A multi-mode decomposition control method for a VSG system based on energy storage state of charge

Abstract

The application belongs to the technical field of wind power generation control, and specifically discloses a VSG system multi-modal decomposition control method based on the state of charge of energy storage, which is performed according to the following steps: S1, obtaining the predicted residual power of energy storage according to the residual capacity of the battery; S2, constructing a nonlinear state-of-charge power system of energy storage to obtain the predicted power of energy storage considering the residual power; S3, defining a convolution network algorithm of multi-modal decomposition control to obtain the feedback power of the modal convolution algorithm; S4, using the difference between the reference output power of energy storage and the feedback power of the modal convolution control algorithm as the reference input power of the VSG active control loop; S5, obtaining the change amount of the reference input power of the VSG active control loop of energy storage, and further obtaining the steady-state equation of the VSG system applied to energy storage. The application has superior frequency modulation performance, can improve the utilization rate of the energy storage system, and reduce the configuration of the energy storage capacity.

Description

Technical Field

[0001] This invention belongs to the field of wind power generation control technology, and relates to a multimodal decomposition control method for VSG systems based on energy storage state of charge. Background Technology

[0002] Introducing the characteristics of a synchronous generator into a photovoltaic (PV) grid-connected system allows the system to simulate the performance of a synchronous generator in terms of external characteristics. This has led to the development of the basic idea of ​​a Virtual Synchronous Generator (VSG): by introducing an energy storage system into the distributed generation unit (DRG) and employing a corresponding control strategy, a virtual rotational variable is introduced into the DRG, enabling the DRG to exhibit synchronous generator characteristics during grid transients. For example, the paper "Research on Control of PV Grid-Connected System Based on Virtual Synchronous Generator," published in the December 2018 issue of the *Journal of Electrical Engineering* by Bao Guangqing, Tan Hongtao, et al., constructs a PV grid-connected control system equipped with an energy storage system on the DC side, employs a virtual synchronous generator control strategy, and analyzes the strategy and its implementation process.

[0003] By retaining only the energy storage portion at the inverter input of the aforementioned photovoltaic grid-connected system based on a virtual synchronous machine, the following can be obtained: Figure 1 The diagram shows an energy storage VSG system. An energy storage virtual synchronous generator (energy storage VSG control strategy) is connected in parallel with the energy storage to the inverter, and the energy from the energy storage system is exchanged with the grid through the inverter. On the DC side, through power coordination control, the energy storage smooths out fluctuations in the grid or load and maintains a constant DC bus voltage. At this time, the grid-connected inverter implements VSG control on the AC side. However, this control structure heavily relies on the performance of the energy storage system and requires high reliability. Once the energy storage system fails or its capacity is insufficient, the energy storage output power, DC bus voltage, etc., will not remain constant, causing the energy storage VSG control system to malfunction. Summary of the Invention

[0004] The purpose of this invention is to provide a multimodal decomposition control method for VSG systems based on the state of charge of energy storage. This method constructs an energy storage charge control law based on the discharge power state of the energy storage battery, and achieves dynamic compensation of energy storage power through multimodal decomposition control. This ensures the boundedness of the closed-loop system signal and the rapid convergence of tracking errors, enabling complementary and coordinated operation of VSG systems applied to energy storage based on multimodal decomposition control, and providing continuous and reliable frequency support capability for the system.

[0005] To achieve the above objectives, the technical solution adopted by this invention is as follows:

[0006] A multimodal decomposition control method for a VSG system based on the state of charge of energy storage is proposed, which proceeds in the following steps:

[0007] S1. Obtain the predicted remaining power of energy storage based on the remaining charge of the battery;

[0008] S2. Construct a nonlinear charge-power state system for energy storage to obtain the predicted energy storage power considering the remaining power.

[0009] S3. Define a convolutional network algorithm for multimodal decomposition control and obtain the feedback power of the modal convolution algorithm;

[0010] S4. Use the difference between the reference output power of the energy storage and the feedback power of the modal convolution control algorithm as the reference input power of the VSG active control loop.

[0011] S5. Obtain the change in reference input power of the active control loop of the energy storage VSG, and further obtain the steady-state equation of the VSG system applied to energy storage.

[0012] As a limitation, step S1 shall be performed in the following order:

[0013] S11, The remaining battery power is expressed as follows:

[0014]

[0015] Where SOC represents the remaining battery capacity, C z Q represents the total battery capacity. f This indicates the amount of electricity discharged by the battery;

[0016] S12. First, calculate the initial state of charge (SOC) of the battery, then use the charge accumulation method to calculate the battery SOC(t).

[0017]

[0018] Among them, i b Indicates the output current of the battery;

[0019] S13. Ignoring the losses of the energy storage converter, we can approximately obtain...

