Parameter optimization method, device and equipment for multiple network-forming converters and medium

Abstract

The invention discloses a parameter optimization method, device and equipment for a plurality of grid-forming type converters and a medium, and relates to the technical field of electrical engineering. Control loop input ends of the plurality of grid-forming type converters comprise additional damping controllers; the method comprises the following steps: carrying out modal analysis on a state space model containing a network-forming converter and an alternating current power grid to obtain characteristic values of all oscillation modes so as to obtain expected damping ratios under all operation conditions; constructing a target parameter optimization model taking maximization of the target expected damping ratio as an optimization target; the target expected damping ratio is the minimum expected damping ratio; based on an optimization target of the target parameter optimization model, an attraction search mechanism, a self-adaptive step length mechanism and an opposite learning mechanism, searching a target parameter combination in a space formed by the parameter combinations of the plurality of network construction type converters; and deploying the target parameter combination to the plurality of network-forming converters. The problem of combined explosion of parameters of a plurality of network-forming converters is solved, and the optimization efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the field of electrical engineering technology, and in particular to a parameter optimization method, device, equipment and medium for multiple grid-type converters. Background Art

[0002] The high penetration of renewable energy sources, such as wind farms and photovoltaics, significantly impacts power system dynamics, including reduced system inertia and damping performance, which in turn impacts power system stability. Grid-connected converters (GCCs) are widely used today, playing a crucial role in emerging power systems by providing inertia and damping support to the grid. However, grids employing GCCs can experience deterioration in system stability due to the strong interaction between the GCC and the AC grid. This interaction is influenced by controller parameters, necessitating optimization of controller parameters to improve system stability.

[0003] However, there are a large number of converters, and there is no clear relationship between the decision variables (i.e., the parameters of the grid-type converter) and the objective function. This is a non-convex nonlinear optimization problem. The controller parameter design faces the problem of combinatorial explosion. In addition, the traditional method has low search efficiency when searching for parameters, and is prone to falling into local optimization, resulting in low optimization accuracy.

[0004] In summary, it can be seen that how to overcome the combinatorial explosion problem of parameters of multiple grid-type converters and improve the optimization efficiency and accuracy are problems to be solved in this field. Summary of the Invention

[0005] In view of this, the present invention aims to provide a parameter optimization method, device, equipment, and medium for multiple grid-type converters, thereby overcoming the combinatorial explosion problem of multiple grid-type converter parameters and improving optimization efficiency and accuracy. The specific solution is as follows: In a first aspect, the present application discloses a parameter optimization method for multiple grid-type converters, wherein the control loop input ends of the multiple grid-type converters include an additional damping controller; the method comprises: Performing a modal analysis on a state space model including the grid-type converter and the AC power grid to obtain eigenvalues ​​of each oscillation mode, and obtaining a desired damping ratio under each operating condition based on the eigenvalues; Based on the expected damping ratio, a target parameter optimization model is constructed with maximizing the target expected damping ratio as the optimization target; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios; Based on the optimization objective, attraction search mechanism, adaptive step size mechanism, and adversarial learning mechanism of the target parameter optimization model, searching for a target parameter combination in a space formed by parameter combinations of a plurality of the grid-type converters; The target parameter combination is deployed to a plurality of the grid-type converters.

[0006] Optionally, the parameters in the parameter combination of each of the grid-type converters are respectively a virtual inertia time constant, a virtual damping coefficient and parameters of an additional damping controller, and the parameters of the additional damping controller include a first time constant, a second time constant and a gain.

[0007] Optionally, performing a modal analysis on a state space model including the grid-type converter and the AC power grid to obtain characteristic values ​​of each oscillation mode, and obtaining an expected damping ratio under each operating condition based on the characteristic values, includes: Constructing a state space model including the grid-type converter and the AC power grid; wherein the state space model includes a state matrix, a control input matrix, and a control output matrix; Decoupling the state space model through linear transformation to obtain a diagonal matrix; wherein the diagonal matrix contains the eigenvalues ​​of each oscillation mode; The damping ratio of each oscillation mode is obtained by using the characteristic value of each oscillation mode, and the expected damping ratio under each operating condition is obtained based on the damping ratio of each oscillation mode.

[0008] Optionally, constructing a target parameter optimization model based on the expected damping ratio with maximizing the target expected damping ratio as the optimization goal includes: Constructing, based on the desired damping ratio, an objective function with maximizing the target desired damping ratio as an optimization objective, a plurality of parameter range constraints of the grid-type converter, a probabilistic stability constraint, a desired damping ratio constraint, and a target parameter optimization model of the state space model; The target parameter optimization model is: ; in, is the expected damping ratio, L is the decision variable set containing the parameter combinations of each grid-type converter, is the set of weakly damped oscillation modes, 、 are the state vector and output vector of the grid-type converter respectively, is the input vector composed of active power and reactive power reference values, A is the state matrix, B is the control input matrix, C is the control output matrix, is the virtual inertia time constant, 、 are the minimum and maximum values ​​of the virtual inertia time constant, is the virtual damping coefficient, 、 are the minimum and maximum values ​​of the virtual damping coefficient, is the first time constant, 、 are the minimum and maximum values ​​of the first time constant, respectively. is the second time constant, 、 are the minimum and maximum values ​​of the second time constant respectively, K is the gain of the additional damping controller, 、 are the minimum and maximum gains of the additional damping controller, n is the nth oscillation mode, is a probability stability indicator.

[0009] Optionally, searching for a target parameter combination in a space formed by parameter combinations of a plurality of the grid-connected converters based on the optimization objective, the attraction search mechanism, the adaptive step size mechanism, and the adversarial learning mechanism of the target parameter optimization model includes: Initializing a parameter combination of the grid-type converter, determining the parameter combination of the grid-type converter as an individual, determining the target expected damping ratio of each individual, and determining the target expected damping ratio of the individual as an individual attribute; Initializing an adaptive step size control coefficient and an attraction coefficient, and setting a boundary value of the adaptive step size control coefficient, a boundary value of the attraction coefficient, a maximum number of iterations, and the number of individuals; Based on the attraction search mechanism, the adaptive step size mechanism and the adversarial learning mechanism, a target individual with the largest individual attribute is found in the space composed of all the individuals, and the target individual is determined as a target parameter combination.

