Region flexible device based on network construction type SVG (static var generator) and control method thereof

Abstract

The invention discloses a transformer area flexible device based on a network construction type SVG (static var generator) and a control method thereof. The device comprises a distributed photovoltaic device, a multi-active bridge converter, the network construction type SVG, a parallel compensator and an energy storage device, three converters on the power supply side of the multi-active bridge converter and one converter on the power grid side form three direct current links; establishing a direct current port voltage model of each converter; based on the DC port voltage model, establishing a decoupling control model of the device; performing iterative solution on the decoupling control model of the device by adopting an improved extreme learning machine method to obtain a phase-shift ratio of each direct-current port; and determining a modulation voltage phase of each converter based on the phase-to-shift ratio of each DC port so as to control the multi-active bridge converter. Energy exchange between photovoltaic and energy storage is supported through a multi-active bridge converter, and energy interaction between a power grid and a load is realized through an SVG (Static Var Generator).

Description

Technical Field

[0001] This invention relates to the field of power grid data processing technology, and in particular to a flexible distribution transformer (SVT) device that can integrate distributed photovoltaic and energy storage and its decoupling control method. Background Technology

[0002] In the power system field, ensuring power quality is crucial, and the grid-connected variable transformer substation (SVG) has emerged as a key piece of equipment. Traditional SVGs typically have a complex structure, with the presence of low-frequency isolation transformers not only increasing the size and weight of the equipment but also limiting its power density and flexibility to some extent. With the continuous development of power electronics technology, the need for performance optimization of SVGs is becoming increasingly urgent.

[0003] In recent years, SVG topologies based on multi-active bridge (MAB) converters have gradually attracted attention. This novel topology eliminates the bulky low-frequency isolation transformer and provides more power electronic interfaces for renewable energy integration, demonstrating significant application potential. However, MAB converters inherently possess nonlinear and strong cross-coupling characteristics, making the control of their DC-link voltages extremely difficult. Especially in achieving independent control of multiple DC-link voltages, traditional control methods face significant challenges, consequently affecting the maximum power point tracking (MPPT) performance of the photovoltaic array and making it difficult to maintain stable system operation and good performance under large disturbances. Currently, there is no method to solve the problem of independent control of DC-link voltages in multi-active bridge converters. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a flexible distribution area device based on a grid-type SVG and its control method. The flexible distribution area device is established based on a grid-type SVG with a multi-active bridge converter (MAB). Photovoltaic panels, energy storage systems, and the MAB are integrated on the DC side of the device. Photovoltaic, energy storage, and DC loads are connected to the DC ports of the MAB via DC capacitors. The MAB is then connected to the DC side of the grid-type SVG via DC capacitors. The AC side of the SVG is connected to the power grid through a filter. This supports direct energy exchange between photovoltaic and energy storage systems via the MAB, and also enables energy interaction between the SVG and the power grid and loads. Furthermore, it achieves precise and independent control of the three DC link voltages of the MAB converter.

[0005] The present invention adopts the following technical solution.

[0006] This invention proposes a flexible transformer substation device based on a grid-type SVG, comprising:

[0007] Distributed photovoltaic (PV) systems, multi-active bridge converters (MPB), grid-type SVG (Static Var Generator), parallel compensators, and energy storage devices; distributed PV systems and energy storage devices are connected to the power supply side of the MPB, the grid side of the MPB is connected to the grid-type SVG, and the grid-type SVG is connected to the grid via the parallel compensator.

[0008] The multi-active bridge converter includes: a high-frequency isolation transformer, n converters on the power supply side, and m converters on the grid side. The DC ports of the first to (n-1)th converters on the power supply side of the multi-active bridge converter are all connected to distributed photovoltaics, the DC port of the nth converter is connected to an energy storage device, and the AC ports of the first to nth converters are all connected to the primary winding of the high-frequency isolation transformer. The DC ports of the first to mth converters on the grid side of the multi-active bridge converter are connected to a grid-type SVG, and the AC ports of the first to mth converters are all connected to the secondary winding of the high-frequency isolation transformer.

[0009] Based on a high-frequency isolation transformer, three converters on the power supply side of the multi-active bridge converter and one converter on the grid side form three DC links; the DC ports of the three power supply-side converters participating in forming the DC links are all connected to photovoltaics, and the DC port of the one grid-side converter participating in forming the DC links is connected to a grid-type SVG.

[0010] This invention also proposes a control method for a flexible transformer substation based on a grid-type SVG, comprising:

[0011] Based on the topology of the flexible transformer substation, the DC bus capacitance and DC bus voltage of each converter on the power supply side, the equivalent connection inductance and phase shift ratio between the two DC ports on the power supply side, the switching frequency of the power supply side converter, the gain constant of the grid side converter constituting the DC link, and the phase shift ratio of each DC port on the power supply side are obtained.