[0020] U b ·i b =U dc ·i d =P ③

[0021] Among them, U b Indicates the input voltage of the battery converter interface; U dc Indicates the DC bus voltage, i d P represents the output current of the battery converter interface, and P represents the output power of the battery converter interface.

[0022] S14. Combining equations ② and ③, we obtain the formula for calculating the SOC of the battery.

[0023]

[0024] S15. By controlling the output current of the converter, the state of charge (SOC) of the battery is changed.

[0025]

[0026] Where λ represents the loss coefficient, e represents the natural constant, and t represents the discharge time;

[0027] S16. Further, the predicted surplus power of energy storage is expressed as:

[0028]

[0029] Among them, K t denoted as the energy storage remaining power adjustment coefficient, and s represents the Laplace operator.

[0030] As a further limitation, step S2 is performed in the following order:

[0031] S21. Construct a nonlinear charged power state system f(P) for energy storage.

[0032]

[0033] in, This indicates the state of the energy storage system at time t when it is charged. Let represent the average energy, φ represent the unknown nonlinear smooth decreasing function, g represent the radial basis function, D represent the input energy error, and x1 represent the initial value of the current factor x, where |x1| ≥ 0.5. This represents the estimated charge of the energy storage system.

[0034] Consider the input energy error model

[0035] D(i b ) = ci b +d ⑧

[0036] Where c represents the system control input energy coefficient, and d represents the energy balance coefficient;

[0037]

[0038]

[0039] Where c1>0 represents the iteration slope of the energy function on the negative half-axis; c2>0 represents the iteration slope of the energy function on the positive half-axis; d1<0 and d2>0 both represent energy balance compensation factors.

[0040] S22, Define the radial basis function g as

[0041]

[0042] Among them, g0 T Let g denote the initial value of the transpose of the radial basis function g, fg(x) denote the approximation error function, fg(x) ≤ ε, ε denote the stability margin constant, and R(i b ) represents the odd function components of the radial basis; and has

[0043]

[0044] Where α and β represent adaptive correction parameters, μ represents the center of the odd function, and v represents the width of the odd function;

[0045] S23. An unknown nonlinear smooth decreasing function φ maintains the energy storage system in a stable state over a fast finite time, i.e.

[0046]

[0047] Where ρ represents the asymptotic control value, and ω0 represents the virtual angular frequency of the energy storage system;

[0048] The energy storage charge power current follower function is expressed as follows:

[0049]

[0050] S24, simultaneous formula ⑦ Then the stored energy charge power P b for

[0051]

[0052] S25, The reactive power-voltage regulation equation for the energy storage VSG system is:

[0053]

[0054] Among them, K s K represents the reactive equivalent inertia coefficient. v Represents the reactive power-voltage regulation coefficient, ΔU ref Q represents the reactive power-voltage regulation. ref Q represents the reference reactive power of a VSG system used for energy storage. e U represents the reactive power output of a VSG system used in energy storage. d U represents the effective value of the voltage on the d-axis in the dq coordinate system. vd U represents the virtual internal potential along the d-axis in the dq coordinate system. ωd Indicates the feedback voltage of the remaining energy storage power;

[0055] U ωd =p n / I d

[0056] Among them, I d p represents the effective value of the d-axis current in the dq coordinate system. n This represents the feedback power of the modal convolution control algorithm;

[0057] S26, simultaneous formula⑥ The predicted energy storage power, taking into account the remaining power, is:

[0058]

[0059] As a further limitation, step S3 is performed in the following order:

[0060] S31. Define a convolutional network algorithm controlled by multimodal decomposition:

[0061]

[0062] Where, ξ i Let ξ1 represent the error variable. i The initial value, p1 represents p n The initial value of v i-1 p represents the modal control law. i Represents the control coefficient of the i-th mode decomposition;

[0063] S32, Formula Combining the energy storage charge power formula (⑦), we get

[0064]

[0065] S33. Construct the following convolutional network function.

[0066]

[0067] Where, δ s This represents the virtual power angle output by the VSG system used for energy storage; k represents the frequency reference value. ω Indicates convolution gain;

[0068] Pair Differentiating gives

[0069]

[0070] in, φ1 represents the initial value of the convolution function. This represents the average error of the angular frequency ω1 of the VSG system applied to energy storage. express The initial value;

[0071] Combined formula Approximating the convolution function with a radial basis function, we can obtain

[0072]

[0073] where p1 = [p 2 1, p n , fh(p1)<u1 represents the acceleration error, and u1 represents the minimum acceleration factor;

[0074] S34.联立公式 Construct the minimum modal norm

[0075]

[0076] where η = ||φ|| ξ / 2 represents the norm coefficient, and R T represents the transpose of the odd function component of the radial basis;

[0077] S35.构造模态控制律v i

[0078]

[0079] where b represents the modal adaptation constant, b > 0, and further simplification gives

[0080]

[0081] S36.设计自适应模态控制律

[0082]

[0083] where h represents the modal decomposition factor, h ∈ (0.5, 1);

[0084] S37.联立公式 to obtain the feedback power of the modal convolution algorithm

[0085]

[0086] where k vsg represents the modal weighting coefficient.