[0010] Optionally, the method of searching for a target individual with the maximum individual attribute in the space composed of all individuals based on the attraction search mechanism, the adaptive step size mechanism, and the adversarial learning mechanism includes: Randomly determine a current individual from all the individuals, and determine whether the current individual has the greatest individual attribute among all the individuals; If the current individual has the largest individual attribute among all the individuals, the current individual is determined as a candidate individual; If the current individual is not the individual with the largest individual attribute among all the individuals, randomly determine a reference individual from all the individuals, and determine whether the first individual attribute of the current individual is less than the second individual attribute of the reference individual; If the first individual attribute is less than the second individual attribute, updating an adaptive step length control coefficient according to the attraction and the Euclidean distance between the current individual and the reference individual, and determining a new current individual based on the adaptive step length control coefficient; Determine whether all the individuals have been traversed. If not, jump back to the step of determining whether the current individual has the largest individual attribute among all the individuals. If so, determine the current individual as a candidate individual. A target individual is determined based on the candidate individuals and the adversarial learning mechanism.

[0011] Optionally, determining a target individual based on the candidate individuals and the adversarial learning mechanism includes: Randomly move the candidate individual at its position to obtain a moved individual, and determine the individual at the opposite position of the moved individual; If the individual attribute of the moved individual is greater than the individual attribute of the individual at the opposite position and all the individuals have not been traversed yet, the moved individual is determined as the new current individual, and the process jumps again to the step of determining whether the current individual has the largest individual attribute among all the individuals; If the individual attribute of the moved individual is greater than the individual attribute of the individual at the opposite position and all the individuals have been traversed, the moved individual is determined as the target individual; If the individual attribute of the moved individual is less than the individual attribute of the individual at the opposite position and all the individuals have not been traversed yet, the individual at the opposite position is determined as the new current individual, and the process jumps back to the step of determining whether the current individual has the largest individual attribute among all the individuals; If the individual attribute of the moved individual is less than the individual attribute of the individual at the opposite position and all the individuals have been traversed, the individual at the opposite position is determined as the target individual.

[0012] In a second aspect, the present application discloses a parameter optimization device for multiple grid-type converters, wherein the control loop input ends of the multiple grid-type converters include an additional damping controller; the device includes: a damping ratio acquisition module, configured to perform modal analysis on a state space model including the grid-type converter and the AC power grid to obtain characteristic values ​​of each oscillation mode, and to obtain a desired damping ratio under each operating condition based on the characteristic values; A model construction module is used to construct a target parameter optimization model based on the expected damping ratio, with maximizing the target expected damping ratio as the optimization goal; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios; a parameter optimization module for searching for a target parameter combination in a space formed by parameter combinations of the plurality of grid-type converters based on the optimization objective, the attraction search mechanism, the adaptive step size mechanism, and the adversarial learning mechanism of the target parameter optimization model; An optimal parameter deployment module is used to deploy the target parameter combination to multiple grid-type converters.

[0013] In a third aspect, the present application discloses an electronic device, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the parameter optimization method of multiple grid-type converters disclosed above.

[0014] In a fourth aspect, the present application discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the steps of the parameter optimization method of multiple grid-type converters disclosed above are implemented.

[0015] The beneficial effects of the present application are as follows: the control loop input end of the multiple grid-type converters of the present application includes an additional damping controller; the method includes: performing modal analysis on a state space model including the grid-type converter and the AC power grid to obtain the characteristic values ​​of each oscillation mode, and obtaining the expected damping ratio under each operating condition based on the characteristic values; constructing a target parameter optimization model based on the expected damping ratio with maximizing the target expected damping ratio as the optimization target; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios; based on the optimization target, attraction search mechanism, adaptive step size mechanism and adversarial learning mechanism of the target parameter optimization model, searching for a target parameter combination in the space composed of parameter combinations of multiple grid-type converters; and deploying the target parameter combination to multiple grid-type converters. It can be seen that the present application adds an additional damping controller to the input end of the control loop of the grid-type converter, which can suppress the oscillation of the converter and improve the stability of the system. The modal analysis of the state space model is combined to accurately quantify the expected damping ratio of each oscillation mode, and then construct an optimization model with maximizing the target expected damping ratio as the core. The target expected damping ratio is the minimum expected damping ratio among all the expected damping ratios. The attractive search mechanism and the adaptive step size mechanism are used to efficiently search for the optimal value in the high-dimensional space of multiple converter parameter combinations, which can greatly improve the optimization efficiency. The adversarial learning mechanism can explore more diverse search space areas and avoid local optimality. In other words, it overcomes the combinatorial explosion problem of traditional methods in high-dimensional nonlinear optimization. Through the balance between global search and local tuning, it quickly converges to the optimal parameter combination. The optimal parameter combination enables the system to maintain excellent damping characteristics under multiple operating conditions, effectively suppressing power oscillations. Finally, through the deployment of the target parameter combination, the stability of the power system is systematically improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work.

[0017] Figure 1 This is a flow chart of a parameter optimization method for multiple grid-type converters disclosed in this application; Figure 2 A specific control diagram of a grid-type converter including an additional damping controller disclosed in this application; Figure 3 A specific grid-type converter control diagram disclosed in this application; Figure 4 A specific iterative optimization flow chart disclosed in this application; Figure 5 This is a specific individual location update schematic diagram disclosed in this application; Figure 6 This is a schematic diagram of a specific fitness convergence curve disclosed in this application; Figure 7 Schematic diagram of the effect of different controller parameter combinations on rotor angle oscillation of a specific grid-type converter disclosed in this application; wherein (a) is a graph showing the rotor angle response curve of generator No. 1 under different controller parameter combinations, (b) is a graph showing the rotor angle response curve of generator No. 5 under different controller parameter combinations, and (c) is a graph showing the rotor angle response curve of generator No. 10 under different controller parameter combinations; Figure 8 Schematic diagram showing the effects of different controller parameter combinations on power oscillations on a transmission line for a specific grid-connected converter disclosed in this application; (a) is a graph showing the active power response curves for lines 01-39 under different controller parameter combinations, (b) is a graph showing the active power response curves for lines 06-07 under different controller parameter combinations, and (c) is a graph showing the active power response curves for lines 23-24 under different controller parameter combinations. Figure 9 This is a schematic structural diagram of a parameter optimization device for multiple grid-type converters disclosed in this application; Figure 10 This is a structural diagram of an electronic device disclosed in this application. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] The high penetration of renewable energy sources, such as wind farms and photovoltaics, significantly impacts power system dynamics, including reduced system inertia and damping performance, which in turn impacts power system stability. Grid-connected converters (GCCs) are widely used today, playing a crucial role in emerging power systems by providing inertia and damping support to the grid. However, grids employing GCCs can experience deterioration in system stability due to the strong interaction between the GCC and the AC grid. This interaction is influenced by controller parameters, necessitating optimization of controller parameters to improve system stability.

