Virtual synchronization micro-grid adaptive secondary frequency modulation control method matched with load fluctuation

Abstract

The invention provides a virtual synchronization micro-grid adaptive secondary frequency modulation control method matched with load fluctuation, and the method comprises the steps: building an inverter island operation model based on VSG control, and generating a transfer function; determining stable operation conditions of the system according to the transfer function; and according to stable operation conditions of the system, optimizing values of a droop coefficient K [omega], a virtual inertia J, a virtual damping D, a proportionality coefficient Kp and an integral coefficient Ki according to change of a switching load by utilizing a discrete particle swarm algorithm, and controlling operation of the VSG. The droop coefficient K omega, the virtual damping D, the virtual inertia J, the proportionality coefficient Kp and the integral coefficient Ki in the secondary frequency modulation controller can be automatically adjusted according to the size of the switching load, and the problem that the system frequency exceeds the safe operation range in the adjusting process due to the fact that the switching load of the system is too large is solved.

Description

Technical Field

[0001] The invention belongs to the field of microgrids in power systems and relates to adaptive control of secondary frequency modulation of microgrids. Background Art

[0002] In recent years, with the rapid development of new energy technologies, microgrids, as an intelligent, flexible, and renewable energy system, have received extensive attention and research. One of the core tasks of microgrids is to maintain the frequency stability of the system through secondary frequency control.

[0003] In the field of secondary frequency regulation control of microgrids, scholars have made many important studies. In the paper "Micropower Frequency Control Strategy Based on the Principle of Secondary Frequency Regulation of Power System", Xu Cheng et al. introduced the PI link in the frequency regulation unit of the traditional droop control strategy to achieve zero-difference control of frequency. However, the PI controller still needs to work when the power is stable or the fluctuation is very small. Since the PI controller has a delay effect, the frequency and power regulation speed is reduced. At the same time, no specific analysis is carried out on the PI controller, and its impact on the dynamic performance of the system and the selection of parameters are not involved; Li Bin et al. selected VSG control to replace the droop control in the paper "Microgrid Inverter Secondary Frequency Regulation Scheme Based on Improved Virtual Synchronous Generator Algorithm", which improves the inertia of the system and improves the secondary frequency regulation control strategy of the VSG algorithm. Integrators are introduced in the damping link and combined with the damping torque to form a PI controller to complete the attenuation control of the frequency disturbance component, provide frequency support for the microgrid, and achieve zero-difference control of the frequency during load disturbances, but in the dynamic response of the system frequency. Less consideration has been given to the characteristics; Chen Laijun et al. designed an adaptive adjustment algorithm for rotational inertia and damping coefficient in the document "Optimal Control of Transient Response of Virtual Synchronous Generators Based on Parameter Adaptive Adjustment" to effectively suppress the fluctuations of power and frequency in the transient process; Song Qiong et al. proposed an adaptive control of rotational inertia to improve active power and frequency characteristics in the document "Improved Adaptive Control of Rotational Inertia of Virtual Synchronous Generators in Multi-Microsource Independent Microgrids". The algorithm dynamically adjusts the rotational inertia according to the frequency deviation to improve the stability of the system; Zhang Weiming designed a control algorithm for dynamically adjusting the rotational inertia in the document "Island Mode Adaptive Rotational Inertia VSG Inverter Power Supply Control Strategy". This strategy enables the system to have a faster response speed and reduces the frequency overshoot during load disturbances. However, the VSG adaptive rotational inertia primary frequency modulation control strategy proposed in this document has frequency deviation during load disturbances and cannot guarantee the power quality when it operates in off-grid mode. In summary, the existing virtual synchronous microgrid secondary frequency regulation technology that introduces a PI controller will cause short-term overshooting during the dynamic adjustment of the system frequency when facing large disturbance load switching, thereby affecting the stable operation of the system. Summary of the invention

[0004] To solve the above problems, the present invention proposes a virtual synchronous microgrid secondary frequency regulation control method that can adapt to the size of the switched load, and can automatically adjust the droop coefficient K in the secondary frequency regulation controller according to the size of the switched load. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i , which solved the problem that the system frequency exceeded the safe operating range during the regulation process due to excessive load switching.

[0005] The adaptive control method proposed in the present invention adjusts the droop coefficient, virtual damping, virtual inertia, proportional coefficient and integral coefficient of the virtual synchronous generator (VSG) in real time according to the load switching size to improve the dynamic regulation process of the system frequency.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is a virtual synchronous microgrid adaptive secondary frequency modulation control method matching load fluctuations, comprising:

[0007] Establish the inverter island operation model based on VSG control and generate the transfer function;

[0008] According to the transfer function, determine the stable operation conditions of the system;

[0009] According to the stable operation conditions of the system, the discrete particle swarm algorithm is used to calculate the droop coefficient K according to the change of the switching load. ω , virtual inertia J, virtual damping D, proportional coefficient K p , integral coefficient K i Optimize values ​​and control VSG operation.

[0010] The inverter island operation model based on VSG control is established to generate a transfer function, specifically: after the proportional integral PI regulator is introduced into the traditional VSG power frequency control loop, its rotor mechanical equation is shown as follows:

[0011]

[0012] In the above formula: P ref is the active power reference value; K ω is the active frequency droop coefficient; Δω is the electrical angular velocity difference; ω0 is the rated angular frequency; ω is the angular frequency output by the virtual synchronous generator; P e is the active power output by the virtual synchronous generator; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; θ is the internal potential phase angle of the virtual synchronous generator; K p , K i They are the proportional coefficient and integral coefficient of the secondary frequency modulation respectively;

[0013] The transfer function between the output angular frequency and active power of the improved system is derived using the Mason formula, and the improved Δω and ΔP are obtained. e The transfer function G' ω-P (s), as shown in the following formula:

[0014]

[0015] In the above formula, Δω is the electrical angular velocity difference; ΔP e is the active output difference; s is the frequency in the Laplace transformation; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; ω0 is the rated angular frequency; ω is the angular frequency of the virtual synchronous generator output; K p , K i are the proportional coefficient and integral coefficient of secondary frequency modulation respectively; K' VSG is the closed-loop gain of the closed-loop origin function; ξ' VSG is the damping ratio of the closed-loop origin function; ω' nVSG is the undamped natural oscillation angular frequency of the closed-loop origin function.