[0087] As a further limitation, step S4 is carried out in the following order of steps:

[0088] S41. In the active - frequency control link of the VSG system applied to energy storage, the virtual rotor motion equation is expressed as

[0089]

[0090] Where, ω vsg P represents the virtual angular velocity of a VSG system used for energy storage. ref H represents the reference input power of the active power control loop in a VSG system used for energy storage. v This represents the inertia coefficient of a VSG system applied to energy storage, and has...

[0091] P ref =P * -P n

[0092] Among them, P n P represents the feedback power of the modal convolution control algorithm. * Indicates the reference output power of the energy storage system;

[0093] S42, through virtual damping element D v Obtain the damping power

[0094] P d =D v (ω vsg +ω n )

[0095] Where, ω n Represents the angular frequency of the grid connection point voltage.

[0096] S43, The active power output of the energy storage after adopting VSG control is obtained as follows:

[0097]

[0098] Among them, U ref U0 represents the voltage amplitude of the VSG obtained through the virtual excitation link in the reactive power-voltage control of the energy storage control system, Z represents the grid connection point voltage of the energy storage system, and Z represents the grid connection impedance of the energy storage control system.

[0099] As a further limitation, step S5 is performed in the following order:

[0100] S51, When the initial angular velocity of VSG is ω vsg At the equilibrium point, the equation is... Linearizing the rotor motion equations, we obtain the small perturbation equations as follows:

[0101]

[0102] Where, δ s0 This represents the initial value of the VSG power angle;

[0103] S52. Ignoring the response delay and friction loss of the energy storage battery, obtain the change in reference input power of the active power control loop of the VSG system applied to energy storage.

[0104]

[0105] Joint Mode have to

[0106]

[0107] Where, k m Represents the control parameters for multimodal decomposition, and has

[0108]

[0109] Where T represents the time constant, with a value of (0.2, 1);

[0110] S53, Further Resolution The steady-state equations of the VSG system applied to energy storage are obtained.

[0111]

[0112] The present invention, by adopting the above-described technical solution, achieves the following technical advancements compared to existing technologies:

[0113] (1) The VSG control strategy designed in this invention constructs the energy storage charge control law based on the discharge power state of the energy storage battery. Through multimodal decomposition control, it can realize dynamic compensation of energy storage power to ensure the boundedness of the closed-loop system signal and the rapid convergence of tracking error. It realizes the complementary and coordinated operation of the VSG system applied to energy storage based on multimodal decomposition control, and provides the system with continuous and reliable frequency support capability.

[0114] (2) In this invention, the odd function μ of the radial basis is always greater than the adaptive correction parameter β, and the norm η of the modal control law is a positive real number. Therefore, the energy storage power of the SOC feedback can guarantee fast response characteristics and smooth output effect; the constructed multimodal convolutional network error variable ξ i The adaptive modal control law is approximated by radial basis functions. To prevent secondary frequency disturbances to the system, and to ensure that the adaptive constant α takes a value on the positive half-axis, the multimodal decomposition control parameter k is adjusted. m The selected value can provide more frequency regulation capacity during the frequency regulation phase of the energy storage VSG to meet the capacity requirements of the system frequency and prevent the system frequency from deteriorating sharply, which could lead to safety and stability problems.

[0115] (3) Because traditional energy storage VSG systems do not take into account their own SOC, their suppression effect on frequency changes is poor. Under load disturbances, the system frequency changes drastically and the steady-state frequency deviation is also large during the steady-state recovery phase. When the power is exhausted in the later stage, it causes repeated frequency fluctuations, resulting in a large drop in transient frequency. At the same time, it causes the remaining energy storage to over-output, affecting the energy storage life. Therefore, the traditional VSG system used for energy storage has a weak ability to participate in system frequency regulation. The energy storage control method designed based on this invention has the best frequency response effect. It fully considers the system's primary frequency regulation requirements and the energy storage's own need to maintain SOC. It plays an effective suppression role in the early stage of frequency disturbance. As the system frequency recovers, the energy storage enters its own state of charge maintenance state, so that its own SOC is also at a high level. There is no overcharging or over-discharging phenomenon, and the energy storage unit's SOC level under low state of charge is effectively maintained.

