Particle swarm and nonlinear programming hybrid optimization vsg adaptive control method and system

Abstract

The application discloses a kind of particle swarm and nonlinear programming hybrid optimization VSG adaptive control method and system, it is related to virtual synchronous generator control technical field, including acquisition filter inductance current and capacitor voltage signal, output instantaneous active power and reactive power, construct the small signal dynamic model of VSG, establish parameter mapping relationship;Based on parameter mapping relationship combining power system second-order oscillation mode stability criterion to set constraint range;According to constraint range, construct hybrid optimization algorithm, output steady-state data;Construct double-dimension threshold criterion system, through governor and controller collaborative generation voltage reference value;Voltage reference value is compared with actual output voltage, by PI control and PWM modulation drive inverter output.The method described in the application improves the precision of parameter optimization, guarantees the stability of system, and improves the calculation efficiency of algorithm.

Description

Technical Field

[0001] This invention relates to the field of virtual synchronous generator control technology, specifically to a hybrid optimization VSG adaptive control method and system combining particle swarm optimization and nonlinear programming. Background Technology

[0002] Virtual synchronous generators (VSGs), as a core technology for new energy grid integration, effectively improve the frequency stability and dynamic response of grids with a high proportion of renewable energy by simulating the inertia and damping characteristics of traditional synchronous generators. The virtual inertia J and damping coefficient D are key parameters determining VSG performance: J affects the system's inertial support capability, and D determines the suppression effect of frequency fluctuations. However, existing VSG control methods generally employ fixed parameters or offline optimization strategies, making it difficult to adapt to complex operating conditions such as sudden changes in grid load and intermittent new energy output, resulting in insufficient system performance in areas such as frequency deviation suppression and oscillation decay rate.

[0003] Currently, research on VSG parameter optimization mainly focuses on single intelligent algorithms or traditional optimization methods. For example, the Particle Swarm Optimization (PSO) algorithm is used in parameter optimization due to its global search capability, but it suffers from premature convergence and insufficient search accuracy in the later stages. While Nonlinear Programming (NLP) algorithms can utilize gradient information for local fine-tuning, they rely on initial values ​​and have low efficiency in solving high-dimensional non-convex problems. Furthermore, existing technologies lack effective dynamic adaptive mechanisms, failing to automatically trigger parameter optimization based on real-time grid operating conditions (such as frequency deviation and power fluctuations). This leads to a disconnect between the optimization process and actual operating conditions, making precise control difficult to achieve in engineering applications.

[0004] In practical engineering, VSG parameter optimization needs to simultaneously consider frequency stability, system oscillation suppression, and parameter physical constraints. Traditional single algorithms struggle to balance these multiple objectives. Furthermore, high-proportion renewable energy power grids place higher demands on the real-time performance and robustness of VSGs, necessitating a hybrid optimization algorithm that combines global search and local refinement with dynamic adaptive capabilities to achieve online real-time optimization of J and D, thereby improving the stability and reliability of power grid operation. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by this invention is that existing VSG parameter optimization methods suffer from premature convergence, insufficient search accuracy in the later stages, and low solution efficiency, as well as the problem of how to achieve an effective dynamic adaptive mechanism.

[0007] To address the aforementioned technical problems, this invention provides the following technical solution: a hybrid optimization method for VSG (Variable Reactive Power Grid) control using particle swarm optimization and nonlinear programming, comprising: acquiring filtered inductor current and capacitor voltage signals; outputting instantaneous active and reactive power; constructing a small-signal dynamic model of the VSG; establishing parameter mapping relationships; setting constraint ranges based on the parameter mapping relationships and the second-order oscillation mode stability criterion of the power system; constructing a hybrid optimization algorithm based on the constraint range; outputting steady-state data; constructing a two-dimensional threshold criterion system; generating voltage reference values ​​through the collaboration of the speed controller and the speed governor; comparing the voltage reference values ​​with the actual output voltage; and driving the inverter output through PI control and PWM modulation.