[0016] The conditions for determining the stable operation of the system according to the transfer function are specifically:

[0017] According to Routh criterion, the constraints for stable operation of the system are as follows:

[0018]

[0019] Where: K ω is the active frequency droop coefficient in the outer virtual speed regulator control loop; D is the virtual damping coefficient of the virtual synchronous generator in the outer virtual rotor motion control loop; J is the virtual moment of inertia of the virtual synchronous generator in the outer virtual rotor motion control loop; K p is the proportional coefficient of the secondary frequency modulation in the outer virtual rotor motion control loop; K i is the integral coefficient of the secondary frequency modulation in the outer virtual rotor motion control loop;

[0020] At the same time, the virtual inertia J and virtual damping D should satisfy the following constraints:

[0021]

[0022] Among them, ΔP is the inverter output active power deviation; Δω max The maximum deviation of the inverter output angular frequency; ω n is the rated angular frequency of VSG; E n is the effective value of VSG output voltage; f cp is the cut-off frequency of the active power control loop; Xs is the inductive component of the inverter output impedance.

[0023] According to the stable operation conditions of the system, the discrete particle swarm algorithm is used to calculate the droop coefficient K according to the change of the switching load. ω , virtual inertia J, virtual damping D, proportional coefficient K p , integral coefficient K i Optimize the value and control the operation of VSG, specifically: the angular frequency deviation Δω, the angular frequency deviation change rate dΔω / dt and the angular frequency deviation second-order change rate dt of VSG at the kth moment, k-1th moment and k-2th moment 2 Δω / dt 2 As input, the particle swarm algorithm is used to calculate the droop coefficient K of VSG. ω , virtual moment of inertia J, virtual damping coefficient D, proportional coefficient K p , integral coefficient K i Perform optimization iterations to minimize the absolute value of the difference between the actual system frequency and the rated frequency. ω , virtual moment of inertia J, virtual damping coefficient D, proportional coefficient K p , integral coefficient K i The optimization result is used as output to control the VSG operation and realize the dynamic performance optimization of the VSG output frequency.

[0024] The particle swarm algorithm is used to calculate the droop coefficient K of VSG. ω , virtual moment of inertia J, virtual damping coefficient D, proportional coefficient K p , integral coefficient K i Perform optimization iterations, specifically:

[0025] Step 1: Initialize the parameters of the particle swarm algorithm, including particle size N = 5; particle swarm dimension D = 5, each dimension corresponds to five parameters: droop coefficient, virtual damping, virtual inertia, proportional coefficient and integral coefficient; the particle search speed range is [0, 0.1]; use the rand function to randomly generate particle positions and velocities that meet the constraints mentioned above;

[0026] Step 2: Set the fitness function expression; determine the mathematical model of the system. Assume it is a linear steady-state system with a single input and single output and describe it with a difference equation as follows:

[0027] ΔP(k)-ΔP(k-1)=(J+K ω +Dω0+K p +K i )ω0Δω(k)-(2J+K ω +Dω0+K p )ω0Δω(k-1)-Jω0Δω(k-2)

[0028] Where ΔP(k), ΔP(k-1) are the deviations of the system active power at time k and k-1; Δω(k), Δω(k-1) and Δω(k-2) are the deviations of the system output angular frequency at time k, k-1 and k-2; K ω is the droop coefficient; D is the virtual damping; J is the virtual inertia; K p is the proportionality coefficient; K i is the integration coefficient;

[0029] The absolute value of the inverse of the angular frequency deviation output by the system at the kth moment is used as the fitness function, as shown in the following formula:

[0030]

[0031] Step 3: Take the parameter value in each dimension of the particle position at the kth moment as the droop coefficient K ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The initial value of is obtained, and the initial fitness value of each particle at the kth moment is calculated according to the fitness function expression. The position of each particle at the kth moment is taken as the local optimal solution of each particle for the first iteration, and the position of the particle with the largest fitness value among the five particles at the kth moment is taken as the global optimal solution for the first iteration.

[0032] Step 4: Calculate the global optimal solution after the five particles complete the iteration at the kth moment; update the speed and position of the i-th particle after each iteration at the kth moment according to the following formula, and calculate the fitness value of each particle at the kth moment after each iteration. The position with the largest fitness value of each particle during the 5 iterations is taken as the local optimal solution of the particle at the kth moment, and the position with the largest fitness value of all particles during the 5 iterations is taken as the global optimal solution at the kth moment. Finally, the parameter value of each dimension in the global optimal solution at the kth moment is taken as the droop coefficient K at the kth moment. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The value of

[0033]

[0034] Where: is the position of the i-th particle at time k at the t-th iteration; is the velocity of the i-th particle at the t-th iteration at time k; is the local optimal position of the i-th particle at the t-th iteration at time k; is the global optimal position at the tth iteration at time k; w is the inertia weight value; c1=c2 is the acceleration coefficient; rand() is a random function used to generate a random number between 0 and 1;

[0035] Step 5: Determine whether the simulation is finished. If the simulation is finished, stop the optimization. If the simulation is not finished, update the time and set k=k+1.

[0036] Step 6: Update the droop coefficient K at time k+1 according to the process described in steps 3 to 5. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The value of .

[0037] The beneficial effects of the present invention are:

[0038] 1. The present invention proposes a virtual synchronous microgrid secondary frequency regulation control method that can adapt to the size of the switched load. Considering that the traditional method of introducing a PI controller to solve the secondary frequency regulation will cause the system frequency to exceed the limit for a short time during the change process when the switched load is different, the present invention introduces a particle swarm algorithm into the power frequency control loop to optimize the control parameters under different working conditions, thereby avoiding the problem of the system frequency exceeding the limit and improving the system stability.

[0039] 2. During the secondary frequency regulation process of the microgrid, when the size of the switched load changes, the present invention optimizes the control parameters under different working conditions. Compared with the traditional secondary frequency regulation control under fixed parameters, it can reduce the probability of frequency exceeding the limit, reduce the overshoot in the regulation process, and shorten the regulation time. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is the schematic diagram of VSG primary frequency modulation.

[0041] Figure 2 This is the block diagram of traditional VSG active frequency control.

[0042] Figure 3 This is the traditional VSG step response curve in the time domain.

[0043] Figure 4 This is the block diagram of VSG active frequency control with the introduction of PI regulator.

[0044] Figure 5 This is the step response curve of VSG after the PI regulator is introduced in the time domain.

[0045] Figure 6 This is the structural block diagram of the microgrid based on VSG control.

[0046] Figure 7 This is the block diagram of VSG outer loop power control.

[0047] Figure 8 This is the voltage and current dual closed-loop control block diagram.