[0020] However, there are a large number of converters, and there is no clear relationship between the decision variables (i.e., the parameters of the grid-type converter) and the objective function. This is a non-convex nonlinear optimization problem. The controller parameter design faces the problem of combinatorial explosion. In addition, the traditional method has low search efficiency when searching for parameters, and is prone to falling into local optimization, resulting in low optimization accuracy.

[0021] To this end, the present application provides a parameter optimization solution for multiple meshed converters, which overcomes the combinatorial explosion problem of the parameters of multiple meshed converters and improves the optimization efficiency and accuracy.

[0022] See also Figure 1 As shown, an embodiment of the present application discloses a parameter optimization method for multiple grid-type converters, wherein the control loop input ends of the multiple grid-type converters include an additional damping controller; the method includes: Step S11: performing modal analysis on a state space model including the grid-type converter and the AC power grid to obtain characteristic values ​​of each oscillation mode, and obtaining a desired damping ratio under each operating condition based on the characteristic values.

[0023] When using a grid-controlled converter, the converter has characteristics similar to those of a synchronous generator and may participate in oscillations. In order to suppress oscillations and improve stability, for example Figure 2 A specific control diagram of a grid-type converter including an additional damping controller is shown, in which an additional damping controller is added to the input end of the grid-type control loop.

[0024] In this embodiment, the parameters in the parameter combination of each grid-type converter are respectively a virtual inertia time constant, a virtual damping coefficient and parameters of an additional damping controller, and the parameters of the additional damping controller include a first time constant, a second time constant and a gain.

[0025] like Figure 2 As shown, is the initial frequency, To adjust the frequency, is the system output frequency, P ref is the active power reference value, P v is the actual value, is the phase difference, is the virtual inertia time constant, is the virtual damping coefficient. The additional damping controller introduces three parameters, namely the first time constant , the second time constant And the gain K of the additional damping controller, so in addition to the need to optimize and In addition, the parameters of the additional damping controller (i.e., the first time constant, the second time constant, and the gain) need to be optimized to ensure that the system damping is improved and the system stability is enhanced. In other words, 、 、 、 , K is a set of parameter combinations.

[0026] In this embodiment, the state space model including the grid-type converter and the AC power grid is subjected to modal analysis to obtain the eigenvalues ​​of each oscillation mode, and the expected damping ratio under each operating condition is obtained based on the eigenvalues, including: constructing a state space model including the grid-type converter and the AC power grid; wherein the state space model includes a state matrix, a control input matrix and a control output matrix; decoupling the state space model through linear transformation to obtain a diagonal matrix; wherein the diagonal matrix includes the eigenvalues ​​of each oscillation mode; obtaining the damping ratio of each oscillation mode using the eigenvalues ​​of each oscillation mode, and obtaining the expected damping ratio under each operating condition based on the damping ratio of each oscillation mode.

[0027] For example Figure 3 The control diagram of a specific grid-type converter is shown. The grid-type converter adjusts its active power output through the power angle signal at the converter's AC terminal to ensure synchronization with the AC grid. Based on the control loop of the grid-type converter and the AC grid topology, a state space model of the entire system, including the grid-type converter and the AC grid, is established. The state space model can be expressed as: ; Where A, B, and C are the state matrix, control input matrix, and output matrix, respectively; ∆u is the input vector consisting of active and reactive power reference values; ∆x and ∆y are the state vector and output vector of the grid-connected converter, respectively.

[0028] Based on the above state space model, it is linearly transformed ,in is the right eigenvector The matrix composed of , eliminates the cross coupling between state variables, and decouples the state space model through linear transformation, and finally transforms the original state space in the state space model into: ; Where z is the new state vector obtained by linear transformation, Λ is a diagonal matrix whose elements are the eigenvalues ​​of A, that is, the diagonal matrix contains the eigenvalues ​​of each oscillation mode, B′ is the controllability matrix, and C′ is the observability matrix. In other words, the state space model is decoupled by linear transformation to obtain a transformed state space model containing the diagonal matrix, the controllability matrix, and the observability matrix. Through modal analysis, the eigenvalue of each oscillation mode can be obtained, which is denoted as , represents the nth modal eigenvalue, where is the real part of the eigenvalue of each oscillation mode, is the imaginary part of the eigenvalue of each oscillation mode.

[0029] Use the eigenvalues ​​of each oscillation mode to obtain the damping ratio of each oscillation mode , the formula is as follows: ; Furthermore, based on the damping ratio of each oscillation mode, the expected damping ratio under each operating condition is obtained. , the formula is as follows: ; where N sample Represents the total number of random scenarios, i.e., different operating conditions of the power system.

[0030] Step S12: constructing a target parameter optimization model based on the expected damping ratio with maximizing the target expected damping ratio as the optimization target; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios.

[0031] In this embodiment, the target parameter optimization model with maximizing the target expected damping ratio as the optimization goal is constructed based on the expected damping ratio, including: constructing an objective function with maximizing the target expected damping ratio as the optimization goal, a plurality of parameter range constraints of the grid-type converter, a probabilistic stability constraint, an expected damping ratio constraint, and a target parameter optimization model of the state space model based on the expected damping ratio; The target parameter optimization model is: ; in, is the expected damping ratio, L is the decision variable set containing the parameter combinations of each grid-type converter, is the set of weakly damped oscillation modes, 、 are the state vector and output vector of the grid-type converter respectively, is the input vector composed of active power and reactive power reference values, A is the state matrix, B is the control input matrix, C is the control output matrix, is the virtual inertia time constant, 、 are the minimum and maximum values ​​of the virtual inertia time constant, is the virtual damping coefficient, 、 are the minimum and maximum values ​​of the virtual damping coefficient, is the first time constant, 、 are the minimum and maximum values ​​of the first time constant, respectively. is the second time constant, 、 are the minimum and maximum values ​​of the second time constant respectively, K is the gain of the additional damping controller, 、 are the minimum and maximum gains of the additional damping controller, n is the nth oscillation mode, is a probability stability indicator.

