A method for configuring optimal virtual inertia support nodes of a wind farm access system

Abstract

This invention relates to the field of power system stability node configuration technology, and particularly to a method for configuring optimal virtual inertia support nodes in a wind farm access system. The method includes: establishing a mathematical model of a fully variable-current wind turbine with integrated virtual inertia support, and setting stability constraints on the corresponding virtual inertia coefficients of each node in the target wind farm; establishing a stability objective function for each node based on the stability constraints, and performing damping ratio sensitivity analysis of virtual inertia support and key modes within the range of the stability constraints; performing node ranking optimization calculations based on the damping ratio sensitivity analysis results, obtaining stability optimization results for each node based on sensitivity ranking, and determining the optimal virtual inertia support node configuration for the target wind farm based on the optimization results. This invention, through sensitivity screening combined with intelligent algorithm optimization, provides the optimal access node for the system, ensuring system stability and providing the optimal economic site selection for the system.

Description

Technical Field

[0001] This invention relates to the field of power system stability node configuration technology, and in particular to a method for configuring optimal virtual inertia support nodes in a wind farm access system. Background Technology

[0002] Wind power, as a clean and renewable energy source, continues to increase its share in the global power system. However, as traditional synchronous turbines are replaced by wind power and other power electronic devices, the system's equivalent inertia decreases. With the increasing penetration rate of wind power, the system's equivalent inertia will decrease, leading to a rise in the rate of frequency fluctuation and a decline in stability. Therefore, reliable virtual inertia support is needed when considering the integration of renewable energy systems into the power grid.

[0003] To compensate for the lack of inertia, virtual inertia control technology modifies the control strategy of wind turbine converters, enabling them to dynamically adjust output power based on the system frequency change rate, simulating the inertial response characteristics of synchronous generator sets. However, existing solutions for supporting virtual inertia systems lack the selection of supporting nodes, leading to oscillations and instability after some nodes are connected, and are also uneconomical. This threatens power system security and increases power system integration costs, necessitating new methods to ensure system stability and reduce integration costs. Summary of the Invention

[0004] The purpose of this invention is to overcome the problems of oscillation instability and lack of economic efficiency caused by the operation of some nodes in the wind farm access system in the prior art. It provides a method for configuring the optimal virtual inertia support node in the wind farm access system. By combining the judgment results of sensitivity screening and intelligent algorithm optimization, the optimal access node of the system is given. The above measures ensure system stability and provide the optimal economic site selection for the system.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] A method for configuring optimal virtual inertia support nodes in a wind farm grid connection system includes:

[0007] A mathematical model of a fully variable-flow wind turbine with integrated virtual inertia support is established, and stability constraints are set for the corresponding virtual inertia coefficients of each node in the target wind farm.

[0008] Based on stability constraints, establish the stability objective function corresponding to each node, and perform virtual inertia support and damping ratio sensitivity analysis of key modes within the range of the stability constraints.

[0009] Based on the damping ratio sensitivity analysis results, node sorting optimization calculations are performed to obtain the stability optimization results of each node based on sensitivity sorting, and the optimal virtual inertia support node configuration of the target wind farm is determined based on the optimization results.

[0010] Optionally, the mathematical model of the fully variable flow fan is:

[0011]

[0012] Where, x p4 Let K be the state variable of the DC link loop, where t is time and K is the value of K. pi4 V is the integral coefficient of the PI controller. pdc V is the capacitor voltage on the DC link side. pdcref J is the reference value for the capacitor voltage on the DC link side. pll Let ω be the inertial time constant of the PLL, s be the integral operator, and ω be the inertial time constant. pll D is the frequency tracking coefficient of the PLL. pll ω is the damping coefficient of the PLL. pllref This is the frequency coefficient of the PLL.

[0013] Optionally, the stability constraint is:

[0014]

[0015] Among them, J PLLi Let D be the inertial time constant of PLL node i. PLLi J is the damping coefficient of PLL node i. PLLimax J PLLimin J PLLi The upper and lower limits of the requirements, D PLLimax D PLLimin D respectively PLLi The upper and lower limits are required.

[0016] Optionally, the stability objective function is:

[0017]

[0018] Among them, DR i DR represents the damping ratio of the updated key modes. iori σ represents the damping ratio of the key mode of the original system; σ and ω represent the real and imaginary parts of the characteristic roots of the key mode, respectively.

