high solution accuracy.

Description

Technical Field

[0001] This invention belongs to the field of wind power generation technology, and specifically relates to a method for optimizing the allocation of virtual inertia in wind farms based on an improved particle swarm optimization algorithm. Background Technology

[0002] As the penetration rate of new energy sources such as wind power in the power system gradually increases, the equivalent inertia of the system is continuously weakened, the grid's inertia support capacity is significantly reduced, and the system stability faces a great challenge. To address this issue, wind power virtual inertia control strategies are an effective means (for example, Chinese patent document CN111384730A discloses a method for determining wind turbine virtual inertia control parameters), which provides inertia support to the grid by introducing a system frequency differential element into the power generation control. However, how to coordinate the virtual inertia distribution among wind turbines and fully utilize the virtual inertia of wind power to provide inertia support to the grid remains a problem.

[0003] Therefore, the technical problem that this invention needs to solve is: how to improve the particle swarm optimization algorithm and apply it to the virtual inertia optimization allocation, so as to coordinate the virtual inertia support capabilities of each wind farm and maintain the stability of the system frequency. Summary of the Invention

[0004] In view of the technical problems existing in the background technology, the wind farm virtual inertia optimization allocation method based on the improved particle swarm algorithm provided by the present invention can obtain the optimal allocation scheme of virtual inertia for each wind farm under the premise of ensuring that the grid inertia is at a safe level and the wind turbine operates stably. The algorithm has strong convergence ability and high solution accuracy.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] A method for optimizing the allocation of virtual inertia in wind farms based on an improved particle swarm optimization algorithm, comprising the following steps:

[0007] Step 1: Solve for the critical inertia based on system frequency security constraints: Solve for the critical inertia of the power grid based on the maximum frequency deviation constraint and the maximum frequency change rate constraint. H min ;

[0008] Step 2: Solve for the constraints: The constraints include the grid inertia level constraint, the wind turbine virtual inertia support capacity constraint, and the system frequency stability constraint.

[0009] Step 3: Establish the optimized allocation model: Calculate the maximum frequency deviation Δ of the system. f max Minimize the virtual inertia compensation objective of the wind farm as the optimization objective. H wfAs the optimization object, the constraints of grid inertia level, wind turbine virtual inertia support capacity and system frequency stability are used as constraints to establish an optimization allocation model;

[0010] Step 4: Solve the model using an improved particle swarm optimization algorithm: Iteratively solve the model using an improved particle swarm optimization algorithm to obtain the optimal allocation scheme of virtual inertia for each wind farm, coordinate the virtual inertia support capacity of each wind farm, and thus compensate the grid inertia to above the critical inertia.

[0011] Preferably, in step 1, the power system is considered as a whole, and its equivalent rotor motion equation can be expressed as:

[0012] (1)

[0013] In the formula, H Δ represents the equivalent inertia of the power grid. f For power grid frequency deviation, D For the equivalent unit damping, Δ P m Δ represents the increase in total mechanical power. P L This represents the increase in total load power.

[0014] The rate of change of system frequency can be obtained from equation (1):

[0015] (2)

[0016] In the formula, time 0 is the time when the frequency disturbance occurs;

[0017] The maximum frequency deviation of the system after the disturbance is:

[0018] (3)

[0019] In the formula, The equivalent unit damping ratio; ω n The system angular frequency; α These are coefficients generated when deriving the expression for the maximum frequency deviation. t max This is the moment when the maximum frequency deviation occurs; K For generator speed governor gain; Δ P This represents the total active power increment of the system, and this value is related to the grid inertia.

[0020] dΔ is obtained from equations (2) and (3) respectively. f / d t | max and Δ f max-c The power grid inertia under constraints, and then taking the larger of the two values ​​as the critical inertia:

[0021] (4)

[0022] In the formula: H RoCoF and H Δf Corresponding to dΔ f / d t | max and Δ f max-c Critical inertia value under constraints.