[0008] As a preferred embodiment of the particle swarm optimization and nonlinear programming hybrid optimization VSG adaptive control method described in this invention, the construction of the small-signal dynamic model of the VSG includes deriving the second-order transfer function of the active power control loop, and establishing the parameter mapping relationship between the system damping ratio and the second-order natural angular frequency and the moment of inertia and damping coefficient based on the pole placement characteristics of the second-order transfer function.

[0009] As a preferred embodiment of the particle swarm optimization and nonlinear programming hybrid optimization VSG adaptive control method described in this invention, the set constraint range includes, based on parameter mapping relationship combined with the second-order oscillation mode stability criterion of power system, including damping ratio coefficient and dynamic response constraint, simultaneous parameter space constraint equations of moment of inertia and damping coefficient, and the range of values ​​for output constraint.

[0010] As a preferred embodiment of the hybrid optimization VSG adaptive control method of particle swarm optimization and nonlinear programming described in this invention, the construction of the hybrid optimization algorithm includes: based on the range of values ​​of moment of inertia and damping coefficient, using a hybrid optimization algorithm of particle swarm optimization and sequential quadratic programming, combining the global search capability of the particle swarm algorithm and the local fine-tuning capability of the nonlinear programming algorithm, and using the parameters under steady-state conditions as optimization variables for multi-objective optimization.

[0011] As a preferred embodiment of the hybrid optimization VSG adaptive control method of particle swarm optimization and nonlinear programming described in this invention, the parameters under steady-state conditions as optimization variables include the moment of inertia, damping coefficient, and angular velocity change rate threshold under steady-state conditions. A VSG mathematical model is constructed, the parameter optimization range is set, and a multi-objective fitness function is defined. Weight coefficients are initialized, and then the PSO algorithm is initialized. PSO parameters are set, and an initial particle swarm output fitness value is generated in the parameter space. During the PSO iterative optimization stage, particle velocity and position are updated, a new fitness is output, and the optimal solution is updated. SQP local optimization is triggered, a quadratic programming subproblem is constructed, the Hessian matrix is ​​updated using the BFGS method, and the optimal search direction is solved. Whether to re-optimize is determined based on frequency deviation, power fluctuation, or time interval. When the termination condition is met, the optimal parameters are output and the process ends.

[0012] As a preferred embodiment of the VSG adaptive control method of hybrid optimization of particle swarm optimization and nonlinear programming described in this invention, the PSO algorithm initialization includes: initializing the parameters of the particle swarm optimization algorithm, including population size, maximum number of iterations, inertia weight range and learning factor; generating an initial particle swarm in the parameter space and initializing the particle velocity and position; outputting the fitness value of each particle; and recording the individual optimal solution and the global optimal solution.

[0013] As a preferred embodiment of the VSG adaptive control method of hybrid optimization of particle swarm optimization and nonlinear programming described in this invention, the triggering of SQP local optimization includes: determining whether the SQP triggering condition is met according to the POS optimization process, such as the number of PSO iterations reaching an integer multiple of a preset value or the improvement of the global optimal solution stagnating during continuous iteration; if the triggering condition is met, SQP local optimization is performed using the global optimal solution output by POS optimization as the initial value.

[0014] As a preferred embodiment of the particle swarm optimization and nonlinear programming hybrid optimization VSG adaptive control method described in this invention, the collaboratively generated voltage reference value includes: constructing a two-dimensional threshold criterion system based on the output steady-state optimal moment of inertia and damping coefficient and a parameter threshold for distinguishing the magnitude of the change in angular velocity rate; determining the real-time moment of inertia and damping coefficient; and generating the voltage setpoint through the collaborative generation of the virtual speed controller and the excitation controller.

[0015] As a preferred embodiment of the hybrid optimization VSG adaptive control method of particle swarm optimization and nonlinear programming described in this invention, the output of the drive inverter includes: based on the generated voltage setpoint, as the voltage outer loop reference value, compared with the actual output voltage, a current inner loop reference value is generated by the voltage PI controller, then compared with the actual output current and a modulation signal is generated by the current PI controller, a three-phase modulation wave is output after dq inverse transformation, a PWM signal is generated by comparing with a high-frequency triangular carrier wave, and amplified by the drive circuit to control the on / off state of the three-phase inverter full-bridge switch.