[0012] Establish DC port voltage models for the three power-side converters and one grid-side converter that constitute three DC links; establish DC port voltage models for each power-side converter that does not constitute a DC link.

[0013] Based on the DC port voltage model, a decoupling control model for the device is established.

[0014] An improved extreme learning machine method is used to iteratively solve the decoupling control model of the device to obtain the phase shift ratio of each DC port; based on the phase shift ratio of each DC port, the modulation voltage phase of each converter is determined to control the multi-active bridge converter.

[0015] The DC port voltage models of the three power-side converters and one grid-side converter constituting the three DC links are as follows:

[0016]

[0017] In the formula, C τx v τx Let f be the capacitance value of the DC bus capacitor and the DC bus voltage of the converter τx. The values ​​of τx and τ′x are both one of s1, s2, s3, and e1, and the values ​​of τx and τ′x are not the same. s K is the switching frequency of the power supply converter. e1 D is the gain constant of the grid-side converter that constitutes the DC link. τxτ′x For the shift ratio between DC ports τx and τ′x, L τxτ′x K is the equivalent connection inductance between DC ports τx and τ′x. τx and K τ′x The phase shift ratios of DC ports sx and sy are respectively, i′ τx Let i be the DC current output by the converter at port τx. pvx Let x be the output current of the xth distributed photovoltaic system.

[0018] The DC port voltage models of the individual power-side converters that do not form a DC link are as follows:

[0019]

[0020] In the formula, C sx v sx Let D be the capacitance value of the DC bus capacitor and the DC bus voltage of the converter sx on the power supply side. sxsy Compared to the shift between DC ports sx and sy, L sxsy i is the equivalent connection inductance between DC ports sx and sy. pvx Let x be the output current of the xth distributed photovoltaic system.

[0021] Based on the DC port voltage model, a decoupling control model for the device is established, including:

[0022] Based on the DC port voltage model, the phase shift parameter of the DC port is introduced to establish a coupling term between the DC port voltage and the phase shift ratio. The coupling term is used as the output of the decoupling control model, and the current is used as the input of the decoupling control model. Based on the input and output of the decoupling control model, the decoupling control model of the device is established.

[0023] Based on the DC port voltage model, by introducing the phase offset parameter of the DC port, the following relationship is obtained for the n DC port voltages on the power supply side:

[0024]

[0025] In the formula, δ s1 δ s2 , ..., δ snLet A be the phase offset parameters of the n DC ports on the power supply side, and let A be the first coupling matrix. s1 v s2 ... v s(n-1) v sn Let C be the DC bus voltage of the n converters on the power supply side. s1 C s2 ... C s(n-1) C sn For the DC bus capacitors of the n converters on the power supply side;

[0026] The first coupling matrix is ​​as follows:

[0027]

[0028] in, x,y=1,2,3,…n.

[0029] Using current as the input to the decoupled control model, the following expression is satisfied:

[0030]

[0031] In the formula, i s1 i s2 ... s(n-1) i sn Let B be the input current of the n converters on the power supply side, and B be the second coupling matrix.

[0032] The second coupling matrix is ​​as follows:

[0033]

[0034] Among them, b xy =a xy ·v sy , x, y = 1, 2, 3, ... n.

[0035] The decoupling control model of the device satisfies the following expression:

[0036]

[0037] An improved extreme learning machine method is used to iteratively solve the decoupling control model of the device, obtaining the phase shift ratio of each DC port, including:

[0038] An iterative model of the inverse function of the output term of the decoupled control model is established using the extreme learning machine method; wherein, the initial values ​​of the connection weights from the hidden layer nodes to the output in the iterative model are optimized using the swarm intelligence optimization method.

[0039] The beneficial effects of this invention are as follows, compared with the prior art, at least including the following: This invention proposes a flexible distribution area device that integrates distributed photovoltaic and energy storage grid-type SVG. The distributed photovoltaic and energy storage devices are connected to the power grid through the power supply side of a multi-active bridge converter, the grid-type SVG, and a parallel compensator, forming a PV-SVG-MAB system; furthermore, based on a high-frequency isolation transformer, the three converters on the power supply side of the multi-active bridge converter and the one converter on the grid side constitute three DC links; simultaneously, by using a high-frequency isolation transformer to replace the traditional low-frequency isolation transformer, the size and weight of the equipment are significantly reduced, installation is simpler, and spatial adaptability is enhanced. At the same time, the multi-interface characteristics of the MAB converter facilitate the access of renewable energy, greatly improving the system's flexibility and scalability, making it perform excellently in distributed energy application scenarios and better adapting to complex and ever-changing energy supply and electricity demand environments.