[0032] The converter optimization design aims to improve the damping performance of the system and suppress oscillations better and faster. Therefore, the system is selected to maximize the minimum expected damping ratio as the objective function. That is, based on the expected damping ratio, an objective function is constructed with maximizing the target expected damping ratio as the optimization goal. The target expected damping ratio is the minimum expected damping ratio among all the expected damping ratios. The objective function is: ; Where, is the expected damping ratio, L is the decision variable set containing the parameter combinations of each grid-type converter, that is, L=[H v1 ,Dv1 ,T b1 ,T a1 ,K1,···,H vNgfm ,D vNgfm ,T bgfm ,T aNgfm ,K Ngfm ] T , Ngfm represents the number of converters, is the weakly damped oscillation mode set (ie, COMs set), which refers to the vibration mode with an expected damping ratio less than 0.1 under random working conditions, that is, , where S om is the set of all oscillation modes.

[0033] Constraints are constructed so that the parameters of the grid-type converter and the system operating state are limited to a certain range. More importantly, the probabilistic stability of the system is guaranteed to ensure the stability of the system under random operating conditions. The constraints include parameter range constraints of multiple grid-type converters, probabilistic stability constraints, and expected damping ratio constraints. The parameter range constraints of multiple grid-type converters are specifically: ; Where, is the virtual inertia time constant, 、 are the minimum and maximum values ​​of the virtual inertia time constant, is the virtual damping coefficient, 、 are the minimum and maximum values ​​of the virtual damping coefficient, is the first time constant, 、 are the minimum and maximum values ​​of the first time constant, respectively. is the second time constant, 、 are the minimum and maximum values ​​of the second time constant respectively, K is the gain of the additional damping controller, 、 are the minimum and maximum gains of the additional damping controller, respectively, and n is the nth oscillation mode.

[0034] Probabilistic stability P uns Evaluate the stability of the system in random scenarios. According to the law of large numbers and the central limit theorem, the probabilistic stability index P uns It can be expressed as: ; in and is the stable state of the kth random scenario and all Nsample Expected value of the scenario, N sample Represents the total number of random scenarios, i.e., different operating conditions of the power system. uns,k Calculation yields: ; in is the real part of the eigenvalue of each oscillation mode, and m represents an oscillation mode in a random scenario.

[0035] In order to eliminate weakly damped oscillation modes, the parameters of GFM-ESs should ensure that all expected damping ratios of the oscillation modes are greater than 0.05, that is, the expected damping ratio constraint is: ; in, is the desired damping ratio, is a set of weakly damped oscillation modes, and n is the nth oscillation mode.

[0036] The target parameter optimization model is constructed by combining the objective function, constraints and state space model. The target parameter optimization model is specifically as follows: ; in, is the expected damping ratio, L is the decision variable set containing the parameter combinations of each grid-type converter, is the set of weakly damped oscillation modes, 、 are the state vector and output vector of the grid-type converter respectively, is the input vector composed of active power and reactive power reference values, A is the state matrix, B is the control input matrix, C is the control output matrix, is the virtual inertia time constant, 、 are the minimum and maximum values ​​of the virtual inertia time constant, is the virtual damping coefficient, 、 are the minimum and maximum values ​​of the virtual damping coefficient, is the first time constant, 、 are the minimum and maximum values ​​of the first time constant, respectively. is the second time constant, 、 are the minimum and maximum values ​​of the second time constant respectively, K is the gain of the additional damping controller, 、 are the minimum and maximum gains of the additional damping controller, n is the nth oscillation mode, is a probability stability indicator.

[0037] Step S13: based on the optimization target, attraction search mechanism, adaptive step size mechanism and adversarial learning mechanism of the target parameter optimization model, searching for a target parameter combination in a space formed by parameter combinations of a plurality of the grid-connected converters.

[0038] In this embodiment, the optimization target, attraction search mechanism, adaptive step size mechanism and adversarial learning mechanism based on the target parameter optimization model search for a target parameter combination in the space composed of parameter combinations of multiple meshed inverters, including: initializing the parameter combination of the meshed inverter, and determining the parameter combination of the meshed inverter as an individual, determining the target expected damping ratio of each individual, and determining the target expected damping ratio of the individual as an individual attribute; initializing the adaptive step size control coefficient and the attraction coefficient, and setting the boundary value of the adaptive step size control coefficient, the boundary value of the attraction coefficient, the maximum number of iterations and the number of individuals; based on the attraction search mechanism, the adaptive step size mechanism and the adversarial learning mechanism, searching for the target individual with the maximum individual attribute in the space composed of all the individuals, and determining the target individual as the target parameter combination.

[0039] In the parameter optimization design of a grid-type converter, the parameter design of the grid-type converter faces the problem of combinatorial explosion. There is no clear relationship between the decision variables (converter parameters) and the objective function, which is a black box problem. This embodiment uses the maximization of the target expected damping ratio as the objective function, optimizes the minimum expected damping ratio through an optimization algorithm, and finds the optimal parameters of the converter. Therefore, the minimum expected damping ratio is defined as a certain attribute of an individual (each individual represents a set of candidate converter parameter combinations). In other words, the parameter combination of the grid-type converter is initialized and the parameter combination of the grid-type converter is determined as an individual L. The target expected damping ratio of each individual is determined, and the target expected damping ratio of the individual is determined as an individual attribute. The formula of individual attribute I is: ; Among them, I represents a certain attribute of the individual.

[0040] In order to improve the search efficiency, a and Parameter a is a random control parameter that determines the random movement step size of the individual. It should be set to a larger value initially to encourage exploration, and then gradually reduced over time to focus on finding the optimal value within a certain space. The attraction coefficient controls the rate at which individual attraction decreases with distance, starts from a low value to promote exploration, and gradually increases over time, focusing on local search, that is, setting the boundary value of the adaptive step size control coefficient. and a can be expressed as: ; Wherein, the subscript “0” indicates the initial value; t indicates the number of iterations in the program; L max , L min is the position boundary composed of the parameters of the network converter, thus limiting the number of individuals; d a is a decrement factor whose purpose is to reduce random motion as the number of iterations increases; for The maximum value of N iter is the maximum number of iterations, that is, initializing the adaptive step size control coefficient and the attraction coefficient, and the boundary value of the attraction coefficient.