[0048] Fig. 9 This is the control block diagram of the particle swarm algorithm.

[0049] Fig.10 The flowchart of parameter adaptive control based on discrete particle swarm algorithm is shown in Figure 2.

[0050] Fig.11 It is the system frequency variation curve.

[0051] Fig.12 It is the adaptive change curve of droop coefficient.

[0052] Fig.13 It is the adaptive change curve of virtual damping.

[0053] Fig.14 It is the adaptive change curve of virtual inertia.

[0054] Fig.15 It is the adaptive change curve of the proportional parameter.

[0055] Fig.16 is the adaptive change curve of the integral parameter. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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 creative work are within the scope of protection of the present invention.

[0057] The present invention aims at solving the problem that the dynamic change curve of the system frequency reaches a short-time over-limit problem in the secondary frequency regulation process of the microgrid by fixing the control parameters and not considering the change of the switching load size during the secondary frequency regulation process. A virtual synchronous microgrid secondary frequency regulation strategy that can adapt to the switching load size is proposed. The characteristic of the strategy is to minimize the absolute value of the difference between the actual frequency and the rated frequency of the system, and to optimize the droop coefficient, virtual damping, virtual inertia, proportional coefficient and integral coefficient by using the particle swarm algorithm, so that the inverter can adaptively control the control parameters according to the change of the switching load at the same time, reduce the probability of the system frequency over-limit, reduce the overshoot in the regulation process, and shorten the regulation time.

[0058] In view of the different load shedding, the present invention adopts the particle swarm algorithm to simultaneously optimize the five quantities of droop coefficient, virtual damping, virtual inertia, proportional coefficient and integral coefficient in the control system, so that the system frequency can meet the requirements of safe operation under different working conditions.

[0059] The present invention first analyzes the frequency characteristics of the system during the primary and secondary frequency regulation of the virtual synchronized microgrid under island operation, and finally proposes a virtual synchronized microgrid secondary frequency regulation strategy that can adapt to the size of the switching load, which is applied to the controller of the distributed power inverter under the island operation of the microgrid. Figure 6 As shown in Figure 1, the topological structure of the microgrid mainly includes: DG power supply, inverter, filter, load and inverter controller module. dc is the DC side DG equivalent power supply; C dc is the DC side filter capacitor; S1 to S6 form an inverter bridge, L f and C f are the filter inductors and capacitors of each phase respectively; R f is the equivalent resistance of the filter inductor and each switching device; u Li ,i Li (i=a,b,c) are the inverter output voltage and current; u i ,i i (i=a, b, c) are the voltage and current output by the inverter after LC filtering; Load1 and Load2 are two local loads.

[0060] The analysis process of the frequency characteristics of the system during the primary and secondary frequency regulation of the virtual synchronized microgrid under island operation is as follows:

[0061] Step 1: Analysis of the frequency characteristics of the system during a frequency modulation process.

[0062] In an isolated microgrid system, the dynamic and static performance of the frequency is an important guarantee for the reliable and stable operation of the system. Usually, a combination of primary frequency regulation control based on the underlying inverter and secondary frequency regulation control based on the central controller (MGCC) is used to maintain the frequency quality of the system. The droop control strategy of the microgrid inverter is equivalent to the primary frequency regulation link in the power system, which belongs to differential regulation. Figure 1 As shown, assuming that the VSG initially works stably at point A, the corresponding initial angular frequency and output active power are (ω1, P1). At a certain moment, the load suddenly increases, and the VSG output electromagnetic power increases to P2. The speed regulator link of the VSG will reduce the system angular frequency to ω2 to ensure the stable operation of the system. At this time, the VSG works at the new stable state point B (ω2, P2).

[0063] In order to quantify the frequency modulation process, the following Figure 2Specific analysis is carried out on the shown VSG power-frequency control block diagram. First, the transfer function between the system angular frequency and active power is derived using Mason's formula, and the transfer function G e (s) between the frequency deviation Δω and the active power deviation ΔP ω-P is obtained as shown in the following formula.

[0064]

[0065] In the formula: Δω is the difference in electrical angular velocity; ΔP e is the difference in active power output; s is the frequency in Laplace transform; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; ω0 is the rated angular frequency; ω is the angular frequency output by the virtual synchronous generator; K VSG is the closed-loop gain of the closed-loop transfer function; T VSG is the time constant of the closed-loop transfer function.

[0066] Then, the step response of the system with an input magnitude of ΔP L is considered, that is, a local load with a magnitude of ΔP L is applied, and the obtained response curve is inverse Laplace transformed to obtain the expression of the step response curve of the system in the time domain as shown in the following formula, where ΔP L = P e - P ref , and the waveform is as shown in the appendix Figure 3 .

[0067]

[0068] In the formula: ΔP L is the active power of the switched load; K ω is the active frequency droop coefficient; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; ω0 is the rated angular frequency.

[0069] Finally, the final value of the system frequency f deviation under the step response of the transfer function G ω-P (s) with a magnitude of ΔP L is calculated using the Lagrange final value theorem as shown in the following formula.

[0070]

[0071] Analyzing from the above formula, when the load is applied, ΔP L > 0, Δf < 0 (i.e., f < f0), that is, applying the load makes the system frequency lower than the rated frequency; when the load is removed, ΔP L < 0, Δf > 0 (i.e., f > f0), that is, removing the load makes the system frequency higher than the rated frequency.

[0072] In summary, regardless of whether the load is added or removed, the primary frequency regulation control of the system by the traditional VSG in off-grid mode is differential control, that is, the frequency deviation cannot be suppressed by adjusting the droop coefficient, virtual inertia and virtual damping alone. According to the national standard GB / T15945-1995, the frequency of the power system should be controlled within the range of (50±0.2) Hz, but when the load fluctuates greatly, that is, ΔP L When it is large, Δf>0.2Hz will occur, causing the system steady-state frequency to exceed the limit. Therefore, when the load fluctuates greatly, it is necessary to introduce secondary frequency modulation control in the VSG power-frequency control loop.

[0073] The second step is to analyze the frequency characteristics of the system during the secondary frequency modulation process.

[0074] In order to achieve frequency error-free regulation, the output power of multiple frequency modulation units in the system can be changed by introducing centralized secondary frequency control of PI regulator in the virtual synchronous machine control strategy. Figure 4 As shown, after the PI regulator is introduced into the traditional VSG power frequency control loop, its rotor mechanical equation is as follows, where K p , K i They are the proportional coefficient and integral coefficient of secondary frequency modulation respectively.