[0116] (4) Traditional VSG control strategies have problems such as insufficient frequency regulation capacity and unsatisfactory frequency regulation performance. Furthermore, after a disturbance occurs, the output power control exceeds the limit, the convergence is poor, and the regulation speed is slow. However, the energy storage output control strategy based on multimodal decomposition control provided by this invention can achieve complementary and coordinated operation of energy storage and VSG control systems. After the VSG regulation ends, the active and reactive power output of the energy storage system can be restored to the initial operating state. Moreover, the DC voltage does not fluctuate significantly during the disturbance, and the AC load is not affected.

[0117] This invention belongs to the field of wind power generation control technology. It has superior frequency regulation performance and can improve the utilization rate of energy storage systems and reduce the energy storage capacity configuration. Attached Figure Description

[0118] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0119] In the attached diagram:

[0120] Figure 1 This is a schematic diagram of the structural principle of the energy storage VSG system in the prior art of this invention;

[0121] Figure 2 This is a schematic diagram of the energy storage VSG frequency modulation control system based on multimodal decomposition control in an embodiment of the present invention.

[0122] Figure 3 This is a comparison curve of the system frequency change response in an embodiment of the present invention;

[0123] Figure 4 This is a comparison curve of the SOC change of the energy storage system during the frequency modulation process in an embodiment of the present invention;

[0124] Figure 5This is a comparison curve of energy storage output depth in an embodiment of the present invention;

[0125] Figure 6 This is a comparison curve of the output active power of the VSG system applied to energy storage in an embodiment of the present invention;

[0126] Figure 7 This is a comparison curve of the reactive power output of the VSG system applied to energy storage in an embodiment of the present invention;

[0127] Figure 8 This is a comparison curve of DC voltage in a VSG system applied to energy storage in an embodiment of the present invention. Detailed Implementation

[0128] The preferred embodiments of the present invention will now be described with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the scope of the invention.

[0129] Example: Multimodal decomposition control method for VSG system based on energy storage state of charge

[0130] This embodiment is performed in the following order:

[0131] S1. Obtain the predicted remaining power of energy storage based on the remaining charge of the battery;

[0132] S2. Construct a nonlinear charge-power state system for energy storage to obtain the predicted energy storage power considering the remaining power.

[0133] S3. Define a convolutional network algorithm for multimodal decomposition control and obtain the feedback power of the modal convolution algorithm;

[0134] S4. Use the difference between the reference output power of the energy storage and the feedback power of the modal convolution control algorithm as the reference input power of the VSG active control loop.

[0135] S5. Obtain the change in reference input power of the active control loop of the energy storage VSG, and further obtain the steady-state equation of the VSG system applied to energy storage.

[0136] like Figure 2 As shown, f(SOC) represents the remaining capacity of the battery. The battery's f(SOC) is calculated using a combination of the charge accumulation method and the open-circuit voltage method, yielding the predicted remaining energy storage power P. a Construct the nonlinear energy storage power state system f(P) and the energy storage power current follower function fal to obtain the energy storage power P. b The feedback power p of the modal convolution algorithm is obtained through the Modal Convolution Network (MCN) algorithm controlled by multimodal decomposition. nAs the feedback power input of the VSG control system, the VSG system based on the energy storage state of charge (SFC) multimodal decomposition control is finally used to realize the complementary and coordinated operation of the VSG system applied to energy storage.

[0137] Step S1 is performed in the following order:

[0138] S11, The remaining battery power is expressed as follows:

[0139]

[0140] Where SOC represents the remaining battery capacity, C z Q represents the total battery capacity. f This indicates the amount of electricity discharged by the battery;

[0141] S12. The SOC of the battery is calculated using a combination of the charge accumulation method and the open-circuit voltage method. The specific steps are as follows: first, calculate the initial state of charge (SOC) of the battery (0), and then calculate the battery SOC (t) using the charge accumulation method.

[0142]

[0143] Among them, i b Indicates the output current of the battery;

[0144] S13. Since the SOC of the energy storage battery pack changes slowly, its terminal voltage can be considered approximately constant. Ignoring the losses of the energy storage converter, we can approximate the following:

[0145] U b ·i b =U dc ·i d =P ③

[0146] Among them, U b Indicates the input voltage of the battery converter interface; U dc Indicates the DC bus voltage, i d P represents the output current of the battery converter interface, and P represents the output power of the battery converter interface.

[0147] S14. Combining equations ② and ③, we obtain the formula for calculating the SOC of the battery.

[0148]

[0149] S15. The SOC of the battery is related to the output current of the converter. The SOC of the battery can be changed by controlling the magnitude of the converter's output current.