[0023] Preferably, the constraint conditions in step 2 are calculated as follows:

[0024] 1) Horizontal constraint of power grid inertia:

[0025] If the power grid includes m Taiwan synchronous generator units and n For a wind farm without a virtual inertial response, the equivalent inertia of the power grid is:

[0026] (5)

[0027] In the formula, H (1) The equivalent inertia of the power grid without virtual inertial response; the subscript 1 indicates that the wind farm is in grid-connected state. H Gi , S Gi The first i The inertia and rated capacity of the synchronous generator set; S WFj For the first j The rated capacity of each wind farm;

[0028] To ensure frequency stability, the actual inertia of the power grid should not be less than the critical inertia. Therefore, during periods when the actual inertia of the power grid is less than the critical inertia, the target for power grid inertia compensation is:

[0029] (6)

[0030] If the power grid includes m Taiwan synchronous generator units and n For a wind farm with virtual inertial response, the amount of inertial compensation that can be provided to the power grid after all wind farms implement virtual inertial control is:

[0031] (7)

[0032] In the formula, Δ H' WF∑ The amount of inertia compensation that can be provided to the power grid after implementing virtual inertial control for all wind farms; H WFj、S WFj The first j The inertia and rated capacity of each wind farm;

[0033] Analysis shows that in order to compensate the grid inertia to the critical inertia, virtual inertia control needs to be implemented in all grid-connected wind farms for compensation. According to equations (6) and (7), we can obtain:

[0034] (8)

[0035] Expanding equation (8), we obtain the constraints for the virtual inertia compensation targets of each wind farm to ensure system frequency security:

[0036] (9)

[0037] 2) Constraints on the virtual inertia support capacity of the wind turbine:

[0038] Given the virtual inertia of the wind turbine:

[0039] (10)

[0040] In the formula, H DFIG =ω 2 nom J DFIG / ( 2P 2 S N () represents the inherent inertial time constant of the wind turbine; ω nom This is the rated angular frequency of the fan;

[0041] According to equation (10), we can obtain H equ The transfer function is:

[0042] (11)

[0043] In the formula, K df , T f , K pT , K iT These are the filtering time constant, inertial control gain, proportional coefficient, and integral coefficient of the speed controller, respectively;

[0044] When control gain K df Initial angular frequency of the blower rotor ωr0 When each of the values ​​is taken to its maximum value, the virtual inertia of the wind turbine at its maximum inertial response capability is obtained:

[0045] (12)

[0046] If wind farm k Equivalent to a single wind turbine unit, the equivalent virtual inertia of a wind farm is the ratio of the unit's total kinetic energy to its total capacity.

[0047] (13)

[0048] In the formula, P , S N , J equ , ω s0 These are the number of pole pairs of the wind turbine, rated capacity, virtual moment of inertia, and initial synchronous angular velocity of the system, respectively.

[0049] The virtual inertia of the wind farm at its maximum inertial response capacity is calculated according to equations (12) and (13). H WF,max ;

[0050] To ensure that wind farms have sufficient virtual inertial response capability for inertia compensation, the virtual inertia of each wind farm should satisfy equation (14) when being allocated:

[0051] (14)

[0052] 3) System frequency stability constraints:

[0053] The frequency deviation expression for the SFR model of a single-machine system frequency response is:

[0054] (15)

[0055] In the formula, ω r The equivalent unit damping angular frequency; φ These are the coefficients that appear when deriving the frequency deviation expression;

[0056] Differentiating equation (15):

[0057] (16)

[0058] in, , T R The generator reheat time constant;

[0059] Because the time of occurrence of the system's maximum frequency deviation corresponds to dΔf ( t ) / d t =0 time, so the solution is:

[0060] (17)

[0061] Substituting equation (17) into equation (15), we obtain the expression for the maximum frequency deviation of the system:

[0062] (18)

[0063] To ensure that the maximum frequency deviation of the system meets safety requirements, the system frequency stability constraint is as follows:

[0064] (19)

[0065] In the formula: Δ f max-c This is the safe value for the maximum frequency deviation of the system.

[0066] Preferably, to ensure the frequency stability of the power grid, this model uses the maximum frequency deviation Δ of the system. f max Minimize as the optimization objective; based on equations (9), (14) and (19), the following wind farm virtual inertia optimization allocation model is derived:

[0067] (20)

[0068] (twenty one)

[0069] (twenty two)

[0070] (twenty three)

[0071] In the formula: Δ f max It is the maximum frequency deviation of the system, used to characterize the stability of the system frequency; H WF,max This represents the upper limit of the virtual inertia of a wind farm.