[0016] Another objective of this invention is to provide a hybrid optimization VSG adaptive control system that combines particle swarm optimization and nonlinear programming. This system can set the constraint range based on parameter mapping relationships and the stability criterion of the second-order oscillation mode of the power system, thus solving the problem of low computational accuracy in current VSG parameter optimization.

[0017] As a preferred embodiment of the VSG adaptive control system based on the hybrid optimization of particle swarm optimization and nonlinear programming described in this invention, the system comprises: an optimization output module, a parameter adjustment module, and an inverter control module. The optimization output module, based on a three-phase full-bridge inverter topology, acquires the current signal of the filter inductor and the voltage signal of the filter capacitor, obtains the instantaneous active and reactive power outputs of the system through Park transform decoupling calculation, constructs a small-signal dynamic model of the VSG control strategy, and derives the second-order transfer function of the active power control loop. Based on the pole placement characteristics of the second-order transfer function, it establishes a parameter mapping relationship between the system damping ratio and the second-order natural angular frequency, and the moment of inertia and damping coefficient. Based on the established parameter mapping relationship, combined with the damping ratio coefficient and dynamic response constraints of the second-order oscillation mode stability criterion of the power system, it simultaneously establishes the parameter space constraint equations for the moment of inertia and damping coefficient, outputs the value range, and constructs a hybrid optimization algorithm of particle swarm optimization and sequential quadratic programming for multi-objective global parameter optimization. The parameter adjustment module... A dual-dimensional threshold criterion system is constructed based on the numerical values ​​output by the global optimization module: defining an angular velocity deviation threshold and establishing a graded threshold for the rate of change of angular velocity, subdividing the angular velocity change and rate of change intervals, determining the real-time moment of inertia and damping coefficient, and generating a voltage setpoint through the synergistic action of a virtual speed controller and an excitation controller; the inverter control module uses the voltage setpoint generated by the parameter adjustment module as the voltage outer loop reference value, compares it with the actual output voltage, generates a current inner loop reference value through a voltage PI controller, compares it with the actual output current, generates a modulation signal through a current PI controller, obtains a three-phase modulation wave through dq inverse transformation, and finally compares it with a high-frequency triangular carrier wave to generate a PWM signal, which is amplified by the drive circuit and controls the on / off switching of the three-phase inverter full-bridge switch. A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a hybrid optimization VSG adaptive control method of particle swarm optimization and nonlinear programming.

[0018] A computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the steps of a hybrid optimization VSG adaptive control method combining particle swarm optimization and nonlinear programming.

[0019] The beneficial effects of this invention are as follows: The particle swarm optimization and nonlinear programming hybrid optimization VSG adaptive control method provided by this invention collects the current and voltage signals of the filtered inductor and capacitor, outputs instantaneous active and reactive power, constructs a small-signal dynamic model of VSG, establishes parameter mapping relationships, improves the accuracy of parameter optimization, and ensures system stability. Based on the parameter mapping relationships and combined with the stability criterion of the second-order oscillation mode of the power system, a constraint range is set to ensure physical feasibility and improve the computational efficiency of the algorithm. A hybrid optimization algorithm is constructed according to the constraint range, outputs steady-state data, and global optimization and local optimization work together to achieve dynamic balance of multiple objectives. A two-dimensional threshold criterion system is constructed, and a voltage reference value is generated through the collaboration of the speed controller and the controller, which dynamically responds and accurately adapts to enhance anti-interference. The voltage reference value is compared with the actual output voltage, and the inverter output is driven by PI control and PWM modulation. Closed-loop control improves accuracy and the dynamic response speed is faster. This invention achieves better results in terms of low computational accuracy, low solution efficiency, and dynamic adaptation. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0021] Figure 1 The first embodiment of the present invention provides an overall flowchart of a hybrid optimization VSG adaptive control method combining particle swarm optimization and nonlinear programming.