[0150]

[0151] Where λ represents the loss coefficient, e represents the natural constant, and t represents the discharge time;

[0152] S16. Further, the predicted surplus power of energy storage is expressed as:

[0153]

[0154] Among them, K t denoted as the energy storage remaining power adjustment coefficient, and s represents the Laplace operator.

[0155] Step S2 is performed in the following order:

[0156] S21. Construct a nonlinear charged power state system f(P) for energy storage.

[0157]

[0158] in, This indicates the state of the energy storage system at time t when it is charged. Let represent the average energy, φ represent the unknown nonlinear smooth decreasing function, g represent the radial basis function, D represent the input energy error, and x1 represent the initial value of the current factor x, where |x1| ≥ 0.5. This represents the estimated charge of the energy storage system.

[0159] Consider the input energy error model

[0160] D(i b ) = ci b +d ⑧

[0161] Where c represents the system control input energy coefficient, and d represents the energy balance coefficient;

[0162]

[0163]

[0164] Where c1>0 represents the iteration slope of the energy function on the negative half-axis; c2>0 represents the iteration slope of the energy function on the positive half-axis; d1<0 and d2>0 both represent energy balance compensation factors.

[0165] S22, The radial basis function g has the ability to approximate any nonlinear function. The radial basis function g is defined as follows:

[0166]

[0167] Among them, g0 T Let g denote the initial value of the transpose of the radial basis function g, fg(x) denote the approximation error function, fg(x) ≤ ε, where ε represents the minimum ideal weight, and R(ib ) represents the odd function components of the radial basis; and has

[0168]

[0169] Where α and β represent adaptive correction parameters, μ represents the center of the odd function, and v represents the width of the odd function;

[0170] S23. An unknown nonlinear smooth decreasing function φ maintains the energy storage system in a stable state over a fast finite time, i.e.

[0171]

[0172] Where ρ represents the asymptotic control value, and ω0 represents the virtual angular frequency of the energy storage system;

[0173] The energy storage charge power current follower function is expressed as follows:

[0174]

[0175] S24, simultaneous formula ⑦ Then the stored energy charge power P b for

[0176]

[0177] S25. The reactive power-voltage control link of the energy storage VSG system simulates the excitation current control method of a synchronous generator to achieve voltage amplitude regulation, possessing excitation regulation inertia. The reactive power-voltage regulation equation is:

[0178]

[0179] Among them, K s K represents the reactive equivalent inertia coefficient. v Represents the reactive power-voltage regulation coefficient, ΔU ref Q represents the reactive power-voltage regulation. ref Q represents the reference reactive power of a VSG system used for energy storage. e U represents the reactive power output of a VSG system used in energy storage. d U represents the effective value of the voltage on the d-axis in the dq coordinate system. vd U represents the virtual internal potential along the d-axis in the dq coordinate system. ωd Indicates the feedback voltage of the remaining energy storage power;

[0180] U ωd =p n / I d

[0181] Among them, I dp represents the effective value of the d-axis current in the dq coordinate system. n This represents the feedback power of the modal convolution control algorithm;

[0182] S26, simultaneous formula⑥ The predicted energy storage power, taking into account the remaining power, is:

[0183]

[0184] Step S3 is performed in the following order:

[0185] S31. Define the multimodal decomposition-controlled convolutional network (MCN) algorithm:

[0186]

[0187] Where, ξ i Let ξ1 represent the error variable. i The initial value, p1 represents p n The initial value of v i-1 p represents the modal control law. i Represents the control coefficient of the i-th mode decomposition;

[0188] S32, Formula Combining the energy storage charge power formula (⑦), we get

[0189]

[0190] S33. Construct the following convolutional network function.

[0191]

[0192] Where, δ s This represents the virtual power angle output by the VSG system used for energy storage; k represents the frequency reference value. ω Indicates convolution gain;

[0193] Pair Differentiating gives

[0194]

[0195] in, φ1 represents the initial value of the convolution function. This represents the average error of the angular frequency ω1 of the VSG system applied to energy storage. express The initial value;

[0196] Combining formulas Radial basis functions approximate convolution functions, yielding...

[0197]

[0198] Among them, p1 = [p 2 1, p n , fh(p1)<u1 represents the acceleration error, and u1 represents the minimum acceleration factor;

[0199] S34. Simultaneously solve the formula Construct the minimum modal norm

[0200]

[0201] Among them, η = ||φ|| ξ / 2 represents the norm coefficient, and R T represents the transpose of the odd function component of the radial basis;

[0202] S35. Construct the modal control law v i

[0203]

[0204] Among them, b represents the modal adaptive constant, b > 0, and further simplification gives

[0205]

[0206] S36. Design the adaptive modal control law

[0207]

[0208] Among them, h represents the modal decomposition factor, h ∈ (0.5, 1);

[0209] S37. Simultaneously solve the formula to obtain the feedback power of the modal convolution algorithm

[0210]

[0211] Among them, k vsg represents the modal weighting coefficient. <id = 662>

[0212] Step S4 is carried out in the following order:

[0213] S41. In the active - frequency control link of the VSG system applied to energy storage, the virtual rotor motion equation is expressed as

[0214]

[0215] Among them, ω vsg represents the virtual angular velocity of the VSG system applied to energy storage, and P refH represents the reference input power of the active power control loop in a VSG system used for energy storage. v This represents the inertia coefficient of a VSG system applied to energy storage, and has...