[0072] Preferably, the particle swarm optimization algorithm is used to solve the optimization problem. It starts with a random solution and obtains the optimal solution after multiple iterations. The operation method is as follows:

[0073] exist D In 3D space, n Each particle forms a population, and the first... i The position and velocity of each particle are x i , v iFirst, calculate the position of each particle. x i The optimal solution for the current individual is obtained by comparing the corresponding fitness values. p i Then, starting from the location of the optimal solution, we can find the global optimal solution. p g During the iteration process, the particle updates itself. x i and v i Find the optimal solution. p i and p g It is also constantly being updated; particles x i and v i The updated formula is as follows:

[0074] (twenty four)

[0075] (25)

[0076] In the formula, v id and x id For the particle's velocity and position; w Inertial weights; d =1,2,…, D ; i =1,2,…, n c1 and c2 are learning factors; r1 and r2 are random numbers in the range [0,1].

[0077] According to equation (9), consider the case where the grid inertia is just compensated to the critical inertia after the wind farm implements virtual inertia control:

[0078] (26)

[0079] If we consider distributing the virtual inertia of wind farms equally, then the target virtual inertia for each wind farm will be equal:

[0080] (27)

[0081] By combining equations (26) and (27), any wind farm can be obtained. k The goal of virtual inertia compensation:

[0082] (28).

[0083] This patent can achieve the following beneficial effects:

[0084] The improved particle swarm optimization algorithm proposed in this invention solves the optimization model, and under the premise of ensuring that the grid inertia is at a safe level and the wind turbines operate stably, the optimal allocation scheme of virtual inertia for each wind farm can be obtained. The algorithm has strong convergence ability and high solution accuracy. On the other hand, by applying the virtual inertia allocation method proposed in this invention, the minimum point of grid frequency drop and RoCoFmax are both controlled within a safe range, the grid has good frequency response characteristics, and the system frequency stability reaches a superior level. Attached Figure Description

[0085] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0086] Figure 1 This is a flowchart of the present invention;

[0087] Figure 2 This invention provides a framework for optimizing virtual inertia allocation.

[0088] Figure 3 The flowchart of the improved particle swarm algorithm of this invention is shown below;

[0089] Figure 4 This is a simulation system diagram of the present invention;

[0090] Figure 5 This invention improves the iterative process of the particle swarm optimization algorithm.

[0091] Figure 6 The figures show the power grid frequency response curves under different virtual inertia allocation schemes according to the present invention.

[0092] Figure 7 The RoCoF curves of the power grid under different virtual inertia allocation schemes according to the present invention are shown. Detailed Implementation

[0093] Example 1:

[0094] Preferred solutions include Figures 1 to 7 As shown, a method for optimizing the allocation of virtual inertia in a wind farm based on an improved particle swarm optimization algorithm includes the following steps:

[0095] Step 1: Solve for the critical inertia based on system frequency security constraints: Solve for the critical inertia of the power grid based on the maximum frequency deviation constraint and the maximum frequency change rate constraint. H min The specific method is as follows:

[0096] Treating the power system as a whole, its equivalent rotor motion equation can be expressed as:

[0097] (1)

[0098] In the formula, HΔ represents the equivalent inertia of the power grid. f For power grid frequency deviation, D For the equivalent unit damping, Δ P m Δ represents the increase in total mechanical power. P L This represents the increase in total load power.

[0099] The rate of change of system frequency can be obtained from equation (1):

[0100] (2)

[0101] In the formula, time 0 is the time when the frequency disturbance occurs;

[0102] The maximum frequency deviation of the system after the disturbance is:

[0103] (3)

[0104] In the formula, The equivalent unit damping ratio; ω n The system angular frequency; α These are coefficients generated when deriving the expression for the maximum frequency deviation. t max This is the moment when the maximum frequency deviation occurs; K For generator speed governor gain; Δ P This represents the total active power increment of the system, and this value is related to the grid inertia.

