Parameter optimization method and system based on multi-virtual synchronous generator parallel grid-connected system

Abstract

The present invention is a parameter optimization method and system based on a multi-virtual synchronous generator parallel grid-connected system, and the method includes the following steps: S1, Lyapunov indirect method criterion for a small signal model of a multi-virtual synchronous generator parallel grid-connected system containing a control link; S2, calculating the damping ratio and real characteristic root according to the small signal state matrix; S3, constructing a dynamic evaluation function according to the stability and dynamic characteristics of the system; S4, assigning weights to the dynamic evaluation function using the entropy weight method; S5, using the quasi-annealing particle swarm algorithm to determine the optimal solution of the evaluation function. The present invention constructs a dynamic performance evaluation function for a multi-virtual synchronous generator parallel grid-connected system to better evaluate the impact of parameters on the system; uses the entropy weight method to construct weights to avoid a complex trial and error process; uses the artificial intelligence algorithm quasi-annealing particle swarm algorithm to solve the dynamic performance evaluation function, which is simpler and more effective, and better prevents the solution process from falling into a local optimal situation.

Description

Technical Field

[0001] The present invention relates to the technical field of parameter optimization, and in particular to a parameter optimization method and system based on a multi-virtual synchronous generator parallel grid-connected system. Background Art

[0002] Distributed power sources are generally connected to the grid through voltage source converters. Under the early power control strategy, voltage source converters hardly participate in grid regulation and cannot provide voltage and frequency support for the grid when the system is disturbed, which reduces the safe and stable operation capability of the grid. In this regard, the virtual synchronous generator control technology is based on the duality of the mathematical model of the voltage source converter and the mechanical equations and electromagnetic equations of the traditional synchronous generator, so that the voltage source converter has inertia and damping characteristics similar to those of the synchronous generator, which can provide frequency and voltage support for the grid and improve the stability and robustness of the power system with a high proportion of distributed power sources.

[0003] The multi-virtual synchronous generator system faces the problem of strong coupling of multiple control loops, presents more complex dynamic characteristics in a wide frequency band, and has many control parameters, which brings great difficulties to parameter optimization. Therefore, it is urgent to analyze the dynamic performance of the multi-virtual synchronous generator grid-connected system in view of the difficulty of debugging and setting the parameters of the multi-virtual synchronous generator small signal model, and optimize the parameters by combining intelligent optimization algorithm. Summary of the invention

[0004] In order to solve the technical problems existing in the prior art, the present invention provides a parameter optimization method and system based on a multi-virtual synchronous generator parallel grid-connected system, which is used to overcome the problem that the control parameters in multiple control loops in the multi-virtual synchronous generator parallel grid-connected system are numerous and difficult to adjust. By constructing a dynamic evaluation function of the system, it is possible to simultaneously perform a global performance evaluation on multiple control parameters of the system and optimize the control parameters. At the same time, in the dynamic evaluation function of multiple objectives, the entropy weight method is used to make it easier to assign weights to the multi-objective functions and avoid a tedious trial and error process.

[0005] The method of the present invention is implemented by the following technical solution: a parameter optimization method based on a multi-virtual synchronous generator parallel grid-connected system, comprising the following steps:

[0006] S1. The Lyapunov indirect method is used to judge the small signal model of the multi-virtual synchronous generator parallel grid-connected system with control links;

[0007] S2, calculating the damping ratio and real characteristic roots according to the small signal state matrix;

[0008] S3, construct a dynamic evaluation function according to the stability and dynamic characteristics of the system;

[0009] S4, using entropy weight method to assign weights to dynamic evaluation functions;

[0010] S5. Use the quasi-annealing particle swarm algorithm to determine the optimal solution of the evaluation function. If the quasi-annealing particle swarm algorithm determines convergence, the optimized control parameters are output. Otherwise, a new set of parameters is obtained and the step S1 is re-entered.

[0011] The system of the present invention adopts the following technical solutions to achieve: a parameter optimization system based on a multi-virtual synchronous generator parallel grid-connected system, comprising:

[0012] Criteria module: used to perform Lyapunov indirect method criterion on the small signal model of multi-virtual synchronous generator parallel grid-connected system with control link;

[0013] Damping ratio and real characteristic root acquisition module: calculate the damping ratio and real characteristic root through the small signal state matrix;

[0014] Dynamic evaluation function construction module: constructs a dynamic evaluation function based on the stability and dynamic characteristics of the system;

[0015] Dynamic evaluation function weight acquisition module: assigning dynamic evaluation function weights by using entropy weight method;

[0016] The optimal solution judgment module of the evaluation function: the optimal solution of the evaluation function is judged by using the quasi-annealing particle swarm algorithm.