[0019] Optionally, performing the damping ratio sensitivity analysis includes:

[0020]

[0021] Among them, S i This provides bias sensitivity to support the key mode damping ratio and virtual inertia after the current node is connected. P represents the critical mode damping ratio of the current access node. pllVIiProvides supporting power for the current virtual inertia, where λi is the key modal eigenvalue, |λi| is the modulus of the key modal eigenvalue, and σ i ω represents the real part of the current key modal eigenvalue. i This represents the imaginary part of the current key modal eigenvalues. The left eigenvector corresponding to the eigenvalue coefficient matrix, u i The right eigenvector of the corresponding eigenvalue coefficient matrix, A SOi Let represent the eigenvalue coefficient matrix corresponding to the current access point, where i is the corresponding access node number, Re represents the decay rate of the key mode, and IM represents the oscillation frequency of the key mode.

[0022] Optionally, based on the damping ratio sensitivity analysis results, node ranking optimization calculations are performed to obtain the stability optimization results of each node based on the sensitivity ranking, including:

[0023] S5.1: Initialize the particle swarm size, dimension, cross-sectional probability, cross-sectional probability, sensitivity factor, and number of iterations;

[0024] S5.2: Execute the cross-sectional algorithm to obtain the offspring particle swarm;

[0025] S5.3: Based on the parameters of the offspring particle swarm, determine whether each node has reached the preset damping ratio threshold. If any node has reached the preset damping ratio threshold, then execute S5.5; if no node has reached the preset damping ratio threshold, then calculate the sensitivity factor of each node, sort them, and then execute S5.4.

[0026] S5.4: Execute the next cross-cutting algorithm on the sorted nodes to obtain the updated offspring particle swarm, and then execute S5.3;

[0027] S5.5: Proceed to the next iteration and determine if the number of iterations is greater than the maximum number of iterations. If yes, proceed to S5.6; otherwise, proceed to S5.3.

[0028] S5.6: End the iteration and output the globally optimal particle after the iteration.

[0029] Alternatively, the method for sorting the sensitivity factors of each node is as follows:

[0030]

[0031] Among them, S i The bias sensitivity provides supporting power for the critical mode damping ratio and virtual inertia after the current node is connected, where i is the corresponding node number, i = 1, ..., n, and n is the number of nodes that the target wind farm can connect to the system. S n =S imax This indicates that the number of nodes required for a given node to compute with higher sensitivity compared to other nodes, S n ≈Sn-1 ≈S n-2 This indicates the number of nodes with the same sensitivity that were extracted.

[0032] Optionally, in S5.4, when performing the next cross-multiplication algorithm on the sorted nodes, this is achieved by increasing the weight of high-sensitivity nodes. The method for increasing the weight of high-sensitivity nodes is as follows:

[0033]

[0034] Among them, MS hc (o,d)o * It is the updated horizontal crossover sub-individual, where η is the weight coefficient, and MS hc (o,d) represents the horizontally crossed sub-individuals before the update, S o S represents the sensitivity of the highly sensitive node o after sorting. j Let be the sensitivity of node j after sorting, and n be the number of nodes that the target wind farm can connect to the system.

[0035] The beneficial effects of this invention are as follows:

[0036] This invention optimizes virtual inertia support modally by performing modal interaction between the wind turbine system and the turbine unit. It determines which node in the system is most suitable for virtual inertia optimization support, sets the same target stability margin for wind turbine connection stability optimization at each node in the system, and uses an intelligent algorithm of sensitivity ranking optimization to find the optimal virtual inertia support node in the system. This can provide stability optimization in actual systems and also provides an economical site selection method. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a flowchart of an optimal virtual inertia support node configuration method for a wind farm access system according to an embodiment of the present invention;

[0039] Figure 2 This is a flowchart of the cross-sectional algorithm based on virtual inertia damping ratio sensitivity node sorting optimization in an embodiment of the present invention.

[0040] Figure 3 This is a block diagram of the network-side control system for adding virtual inertia support to the DC link side, according to an embodiment of the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] This embodiment provides a method for configuring optimal virtual inertia support nodes in a wind farm access system, including:

[0044] A mathematical model of a fully variable-flow wind turbine with integrated virtual inertia support is established, and stability constraints are set for the corresponding virtual inertia coefficients of each node in the target wind farm.

[0045] Based on the stability constraints, establish the stability objective function for each node, and perform virtual inertia support and damping ratio sensitivity analysis of key modes within the range of the stability constraints.