[0105] dΔ is obtained from equations (2) and (3) respectively. f / d t | max and Δ f max-c The power grid inertia under constraints, and then taking the larger of the two values ​​as the critical inertia:

[0106] (4)

[0107] In the formula: H RoCoF and H Δf Corresponding to dΔ f / d t | max and Δ f max-c Critical inertia value under constraints.

[0108] Step 2: Solve for the constraints: The constraints include grid inertia level constraints, wind turbine virtual inertia support capacity constraints, and system frequency stability constraints; the constraint calculations are as follows:

[0109] 1) Horizontal constraint of power grid inertia:

[0110] If the power grid includes m Taiwan synchronous generator units and n For a wind farm without a virtual inertial response, the equivalent inertia of the power grid is:

[0111] (5)

[0112] In the formula, H (1) The equivalent inertia of the power grid without virtual inertial response; the subscript 1 indicates that the wind farm is in grid-connected state. H Gi , S Gi The first i The inertia and rated capacity of the synchronous generator set; S WFj For the first j The rated capacity of each wind farm;

[0113] To ensure frequency stability, the actual inertia of the power grid should not be less than the critical inertia. Therefore, during periods when the actual inertia of the power grid is less than the critical inertia, the target for power grid inertia compensation is:

[0114] (6)

[0115] If the power grid includes m Taiwan synchronous generator units and n For a wind farm with virtual inertial response, the amount of inertial compensation that can be provided to the power grid after all wind farms implement virtual inertial control is:

[0116] (7)

[0117] In the formula, Δ H' WF∑ The amount of inertia compensation that can be provided to the power grid after implementing virtual inertial control for all wind farms; H WFj 、S WFj The first j The inertia and rated capacity of each wind farm;

[0118] Analysis shows that in order to compensate the grid inertia to the critical inertia, virtual inertia control needs to be implemented in all grid-connected wind farms for compensation. According to equations (6) and (7), we can obtain:

[0119] (8)

[0120] Expanding equation (8), we obtain the constraints for the virtual inertia compensation targets of each wind farm to ensure system frequency security:

[0121] (9)

[0122] 2) Constraints on the virtual inertia support capacity of the wind turbine:

[0123] Given the virtual inertia of the wind turbine:

[0124] (10)

[0125] In the formula, H DFIG =ω 2 nom J DFIG / ( 2P 2 S N () represents the inherent inertial time constant of the wind turbine; ω nom This is the rated angular frequency of the fan;

[0126] According to equation (10), we can obtain H equ The transfer function is:

[0127] (11)

[0128] In the formula, K df , T f , K pT , K iT These are the filtering time constant, inertial control gain, proportional coefficient, and integral coefficient of the speed controller, respectively;

[0129] As can be seen from equation (11), influencing H equ Among the many parameters, H DFIG , ω nom , T f , K pT , K iT For fixed values, ω s0 It remains approximately unchanged at steady state, which determines... H equ The parameter is the control gain. K df Initial angular frequency of the blower rotor ω r0Therefore, during the inertial response phase, two factors determine the magnitude of the wind turbine's inertial response capability: control gain. K df Initial angular frequency of the blower rotor ω r0 ;in ω r0 Determined based on the real-time wind speed of the fan. K df The size is set manually;

[0130] Then when the control gain K df Initial angular frequency of the blower rotor ω r0 When each of the values ​​is taken to its maximum value, the virtual inertia of the wind turbine at its maximum inertial response capability is obtained:

[0131] (12)

[0132] It should be noted that in equation (12), the control gain... K df It is a time-varying value, and the maximum control gain can be obtained through trial and error in simulation. K df,max Based on the predicted maximum wind speed during that period, the maximum initial angular frequency of the wind turbine rotor can be obtained. ω r0,max ;

[0133] If wind farm k Equivalent to a single wind turbine unit, the equivalent virtual inertia of a wind farm is the ratio of the unit's total kinetic energy to its total capacity.

[0134] (13)

[0135] In the formula, P , S N , J equ , ω s0 These are the number of pole pairs of the wind turbine, rated capacity, virtual moment of inertia, and initial synchronous angular velocity of the system, respectively.