[0017] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0018] 1. The present invention can better evaluate the influence of parameters on the system by constructing a dynamic performance evaluation function for a multi-virtual synchronous generator parallel grid-connected system; at the same time, the entropy weight method is used to construct weights to avoid a complicated trial-and-error process; the artificial intelligence algorithm quasi-annealing particle swarm algorithm is used to solve the dynamic performance evaluation function, which is simpler and more effective, and at the same time better prevents the solution process from falling into a local optimal situation.

[0019] 2. The present invention provides an effective optimization method for optimizing control parameters in a multi-virtual synchronous generator parallel grid-connected system, which is beneficial to the promotion and application of the multi-virtual synchronous generator parallel grid-connected system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flow chart of the method of the present invention;

[0021] Figure 2 is a control topology schematic diagram of a multi-virtual synchronous generator parallel grid-connected system of the present invention;

[0022] Figure 3 It is a comparison diagram of extreme values ​​before and after optimization of the multi-virtual synchronous generator parallel grid-connected system of the present invention. DETAILED DESCRIPTION

[0023] The present invention is further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0024] Example

[0025] like Figure 1 As shown, the parameter optimization method of the multi-virtual synchronous generator parallel grid-connected system in this embodiment includes the following steps:

[0026] S1. The Lyapunov indirect method is used to judge the small signal model of the multi-virtual synchronous generator parallel grid-connected system with control links;

[0027] S2, calculating the damping ratio and real characteristic roots according to the small signal state matrix;

[0028] S3, construct a dynamic evaluation function according to the stability and dynamic characteristics of the system;

[0029] S4. Use the entropy weight method to assign weights to dynamic evaluation functions to avoid complex trial and error;

[0030] S5. Use the quasi-annealing particle swarm algorithm to determine the optimal solution of the evaluation function. If the quasi-annealing particle swarm algorithm determines convergence, the optimized control parameters are output. Otherwise, a new set of parameters is obtained and the step S1 is re-entered.

[0031] In this embodiment, in step S1, the state variables of the multi-virtual synchronous generator parallel grid-connected system are grouped and numbered, the state space equations are combined, and linear expansion is performed to obtain a small signal dynamic model:

[0032]

[0033] in, is the differential of the system state variable after linearization, Δy is the output variable of the system after linearization, A sys is the state matrix of the small signal, B, C and D are the simplified matrices in the relevant state equations, u is the input of the controller, x is the state variable of the system, and y is the output of the system.

[0034] In this embodiment, the Lyapunov indirect method is used in step S1 to determine whether the asymptotic stability is maintained under the current parameters: if Re(λ) < 0, the system satisfies the small disturbance stability, otherwise, the system does not satisfy the small disturbance stability; where |λE-A sys |=0, where Re(λ) is the real part of the eigenvalue in complex coordinates, λ is the eigenvalue of the state matrix, and E is the unit matrix.

[0035] In this embodiment, the damping ratio calculation formula in step S2 is as follows:

[0036]

[0037] Where ζ is the damping ratio, α is the real part of the characteristic root, and β is the imaginary part of the characteristic root.

[0038] In this embodiment, in step S3, a dynamic performance evaluation function is constructed according to the characteristics of multiple virtual synchronous generators:

[0039]

[0040]

[0041] Where, f is the objective function, n is the number of real parts of characteristic roots, k is the counting constant, m is the number of damping ratios, i is the total number of virtual synchronous generators, and α k is the real part of the kth characteristic root, α * is the target value of the real part of the characteristic root, υ k is the weight of the kth α, κ k is the weight of the kth ζ, ζ best is the optimal damping ratio, ζ k is the damping ratio of the kth characteristic root, ζ Δk is the equivalent damping ratio of multiple virtual synchronous generators, j is an imaginary number, β k is the imaginary part of the characteristic root, D i is the i-th damping coefficient of the virtual synchronous generator, K fi is the active power droop coefficient of the virtual synchronous generator, J i is the inertia of the virtual synchronous generator, X i is the impedance of the ith link line, ω0 is the rated angular frequency, V0 and V g are the amplitude of the inverter output voltage and the amplitude of the common coupling voltage point, respectively, and ε is the weight of the equivalent damping of the multi-virtual synchronous generator. Specifically, ζ best It can be determined according to actual project requirements.