[0022] Figure 2 The flowchart of the hybrid algorithm of particle swarm optimization and nonlinear programming for the first embodiment of the present invention is shown.

[0023] Figure 3 The diagram below illustrates the structural principle of a VSG parameter adaptive system based on an algorithm-improved hybrid optimization VSG adaptive control method that combines particle swarm optimization and nonlinear programming, as provided in the second embodiment of the present invention. Detailed Implementation

[0024] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0025] Example 1, referring to Figures 1-2 As an embodiment of the present invention, a hybrid optimization VSG adaptive control method combining particle swarm optimization and nonlinear programming is provided, comprising:

[0026] S1: Acquire the current and voltage signals of the filter inductor and capacitor, output instantaneous active and reactive power, construct a small-signal dynamic model of VSG, and establish parameter mapping relationships.

[0027] Furthermore, the construction of the small-signal dynamic model of VSG includes deriving the second-order transfer function of the active power control loop, and establishing the parameter mapping relationship between the system damping ratio and the second-order natural angular frequency and the moment of inertia and damping coefficient based on the pole placement characteristics of the second-order transfer function.

[0028] It should be noted that, based on the three-phase full-bridge inverter topology, the filter inductor current signal and filter capacitor voltage signal are collected. The instantaneous active power P and reactive power Q of the system output are obtained through Park transformation (abc-dq coordinate transformation) decoupling calculation. Then, a small-signal dynamic model of the VSG is constructed, and the parameter mapping relationship between the system damping ratio and second-order natural angular frequency and the moment of inertia and damping coefficient is established, expressed as:

[0029]

[0030] Where ξ is the damping ratio, ω n Let E be the second-order natural angular frequency, E be the input port voltage, U be the output port voltage, J be the moment of inertia, ω0 be the rated angular velocity, Z be the equivalent impedance, and K be the second-order natural angular frequency. ω denoted as the power-frequency droop factor, and D as the damping factor.

[0031] It should also be noted that by collecting electrical signals and performing decoupling calculations, a small-signal dynamic model of the VSG is constructed, and a parameter mapping relationship is established. This solves the problem of mismatch between dynamic response and actual operating conditions caused by fixed parameters in traditional VSG control, achieves the requirements of precise matching and grid stability, ensures adaptive adjustment of parameters, and significantly improves anti-interference capability.

[0032] S2: Set the constraint range based on the parameter mapping relationship and the second-order oscillation mode stability criterion of the power system.

[0033] Furthermore, the constraint range is defined by combining the parameter mapping relationship with the second-order oscillation mode stability criterion of the power system, including the damping ratio coefficient and dynamic response constraints, the parameter space constraint equations of the moment of inertia and damping coefficient are combined, and the range of values ​​for the output constraints is determined.

[0034] It should be noted that, based on establishing the parameter mapping relationship between the system damping ratio and second-order natural angular frequency with the moment of inertia and damping coefficient, and combining the damping ratio coefficient and dynamic response constraints of the power system, the parameter space constraint equations for the moment of inertia and damping coefficient can be simultaneously established to determine the range of values, expressed as:

[0035]

[0036] Among them, J min J is the lower limit of the moment of inertia. max ξ is the upper limit of the moment of inertia. min For the minimum allowable damping ratio, ξ max For the maximum allowable damping ratio, K d For the damping compensation coefficient, t max K is the maximum allowable adjustment time. P This is the power-frequency droop factor.

[0037] It should also be noted that by combining parameter mapping relationships and the stability criteria of second-order oscillation modes of power systems, parameter space constraint equations are established to solve the problem of insufficient stability and lag in dynamic response caused by the lack of theoretical constraints in traditional VSG parameter design. This ensures that the parameters meet the range of system stability, provides clear boundary search for hybrid optimization algorithms, improves computational efficiency, and significantly enhances the dynamics of VSG under complex operating conditions.

[0038] S3: Construct a hybrid optimization algorithm based on the constraint range and output steady-state data.