[0075]

[0076] Where: P ref is the active power reference value; K ω is the active frequency droop coefficient; Δω is the electrical angular velocity difference; ω0 is the rated angular frequency; ω is the angular frequency output by the virtual synchronous generator; P e is the active power output by the virtual synchronous generator; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; θ is the internal potential phase angle of the virtual synchronous generator; K p , K i They are the proportional coefficient and integral coefficient of secondary frequency modulation respectively.

[0077] The following is a quantitative analysis of the VSG secondary frequency modulation process after the introduction of the PI controller. First, the transfer function between the output angular frequency and active power of the improved system is derived using the Mason formula, and the improved Δω and ΔP are obtained. e The transfer function G' ω-P (s), as shown in the following formula.

[0078]

[0079] Where Δω is the electrical angular velocity difference; ΔP eis the active output difference; s is the frequency in the Laplace transformation; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; ω0 is the rated angular frequency; ω is the angular frequency of the virtual synchronous generator output; K p , K i are the proportional coefficient and integral coefficient of secondary frequency modulation respectively; K' VSG is the closed-loop gain of the closed-loop origin function; ξ' VSG is the damping ratio of the closed-loop origin function; ω' nVSG is the undamped natural oscillation angular frequency of the closed-loop origin function.

[0080] At this time, the input size of the system is ΔP L The step response, that is, the input size is ΔP L The local load is then transformed by the reverse pull-type transformation of the obtained response curve to obtain the step response curve expression of the system in the time domain as follows, where ΔP L =P e -P ref .

[0081]

[0082] Where: K' VSG is the closed-loop gain of the closed-loop origin function; ξ' VSG is the damping ratio of the closed-loop origin function; ω' nVSG is the undamped natural oscillation angular frequency of the closed-loop origin function; ΔP L is the active power of the switched load; ω0 is the rated angular frequency.

[0083] Using Lagrange's terminal value theorem, we can calculate the transfer function G' ω-P (s) is of size ΔP L The final value of the system frequency f deviation in the step response is zero, as shown in the following formula.

[0084]

[0085] According to the above analysis, when the system is put into or cut off the load (ΔP L ≠0), the system angular frequency deviation is always zero (Δω≡0, that is, f≡f0), that is, no matter the system is loaded or unloaded, when the system is stable, the system frequency is always equal to the rated frequency, achieving frequency stability.

[0086] Although the frequency error-free control is achieved by introducing the PI regulator, the inventors found that when the microgrid switches on and off large disturbance loads, the fixed secondary frequency regulation controller parameters will cause a short-term over-limit in the dynamic regulation process of the system frequency, as shown in the attached figure. Figure 5 As shown, the peak value f of the system frequency during the recovery processmax As shown in the following formula.

[0087]

[0088] Where: K' VSG is the closed-loop gain of the closed-loop origin function; ξ' VSG is the damping ratio of the closed-loop origin function; ω' nVSG is the undamped natural oscillation angular frequency of the closed-loop origin function; ΔP L is the active power of the switched load; ω0 is the rated angular frequency; t p It is the peak time of the system frequency change process.

[0089] Although the introduction of PI controller solves the problem of secondary frequency regulation of microgrid and can realize the error-free regulation of system frequency, it can be seen from the step response curve expression of the system in the time domain that the peak value in the dynamic frequency regulation process is affected by the droop coefficient K in the VSG strategy. ω , damping coefficient D, characteristic parameter virtual inertia J of VSG, PI control parameters and switching load ΔP L When the load shedding is large, if other control parameters remain unchanged, the peak value of the system frequency dynamic regulation process may exceed the limit. Obviously, fixed secondary frequency regulation controller parameters cannot guarantee that the system frequency always maintains good dynamic performance under various operating conditions. In addition, considering the droop coefficient K in the VSG strategy ω , damping coefficient D, characteristic parameter virtual inertia J of VSG, and PI control parameters are virtual quantities. Therefore, the overshoot of the system frequency change process can be reduced and the dynamic performance can be improved by adaptively controlling the control parameters.

[0090] The controller of the distributed power inverter in the microgrid island operation adopts the virtual synchronous generator (VSG) control strategy. The control block diagram is shown in the attached figure. Figure 7 and attached Figure 8 As shown, it includes an outer loop virtual speed regulator control loop, an outer loop virtual rotor motion control loop, an outer loop virtual synchronous generator excitation control loop, an outer loop three-phase voltage synthesis control loop, an inner loop voltage and current dual closed-loop control loop and a parameter adaptive control loop based on discrete particle swarm optimization (DPSO).

[0091] The outer virtual speed regulator control loop is used to simulate the primary frequency regulation characteristics of the synchronous generator. It collects the angular frequency deviation as the control quantity and realizes the speed regulation function through proportional control. The specific expression is:

[0092]

[0093] Where: P m is the mechanical power output by the virtual prime mover; P ref is the active power reference value; K ω is the active frequency droop coefficient; ω0 is the rated angular frequency; ω is the angular frequency output by the virtual synchronous generator.

[0094] The outer virtual rotor motion control loop is used to simulate the mechanical motion of the synchronous generator rotor. The specific motion equation is expressed as:

[0095]

[0096] Where: P m is the mechanical power output by the virtual prime mover; P e is the active power output by the virtual synchronous generator; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; Δω is the electrical angular velocity difference; ω0 is the rated angular frequency; ω is the angular frequency output by the virtual synchronous generator; θ is the internal potential phase angle of the virtual synchronous generator; K p , K i They are the proportional coefficient and integral coefficient of secondary frequency modulation respectively.

[0097] The outer loop virtual synchronous generator excitation control loop is used to simulate the excitation control loop of the synchronous generator, that is, the primary voltage regulation characteristics of the virtual synchronous generator. Its specific expression is:

[0098] E=E ref +K v (Q ref -Q e )

[0099] Where, E is the potential amplitude inside the virtual synchronous generator; E ref is the rated electromotive force; K v is the reactive voltage droop coefficient; Q ref is the reactive power reference value; Q e is the reactive power output by the virtual synchronous generator.

[0100] The outer three-phase voltage synthesis control loop is used to synthesize the instantaneous value of the virtual internal potential three-phase voltage as the input of the subsequent virtual impedance control. The instantaneous value of the virtual internal potential is calculated by the phase angle θ and amplitude E of the virtual internal potential of the virtual synchronous generator obtained by the outer virtual rotor motion control loop and the outer virtual synchronous generator excitation control loop. The specific expression is:

[0101]

[0102] Where: E is the internal potential amplitude of the virtual synchronous generator; θ is the internal potential phase angle of the virtual synchronous generator; u a is the instantaneous value of the potential a phase voltage in the virtual synchronous generator; u b is the instantaneous value of the potential b-phase voltage in the virtual synchronous generator; u c It is the instantaneous value of the c-phase voltage in the virtual synchronous generator.