[0042] In this embodiment, in step S4, the entropy weight method is used to solve the weight of the real part α of the characteristic root and the damping ratio ζ in the dynamic evaluation function. The specific calculation formula is as follows:

[0043]

[0044]

[0045]

[0046]

[0047]

[0048]

[0049] Among them, b op is the correlation coefficient of the real part of the characteristic root, d op is the correlation coefficient of the damping ratio, α op is the real part of the characteristic root, ζ op is the damping ratio, τ p is the information entropy of the real part of the characteristic root, e p is the information entropy of the damping ratio, υ p is the weight of the pth α, and κ p is the weight of the pth ζ.

[0050] In this embodiment, in step S5, the particle swarm algorithm is used to solve the dynamic evaluation function, the control parameter combination is set to particles within the feasible domain, and the direction and speed are iteratively updated and calculated. The specific process is as follows:

[0051]

[0052] Among them, v id is the velocity of the particle, w is the weight of the velocity, r1 and r2 are random numbers between (0,1), and x id is the current position of the particle, c1 and c2 are learning factors, and p id is the optimal position that the current particle has experienced, p gd is the optimal position experienced by the global particle swarm, and the weight update formula is:

[0053] w(t)=w 1e -t×(w 1e -w 1s ) / MaxIteration (12)

[0054] Among them, w(t) is the weight value, w 1e is the maximum value of weight, w 1s is the minimum value of the weight, MaxIteration is the maximum number of iterations, and t is the current iteration number.

[0055] Specifically, to prevent the optimization process from falling into the local optimal solution, a small disturbance is randomly added to the initial solution to calculate the new state value. If the new state is better, it is accepted; if the new state is not good, a certain probability P is used to choose whether to accept the state; this criterion defines the condition of changing from state i to state i at a certain temperature T. sa Transition to state j sa The probability P T ij as follows:

[0056]

[0057] Among them, f(j sa ) represents the next state, f(i sa ) is the current state, B z is the Boltzmann constant and T is the temperature coefficient.

[0058] like Figure 2 The control topology of the virtual synchronous generator parallel grid-connected system is shown in Figure 2. f and C f are the filter capacitor and inductor of the VSG output port, r f is the equivalent resistance of the VSG filter inductor, r c and L c is the resistance and inductance of the equivalent connection line between VSG and PCC, e abc 、i labc is the port voltage and current of VSG, u oabc 、i oabc is the output voltage and current of VSG, u bABC is the PCC point voltage, r g and L g The equivalent resistance and equivalent inductance of the transmission line and transformer are obtained according to the system topology, and then the small signal general model including the control link can be obtained, and then the parameters can be optimized.

[0059] In order to verify the parameter optimization analysis method proposed in the present invention, a model of three virtual synchronous generators connected in parallel and grid-connected is built in the Matlab / Simulink environment. The dimension of the quasi-annealing particle swarm algorithm is set to 20, the temperature T is set to 90, and the Boltzmann constant B is set to 1. Z is set to 1, the maximum number of iterations is set to 100, the learning factor is set to 0.2, and the optimal damping is selected as ζ best =0.8, α*=40, ε=0.1. like Figure 3 As shown in the figure, it is the trend of extreme value change after optimization. Each group of conjugate characteristic roots is recorded as a group of modes. The change trend in the figure matches the set parameter value.

[0060] Based on the same inventive concept, the present invention provides a parameter optimization system based on a multi-virtual synchronous generator parallel grid-connected system, comprising:

[0061] Criteria module: used to perform Lyapunov indirect method criterion on the small signal model of multi-virtual synchronous generator parallel grid-connected system with control link;

[0062] Damping ratio and real characteristic root acquisition module: calculate the damping ratio and real characteristic root through the small signal state matrix;

[0063] Dynamic evaluation function construction module: constructs a dynamic evaluation function based on the stability and dynamic characteristics of the system;

[0064] Dynamic evaluation function weight acquisition module: assigning dynamic evaluation function weights by using entropy weight method;

[0065] The optimal solution judgment module of the evaluation function: the optimal solution of the evaluation function is judged by using the quasi-annealing particle swarm algorithm.