[0041] For example Figure 4 The specific iterative optimization flow chart shown in FIG. 1 is based on the attraction search mechanism, the adaptive step size mechanism and the adversarial learning mechanism to find the target individual with the largest individual attribute in the space composed of all individuals, and determine the target individual as the target parameter combination; in the process of finding the target individual with the largest individual attribute in the space composed of all individuals, it is necessary to continuously update the individual position and compare the individual fitness after the position update to obtain the optimal fitness, that is, the optimal solution. If the difference in accuracy with the previous iterative fitness is a preset accuracy threshold or the maximum number of iterations has been reached, the iterative optimization ends and the optimal solution, that is, the optimal parameters, is output. If the maximum number of iterations has not been reached and the difference in accuracy with the previous iterative fitness is not a preset accuracy threshold, iterative optimization is still required, that is, in addition to setting a maximum number of iterations N iter After each iteration, the damping ratio of the updated position is compared with that of the previous position. If the difference in damping ratio is less than 1e-3, the iteration is terminated. Therefore, the optimization may be completed before the maximum number of iterations is reached, which improves the optimization efficiency of the algorithm.

[0042] In this embodiment, the method of searching for a target individual with the largest individual attribute in the space formed by all the individuals based on the attraction search mechanism, the adaptive step size mechanism, and the adversarial learning mechanism includes: randomly determining a current individual from all the individuals, and determining whether the current individual is the individual with the largest individual attribute among all the individuals; if the current individual is the individual with the largest individual attribute among all the individuals, determining the current individual as a candidate individual; if the current individual is not the individual with the largest individual attribute among all the individuals, randomly determining a reference individual from all the individuals, and determining whether a first individual attribute of the current individual is less than a second individual attribute of the reference individual; if the first individual attribute is less than the second individual attribute, updating an adaptive step size control coefficient based on the attraction and Euclidean distance between the current individual and the reference individual, and determining a new current individual based on the adaptive step size control coefficient; determining whether all the individuals have been traversed, and if not, jumping back to the step of determining whether the current individual is the individual with the largest individual attribute among all the individuals, and if so, determining the current individual as a candidate individual; and determining a target individual based on the candidate individuals and the adversarial learning mechanism.

[0043] For example Figure 5 A specific individual position update diagram is shown, where the current individual L is randomly determined from all individuals. i , and judge whether the current individual is the individual with the largest individual attribute among all individuals; if the current individual is the individual with the largest individual attribute among all individuals, the current individual is determined as a candidate individual; if the current individual is not the individual with the largest individual attribute among all individuals, a reference individual L is randomly determined from all individuals j , and judge whether the first individual attribute of the current individual is less than the second individual attribute of the reference individual; if the first individual attribute is less than the second individual attribute, then update the adaptive step length control coefficient according to the attraction and Euclidean distance between the current individual and the reference individual, and determine the new current individual based on the adaptive step length control coefficient, that is, when L j Individual attribute ratio of position L i The individual attributes of the position should be good, that is, I(L j )>I(L i ), indicating L j The damping performance of the power system at the location is better, so L i The individual at position L is attracted and moves towards L with a small random step. j Position individual moves, position L i and L j The attraction between individuals is expressed as: ; in, is the position L iand L j The attraction between individuals (this is equivalent to the converter parameter combination i), the reference individual does not include individuals in opposite positions; b0 is the initial attraction; d ij Indicates position L i and L j The Euclidean (geometric) distance between: The Euclidean distance formula is: ; N dim is the dimension of the problem, where N dim =Number of converters in the system Ngfm×number of system grid-type converter parameters to be optimized.

[0044] According to the attraction between individuals β ij and distance d ij , the optimization algorithm will move a certain step length L in each iteration i,step (i.e., adaptive step size control coefficient) updates the position of the individual. The formula for the adaptive step size control coefficient is: ; Here, rand is a uniformly distributed random number between [0,1].

[0045] According to the step length L i,step , update the position L i : ; In this way, the individual update is completed when the first individual attribute is less than the second individual attribute. Next, it is determined whether all individuals have been traversed. If not, it jumps back to the step of determining whether the current individual has the largest individual attribute among all individuals, that is, continuing the iterative optimization. If so, the current individual is determined as a candidate individual.

[0046] If the current individual is the individual with the largest individual attribute among all individuals, or after individual update, all individuals have been traversed to obtain candidate individuals, the target individual is determined based on the candidate individuals and the adversarial learning mechanism.

[0047] In this embodiment, the target individual is determined based on the candidate individuals and the opposition learning mechanism, including: randomly moving at the position of the candidate individual to obtain the moved individual, and determining the individual at the opposite position of the moved individual; if the individual attribute of the moved individual is greater than the individual attribute of the individual at the opposite position and all the individuals have not been traversed yet, the moved individual is determined as the new current individual, and the process of determining whether the current individual is the individual with the largest individual attribute among all the individuals is jumped again; if the individual attribute of the moved individual is greater than the individual attribute of the individual at the opposite position, If the individual attributes of the individual after the movement are less than the individual attributes of the individual at the opposite position and all the individuals have not been traversed yet, the individual at the opposite position is determined as the new current individual, and the process jumps again to the step of determining whether the current individual is the individual with the largest individual attributes among all the individuals; if the individual attributes of the individual after the movement are less than the individual attributes of the individual at the opposite position and all the individuals have been traversed yet, the individual at the opposite position is determined as the target individual.

[0048] Since the candidate individuals are the individuals with the best individual attributes, they will not be attracted by other individuals. Therefore, in the process of determining the target individuals based on the candidate individuals, the individuals are updated by random movement. That is, the individual update formula based on the candidate individuals is: ; In order to explore more diverse search space areas and avoid local optimality, adversarial learning is introduced into the optimization algorithm, that is, the adversarial learning mechanism is introduced. In a specific iteration, individuals with poor attributes will not only be attracted by individuals with good attributes, but will also move in the opposite direction. For a given position L i , the boundary is L min and L max , opposite position L i,oppo Calculated as: ; Among them L min and L max The individual position boundary is composed of the maximum and minimum values ​​of the GFM-ESs parameters. The individual L calculated based on the individual update formula of the candidate individual i , that is, individual L after moving i , move the individual L i The individual attributes of the individual L in the opposite position i,oppoIf the individual attributes of the individual after the move are greater than the individual attributes of the individual at the opposite position, the individual after the move is determined as the new current individual. If the individual attributes of the individual after the move are less than the individual attributes of the individual at the opposite position, the individual at the opposite position is determined as the new current individual, and if all individuals have been traversed, the new current individual is determined as the target individual. That is to say, if the individual attributes of the individual after the move are greater than the individual attributes of the individual at the opposite position and all individuals have not been traversed, the individual after the move is determined as the new current individual, and the process jumps again to determine whether the current individual is among all individuals. The step of determining the individual with the largest individual attribute; if the individual attribute of the individual after moving is greater than the individual attribute of the individual at the opposite position and all individuals have been traversed, the individual after moving is determined as the target individual; if the individual attribute of the individual after moving is less than the individual attribute of the individual at the opposite position and all individuals have not been traversed, the individual at the opposite position is determined as the new current individual, and the process jumps again to the step of determining whether the current individual is the individual with the largest individual attribute among all individuals; if the individual attribute of the individual after moving is less than the individual attribute of the individual at the opposite position and all individuals have been traversed, the individual at the opposite position is determined as the target individual.