[0136] The virtual inertia of the wind farm at its maximum inertial response capacity is calculated according to equations (12) and (13). H WF,max ;

[0137] To ensure that wind farms have sufficient virtual inertial response capability for inertia compensation, the virtual inertia of each wind farm should satisfy equation (14) when being allocated:

[0138] (14)

[0139] 3) System frequency stability constraints:

[0140] The frequency stability of power systems is usually quantitatively described by the maximum frequency deviation and RoCoF. The maximum frequency deviation is a key indicator reflecting the dynamic characteristics of the frequency and a key factor in determining the minimum frequency drop. Therefore, this paper uses the maximum frequency deviation to characterize the frequency stability of the system.

[0141] The frequency deviation expression for the System Frequency Response (SFR) model of a single-machine system is:

[0142] (15)

[0143] In the formula, ω r The equivalent unit damping angular frequency; φ These are the coefficients that appear when deriving the frequency deviation expression;

[0144] Differentiating equation (15):

[0145] (16)

[0146] in, , T R The generator reheat time constant;

[0147] Because the time of occurrence of the system's maximum frequency deviation corresponds to dΔ f ( t ) / d t =0 time, so the solution is:

[0148] (17)

[0149] Substituting equation (17) into equation (15), we obtain the expression for the maximum frequency deviation of the system:

[0150] (18)

[0151] To ensure that the maximum frequency deviation of the system meets safety requirements, the system frequency stability constraint is as follows:

[0152] (19)

[0153] In the formula: Δ f max-c This is the safe value for the maximum frequency deviation of the system.

[0154] Step 3: Establish the optimized allocation model: Calculate the maximum frequency deviation Δ of the system. fmax Minimize the virtual inertia compensation objective of the wind farm as the optimization objective. H wf As the optimization object, the constraints of grid inertia level, wind turbine virtual inertia support capacity and system frequency stability are used as constraints to establish an optimization allocation model;

[0155] To ensure the frequency stability of the power grid, this model uses the maximum frequency deviation Δ of the system. f max Minimize as the optimization objective; based on equations (9), (14) and (19), the following wind farm virtual inertia optimization allocation model is derived:

[0156] (20)

[0157] (twenty one)

[0158] (twenty two)

[0159] (twenty three)

[0160] In the formula: Δ f max It is the maximum frequency deviation of the system, used to characterize the stability of the system frequency; H WF,max This represents the upper limit of the virtual inertia of a wind farm.

[0161] Equation (23) represents the virtual inertia constraint for each wind farm, taking into account the limitations of the wind turbine's own characteristics, such as converter capacity limitations and speed limitations, which can ensure the stable operation of the wind turbine; this model will consider the system Δ f max The optimization objective is set to ensure that the grid inertia is at a safe level and the wind turbine operates stably through constraints. Ultimately, the frequency stability of the system is maximized by optimizing the allocation of the virtual inertia of the wind farm.

[0162] Step 4: Solve the model using an improved particle swarm optimization algorithm: Iteratively solve the model using an improved particle swarm optimization algorithm to obtain the optimal allocation scheme of virtual inertia for each wind farm, coordinate the virtual inertia support capacity of each wind farm, and thus compensate the grid inertia to above the critical inertia.

[0163] Particle Swarm Optimization (PSO) is a classic algorithm for solving optimization problems. It starts with a random solution and obtains the optimal solution after multiple iterations. A detailed description follows: D In 3D space, n Each particle forms a population, and the first... i The position and velocity of each particle are xi , v i First, calculate the position of each particle. x i The optimal solution for the current individual is obtained by comparing the corresponding fitness values. p i Then, starting from the location of the optimal solution, we can find the global optimal solution. p g During the iteration process, the particle updates itself. x i and v i Find the optimal solution. p i and p g It is also constantly being updated; particles x i and v i The updated formula is as follows:

[0164] (twenty four)

[0165] (25)

[0166] In the formula, v id and x id For the particle's velocity and position; w Inertial weights; d =1,2,…, D ; i =1,2,…, n c1 and c2 are learning factors; r1 and r2 are random numbers in the range [0,1].