[0066] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A parameter optimization method based on a multi-virtual synchronous generator parallel grid-connected system, characterized in that: The following steps are involved: S1. The Lyapunov indirect method is used to judge the small signal model of the multi-virtual synchronous generator parallel grid-connected system with control links; S2, calculating the damping ratio and real characteristic roots according to the small signal state matrix; S3, construct a dynamic evaluation function according to the stability and dynamic characteristics of the system; S4, using entropy weight method to assign weights to dynamic evaluation functions; S5, using the quasi-annealing particle swarm algorithm to determine the optimal solution of the dynamic evaluation function, if the quasi-annealing particle swarm algorithm determines convergence, then output the optimized control parameters, otherwise, after obtaining a set of new parameters, re-enter step S1; In step S1, the state variables of the multi-virtual synchronous generator parallel grid-connected system are grouped and numbered, the state space equations are combined, and linear expansion is performed to obtain the small signal dynamic model: in, is the differential of the state variable of the linearized system, Δy is the output variable of the linearized system, A sys is the state matrix of the small signal, B, C and D are the simplified matrices in the relevant state equations, Δu is the input of the linearized controller, and Δx is the state variable of the linearized system; In step S1, the Lyapunov indirect method is used to determine the stability under the current parameters: if Re(λ)<0, the system satisfies the small disturbance stability; where |λE-A sys |=0, where Re(λ) is the real part of the eigenvalue in complex coordinates, λ is the eigenvalue of the state matrix, and E is the identity matrix; The damping ratio calculation formula in step S2 is as follows: Among them, ζ is the damping ratio, α is the real part of the characteristic root; β is the imaginary part of the characteristic root; In step S3, a dynamic evaluation function is constructed according to the characteristics of multiple virtual synchronous generators: Where, f is the objective function, n is the number of real parts of characteristic roots, k is the counting constant, m is the number of damping ratios, i is the total number of virtual synchronous generators, and α k is the real part of the kth characteristic root, α * is the target value of the real part of the characteristic root, υ k is the weight of the kth α, κ k is the weight of the kth ζ, ζ best is the optimal damping ratio, ζ k is the damping ratio of the kth eigenvalue, ζ Δk is the equivalent damping ratio of multiple virtual synchronous generators, j is an imaginary number, β k is the imaginary part of the characteristic root, D i is the i-th damping coefficient of the virtual synchronous generator, K fi is the active power droop coefficient of the virtual synchronous generator, J i is the inertia of the virtual synchronous generator, X i is the impedance of the ith link line, ω0 is the rated angular frequency, V0 and V g are the amplitude of the inverter output voltage and the amplitude of the common coupling voltage point, respectively, and ε is the weight of the equivalent damping of the multi-virtual synchronous generator; In step S5, the particle swarm algorithm is used to solve the dynamic evaluation function, the control parameter combination is set as particles within the feasible domain, and the direction and speed are iteratively updated and calculated. The specific process is as follows: Among them, v id is the velocity of the particle, w is the weight of the velocity, r1 and r2 are random numbers between (0,1), and x id is the current position of the particle, c1 and c2 are learning factors, and p id is the optimal position that the current particle has experienced, p gd is the optimal position experienced by the global particle swarm; the weight update formula is: in(t)=in 1e -t×(in 1e -In 1s ) / MaxIteration (12) Among them, w(t) is the weight value, w 1e is the maximum value of weight, w 1s is the minimum value of the weight, MaxIteration is the maximum number of iterations, and t is the current iteration number.

2. The parameter optimization method based on the multi-virtual synchronous generator parallel grid-connected system according to claim 1 is characterized in that: In step S4, the entropy weight method is used to solve the weights of the real part α of the characteristic root and the damping ratio ζ in the dynamic evaluation function. The specific calculation formula is as follows: Among them, b op is the correlation coefficient of the real part of the characteristic root, d op is the correlation coefficient of the damping ratio, α op is the real part of the characteristic root, ζ op is the damping ratio, τ p is the information entropy of the real part of the characteristic root, e p is the information entropy of the damping ratio, υ p is the weight of the pth α, and κ p is the weight of the pth ζ.

3. A parameter optimization system based on the parameter optimization method based on the multi-virtual synchronous generator parallel grid-connected system according to claim 1, characterized in that: include: Criteria module: used to perform Lyapunov indirect method criterion on the small signal model of multi-virtual synchronous generator parallel grid-connected system with control link; Damping ratio and real characteristic root acquisition module: calculate the damping ratio and real characteristic root through the small signal state matrix; Dynamic evaluation function construction module: constructs a dynamic evaluation function based on the stability and dynamic characteristics of the system; Dynamic evaluation function weight acquisition module: assigning dynamic evaluation function weights by using entropy weight method; The optimal solution judgment module of the evaluation function: the optimal solution of the dynamic evaluation function is judged by using the quasi-annealing particle swarm algorithm. If the quasi-annealing particle swarm algorithm judges convergence, the optimized control parameters are output. Otherwise, a new set of parameters is obtained and the judgment module is re-entered.

Images