[0039] Furthermore, the construction of the hybrid optimization algorithm includes, based on the range of values ​​for moment of inertia and damping coefficient, a hybrid optimization algorithm based on particle swarm optimization and sequential quadratic programming, combining the global search capability of particle swarm optimization and the local fine-tuning capability of nonlinear programming algorithm, using parameters under steady-state conditions as optimization variables for multi-objective optimization.

[0040] The hybrid algorithm combining particle swarm optimization and nonlinear programming uses a weighted function of steady-state frequency offset, settling time, and power oscillation as the objective function, which is expressed as:

[0041] F=α·Δf ss +β·t s +γ·∫|ΔP|dt

[0042] Where F is the objective function, α is the weighting coefficient for steady-state frequency offset, and Δf ss For steady-state frequency deviation, β is the weighting coefficient for settling time, and t s To adjust the time, γ is the weighting coefficient of the power oscillation, ΔP is the power oscillation, and dt is the integral variable.

[0043] It should be noted that the parameters used as optimization variables under steady-state conditions include the moment of inertia, damping coefficient, and angular velocity change rate threshold under steady-state conditions. A VSG mathematical model is constructed, the parameter optimization range is set, and a multi-objective fitness function is defined. The weight coefficients are initialized, and then the PSO algorithm is initialized. The PSO parameters are set, and an initial particle swarm output fitness value is generated in the parameter space. During the PSO iterative optimization phase, the particle velocity and position are updated, a new fitness is output, and the optimal solution is updated. SQP local optimization is triggered, a quadratic programming subproblem is constructed, the Hessian matrix is ​​updated using the BFGS method, and the optimal search direction is solved. Whether to re-optimize is determined based on frequency deviation, power fluctuation, or time interval. When the termination condition is met, the optimal parameters are output and the process ends.

[0044] The VSG mathematical model is constructed by including inertial elements, damping elements, and droop control equations, setting the parameter optimization range, defining a multi-objective fitness function, and initializing weight coefficients.

[0045] It should also be noted that the PSO algorithm initialization includes initializing the parameters of the particle swarm optimization algorithm, including the population size, maximum number of iterations, inertia weight range and learning factor, generating an initial particle swarm in the parameter space, initializing the particle velocity and position, outputting the fitness value of each particle, and recording the individual optimal solution and the global optimal solution.

[0046] Iterative optimization is performed, the inertia weights are updated, the velocity and position of each particle are adjusted and updated, and boundary treatment is performed to ensure that the moment of inertia and damping coefficient are within the constraints. The fitness value of the updated particles is calculated, and the individual optimal solution and the global optimal solution are updated.

[0047] It should also be noted that triggering SQP local optimization includes determining whether the SQP triggering conditions are met based on the POS optimization process, such as the number of PSO iterations reaching an integer multiple of the preset value or the improvement of the global optimal solution stagnating during continuous iterations. If the triggering conditions are met, SQP local optimization is performed using the global optimal solution output by POS optimization as the initial value.

[0048] Based on the optimized solution, a quadratic programming subproblem is constructed. The Hessian matrix is ​​iteratively updated using the BFGS method to find the optimal search direction. The solution vector is updated through search compensation, and the convergence condition of the SQP is checked. If it is met, the optimal solution is output; otherwise, the iteration continues. The optimized parameters are applied to the VSG controller, and feedback data such as grid frequency and power are collected in real time to determine whether re-optimization is needed. If the frequency deviation exceeds the standard, the power fluctuation exceeds the limit, or the preset time interval is reached, the optimization process is restarted. When the termination condition is met, the maximum number of iterations is reached, or the fitness improvement stagnates, the final optimized parameters are output, and the parameter adaptive control process ends. Figure 2The diagram illustrates the overall process of the hybrid particle swarm optimization and nonlinear programming algorithm described above.

[0049] It should also be noted that by constructing a VSG mathematical model, setting the parameter optimization range, updating and iterating the parameters, and completing the parameter adaptive control process, the problem that traditional fixed parameters are not suitable for complex power grid conditions can be solved. This enables dynamic adaptive optimization of VSG parameters, significantly improving the requirements for rapid response and oscillation suppression in dynamic balance, and enhancing the system's anti-interference and robustness.