[0167] The accuracy of the solution obtained by the basic particle swarm optimization algorithm is not necessarily proportional to the number of iterations and the size of the particle swarm. That is, the larger the number of iterations and the larger the size of the particle swarm, the higher the accuracy of the solution is not necessarily obtained. This is directly related to the initialization of particles as random solutions, which has a significant impact on the accuracy of the solution and the number of iterations. To address this issue, this paper considers obtaining high-quality particles by averaging the virtual inertia of the wind farm during the initialization of particle settings, and then iteratively solving for the optimal particles based on these high-quality particles, which can improve the solution performance of the algorithm.

[0168] According to equation (9), consider the case where the grid inertia is just compensated to the critical inertia after the wind farm implements virtual inertia control:

[0169] (26)

[0170] If we consider distributing the virtual inertia of wind farms equally, then the target virtual inertia for each wind farm will be equal:

[0171] (27)

[0172] By combining equations (26) and (27), any wind farm can be obtained. k Virtual inertia compensation target (high-quality initial particle position):

[0173] (28)

[0174] The flowchart of the improved particle swarm optimization algorithm of this invention is as follows: Figure 3 As shown in the figure, the algorithm has the following steps:

[0175] Step (1): Calculate the position of the high-quality initial particle according to equation (28). H WFk Then, the particle swarm is initialized, including the population size, maximum number of iterations, particle position, and particle velocity.

[0176] Step (2): Calculate the particle fitness value (Δ) based on the initialization data. f max The position and velocity of the particles are compared and replaced with the obtained individual optimal value of the particles, and then compared and replaced with the global optimal value of the particle swarm. Finally, the position and velocity of the particles are updated according to equations (24) and (25).

[0177] Step (3): Determine whether the termination condition is met. If Δ f max If the minimum value meets the accuracy requirement or the number of iterations reaches the upper limit, the optimal value is output; otherwise, return to step two to continue iterative calculation.

[0178] The proposed method for optimizing the allocation of virtual inertia in wind farms based on an improved particle swarm optimization algorithm was validated through simulation examples.

[0179] In the Matlab / simulink environment, a system was established. Figure 4 The simulation system incorporates wind farms within the New England 10-unit, 39-node system. The entire simulation system includes 10 synchronous generator units and 3 wind farms.

[0180] The simulation parameters are as follows: Doubly fed wind turbine parameters: Rated voltage V n =575V, rated power P n =1.5MW, stator resistance R s =0.023pu, stator inductance Ls =0.18pu, rotor resistance R r =0.016pu, rotor inductance L r =0.16pu, magnetizing inductance L m =2.9 pu, inherent inertial time constant H DFIG =5.29s, speed controller integral coefficient K i =0.6. Rated angular velocity ω nom =157.08rad / s, rated wind speed V wN =11.7m / s, converter time constant τ =0.02s. Generator parameters are shown in Table 1.

[0181]

[0182] Table 1 Generator Parameters in the Simulation System

[0183] The simulation project includes: 1) Solving the model using an improved particle swarm optimization algorithm, which iterates 31 times to obtain the optimal virtual inertia allocation scheme for each wind farm. This project... Figure 5 Verification; 2) Compare the effects of power grid inertia compensation under different virtual inertia allocation schemes, and obtain the system frequency response curves and RoCoF curves under the proposed scheme and the average allocation scheme. This project passed Figures 6-7 verify.

[0184] Depend on Figure 5 As can be seen, the improved particle swarm optimization algorithm achieves optimal fitness after 31 iterations, with an optimal value of 0.88 Hz. The optimal allocation of virtual inertia for the wind farm after iterative solving is as follows: H WF1 =3.92s, H WF2 =4.61s, H WF3 =3.77s.

[0185] from Figures 6-7It can be seen that the allocation scheme proposed in this invention coordinates the inertia support capabilities of each wind farm, limiting the lowest frequency drop (49.12Hz) to within the 49Hz safety threshold. However, the allocation method in Scheme 2 causes the frequency to drop to 49.08Hz, deepening the lowest frequency point. This is because some wind turbines' inertial response capabilities do not meet the allocated inertia compensation target. When the turbine speed drops below the minimum speed of 0.7 pu, the speed protection module triggers a protection action, and the turbine exits virtual inertial response. Furthermore, the maximum RoCoF value using Scheme 2 is just limited to the -0.5Hz / s safety threshold. However, using the allocation scheme proposed in this invention, by coordinating the inertia support capabilities of each wind farm, the maximum RoCoF value is -0.47Hz / s, achieving optimal system frequency stability and ensuring the safe and stable operation of the power grid.