[0046] Based on the damping ratio sensitivity analysis results, node sorting optimization calculations are performed to obtain the stability optimization results of each node based on sensitivity sorting, and the optimal virtual inertia support node configuration of the target wind farm is determined based on the optimization results.

[0047] Specifically, in this embodiment, the stability of the coupled model of the dynamic interaction between the wind turbine system and the turbine unit combination is verified, and modal optimization of the original wind turbine is performed using virtual inertia support. Under the condition that the same stability margin is set for each wind turbine node connected to the system, a method for sensitivity ranking optimization of nodes is proposed to find the optimal virtual inertia support node configuration most suitable for the wind farm's connection to the system. For example... Figure 1 As shown, the specific steps include:

[0048] S1: Establish a mathematical model of a fully variable flow fan with integrated virtual inertia support;

[0049] S2: Power system stability constraints based on setting the corresponding virtual inertia coefficients of each node according to the damping ratio function;

[0050] S3: Establish the stability objective function for each node based on the stability constraints;

[0051] S4: Based on stability constraints, perform sensitivity analysis of virtual inertia support and damping ratio of key modes within the constraint range;

[0052] S5: Based on the virtual inertia damping ratio sensitivity analysis, an intelligent algorithm (vertical and horizontal cross algorithm) based on the virtual inertia damping ratio sensitivity node sorting optimization is adopted.

[0053] S6: The intelligent algorithm calculates the stability optimization results of each node based on sensitivity ranking within the algorithm range;

[0054] S7: Determine the optimal virtual inertia support node configuration for the wind farm based on the optimization results.

[0055] Specifically, the steps for establishing the mathematical model of the fully variable flow wind turbine with integrated virtual inertia support in step S1 are as follows:

[0056] S1.1: Its virtual inertia support additional power and damping support additional power are as follows:

[0057]

[0058] ΔP D =-D s (f-f0)(2);

[0059] In the formula, H S Where f is the system's inertial constant, f0 is the specified frequency, f is the measurement frequency, and D is the system's inertial constant. s It is the damping coefficient of the system.

[0060] S1.2: To achieve the purpose of virtual inertia support, adding equations (1) and (2) yields:

[0061]

[0062] In the formula, ΔP A This represents the total additional virtual inertia control power.

[0063] S1.3: In FCWG, its frequency change tracking can be represented by the frequency change of the phase-locked loop (PLL), so equation (3) can be transformed into:

[0064]

[0065] In the formula, ΔP pllVI It concerns the power supported by the PLL virtual inertia, J pll It is the inertial time constant of the PLL, ω pll It relates to the frequency tracking coefficient of the PLL, ω pllref It is the specified PLL frequency coefficient, D pll is the damping coefficient with respect to the PLL.

[0066] S1.4: The steps for integrating virtual inertia support into the DC link of the full converter wind turbine are as follows:

[0067] The state equation for the DC link capacitor side is as follows:

[0068]

[0069] P out =ΔP pllVI +P outori (6);

[0070] In the formula, P in This represents the active power input to the GSC, P. outori This represents the active power output of the original system GSC, C. p For the DC link side capacitor, V pdc This is the capacitor voltage on the DC link side.

[0071] The state variable equations after integrating virtual inertia coefficients in the DC link loop are as follows:

[0072]

[0073] In the formula, x p4 These are the state variables of the DC link loop. Network-side control with virtual inertia support added to the DC link side is as follows: Figure 3 As shown.

[0074] Specifically, in step S2, the power system stability constraints are set based on the damping ratio function to determine the corresponding virtual inertia coefficients of each node. The stability constraints are as follows:

[0075]

[0076] In the formula, J PLLi D PLLi The upper and lower limits are required to ensure that the damping ratio of the node is not negative.

[0077] Specifically, in step S3, a stability objective function is established for each node based on the stability constraints. The objective function is as follows:

[0078]

[0079] In the formula, DR i DR represents the damping ratio of the updated key modes. iori σ represents the damping ratio of the key mode of the original system; σ and ω represent the real and imaginary parts of the characteristic roots of the key mode.