[0050] S4: Construct a two-dimensional threshold criterion system and generate voltage reference values ​​through the collaboration of speed regulator and controller.

[0051] Furthermore, the collaboratively generated voltage reference value includes a two-dimensional threshold criterion system based on the output steady-state optimal moment of inertia and damping coefficient, as well as a parameter threshold that distinguishes the magnitude of changes in the rate of change of angular velocity. This system determines the real-time moment of inertia and damping coefficient, and generates the voltage setpoint through the collaborative generation of the virtual speed controller and excitation controller.

[0052] It should be noted that a two-dimensional threshold criterion system is constructed, defining an angular velocity deviation threshold and establishing a graded threshold for the rate of change of angular velocity. This subdivides the angular velocity change and rate of change ranges, determines the real-time moment of inertia and damping coefficient, and adaptively adjusts them as follows:

[0053]

[0054]

[0055] Where sign(x) is the sign function, x is any real number, J is the moment of inertia, J0 is the moment of inertia when the system is stable, arctan is the arctangent function, and Δω is... Let dω be the differential of angular velocity, dt be the differential of time, D be the damping coefficient, and D0 be the damping coefficient when the system is stable. min D is the minimum damping coefficient. max This is the maximum damping coefficient.

[0056] It should also be noted that by constructing a dual-dimensional threshold criterion system, and coordinating the virtual speed controller and excitation controller to generate voltage reference values, the problem of dynamic response hysteresis caused by fixed parameters in traditional VSG control is solved. This enables precise differentiation of disturbance levels using dual thresholds, and combined with the arctangent function, parameter transitions are achieved without jumps, significantly improving frequency stability and oscillation suppression.

[0057] S5: By comparing the voltage reference value with the actual output voltage, the inverter output is driven by PI control and PWM modulation.

[0058] Furthermore, the inverter output includes: based on the generated voltage setpoint, which serves as the voltage outer loop reference value, it is compared with the actual output voltage and then used by the voltage PI controller to generate the current inner loop reference value. This value is then compared with the actual output current and used by the current PI controller to generate a modulation signal. After dq inverse transformation, a three-phase modulation wave is output. This wave is compared with a high-frequency triangular carrier wave to generate a PWM signal. After being amplified by the drive circuit, the PWM signal controls the on / off state of the three-phase inverter full-bridge switch.

[0059] It should be noted that by using a dual closed-loop PI control structure, the voltage reference value is compared with the actual value to generate a modulation signal, which solves the problems of insufficient accuracy and dynamic response lag in traditional open-loop control, realizes voltage-current dual-loop coordinated control, improves the tracking accuracy and waveform quality of output voltage and current, and enhances the requirements of power grid power quality.

[0060] Example 2, refer to Figure 3 As an embodiment of the present invention, a hybrid optimization VSG adaptive control system combining particle swarm optimization and nonlinear programming is provided, including an optimization output module, a parameter adjustment module, and an inverter control module.

[0061] S6: The optimized output module 100 is used to collect the current signal of the filter inductor and the voltage signal of the filter capacitor based on the three-phase full-bridge inverter topology, obtain the instantaneous active power and reactive power of the system output through Park transform decoupling calculation, construct a small-signal dynamic model of the VSG control strategy, and derive the second-order transfer function of the active power control loop; based on the pole placement characteristics of the second-order transfer function, establish the parameter mapping relationship between the system damping ratio and the second-order natural angular frequency and the moment of inertia and damping coefficient; based on the established parameter mapping relationship, combined with the damping ratio coefficient of the second-order oscillation mode stability criterion of the power system and the dynamic response constraint, simultaneously solve the parameter space constraint equation of the moment of inertia and the damping coefficient, output the value range, construct a hybrid optimization algorithm of particle swarm optimization and sequential quadratic programming, and perform multi-objective global parameter optimization.