[0216] P ref =P * -P n

[0217] Among them, P n P represents the feedback power of the modal convolution control algorithm. * Indicates the reference output power of the energy storage system;

[0218] S42, through virtual damping element D v Obtain the damping power

[0219] P d =D v (ω vsg +ω n )

[0220] Where, ω n Represents the angular frequency of the grid connection point voltage.

[0221] S43, The active power output of the energy storage after adopting VSG control is obtained as follows:

[0222]

[0223] Among them, U ref U0 represents the voltage amplitude of the VSG obtained through the virtual excitation link in the reactive power-voltage control of the energy storage control system, Z represents the grid connection point voltage of the energy storage system, and Z represents the grid connection impedance of the energy storage control system.

[0224] Step S5 is performed in the following order:

[0225] S51, When the initial angular velocity of VSG is ω vsg At the equilibrium point, the equation is... Linearizing the rotor motion equations, we obtain the small perturbation equations as follows:

[0226]

[0227] Where, δ s0 This represents the initial value of the VSG power angle;

[0228] S52. Ignoring the response delay and friction loss of the energy storage battery, obtain the change in reference input power of the energy storage VSG active power control loop.

[0229]

[0230] Joint Mode have to

[0231]

[0232] Where, k m Represents the control parameters for multimodal decomposition, and has

[0233]

[0234] Where T represents the time constant, with a value of (0.2, 1);

[0235] S53, Further Resolution The steady-state equations of the VSG system applied to energy storage are obtained.

[0236]

[0237] From the formula Analysis shows that the energy storage output control strategy based on multimodal decomposition control provided in this embodiment can achieve complementary and coordinated operation of the energy storage and VSG control systems. The odd function μ of the radial basis function is constantly greater than the adaptive correction parameter β, and the norm η of the modal control law is a positive real number. Therefore, the energy storage power fed back by the SOC can guarantee fast response characteristics and smooth output performance. The constructed multimodal convolutional network error variable ξ... i The adaptive modal control law is approximated by radial basis functions. To prevent secondary frequency disturbances to the system, and to ensure that the adaptive constant α takes a value on the positive half-axis, the multimodal decomposition control parameter k is adjusted. m The selected value can provide more frequency regulation capacity during the frequency regulation phase of the energy storage VSG to meet the system frequency capacity requirements and prevent the system frequency from deteriorating rapidly, which could lead to safety and stability issues.

[0238] A virtual synchronous motor simulation model of the energy storage system was built in MATLAB / Simulink simulation software. The energy storage battery consists of 12 3.2V / 200Ah batteries connected in series to form a 38.4V / 200Ah low-voltage battery pack. The measured open-circuit voltage is 39.4V. Its bidirectional DC / DC link maintains the DC bus voltage stability, and the system output frequency is 50Hz. The time constant T of the energy storage power prediction control link is set to 0.08s, and the time constant T2 of the VSG output power feedback control link is set to 0.05s.

[0239] To illustrate the effectiveness of the control method provided in this embodiment, a load disturbance of 0.015 pu was introduced under grid operating conditions. The frequency adjustment curves under the control method provided in this embodiment (MCN-VSG) and the traditional control method (VSG) are shown below. Figure 3As shown, the variation curves of the energy storage system with an initial energy storage value SOC = 0.9 under the control method (MCN-VSG) and the traditional control method (VSG) provided in this embodiment are as follows. Figure 4 As shown, the energy storage output depth comparison curve is as follows: Figure 5 As shown.