[0186] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be defined as the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A method for optimizing the allocation of virtual inertia in wind farms based on an improved particle swarm optimization algorithm, characterized in that... Includes the following steps: Step 1: Solve for the critical inertia based on system frequency security constraints: Solve for the critical inertia of the power grid based on the maximum frequency deviation constraint and the maximum frequency change rate constraint. H min ; Step 2: Solve for the constraints: The constraints include the grid inertia level constraint, the wind turbine virtual inertia support capacity constraint, and the system frequency stability constraint. Step 3: Establish the optimized allocation model: Calculate the maximum frequency deviation Δ of the system. f max Minimize the virtual inertia compensation objective of the wind farm as the optimization objective. H wf As the optimization object, the constraints of grid inertia level, wind turbine virtual inertia support capacity and system frequency stability are used as constraints to establish an optimization allocation model; Step 4: Solve the model using an improved particle swarm optimization algorithm: Iteratively solve the model using an improved particle swarm optimization algorithm to obtain the optimal allocation scheme of virtual inertia for each wind farm, coordinate the virtual inertia support capacity of each wind farm, and thus compensate the grid inertia to above the critical inertia.

2. The wind farm virtual inertia optimization allocation method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that: In step 1, the power system is treated as a whole, and its equivalent rotor motion equation can be expressed as: (1)； In the formula, H Δ represents the equivalent inertia of the power grid. f For power grid frequency deviation, D For the equivalent unit damping, Δ P m Δ represents the increase in total mechanical power. P L This represents the increase in total load power. The rate of change of system frequency can be obtained from equation (1): (2)； In the formula, time 0 is the time when the frequency disturbance occurs; The maximum frequency deviation of the system after the disturbance is: (3)； In the formula, The equivalent unit damping ratio; ω n The system angular frequency; α These are coefficients generated when deriving the expression for the maximum frequency deviation. t max This is the moment when the maximum frequency deviation occurs; K For generator speed governor gain; Δ P This represents the total active power increment of the system, and this value is related to the grid inertia. dΔ is obtained from equations (2) and (3) respectively. f / d t | max and Δ f max-c The power grid inertia under constraints, and then taking the larger of the two values ​​as the critical inertia: (4)； In the formula: H RoCoF and H Δf Corresponding to dΔ f / d t | max and Δ f max-c Critical inertia value under constraints, Δ f max-c This is the safe value for the maximum frequency deviation of the system.