[0080] Specifically, in step S4, based on stability constraints, a sensitivity analysis of the virtual inertia support and the damping ratio of the key modes is performed within the constraint range. The sensitivity analysis calculation formula is as follows:

[0081]

[0082] In the formula, S i The sensitivity of node i is specifically the bias sensitivity that provides supporting power for the critical mode damping ratio and virtual inertia after the current node is connected. P represents the critical modal damping ratio at the current access point. pllVIi Provides supporting power for the current virtual inertia, where λi is the key modal eigenvalue, |λi| is the modulus of the key modal eigenvalue, and σ i ω represents the real part of the current key modal eigenvalue. i This represents the imaginary part of the current key modal eigenvalues. The left eigenvector corresponding to the eigenvalue coefficient matrix, u i The right eigenvector of the corresponding eigenvalue coefficient matrix, A SOi This is the eigenvalue coefficient matrix corresponding to the current access point, where i refers to the corresponding access point number, Re represents the decay rate of the key mode, and IM represents the oscillation frequency of the key mode.

[0083] Specifically, such as Figure 2 As shown, in step S5, based on the virtual inertia damping ratio sensitivity analysis, an intelligent algorithm (vertical and horizontal cross algorithm) based on virtual inertia damping ratio sensitivity node sorting optimization is adopted. The steps are as follows:

[0084] S5.1: Initialize the particle swarm size, dimension, cross-sectional probability, cross-sectional probability, sensitivity factor, and number of iterations;

[0085] S5.2: The cross-sectional algorithm is used to obtain the offspring particle swarm;

[0086] S5.3: Based on the currently generated particle swarm parameters, determine whether the node has reached the required damping ratio threshold. If yes, execute S5.5; otherwise, calculate the sensitivity factor of each node based on the currently generated particle swarm parameters, sort the sensitivity factors of each node, and execute S5.4.

[0087] S5.4: Perform the next cross-cutting operation on the sorted nodes to obtain the updated offspring particle swarm, then execute S5.3.

[0088] S5.5: Proceed to the next iteration and determine if the number of iterations is greater than the maximum number of iterations. If yes, proceed to S5.6; otherwise, proceed to S5.3.

[0089] S5.6: End the iteration and output the globally optimal particle after the iteration.

[0090] Specifically, in step S5.3, the formula for calculating the sensitivity ranking is as follows:

[0091]

[0092] In the formula, n refers to the number of nodes that the wind farm can connect to the system, and S n =S imax This refers to the requirement that the number of nodes acquired must be such that the computational sensitivity of that node is higher than that of other nodes, where S... n ≈S n-1 ≈S n-2 This refers to the number of nodes with the same sensitivity that were extracted. The basis for extracting these nodes is that their sensitivities are similar.

[0093] Specifically, the steps following the sensitivity ranking in step S5.3 are as follows:

[0094] S5.3.1: Calculate the sensitivity S of each node. i ;

[0095] S5.3.2: Filtering high-sensitivity nodes;

[0096] S5.3.3: Extract nodes with roughly the same high sensitivity;

[0097] Specifically, in step S5.4, the next cross-cutting algorithm is performed on the sorted nodes to obtain the updated offspring particle swarm. Step S5.3 then proceeds with the following steps:

[0098] Based on the ranking results, high-sensitivity nodes are selected, and their weights are increased to perform cross-crossing coefficients, accelerating local convergence. The formula for increasing the weights of high-sensitivity nodes is as follows:

[0099]

[0100] Among them, MS hc (o,d)o * It is the updated horizontal crossover sub-individual, where η is the weight coefficient, and MS hc (o, d) represents the horizontally crossed sub-individuals before the update, where o is the number of high-sensitivity nodes, and S is the number of nodes. o S represents the sensitivity of the highly sensitive node o after sorting. j Let be the sensitivity of node j after sorting, and n be the number of nodes that the target wind farm can connect to the system.

[0101] This embodiment uses modal interaction between the wind turbine system and the turbine unit to perform virtual inertia support modal optimization. It determines which node in the system is most suitable for virtual inertia optimization support, sets the same target stability margin for wind turbine connection stability optimization for each node in the system, and uses an intelligent algorithm of sensitivity ranking optimization to find the optimal virtual inertia support node in the system. This can provide stability optimization in the actual system and also provide an economical site selection method.