[0103] The inner loop voltage and current dual closed-loop control loop is designed to quickly and accurately track the voltage reference value output by the outer loop control strategy. An inner loop controller needs to be added to improve the power quality of the VSG. For voltage source VSG, the inner loop usually adopts a voltage and current dual closed-loop control strategy. The voltage loop adopts PI control, which is mainly responsible for providing the current reference value to the current inner loop. The specific method is to convert the instantaneous voltage signal output by the outer loop to the dq rotating coordinate system, and then use the PI controller to track the error of the transformed d-axis and q-axis voltage components. The current loop also adopts PI control, which is mainly used to track the reference current output by the voltage loop, and obtain the modulation voltage of the SPWM wave through the PI controller. The specific control block diagram is shown in the attached figure. Figure 8 shown.

[0104] in, The voltage loop reference value obtained by virtual impedance control; u od 、u oq is the dq axis component of the voltage output by the VSG; the feedforward quantity i od 、i oq is the dq axis component of the current output by the VSG; C f is the filter capacitance value of the LC filter; i * ld , It is the reference value output by the voltage loop control to the subsequent current loop; i ld 、i lq is the dq axis component of the filter inductor current of the virtual synchronous machine LC filter; the feedforward quantity u od 、u oq is the dq axis component of the voltage output by the virtual synchronous machine; L f is the filter inductance value of the LC filter; the final output result It is a modulation signal for SPWM drive control, which finally generates a drive signal for each IGBT switching device after SPWM modulation.

[0105] The control parameters in the outer virtual speed regulator control loop and the outer virtual rotor motion control loop need to be kept in a stable state at all times when taking values. Therefore, the droop coefficient K ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K iThe five control parameters are constrained accordingly. After calculation using the Routh criterion, the specific constraint relationship between the five control parameters is as follows:

[0106]

[0107] Where: K ω is the active frequency droop coefficient in the outer virtual speed regulator control loop; D is the virtual damping coefficient of the virtual synchronous generator in the outer virtual rotor motion control loop; J is the virtual moment of inertia of the virtual synchronous generator in the outer virtual rotor motion control loop; K p is the proportional coefficient of the secondary frequency modulation in the outer virtual rotor motion control loop; K i It is the integral coefficient of the secondary frequency modulation in the outer loop virtual rotor motion control loop.

[0108] In addition, regarding the selection method of virtual inertia J and virtual damping D in the virtual synchronous generator control strategy, it can be seen that:

[0109]

[0110] Where: ΔP is the inverter output active power deviation; Δω max The maximum deviation of the inverter output angular frequency; ω n is the rated angular frequency of VSG; E n is the effective value of VSG output voltage; f cp is the cut-off frequency of the active power control loop; X s is the inductive component of the inverter output impedance.

[0111] The parameter adaptive control loop based on discrete particle swarm algorithm is used to adjust the droop coefficient K in the outer ring virtual speed regulator control loop and the outer ring virtual rotor motion control loop according to the size of the switched load. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i Perform adaptive control, such as Fig. 9 As shown, by calculating the angular frequency deviation Δω, the angular frequency deviation change rate dΔω / dt, and the angular frequency deviation second-order change rate d 2 Δω / dt 2 As an input variable, the droop coefficient K ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The adaptive variable is used as the output variable. The adaptive setting is to adjust the droop coefficient K in each sampling period. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K iThe adaptive control based on particle swarm algorithm is to correspond each particle to a set of parameters to be controlled, and perform multiple iterations at sampling time k according to the changes of the system at time k-1, k-2, and k-3. In each iteration, the control amount corresponding to each particle is calculated, and the corresponding system output and its fitness value are calculated. According to the fitness value, the individual historical optimal particle and the global optimal particle are selected. After the iteration, the global optimal particle obtained is used as the control parameter of the next moment k+l. In this way, the parameters to be controlled will change continuously with the change of the system state. The parameter adaptive control process based on discrete particle swarm algorithm is as follows Fig.10 As shown, the specific control steps are as follows:

[0112] Step 1: Initialize the parameters of the particle swarm algorithm, including particle size N = 5; total number of algorithm iterations G = 5; particle swarm dimension D = 5, each dimension corresponds to five parameters: droop coefficient, virtual damping, virtual inertia, proportional coefficient and integral coefficient; the particle search speed range is [0, 0.1]; use the rand function to randomly generate particle positions and velocities that meet the constraints mentioned above.

[0113] Step 2: Set the fitness function expression. Determine the mathematical model of the system and assume it is a linear steady-state system with a single input and single output, which can be described by a difference equation as follows:

[0114] ΔP(k)-ΔP(k-1)=(J+K ω +Dω0+K p +K i )ω0Δω(k)-(2J+K ω +Dω0+K p )ω0Δω(k-1)-Jω0Δω(k-2)

[0115] Where ΔP(k), ΔP(k-1) are the deviations of the system active power at time k and k-1; Δω(k), Δω(k-1) and Δω(k-2) are the deviations of the system output angular frequency at time k, k-1 and k-2; K ω is the droop coefficient; D is the virtual damping; J is the virtual inertia; K p is the proportionality coefficient; K i is the integration coefficient.

[0116] The absolute value of the inverse of the angular frequency deviation output by the system at the kth moment is used as the fitness function, as shown in the following formula:

[0117]

[0118] Step 3: Take the parameter value in each dimension of the particle position at the kth moment as the droop coefficient K ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient Ki The initial value of is taken, and the initial fitness value of each particle at the kth moment is calculated according to the fitness function expression. The position of each particle at the kth moment is taken as the local optimal solution of each particle for the first iteration, and the position of the particle with the largest fitness value among the five particles at the kth moment is taken as the global optimal solution for the first iteration.

[0119] Step 4: Calculate the global optimal solution after 5 iterations of 5 particles at the kth moment. Update the speed and position of the ith particle after each iteration at the kth moment according to the following formula, and calculate the fitness value of each particle at the kth moment after each iteration. The position with the largest fitness value of each particle during the 5 iterations is taken as the local optimal solution of the particle at the kth moment, and the position with the largest fitness value of all particles during the 5 iterations is taken as the global optimal solution at the kth moment. Finally, the parameter value of each dimension in the global optimal solution at the kth moment is taken as the droop coefficient K at the kth moment. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The value of .