[0049] Step S14: deploying the target parameter combination to the plurality of grid-connected converters.

[0050] After obtaining the optimal parameter combination of multiple meshed converters, the target parameter combination is deployed to the multiple meshed converters. At this time, the system stability of the multiple meshed converters is better.

[0051] This embodiment uses the parameters of the grid-type converter of multiple additional damping controllers as the object, constructs the maximum value of the minimum expected damping ratio as the objective function and the constraint conditions, optimizes the objective function through the optimization algorithm, and obtains the optimized parameters of the converter. The key point is that the maximum value of the minimum expected damping ratio is used as the objective function, and the optimization algorithm is used to optimize the objective function to obtain the parameters of the grid-type converter of multiple additional damping controllers (virtual inertia time constant H v , virtual damping coefficient D v , time constant T b 、T a , additional damping controller gain K) optimization results. Adaptive parameters and adversarial learning mechanism are used to explore more diverse search space areas and avoid local optimality. Adaptive parameters, scaling a and , which greatly improves the efficiency of the algorithm. The converter with additional damping controller introduces a time constant T b 、T a, damping controller gain K. The maximum value of the minimum expected damping ratio is taken as the objective function, the minimum expected damping ratio is defined as an individual attribute, and adaptive parameters and adversarial learning mechanism are introduced into the optimization algorithm.

[0052] The beneficial effects of the present application are as follows: the control loop input end of the multiple grid-type converters of the present application includes an additional damping controller; the method includes: performing modal analysis on a state space model including the grid-type converter and the AC power grid to obtain the characteristic values ​​of each oscillation mode, and obtaining the expected damping ratio under each operating condition based on the characteristic values; constructing a target parameter optimization model based on the expected damping ratio with maximizing the target expected damping ratio as the optimization target; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios; based on the optimization target, attraction search mechanism, adaptive step size mechanism and adversarial learning mechanism of the target parameter optimization model, searching for a target parameter combination in the space composed of parameter combinations of multiple grid-type converters; and deploying the target parameter combination to multiple grid-type converters. It can be seen that the present application adds an additional damping controller to the input end of the control loop of the grid-type converter, which can suppress the oscillation of the converter and improve the stability of the system. The modal analysis of the state space model is combined to accurately quantify the expected damping ratio of each oscillation mode, and then construct an optimization model with maximizing the target expected damping ratio as the core. The target expected damping ratio is the minimum expected damping ratio among all the expected damping ratios. The attractive search mechanism and the adaptive step size mechanism are used to efficiently search for the optimal value in the high-dimensional space of multiple converter parameter combinations, which can greatly improve the optimization efficiency. The adversarial learning mechanism can explore more diverse search space areas and avoid local optimality. In other words, it overcomes the combinatorial explosion problem of traditional methods in high-dimensional nonlinear optimization. Through the balance between global search and local tuning, it quickly converges to the optimal parameter combination. The optimal parameter combination enables the system to maintain excellent damping characteristics under multiple operating conditions, effectively suppressing power oscillations. Finally, through the deployment of the target parameter combination, the stability of the power system is systematically improved.

[0053] Next, the grid-type converter is integrated into the 39-bus system. The parameters of each grid-type converter (i.e., H v , D v ,,T b , T a and K) are optimized to provide corresponding descriptions for this application.

[0054] (1) Parameter settings: Since there are 5 grid-type converters, the dimension of the optimization problem is N dim = 25. In order to obtain the optimal parameters, 40 individual search parameter spaces are created, and the number of iterations N iter Set it to 1000, and other parameter settings are as follows: Table 1 Parameter settings (2) Simulation analysis: By writing Python code, the Python-PowerFactory automated simulation framework was established to implement and verify the proposed algorithm. After 1000 iterations, the optimal parameters of all grid-type converters were successfully found, denoted as P opt , see Table 2. At the same time, in order to evaluate the effectiveness of the optimal parameters, two other sets of parameters are used for comparison, denoted as Pother1 and Pother2.

[0055] Table 2 Optimization parameters and comparison parameters of GFM-BESs From Table 2, we can see that P opt Compared with the other two sets of parameters, the minimum expected damping ratio is improved through the optimization algorithm, the damping performance is effectively improved, and the stability is improved. At the same time, from the convergence diagram of the minimum expected damping ratio, such as Figure 6 As shown in FIG, a specific fitness convergence curve diagram shows that when the optimal parameters are used, after 1000 iterations, the minimum expected damping ratio of the test system is increased from 6.4% to 9.1%.

[0056] like Figure 7 The schematic diagram of the influence of different controller parameter combinations on rotor angle oscillation of a specific grid-type converter is shown in FIG. Figure 8 The schematic diagram of the influence of different controller parameter combinations of a specific grid-type converter on the power oscillation on the transmission line is shown. Time domain simulation is performed to compare the system dynamic response with the optimal parameters and the other two parameter sets, as shown in Figure 8 As shown by the red curve, the grid-type converter can quickly suppress power oscillations caused by short-circuit faults within 10 seconds, effectively suppressing power oscillations on the transmission line. In contrast, grid-type converters using other parameters (i.e., Pother1 and Pother2) cannot achieve satisfactory damping performance, resulting in long-term oscillations in the active power on the transmission line.

[0057] As can be seen, this embodiment addresses the "black box" optimization problem, optimizing the parameters of multiple grid-type converters to improve the system's damping performance, thereby enhancing system stability. This embodiment utilizes adaptive parameters and an adversarial learning mechanism to expand the algorithm's optimization scope and significantly improve its efficiency.