3. The wind farm virtual inertia optimization allocation method based on the improved particle swarm optimization algorithm according to claim 2, characterized in that: The constraint conditions for step 2 are calculated as follows: 1) Horizontal constraint of power grid inertia: If the power grid includes m Taiwan synchronous generator units and n For a wind farm without a virtual inertial response, the equivalent inertia of the power grid is: (5)； In the formula, H (1) The equivalent inertia of the power grid without virtual inertial response; the subscript 1 indicates that the wind farm is in grid-connected state. H Gi , S Gi The first i The inertia and rated capacity of the synchronous generator set; S WFj For the first j The rated capacity of each wind farm; To ensure frequency stability, the actual inertia of the power grid should not be less than the critical inertia. Therefore, during periods when the actual inertia of the power grid is less than the critical inertia, the target for power grid inertia compensation is: (6)； If the power grid includes m Taiwan synchronous generator units and n For a wind farm with virtual inertial response, the amount of inertial compensation that can be provided to the power grid after all wind farms implement virtual inertial control is: (7)； In the formula, Δ H' WF∑ The amount of inertia compensation that can be provided to the power grid after implementing virtual inertial control for all wind farms; H WFj 、S WFj The first j The inertia and rated capacity of each wind farm; Analysis shows that in order to compensate the grid inertia to the critical inertia, virtual inertia control needs to be implemented in all grid-connected wind farms for compensation. According to equations (6) and (7), we can obtain: (8)； Expanding equation (8), we obtain the constraints for the virtual inertia compensation targets of each wind farm to ensure system frequency security: (9)； 2) Constraints on the virtual inertia support capacity of the wind turbine: Given the virtual inertia of the wind turbine: (10)； In the formula, H DFIG =ω 2 nom J DFIG / ( 2P 2 S N () represents the inherent inertial time constant of the wind turbine; ω nom This is the rated angular frequency of the fan; J equ Represented as virtual moment of inertia, P Indicates the number of pole pairs of the wind turbine. S N Indicates rated capacity, ω s0 Indicates the initial synchronous angular velocity of the system, ω r0 The initial angular frequency of the fan rotor; According to equation (10), we can obtain H equ The transfer function is: (11)； In the formula, K df , T f , K pT , K iT These are the filtering time constant, inertial control gain, proportional coefficient, and integral coefficient of the speed controller, respectively; When control gain K df Initial angular frequency of the blower rotor ω r0 When each of the values ​​is taken to its maximum value, the virtual inertia of the wind turbine at its maximum inertial response capability is obtained: (12)； If wind farm k Equivalent to a single wind turbine unit, the equivalent virtual inertia of a wind farm is the ratio of the unit's total kinetic energy to its total capacity. (13)； In the formula, P , S N , J equ , ω s0 These are the number of pole pairs of the wind turbine, rated capacity, virtual moment of inertia, and initial synchronous angular velocity of the system, respectively. The virtual inertia of the wind farm at its maximum inertial response capacity is calculated according to equations (12) and (13). H WF,max ; To ensure that wind farms have sufficient virtual inertial response capability for inertia compensation, the virtual inertia of each wind farm should satisfy equation (14) when being allocated: (14)； 3) System frequency stability constraints: The frequency deviation expression for the SFR model of a single-machine system frequency response is: (15)； In the formula, ω r The equivalent unit damping angular frequency; φ These are the coefficients that appear when deriving the frequency deviation expression; Differentiating equation (15): (16)； in, , T R The generator reheat time constant; Because the time of occurrence of the system's maximum frequency deviation corresponds to dΔ f ( t ) / d t =0 time, so the solution is: (17)； Substituting equation (17) into equation (15), we obtain the expression for the maximum frequency deviation of the system: (18)； To ensure that the maximum frequency deviation of the system meets safety requirements, the system frequency stability constraint is as follows: (19)； In the formula: Δ f max-c This is the safe value for the maximum frequency deviation of the system.

4. The wind farm virtual inertia optimization allocation method based on the improved particle swarm optimization algorithm according to claim 3, characterized in that: To ensure the frequency stability of the power grid, this model uses the maximum frequency deviation Δ of the system. f max Minimize as the optimization objective; based on equations (9), (14) and (19), the following wind farm virtual inertia optimization allocation model is derived: (20)； (21)； (22)； (23)； In the formula: Δ f max It is the maximum frequency deviation of the system, used to characterize the stability of the system frequency; H WF,max This represents the upper limit of the virtual inertia of a wind farm.

5. The wind farm virtual inertia optimization allocation method based on the improved particle swarm optimization algorithm according to claim 2, characterized in that: The particle swarm optimization algorithm is used to solve the optimization problem. It starts with random solutions and iterates through multiple iterations to obtain the optimal solution. The operation method is as follows: exist D In 3D space, n Each particle forms a population, and the first... i The position and velocity of each particle are x i , v i First, calculate the position of each particle. x i The optimal solution for the current individual is obtained by comparing the corresponding fitness values. p i Then, starting from the location of the optimal solution, we can find the global optimal solution. p g ;exist During the iteration process, the particles update themselves. x i and v i Find the optimal solution. p i and p g It is also constantly being updated; particles x i and v i The updated formula is as follows: (24)； (25)； In the formula, v id and x id For the particle's velocity and position; w Inertial weights; d =1,2,…, D ; i =1,2,…, n c1 and c2 are learning factors; r1 and r2 are random numbers in the range [0,1]. According to equation (9), consider the case where the grid inertia is just compensated to the critical inertia after the wind farm implements virtual inertia control: (26)； If we consider distributing the virtual inertia of wind farms equally, then the target virtual inertia for each wind farm will be equal: (27)； By combining equations (26) and (27), any wind farm can be obtained. k The goal of virtual inertia compensation: (28)。

Images