[0062] S7: The parameter adjustment module 200 is used to construct a two-dimensional threshold criterion system based on the value output by the global optimization module 100: define the angular velocity deviation threshold and establish the angular velocity change rate classification threshold, subdivide the angular velocity change and angular velocity change rate intervals, determine the real-time rotational inertia and damping coefficient, and generate the voltage setpoint through the synergistic effect of the virtual speed controller and the excitation controller.

[0063] In S8, the inverter control module 300 uses the voltage setpoint generated by the parameter adjustment module 200 as the voltage outer loop reference value. After comparing it with the actual output voltage, it generates the current inner loop reference value through the voltage PI controller, compares it with the actual output current, and generates a modulation signal through the current PI controller. After dq inverse transformation, a three-phase modulation wave is obtained. Finally, it is compared with a high-frequency triangular carrier wave to generate a PWM signal. After being amplified by the drive circuit, it controls the on / off state of the three-phase inverter full-bridge switch.

Claims

1. A hybrid optimization VSG adaptive control method combining particle swarm optimization and nonlinear programming, characterized in that, include: Collect the current and voltage signals of the filter inductor and capacitor, output the instantaneous active power and reactive power, construct the small-signal dynamic model of VSG, and establish the parameter mapping relationship; The constraint range is set based on the parameter mapping relationship and the second-order oscillation mode stability criterion of the power system. A hybrid optimization algorithm is constructed based on the constraints, and steady-state data is output. The construction of the hybrid optimization algorithm includes, based on the range of values ​​of moment of inertia and damping coefficient, a hybrid optimization algorithm based on particle swarm optimization and sequential quadratic programming, combining the global search capability of particle swarm optimization and the local fine-tuning capability of nonlinear programming algorithm, and using parameters under steady-state conditions as optimization variables for multi-objective optimization. The parameters under steady-state conditions used as optimization variables include the moment of inertia, damping coefficient, and angular velocity change rate threshold under steady-state conditions. A VSG mathematical model is constructed, the parameter optimization range is set, and a multi-objective fitness function is defined. The weight coefficients are initialized, and then the PSO algorithm is initialized. The PSO parameters are set, and an initial particle swarm output fitness value is generated in the parameter space. During the PSO iterative optimization phase, the particle velocity and position are updated, a new fitness is output, and the optimal solution is updated. SQP local optimization is triggered, a quadratic programming subproblem is constructed, the Hessian matrix is ​​updated using the BFGS method, and the optimal search direction is solved. Whether to re-optimize is determined based on frequency deviation, power fluctuation, or time interval. When the termination condition is met, the optimal parameters are output and the process ends. A two-dimensional threshold criterion system is constructed, and a voltage reference value is generated through the collaboration of the speed controller and the speed regulator. By comparing the voltage reference value with the actual output voltage, the inverter output is driven through PI control and PWM modulation. The dual-dimensional threshold criterion system includes defining angular velocity deviation threshold and angular velocity change rate classification threshold, subdividing angular velocity change and angular velocity change rate ranges, and dynamically adjusting parameters.

2. The VSG adaptive control method based on a hybrid optimization of particle swarm optimization and nonlinear programming as described in claim 1, characterized in that: The construction of the small-signal dynamic model of VSG includes... The second-order transfer function of the active power control loop is derived. Based on the pole placement characteristics of the second-order transfer function, the parameter mapping relationship between the system damping ratio, the second-order natural angular frequency, the moment of inertia, and the damping coefficient is established.

3. The VSG adaptive control method based on a hybrid optimization of particle swarm optimization and nonlinear programming as described in claim 1, characterized in that: The defined constraint range includes, Based on the parameter mapping relationship and the second-order oscillation mode stability criterion of the power system, including the damping ratio coefficient and dynamic response constraints, the parameter space constraint equations of the moment of inertia and damping coefficient are combined to output the range of values ​​for the constraints.

4. The VSG adaptive control method based on a hybrid optimization of particle swarm optimization and nonlinear programming as described in claim 1, characterized in that: The PSO algorithm initialization includes, Initialize the parameters of the particle swarm optimization algorithm, including population size, maximum number of iterations, inertia weight range and learning factor. Generate an initial particle swarm in the parameter space and initialize the particle velocity and position. Output the fitness value of each particle and record the individual optimal solution and the global optimal solution.