[0240] Depend on Figure 3 It is known that traditional energy storage VSG systems, due to their lack of consideration for their own State of Charge (SOC), have poor suppression of frequency changes, resulting in drastic frequency fluctuations under load disturbances and significant steady-state frequency deviations during the steady-state recovery phase. Compared to traditional control methods, the multimodal decomposition control strategy (MCN-VSG) for VSG systems based on the state of charge of energy storage in this embodiment has a maximum frequency deviation of less than 1% and can recover stable operation within 4 seconds after the disturbance occurs. Figure 4 It can be seen that the control strategy in this embodiment achieves frequency stability when the remaining power of the energy storage system reaches 50%, while the traditional control method requires a remaining power of 40%. Based on the optimal frequency response of the energy storage control method in this embodiment, which fully considers the system's primary frequency regulation requirements and the energy storage's need to maintain its state of charge (SOC), it effectively suppresses frequency disturbances in the initial stage. As the system frequency recovers, the energy storage enters a state of charge maintenance state, maintaining its SOC at a relatively high level without overcharging or over-discharging, effectively maintaining the SOC level of the energy storage unit at a low state of charge. Therefore, the traditional control method causes repeated frequency fluctuations and a significant drop in transient frequency when the energy is depleted in the later stages. Figure 5 It can be seen that, compared with traditional control, the output depth of the VSG system multimodal decomposition control strategy based on energy storage state of charge can reach 70%. Therefore, the VSG system applied to energy storage in this embodiment has a strong ability to participate in system frequency regulation.

[0241] Figures 6-8 The figures show the active power comparison curves, reactive power comparison curves, and voltage output comparison curves of the VSG system applied to energy storage under the control method (MCN-VSG) provided in this embodiment and the traditional control method (VSG). Figure 6 It can be seen that the multimodal decomposition control strategy (MCN-VSG) for the VSG system based on the state of charge of energy storage in this embodiment can ensure that the active power output is stably maintained at around 97% during frequency regulation of the energy storage system, with minimal impact on the energy storage system. In contrast, the traditional control strategy achieves an active power output of 99%, which affects the service life of the energy storage system. Figure 7 It is known that traditional energy storage systems experience overshoot when reactive power exceeds the upper limit by 1% during frequency regulation, resulting in excessive output. However, the multimodal decomposition control strategy (MCN-VSG) based on the energy storage state of charge (VSG) system in this embodiment adjusts the system near the rated value, effectively protecting the energy storage system. Figure 8It is known that traditional strategies reduce frequency regulation to 97%, while the multimodal decomposition control strategy (MCN-VSG) based on the state of charge of the energy storage system in this embodiment can ensure that the DC voltage of the energy storage system operates at 2% above the rated value, effectively protecting the impact of the energy storage system on the DC capacitor during frequency regulation. In summary, traditional VSG control strategies suffer from insufficient frequency regulation capacity and unsatisfactory frequency regulation performance, and also exhibit problems such as output power exceeding control limits, poor convergence, and slow regulation speed after disturbances. However, the energy storage output control strategy based on multimodal decomposition control provided in this embodiment can achieve complementary and coordinated operation of the energy storage and VSG control systems. After VSG regulation, the active and reactive power output of the energy storage system can recover to the initial operating state, and the DC voltage does not fluctuate significantly during the disturbance, and the AC load is not affected.

[0242] In summary, the energy storage VSG control strategy based on multimodal decomposition control provided in this embodiment has superior frequency regulation performance and can improve the utilization rate of energy storage systems and reduce the energy storage capacity configuration.