[0102] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for optimal virtual inertia support node configuration of a wind farm integration system, characterized in that, include: A mathematical model of a fully variable-current wind turbine with integrated virtual inertia support is established, and stability constraints are set for the corresponding virtual inertia coefficients of each node in the target wind farm. In the full converter wind turbine model, frequency change tracking can be represented by the frequency change of the phase-locked loop. Virtual inertia support is integrated into the DC link of the full converter wind turbine to establish a mathematical model of the full converter wind turbine with integrated virtual inertia support. The mathematical model for the fully variable flow fan is as follows: ； in, For the state variables of the DC link loop, t For time, The integral coefficient of the PI controller. This is the capacitor voltage on the DC link side. This is a reference value for the capacitor voltage on the DC link side. The inertial time constant of the PLL. s For integration operators, For the frequency tracking coefficient of the PLL, Here is the damping coefficient of the PLL. The frequency coefficient of the PLL; Based on the stability constraints, establish the stability objective function for each node, and perform virtual inertia support and damping ratio sensitivity analysis of key modes within the range of the stability constraints. Based on the damping ratio sensitivity analysis results, node sorting optimization calculations are performed. High-sensitivity nodes are selected based on the sorting results, and their weights are increased. Cross-crossing coefficients are then performed to obtain the stability optimization results of each node based on the sensitivity sorting. Based on the optimization results, the optimal virtual inertia support node configuration for the target wind farm is determined.

2. The method of claim 1, wherein, The stability constraint is as follows: ； in, Let be the inertial time constant of PLL node i. Let i be the damping coefficient of PLL node i. , They are respectively Upper and lower limit requirements, , They are respectively The upper and lower limits are required.

3. The method of claim 1, wherein, The stability objective function is: ； wherein, i represents the damping ratio of the updated key mode, represents the damping ratio of the original system key mode; , respectively represent the real and imaginary parts of the key mode eigenvalue.

4. The method of claim 1, wherein, The damping ratio sensitivity analysis includes: ； ； in, This provides bias sensitivity to support the key mode damping ratio and virtual inertia after the current node is connected. The critical mode damping ratio of the current access node. Provides supporting power for the current virtual inertia. These are key modal eigenvalues. The modulus of the key modal eigenvalues The real part of the current key mode eigenvalue. This represents the imaginary part of the current key modal eigenvalues. The left eigenvector of the corresponding eigenvalue coefficient matrix, The right eigenvector of the corresponding eigenvalue coefficient matrix, Let represent the eigenvalue coefficient matrix corresponding to the current access point, where i is the corresponding access node number, Re represents the decay rate of the key mode, and IM represents the oscillation frequency of the key mode.

5. The method of claim 1, wherein, Based on the damping ratio sensitivity analysis results, node ranking optimization calculations are performed to obtain the stability optimization results of each node based on the sensitivity ranking, including: S5.1: Initialize the particle swarm size, dimension, cross-sectional probability, cross-sectional probability, sensitivity factor, and number of iterations; S5.2: Execute the cross-sectional algorithm to obtain the offspring particle swarm; S5.3: Based on the parameters of the offspring particle swarm, determine whether each node has reached the preset damping ratio threshold. If any node has reached the preset damping ratio threshold, then execute S5.5; if no node has reached the preset damping ratio threshold, then calculate the sensitivity factor of each node, sort them, and then execute S5.

4. S5.4: Execute the next cross-cutting algorithm on the sorted nodes to obtain the updated offspring particle swarm, and then execute S5.3; S5.5: Proceed to the next iteration and determine if the number of iterations is greater than the maximum number of iterations. If yes, proceed to S5.6; otherwise, proceed to S5.

3. S5.6: End the iteration and output the globally optimal particle after the iteration.

6. The method of claim 5, wherein, The method for ranking the sensitivity factors of each node is as follows: ； in, The bias sensitivity provides supporting power for the critical mode damping ratio and virtual inertia after the current node is connected, where i is the corresponding node number, i=1,…,n, and n is the number of nodes that the target wind farm can connect to the system. This indicates that the sensitivity of the number of nodes obtained is higher compared to other nodes. This indicates the number of nodes with the same sensitivity that were extracted.

7. The method of claim 5, wherein, In S5.4, when performing the next cross-multiplication algorithm on the sorted nodes, this is achieved by increasing the weight of high-sensitivity nodes. The method for increasing the weight of high-sensitivity nodes is as follows: ； in, It is the updated horizontally intersecting sub-individual. These are weighting coefficients. These are the horizontally intersecting sub-individuals before the update. The sensitivity of the highly sensitive node o after sorting. Let be the sensitivity of node j after sorting, and n be the number of nodes that the target wind farm can connect to the system.

Images