[0058] See also Figure 9 As shown, an embodiment of the present application discloses a parameter optimization device for multiple grid-type converters, wherein the control loop input ends of the multiple grid-type converters include an additional damping controller; the device includes: a damping ratio acquisition module 11 for performing modal analysis on a state space model including the grid-type converter and the AC power grid to obtain characteristic values ​​of each oscillation mode, and obtaining a desired damping ratio under each operating condition based on the characteristic values; A model construction module 12 is configured to construct a target parameter optimization model based on the expected damping ratio, with maximizing the target expected damping ratio as an optimization target; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios; a parameter optimization module 13 for searching for a target parameter combination in a space formed by parameter combinations of the plurality of grid-type converters based on the optimization objective, the attraction search mechanism, the adaptive step size mechanism, and the adversarial learning mechanism of the target parameter optimization model; The optimal parameter deployment module 14 is configured to deploy the target parameter combination to the plurality of grid-connected converters.

[0059] The beneficial effects of the present application are as follows: the control loop input end of the multiple grid-type converters of the present application includes an additional damping controller; the method includes: performing modal analysis on a state space model including the grid-type converter and the AC power grid to obtain the characteristic values ​​of each oscillation mode, and obtaining the expected damping ratio under each operating condition based on the characteristic values; constructing a target parameter optimization model based on the expected damping ratio with maximizing the target expected damping ratio as the optimization target; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios; based on the optimization target, attraction search mechanism, adaptive step size mechanism and adversarial learning mechanism of the target parameter optimization model, searching for a target parameter combination in the space composed of parameter combinations of multiple grid-type converters; and deploying the target parameter combination to multiple grid-type converters. It can be seen that the present application adds an additional damping controller to the input end of the control loop of the grid-type converter, which can suppress the oscillation of the converter and improve the stability of the system. The modal analysis of the state space model is combined to accurately quantify the expected damping ratio of each oscillation mode, and then construct an optimization model with maximizing the target expected damping ratio as the core. The target expected damping ratio is the minimum expected damping ratio among all the expected damping ratios. The attractive search mechanism and the adaptive step size mechanism are used to efficiently search for the optimal value in the high-dimensional space of multiple converter parameter combinations, which can greatly improve the optimization efficiency. The adversarial learning mechanism can explore more diverse search space areas and avoid local optimality. In other words, it overcomes the combinatorial explosion problem of traditional methods in high-dimensional nonlinear optimization. Through the balance between global search and local tuning, it quickly converges to the optimal parameter combination. The optimal parameter combination enables the system to maintain excellent damping characteristics under multiple operating conditions, effectively suppressing power oscillations. Finally, through the deployment of the target parameter combination, the stability of the power system is systematically improved.

[0060] Furthermore, an embodiment of the present application also provides an electronic device. Figure 10 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content in the diagram should not be considered as any limitation to the scope of application of the present application.

[0061] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Specifically, the device may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 is used to store a computer program, which is loaded and executed by the processor 21 to implement the relevant steps of the parameter optimization method for multiple grid-connected converters performed by the electronic device as disclosed in any of the aforementioned embodiments.

[0062] In this embodiment, the power supply 23 is used to provide operating voltage for various hardware devices on the electronic device; the communication interface 24 can create a data transmission channel between the electronic device and external devices. The communication protocol it follows is any communication protocol that can be applied to the technical solution of this application and is not specifically limited here; the input and output interface 25 is used to obtain external input data or output data to the outside world. Its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0063] Among them, the processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 can be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 21 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 21 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 21 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0064] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or CD, etc. The resources stored thereon include an operating system 221, a computer program 222 and data 223, etc. The storage method can be temporary storage or permanent storage.

[0065] The operating system 221 is used to manage and control the hardware devices and computer programs 222 on the electronic device, enabling the processor 21 to calculate and process the massive amount of data 223 in the memory 22. The operating system 221 can be Windows, Unix, Linux, etc. In addition to including computer programs capable of implementing the parameter optimization method for multiple grid-connected converters performed by the electronic device as disclosed in any of the aforementioned embodiments, the computer programs 222 may further include computer programs capable of performing other specific tasks. Data 223 may include data received by the electronic device from external devices, as well as data collected by its own input / output interface 25.

[0066] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when executed by a processor, the computer program implements the aforementioned method for optimizing the parameters of multiple grid-connected converters. The specific steps of this method can be referred to the corresponding contents disclosed in the aforementioned embodiments and will not be repeated here.

[0067] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from the other embodiments. Reference can be made to the descriptions of the identical or similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and the relevant parts can be referred to the descriptions of the methods.

[0068] Professionals may further appreciate that the units and algorithmic steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application. The steps of the method or algorithm described in conjunction with the embodiments disclosed herein can be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module can be placed in random access memory (RAM), memory, read-only memory (ROM), electrically programmable EPROM (Erasable Programmable Read Only Memory), electrically erasable programmable EEPROM (Electrically Erasable Programmable read only memory), registers, hard disk, removable disk, CD-ROM (Compact Disc Read-Only Memory), or any other form of storage medium known in the technical field.

[0069] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0070] The above is a detailed introduction to the parameter optimization method, device, equipment and medium for multiple grid-type converters provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for general technical personnel in this field, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A parameter optimization method for multiple grid-type converters, characterized in that: The control loop inputs of the plurality of grid-type converters include an additional damping controller; the method comprises: Performing a modal analysis on a state space model including the grid-type converter and the AC power grid to obtain eigenvalues ​​of each oscillation mode, and obtaining a desired damping ratio under each operating condition based on the eigenvalues; Based on the expected damping ratio, a target parameter optimization model is constructed with maximizing the target expected damping ratio as the optimization target; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios; Based on the optimization objective, attraction search mechanism, adaptive step size mechanism, and adversarial learning mechanism of the target parameter optimization model, searching for a target parameter combination in a space formed by parameter combinations of a plurality of the grid-type converters; The target parameter combination is deployed to a plurality of the grid-type converters.

2. The parameter optimization method for multiple grid-type converters according to claim 1, characterized in that: The parameters in the parameter combination of each grid-type converter are respectively a virtual inertia time constant, a virtual damping coefficient and parameters of an additional damping controller, and the parameters of the additional damping controller include a first time constant, a second time constant and a gain.