Claims

1. A multimodal decomposition control method for a VSG system based on energy storage state of charge, characterized in that, Follow these steps in sequence: S1. Obtain the predicted remaining power of energy storage based on the remaining charge of the battery; S2. Construct a nonlinear charge-power state system for energy storage to obtain the predicted energy storage power considering the remaining power. S3. Define a convolutional network algorithm for multimodal decomposition control and obtain the feedback power of the modal convolution algorithm; S4. Use the difference between the reference output power of the energy storage and the feedback power of the modal convolution control algorithm as the reference input power of the VSG active control loop. S5. Obtain the change in reference input power of the active control loop of the energy storage VSG, and further obtain the steady-state equation of the VSG system applied to energy storage. Step S2 is performed in the following order: S21. Constructing a nonlinear charge-power state system for energy storage. in, This indicates the state of the energy storage system at time t when it is charged. This represents the average energy value. Represents an unknown nonlinear smooth decreasing function. Represents radial basis functions. Indicates input energy error. Indicates current factor initial value and , This represents the estimated charge of the energy storage system. Consider the input energy error model in, d represents the system control input energy coefficient, and d represents the energy balance coefficient. in, >0 indicates the slope of the energy function on the negative half-axis; >0 indicates the iterative slope of the energy function on the positive half-axis; d1<0 and d2>0 both represent energy balance compensation factors; S22, Define radial basis functions for in, Represents radial basis functions Transpose initial value, This represents the approximation error function. , Represents the stability margin constant. Denotes the odd functional components of the radial basis; and has in, , Indicates adaptive correction parameters. Indicates the center of the odd function. Indicates the width of the odd function; S23, Unknown Nonlinear Smooth Decreasing Function Maintaining the energy storage system to a stable state within a rapid and finite timeframe, i.e. in, Indicates the asymptotic control value. This represents the virtual angular frequency of the energy storage system; The energy storage charge power current follower function is expressed as follows: S24, simultaneous formula Then the energy storage charge power for S25, The reactive power-voltage regulation equation for the energy storage VSG system is: in, Represents the reactive equivalent inertia coefficient. This represents the reactive power-voltage regulation coefficient. Indicates reactive power-voltage regulation. This represents the reference reactive power of the VSG system used for energy storage. This indicates the reactive power output of the VSG system used in energy storage. This represents the effective value of the voltage along the d-axis in the dq coordinate system. This represents the virtual internal potential along the d-axis in the dq coordinate system. Indicates the feedback voltage of the remaining energy storage power; in, This represents the effective value of the d-axis current in the dq coordinate system. This represents the feedback power of the modal convolution control algorithm; S26, simultaneous formula The predicted energy storage power must take into account the remaining power. ； Step S3 is performed in the following order: S31. Define a convolutional network algorithm controlled by multimodal decomposition: in, Represents the error variable. express initial value, express initial value, Represents the modal control law. Represents the control coefficient of the i-th mode decomposition; S32, Formula Simultaneous Energy Storage Charge Power Formula have to S33. Construct the following convolutional network function. in, This represents the virtual power angle output by the VSG system used for energy storage; Indicates the frequency reference value. Indicates convolution gain; Pair Differentiating gives in, , , This represents the initial value of the convolution function. This represents the angular frequency of a VSG system used for energy storage. The average error, express The initial value; Combining formulas Radial basis functions approximate convolution functions, yielding... in, , Indicates acceleration error, Indicates the minimum acceleration factor; S34, simultaneous formula Constructing the minimum modal norm in Represents the norm coefficients, Represents the transpose of the odd function components of the radial basis; S35, Constructing Modal Control Laws Where b represents the modal adaptive constant, b > 0, further resolved to obtain S36. Design an adaptive modal control law Where h represents the mode decomposition factor, h∈(0.5,1); S37, simultaneous formula The feedback power of the modal convolution algorithm in, Represents the modal weighting coefficients; Step S4 is performed in the following order: S41. The motion equation of the virtual rotor in the active-frequency control link of the VSG system applied to energy storage is expressed as follows: in, This represents the virtual angular velocity used in VSG systems for energy storage. This represents the reference input power of the active power control loop in a VSG system used for energy storage. This represents the inertia coefficient of a VSG system applied to energy storage, and has... in, This represents the feedback power of the modal convolution control algorithm. Indicates the reference output power of the energy storage system; S42, Through virtual damping element Obtain the damping power in, Represents the angular frequency of the grid connection point voltage. S43, The active power output of the energy storage after adopting VSG control is obtained as follows: in, This represents the VSG voltage amplitude obtained through the virtual excitation stage in the reactive power-voltage control of the energy storage control system. This indicates the voltage at the grid connection point of the energy storage system. This indicates the grid connection impedance of the energy storage control system.

2. The multimodal decomposition control method for a VSG system based on energy storage state of charge as described in claim 1, characterized in that, Step S1 is performed in the following order: S11, The remaining battery power is expressed as follows: Here, SOC represents the remaining battery capacity. Indicates the total battery capacity. This indicates the amount of electricity discharged by the battery; S12. First, calculate the initial state of charge (SOC) of the battery, then use the charge accumulation method to calculate the battery SOC(t). in, This indicates the output current of the battery; S13. Ignoring the losses of the energy storage converter, we can approximately obtain... in, This indicates the input voltage of the battery converter interface; Indicates the DC bus voltage. P represents the output current of the battery converter interface, and P represents the output power of the battery converter interface. S14, Combined Forms with formula The formula for calculating the SOC of the battery is obtained. ； S15. By controlling the output current of the converter, the state of charge (SOC) of the battery is changed. in, represents the loss coefficient, e represents the natural constant, and t represents the discharge time; S16. Further, the predicted surplus power of energy storage is expressed as: ； in, This represents the energy storage surplus power adjustment coefficient. This represents the Laplace operator.

3. The multimodal decomposition control method for a VSG system based on energy storage state of charge as described in claim 1, characterized in that, Step S5 is performed in the following order: S51, when the initial angular velocity of VSG is At the equilibrium point, the equation is... Linearizing the rotor motion equations, we obtain the small perturbation equations as follows: in, This represents the initial value of the VSG power angle; S52. Ignoring the response delay and friction loss of the energy storage battery, obtain the change in reference input power of the active power control loop of the VSG system applied to energy storage. Joint ,Mode have to in, Represents the control parameters for multimodal decomposition, and has Where T represents the time constant, with values ​​ranging from 0.2 to 1. S53, Further Resolution The steady-state equations of the VSG system applied to energy storage are obtained. 。

Images