[0120]

[0121] Where: is the position of the i-th particle at time k at the t-th iteration; is the velocity of the i-th particle at the t-th iteration at time k; is the local optimal position of the i-th particle at the t-th iteration at time k; is the global optimal position at the tth iteration at time k; w = 1 is the inertia weight value; c1 = c2 = 0.5 is the acceleration coefficient; rand() is a random function used to generate a random number between 0 and 1;

[0122] Step 5: Determine whether the number of iterations G is greater than or equal to 5. If so, the simulation ends; if not, update the time and set k=k+1.

[0123] Step 6: Update the droop coefficient K at time k+1 according to the process described in steps 3 to 5. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The value of .

[0124] Implementation example:

[0125] 1. Virtually synchronized microgrid structure under island operation

[0126] The basic structure of the microgrid described in the present invention is shown in the attached figure. Figure 6As shown in FIG. , it specifically includes: DG power supply, inverter, filter, load and inverter controller module. dc is the DC side DG equivalent power supply; C dc is the DC side filter capacitor; S1 to S6 form an inverter bridge, L f and C f are the filter inductors and capacitors of each phase respectively; R f is the equivalent resistance of the filter inductor and each switching device; u Li ,i Li (i=a,b,c) are the inverter output voltage and current; u i ,i i (i=a, b, c) are the voltage and current output by the inverter after LC filtering; Load1 and Load2 are two local loads.

[0127] 2. Inverter control strategy in virtual synchronized microgrid under island operation

[0128] The controller of the distributed power inverter in the microgrid island operation adopts the virtual synchronous generator (VSG) control strategy. The control block diagram is shown in the attached figure. Figure 7 and attached Figure 8 As shown, it includes an outer loop virtual speed regulator control loop, an outer loop virtual rotor motion control loop, an outer loop virtual synchronous generator excitation control loop, an outer loop three-phase voltage synthesis control loop, an inner loop voltage and current dual closed-loop control loop and a parameter adaptive control loop based on discrete particle swarm optimization (DPSO).

[0129] 2.1 Outer loop virtual speed regulator control loop

[0130] The outer loop virtual speed regulator control loop is used to simulate the primary frequency regulation characteristics of the synchronous generator. It collects the angular frequency deviation as the control quantity and realizes the speed regulation function through proportional control. The specific expression is:

[0131] P m =P ref -K ω (ω-ω0)

[0132] Where: P m is the mechanical power output by the virtual prime mover; P ref is the active power reference value; K ω is the active frequency droop coefficient; ω0 is the rated angular frequency; ω is the angular frequency output by the virtual synchronous generator.

[0133] 2.2 Outer ring virtual rotor motion control loop

[0134] The outer virtual rotor motion control loop is used to simulate the mechanical motion of the synchronous generator rotor. The specific motion equation is expressed as:

[0135]

[0136] Where: P m is the mechanical power output by the virtual prime mover; P e is the active power output by the virtual synchronous generator; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; Δω is the electrical angular velocity difference; ω0 is the rated angular frequency; ω is the angular frequency output by the virtual synchronous generator; θ is the internal potential phase angle of the virtual synchronous generator; K p , K i They are the proportional coefficient and integral coefficient of secondary frequency modulation respectively.

[0137] 2.3 Outer loop virtual synchronous generator excitation control loop

[0138] The outer loop virtual synchronous generator excitation control loop is used to simulate the excitation control loop of the synchronous generator, that is, the primary voltage regulation characteristics of the virtual synchronous generator. Its specific expression is:

[0139] E=E ref +K v (Q ref -Q e )

[0140] Where, E is the potential amplitude inside the virtual synchronous generator; E ref is the rated electromotive force; K v is the reactive voltage droop coefficient; Q ref is the reactive power reference value; Q e is the reactive power output by the virtual synchronous generator.

[0141] 2.4 Outer three-phase voltage synthesis control loop

[0142] The outer three-phase voltage synthesis control loop is used to synthesize the instantaneous value of the virtual internal potential three-phase voltage as the input of the subsequent virtual impedance control. The instantaneous value of the virtual internal potential is calculated by the phase angle θ and amplitude E of the virtual internal potential of the virtual synchronous generator obtained by the outer virtual rotor motion control loop and the outer virtual synchronous generator excitation control loop. The specific expression is:

[0143]

[0144] Where: E is the internal potential amplitude of the virtual synchronous generator; θ is the internal potential phase angle of the virtual synchronous generator; u a is the instantaneous value of the potential a phase voltage in the virtual synchronous generator; u bis the instantaneous value of the potential b-phase voltage in the virtual synchronous generator; u c It is the instantaneous value of the c-phase voltage in the virtual synchronous generator.

[0145] 2.5 Inner voltage and current double closed-loop control loop

[0146] The inner loop voltage and current dual closed-loop control loop is designed to quickly and accurately track the voltage reference value output by the outer loop control strategy. An inner loop controller needs to be added to improve the power quality of the VSG. For voltage source VSG, the inner loop usually adopts a voltage and current dual closed-loop control strategy. The voltage loop adopts PI control, which is mainly responsible for providing the current reference value to the current inner loop. The specific method is to convert the instantaneous voltage signal output by the outer loop to the dq rotating coordinate system, and then use the PI controller to track the error of the transformed d-axis and q-axis voltage components. The current loop also adopts PI control, which is mainly used to track the reference current output by the voltage loop, and obtain the modulation voltage of the SPWM wave through the PI controller. The specific control block diagram is shown in the attached figure. Figure 8 shown.

[0147] in, The voltage loop reference value obtained by virtual impedance control; u od 、u oq is the dq axis component of the voltage output by the VSG; the feedforward quantity i od 、i oq is the dq axis component of the current output by the VSG; C f is the filter capacitance value of the LC filter; i * ld 、i * lq It is the reference value output by the voltage loop control to the subsequent current loop; i ld 、i lq is the dq axis component of the filter inductor current of the virtual synchronous machine LC filter; the feedforward quantity u od 、u oq is the dq axis component of the voltage output by the virtual synchronous machine; L f is the filter inductance value of the LC filter; the final output result It is a modulation signal for SPWM drive control, which finally generates a drive signal for each IGBT switching device after SPWM modulation.