5. The VSG adaptive control method based on a hybrid optimization of particle swarm optimization and nonlinear programming as described in claim 4, characterized in that: The triggering of SQP local optimization includes... Based on the POS optimization process, determine whether the SQP triggering condition is met: the number of PSO iterations reaches an integer multiple of the preset value or the improvement of the global optimal solution stagnates during continuous iteration. If the triggering condition is met, use the global optimal solution output by POS optimization as the initial value for SQP local optimization.

6. The VSG adaptive control method based on a hybrid optimization of particle swarm optimization and nonlinear programming as described in claim 1, characterized in that: The collaboratively generated voltage reference value includes, Based on the output steady-state optimal moment of inertia and damping coefficient, as well as the parameter threshold for distinguishing the magnitude of changes in the rate of change of angular velocity, a two-dimensional threshold criterion system is constructed to determine the real-time moment of inertia and damping coefficient. The voltage setpoint is generated through the collaboration of the virtual speed controller and the excitation controller.

7. The VSG adaptive control method based on a hybrid optimization of particle swarm optimization and nonlinear programming as described in claim 1, characterized in that: The output of the drive inverter includes, Based on the generated voltage setpoint, it serves as the voltage outer loop reference value. After being compared with the actual output voltage, it generates the current inner loop reference value through the voltage PI controller. Then, it is compared with the actual output current and generates a modulation signal through the current PI controller. After dq inverse transformation, it outputs a three-phase modulation wave, which is compared with a high-frequency triangular carrier wave to generate a PWM signal. After being amplified by the drive circuit, it controls the on / off state of the full-bridge switch of the three-phase inverter.

8. A hybrid optimization VSG adaptive control system based on particle swarm optimization and nonlinear programming, employing the hybrid optimization VSG adaptive control method based on particle swarm optimization and nonlinear programming as described in any one of claims 1 to 7, characterized in that: Includes an optimized output module (100), a parameter adjustment module (200), and an inverter control module (300); The optimized output module (100) is used to collect the current signal of the filter inductor and the voltage signal of the filter capacitor based on the three-phase full-bridge inverter topology, obtain the instantaneous active power and reactive power of the system output through Park transform decoupling calculation, construct the small-signal dynamic model of the VSG control strategy, and derive the second-order transfer function of the active power control loop. Based on the pole placement characteristics of the second-order transfer function, the parameter mapping relationship between the system damping ratio and the second-order natural angular frequency and the moment of inertia and damping coefficient is established. Based on the established parameter mapping relationship, combined with the damping ratio coefficient and dynamic response constraints of the second-order oscillation mode stability criterion of the power system, the parameter space constraint equations of the moment of inertia and damping coefficient are solved simultaneously, the value range is output, and a hybrid optimization algorithm of particle swarm optimization and sequential quadratic programming is constructed to perform global optimization of multi-objective parameters. The parameter adjustment module (200) is used to construct a two-dimensional threshold criterion system based on the value output by the global optimization module (100): define the angular velocity deviation threshold and establish the angular velocity change rate classification threshold, subdivide the angular velocity change and angular velocity change rate range, determine the real-time rotational inertia and damping coefficient, and generate the voltage setpoint through the synergistic effect of the virtual speed controller and the excitation controller. The inverter control module (300) is used to take the voltage setpoint generated by the parameter adjustment module (200) as the voltage outer loop reference value, compare it with the actual output voltage, generate the current inner loop reference value through the voltage PI controller, compare it with the actual output current, generate the modulation signal through the current PI controller, obtain the three-phase modulation wave through dq inverse transformation, and finally compare it with the high-frequency triangular carrier to generate the PWM signal. After being amplified by the drive circuit, it controls the on and off of the three-phase inverter full bridge switch.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the particle swarm optimization and nonlinear programming hybrid optimization VSG adaptive control method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the VSG adaptive control method of the hybrid optimization of particle swarm optimization and nonlinear programming as described in any one of claims 1 to 7.

Images