3. The parameter optimization method for multiple grid-type converters according to claim 1, characterized in that: The performing of modal analysis on a state space model including the grid-type converter and the AC power grid to obtain characteristic values ​​of each oscillation mode, and obtaining a desired damping ratio under each operating condition based on the characteristic values, includes: Constructing a state space model including the grid-type converter and the AC power grid; wherein the state space model includes a state matrix, a control input matrix, and a control output matrix; Decoupling the state space model through linear transformation to obtain a diagonal matrix; wherein the diagonal matrix contains the eigenvalues ​​of each oscillation mode; The damping ratio of each oscillation mode is obtained by using the characteristic value of each oscillation mode, and the expected damping ratio under each operating condition is obtained based on the damping ratio of each oscillation mode.

4. The parameter optimization method for multiple grid-type converters according to any one of claims 1 to 3, characterized in that: The target parameter optimization model is constructed based on the expected damping ratio with maximizing the target expected damping ratio as the optimization goal, including: Constructing, based on the desired damping ratio, an objective function with maximizing the target desired damping ratio as an optimization objective, a plurality of parameter range constraints of the grid-type converter, a probabilistic stability constraint, a desired damping ratio constraint, and a target parameter optimization model of the state space model; The target parameter optimization model is: ； in, is the expected damping ratio, L is the decision variable set containing the parameter combinations of each grid-type converter, is the set of weakly damped oscillation modes, 、 are the state vector and output vector of the grid-type converter respectively, is the input vector composed of active power and reactive power reference values, A is the state matrix, B is the control input matrix, C is the control output matrix, is the virtual inertia time constant, 、 are the minimum and maximum values ​​of the virtual inertia time constant, is the virtual damping coefficient, 、 are the minimum and maximum values ​​of the virtual damping coefficient, is the first time constant, 、 are the minimum and maximum values ​​of the first time constant, respectively. is the second time constant, 、 are the minimum and maximum values ​​of the second time constant respectively, K is the gain of the additional damping controller, 、 are the minimum and maximum gains of the additional damping controller, n is the nth oscillation mode, is a probability stability indicator.

5. The parameter optimization method for multiple grid-type converters according to claim 1, characterized in that: The optimization target, attraction search mechanism, adaptive step size mechanism, and adversarial learning mechanism based on the target parameter optimization model are used to search for a target parameter combination in a space formed by parameter combinations of a plurality of the grid-type converters, including: Initializing a parameter combination of the grid-type converter, determining the parameter combination of the grid-type converter as an individual, determining the target expected damping ratio of each individual, and determining the target expected damping ratio of the individual as an individual attribute; Initializing an adaptive step size control coefficient and an attraction coefficient, and setting a boundary value of the adaptive step size control coefficient, a boundary value of the attraction coefficient, a maximum number of iterations, and the number of individuals; Based on the attraction search mechanism, the adaptive step size mechanism and the adversarial learning mechanism, a target individual with the largest individual attribute is found in the space composed of all the individuals, and the target individual is determined as a target parameter combination.

6. The parameter optimization method for multiple grid-type converters according to claim 5, characterized in that: The method of searching for a target individual with the maximum individual attribute in the space composed of all individuals based on the attraction search mechanism, the adaptive step size mechanism, and the adversarial learning mechanism includes: Randomly determine a current individual from all the individuals, and determine whether the current individual has the greatest individual attribute among all the individuals; If the current individual has the largest individual attribute among all the individuals, the current individual is determined as a candidate individual; If the current individual is not the individual with the largest individual attribute among all the individuals, randomly determine a reference individual from all the individuals, and determine whether the first individual attribute of the current individual is less than the second individual attribute of the reference individual; If the first individual attribute is less than the second individual attribute, updating an adaptive step length control coefficient according to the attraction and the Euclidean distance between the current individual and the reference individual, and determining a new current individual based on the adaptive step length control coefficient; Determine whether all the individuals have been traversed. If not, jump back to the step of determining whether the current individual has the largest individual attribute among all the individuals. If so, determine the current individual as a candidate individual. A target individual is determined based on the candidate individuals and the adversarial learning mechanism.

7. The parameter optimization method for multiple grid-type converters according to claim 6, characterized in that: The determining of the target individual based on the candidate individuals and the adversarial learning mechanism includes: Randomly move the candidate individual at its position to obtain a moved individual, and determine the individual at the opposite position of the moved individual; If the individual attribute of the moved individual is greater than the individual attribute of the individual at the opposite position and all the individuals have not been traversed yet, the moved individual is determined as the new current individual, and the process jumps again to the step of determining whether the current individual has the largest individual attribute among all the individuals; If the individual attribute of the moved individual is greater than the individual attribute of the individual at the opposite position and all the individuals have been traversed, the moved individual is determined as the target individual; If the individual attribute of the moved individual is less than the individual attribute of the individual at the opposite position and all the individuals have not been traversed yet, the individual at the opposite position is determined as the new current individual, and the process jumps back to the step of determining whether the current individual has the largest individual attribute among all the individuals; If the individual attribute of the moved individual is less than the individual attribute of the individual at the opposite position and all the individuals have been traversed, the individual at the opposite position is determined as the target individual.

8. A parameter optimization device for multiple grid-type converters, characterized in that: The control loop input of the plurality of grid-type converters includes an additional damping controller; the device includes: a damping ratio acquisition module, configured to perform modal analysis on a state space model including the grid-type converter and the AC power grid to obtain characteristic values ​​of each oscillation mode, and to obtain a desired damping ratio under each operating condition based on the characteristic values; A model construction module is used to construct a target parameter optimization model based on the expected damping ratio, with maximizing the target expected damping ratio as the optimization goal; the target expected damping ratio is the minimum expected damping ratio among the expected damping ratios; a parameter optimization module for searching for a target parameter combination in a space formed by parameter combinations of the plurality of grid-type converters based on the optimization objective, the attraction search mechanism, the adaptive step size mechanism, and the adversarial learning mechanism of the target parameter optimization model; An optimal parameter deployment module is used to deploy the target parameter combination to multiple grid-type converters.

9. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor is configured to execute the computer program to implement the steps of the parameter optimization method for multiple grid-type converters according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that Used to store a computer program; wherein, when the computer program is executed by a processor, the steps of the parameter optimization method for multiple grid-type converters according to any one of claims 1 to 7 are implemented.

Images