[0148] 3. Parameter adaptive control loop design based on discrete particle swarm algorithm

[0149] 3.1 Parameter adaptive control loop structure based on discrete particle swarm algorithm

[0150] The parameter adaptive control structure based on discrete particle swarm algorithm is as follows Fig. 9As shown, it is mainly used to adjust the droop coefficient K in the outer ring virtual speed regulator control loop and the outer ring virtual rotor motion control loop according to the size of the switched load. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i Perform adaptive control, such as Fig. 9 As shown, by using the angular frequency deviation Δω, the angular frequency deviation change rate dΔω / dt and the angular frequency deviation second-order change rate d 2 Δω / dt 2 As an input variable, the droop coefficient K ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The adaptive variable is used as the output variable. The adaptive setting is to adjust the droop coefficient K in each sampling period. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The adaptive control based on the particle swarm algorithm is to correspond each particle to a set of parameters to be controlled, and perform multiple iterations at the sampling time k according to the changes of the system at the time k-1, k-2, and k-3. In each iteration, the control amount corresponding to each particle is calculated, and the corresponding system output and its fitness value are calculated. According to the fitness value, the individual historical optimal particle and the global optimal particle are selected. After the iteration, the global optimal particle obtained is used as the control parameter of the next time k+l, so that the parameters to be controlled will change continuously with the change of the system state.

[0151] 3.2 Analysis of parameter adaptive control process based on discrete particle swarm algorithm

[0152] The parameter adaptive control process based on discrete particle swarm algorithm is as follows Fig.10 As shown, the specific control steps are as follows:

[0153] Step 1: Initialize the parameters of the particle swarm algorithm, including particle size N = 5; total number of algorithm iterations G = 5; particle swarm dimension D = 5, each dimension corresponds to five parameters: droop coefficient, virtual damping, virtual inertia, proportional coefficient and integral coefficient; the particle search speed range is [0, 0.1]; use the rand function to randomly generate particle positions and velocities that meet the constraints mentioned above.

[0154] Step 2: Set the fitness function expression. Determine the mathematical model of the system and assume it is a linear steady-state system with a single input and single output, which can be described by a difference equation as follows:

[0155] ΔP(k)-ΔP(k-1)=(J+K ω +Dω0+K p +K i)ω0Δω(k)-(2J+K ω +Dω0+K p )ω0Δω(k-1)-Jω0Δω(k-2) where ΔP(k), ΔP(k-1) are the system active power deviations at time k and k-1; Δω(k), Δω(k-1) and Δω(k-2) are the deviations of the system output angular frequency at time k, k-1 and k-2; K ω is the droop coefficient; D is the virtual damping; J is the virtual inertia; K p is the proportionality coefficient; K i is the integration coefficient.

[0156] The absolute value of the inverse of the angular frequency deviation output by the system at the kth moment is used as the fitness function, as shown in the following formula:

[0157]

[0158] Step 3: Take the parameter value in each dimension of the particle position at the kth moment as the droop coefficient K ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The initial value of is taken, and the initial fitness value of each particle at the kth moment is calculated according to the fitness function expression. The position of each particle at the kth moment is taken as the local optimal solution of each particle for the first iteration, and the position of the particle with the largest fitness value among the five particles at the kth moment is taken as the global optimal solution for the first iteration.

[0159] Step 4: Calculate the global optimal solution after 5 iterations of 5 particles at the kth moment. Update the speed and position of the ith particle after each iteration at the kth moment according to the following formula, and calculate the fitness value of each particle at the kth moment after each iteration. The position with the largest fitness value of each particle during the 5 iterations is taken as the local optimal solution of the particle at the kth moment, and the position with the largest fitness value of all particles during the 5 iterations is taken as the global optimal solution at the kth moment. Finally, the parameter value of each dimension in the global optimal solution at the kth moment is taken as the droop coefficient K at the kth moment. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The value of .

[0160]

[0161] Where: is the position of the i-th particle at time k at the t-th iteration; is the velocity of the i-th particle at the t-th iteration at time k; is the local optimal position of the i-th particle at the t-th iteration at time k; is the global optimal position at the tth iteration at time k; w = 1 is the inertia weight value; c1 = c2 = 0.5 is the acceleration coefficient; rand() is a random function used to generate a random number between 0 and 1;

[0162] Step 5: Determine whether the number of iterations G is greater than or equal to 5. If so, the simulation ends; if not, update the time and set k=k+1.

[0163] Step 6: Update the droop coefficient K at time k+1 according to the process described in steps 3 to 5. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The value of .

[0164] 4. Simulation Verification

[0165] In order to verify the feasibility and effectiveness of the control strategy proposed in this invention, Matlab / Simulink simulation software was used to build the Fig. 9 The VSG island operation simulation model is shown in Figure 1, and the main circuit parameters are shown in Table 1.

[0166] Table 1 System simulation parameters

[0167]

[0168] During the simulation, the system initial load power is set to 10kW, the load power is adjusted to 17kW at 0.2s, 27kW at 0.4s, 17kW at 0.6s, and 10kW at 0.8s. The frequency control is performed by using the primary frequency modulation method, the secondary frequency modulation method with fixed PI parameters, and the adaptive control parameter adjustment method proposed in this paper. Fig.11 The comparison results of three frequency modulation methods are given respectively.

[0169] Table 2 Control parameters under three control strategies

[0170]

[0171] By comparing the waveforms of system frequency changes under the three strategies, it can be seen that the strategy adopted by the present invention can adjust the parameter values ​​according to the size of the switched load, thereby achieving the purpose of reducing overshoot, shortening the adjustment time and reducing the number of oscillations at the same time, thereby improving the stability of the system.

Claims

1. A virtual synchronous microgrid adaptive secondary frequency modulation control method for matching load fluctuations, characterized in that: include: Establish the inverter island operation model based on VSG control and generate the transfer function; According to the transfer function, determine the stable operation conditions of the system; According to the stable operation conditions of the system, the discrete particle swarm algorithm is used to calculate the droop coefficient K according to the change of the switching load. ω , virtual inertia J, virtual damping D, proportional coefficient K p , integral coefficient K i Optimize values ​​and control VSG operation.

2. The method for adaptive secondary frequency modulation control of a virtual synchronous microgrid matching load fluctuation according to claim 1 is characterized in that: The inverter island operation model based on VSG control is established to generate a transfer function, specifically: after the proportional integral PI regulator is introduced into the traditional VSG power frequency control loop, its rotor mechanical equation is shown as follows: In the above formula: P ref is the active power reference value; K ω is the active frequency droop coefficient; Δω is the electrical angular velocity difference; ω0 is the rated angular frequency; ω is the angular frequency of the virtual synchronous generator output; P e is the active power output by the virtual synchronous generator; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; θ is the internal potential phase angle of the virtual synchronous generator; K p , K i are the proportional coefficient and integral coefficient of the secondary frequency modulation respectively; The transfer function between the output angular frequency and active power of the improved system is derived using the Mason formula, and the improved Δω and ΔP are obtained. e The transfer function G' ω-P (s), as shown in the following formula: In the above formula, Δω is the electrical angular velocity difference; ΔP e is the active output difference; s is the frequency in Laplace transformation; J is the virtual moment of inertia of the virtual synchronous generator; D is the virtual damping coefficient of the virtual synchronous generator; ω0 is the rated angular frequency; ω is the angular frequency of the virtual synchronous generator output; K p , K i are the proportional coefficient and integral coefficient of secondary frequency modulation respectively; K' VSG is the closed-loop gain of the closed-loop origin function; ξ' VSG is the damping ratio of the closed-loop origin function; ω' nVSG is the undamped natural oscillation angular frequency of the closed-loop origin function.

3. The method for adaptive secondary frequency modulation control of a virtual synchronous microgrid matching load fluctuation according to claim 1 is characterized in that: The conditions for determining the stable operation of the system according to the transfer function are specifically: According to Routh criterion, the constraints for stable operation of the system are as follows: Where: K ω is the active frequency droop coefficient in the outer virtual speed regulator control loop; D is the virtual damping coefficient of the virtual synchronous generator in the outer virtual rotor motion control loop; J is the virtual moment of inertia of the virtual synchronous generator in the outer virtual rotor motion control loop; K p is the proportional coefficient of the secondary frequency modulation in the outer virtual rotor motion control loop; K i is the integral coefficient of the secondary frequency modulation in the outer virtual rotor motion control loop; At the same time, the virtual inertia J and virtual damping D should satisfy the following constraints: Among them, ΔP is the inverter output active power deviation; Δω max The maximum deviation of the inverter output angular frequency; ω n is the rated angular frequency of VSG; E n is the effective value of VSG output voltage; f cp is the cut-off frequency of the active power control loop; X s is the inductive component of the inverter output impedance.

4. The method for adaptive secondary frequency modulation control of a virtual synchronous microgrid matching load fluctuation according to claim 1 is characterized in that: According to the stable operation conditions of the system, the discrete particle swarm algorithm is used to calculate the droop coefficient K according to the change of the switching load. ω , virtual inertia J, virtual damping D, proportional coefficient K p , integral coefficient K i Optimize the value and control the operation of VSG, specifically: the angular frequency deviation Δω, the angular frequency deviation change rate dΔω / dt and the angular frequency deviation second-order change rate dt of VSG at the kth moment, k-1th moment and k-2th moment 2 Δω / dt 2 As input, the particle swarm algorithm is used to calculate the droop coefficient K of VSG. ω , virtual moment of inertia J, virtual damping coefficient D, proportional coefficient K p , integral coefficient K i Perform optimization iterations to minimize the absolute value of the difference between the actual system frequency and the rated frequency. ω , virtual moment of inertia J, virtual damping coefficient D, proportional coefficient K p , integral coefficient K i The optimization result is used as output to control the VSG operation and realize the dynamic performance optimization of the VSG output frequency.

5. The method for adaptive secondary frequency modulation control of a virtual synchronous microgrid matching load fluctuation according to claim 4 is characterized in that: The particle swarm algorithm is used to calculate the droop coefficient K of VSG. ω , virtual moment of inertia J, virtual damping coefficient D, proportional coefficient K p , integral coefficient K i Perform optimization iterations, specifically: Step 1: Initialize the parameters of the particle swarm algorithm, including particle size N = 5; particle swarm dimension D = 5, each dimension corresponds to five parameters: droop coefficient, virtual damping, virtual inertia, proportional coefficient and integral coefficient; the particle search speed range is [0, 0.1]; use the rand function to randomly generate particle positions and velocities that meet the constraints mentioned above; Step 2: Set the fitness function expression; determine the mathematical model of the system. Assume it is a linear steady-state system with a single input and single output and describe it with a difference equation as follows: ΔP(k)-ΔP(k-1)=(J+K ω +Dω0+K p +K i )ω0Δω(k)-(2J+K ω +Dω0+K p )ω0Δω(k-1)-Jω0Δω(k-2) where ΔP(k), ΔP(k-1) are the system active power deviations at time k and k-1; Δω(k), Δω(k-1) and Δω(k-2) are the deviations of the system output angular frequency at time k, k-1 and k-2; K ω is the droop coefficient; D is the virtual damping; J is the virtual inertia; K p is the proportionality coefficient; K i is the integration coefficient; The absolute value of the inverse of the angular frequency deviation output by the system at the kth moment is used as the fitness function, as shown in the following formula: Step 3: Take the parameter value in each dimension of the particle position at the kth moment as the droop coefficient K ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The initial value of is obtained, and the initial fitness value of each particle at the kth moment is calculated according to the fitness function expression. The position of each particle at the kth moment is taken as the local optimal solution of each particle for the first iteration, and the position of the particle with the largest fitness value among the five particles at the kth moment is taken as the global optimal solution for the first iteration. Step 4: Calculate the global optimal solution after the five particles complete the iteration at the kth moment; update the speed and position of the i-th particle after each iteration at the kth moment according to the following formula, and calculate the fitness value of each particle at the kth moment after each iteration. The position with the largest fitness value of each particle during the 5 iterations is taken as the local optimal solution of the particle at the kth moment, and the position with the largest fitness value of all particles during the 5 iterations is taken as the global optimal solution at the kth moment. Finally, the parameter value of each dimension in the global optimal solution at the kth moment is taken as the droop coefficient K at the kth moment. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The value of Where: is the position of the i-th particle at time k at the t-th iteration; is the velocity of the i-th particle at the t-th iteration at time k; is the local optimal position of the i-th particle at the t-th iteration at time k; is the global optimal position at the tth iteration at time k; w is the inertia weight value; c1=c2 is the acceleration coefficient; rand() is a random function used to generate a random number between 0 and 1; Step 5: Determine whether the simulation is finished. If the simulation is finished, stop the optimization. If the simulation is not finished, update the time and set k=k+1. Step 6: Update the droop coefficient K at time k+1 according to the process described in steps 3 to 5. ω , virtual damping D, virtual inertia J, proportional coefficient K p and the integral coefficient K i The value of .

Images