[0040] This invention also establishes a novel decoupling control method based on Extreme Learning Machine (ELM) for decoupling control of flexible transformer substations based on grid-type SVG. Under complex operating conditions and large disturbances, it achieves precise and independent control of the three DC link voltages of the flexible transformer substation based on grid-type SVG that can integrate distributed photovoltaics. This overcomes the shortcomings of traditional control methods in handling MAB nonlinearity and cross-coupling problems, ensures the stability of DC link voltages, significantly improves the accuracy and reliability of MPPT control, enables the photovoltaic array to operate continuously and efficiently near the maximum power point, greatly improves the utilization efficiency of photovoltaic energy, and reduces energy waste. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the topology of the flexible transformer station based on the network-type SVG proposed in this invention;

[0042] Figure 2 This is a flowchart of a control method for a flexible transformer substation based on a grid-type SVG, which is also proposed in this invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0044] This invention proposes a flexible distribution network device based on a grid-type SVG, applied to a photovoltaic grid-type SVG flexible distribution network system (PV-SVG) based on a multi-active bridge converter (MAB), forming a three-phase PV-SVG-MAB structure flexible distribution network device; as shown... Figure 1 As shown, the device includes: a distributed photovoltaic system 10, a multi-active bridge converter 20, a grid-type SVG 30, a parallel compensator 40, and an energy storage device 50; Figure 1 In this multi-active bridge converter, there are: a high-frequency isolation transformer 201, n converters 202-s1, 202-s2, 202-s3, ..., 202-sn on the power supply side, and m converters 203-e1, 203-e2, ..., 203-em on the grid side. The DC ports of the first to (n-1)th converters on the power supply side are all connected to distributed photovoltaic systems, the DC port of the nth converter is connected to an energy storage device, and the AC ports of the first to nth converters are all connected to the primary winding of the high-frequency isolation transformer. The DC ports of the first to mth converters on the grid side are connected to a grid-type SVG, and the AC ports of the first to mth converters are all connected to the secondary winding of the high-frequency isolation transformer. Furthermore, based on the high-frequency isolation transformer, the three converters on the power supply side and the one converter on the grid side form three DC links. In this embodiment, the three converters 202-s1, 202-s2, ..., 202-sn on the power supply side... 2. 202-s 3, together with one grid-side converter 203-e 1, forms three DC links. The number of power-side converters n and the number of grid-side converters m can be arbitrarily configured. The DC ports of the three power-side converters participating in forming the DC links are all connected to photovoltaics. The DC port of the grid-side converter participating in forming the DC links is connected to the grid-type SVG. The DC ports of the remaining power-side converters that do not form DC links are connected to photovoltaics or energy storage devices. The DC ports of the grid-side converters that do not form DC links are short-circuited.

[0045] Figure 1 In the diagram, the equivalent inductance of the power grid is L. g The grid voltage is v Ga v Gb v Gc The grid current is i Sa i Sb i Sc The voltage of the transmission line is v La v Lb v Lc The transmission line current is i La i Lb i Lc The grid-connected line current is i sha i shb i shc The grid-connected voltage is VSa v Sb v Sc Parallel compensators include inductor L sha L shb L shc ;

[0046] The output currents of distributed photovoltaic systems 10⁻¹, 10⁻², 10⁻³, ..., 10⁻(n⁻¹) are i pv1 i pv2 ... pv(n-1) The output current of the energy storage device 50 is i s1n The DC bus capacitors of the converter on the power supply side are C s1 C s2 ... C s(n-1) C sn The DC bus voltages of the converter on the power supply side are respectively v s1 v s2 ... v s(n-1) v sn The input currents of the converters on the power supply side are i s1 i s2 ... s(n-1) i sn The leakage inductance of the DC port of the converter on the power supply side is L. s1 L s2 ... L s(n-1) L sn The DC ports of the power supply converter connected to the photovoltaic are s1, s2, ..., s(n-1), and the DC ports of the power supply converter connected to the energy storage are sn.

[0047] The leakage inductance at the DC port of the m converters on the grid side is L. e1 ... L em The DC bus capacitances of the m converters on the grid side are Cm and Cm respectively. e1 ... C em The DC bus voltages of the m converters on the grid side are respectively v e1 ... v em The input currents of the m converters on the grid side are i e1 ... em The DC ports of the grid-side converter connected to the SVG are e1, e3, ..., em.

[0048] As can be seen, in the device, the power supply side converters 202-s1, 202-s2, ..., 202-s(n-1), 202-sn, together with the grid side converters 202-e1, 202-e2, ..., 202-em, form SVG–MAB.

[0049] The SVG-MAB collects DC power from the photovoltaic and energy storage sides and converts it into high-frequency AC power. After transmission via the bus, the grid-side module converts it into voltage and current waveforms suitable for the grid or load. It also features reactive power compensation, harmonic mitigation, and voltage stabilization. In grid-connected photovoltaic and energy storage scenarios, this device can convert DC power to AC power while ensuring stable and safe system operation.

[0050] Based on the three-phase PV-SVG-MAB structure of the flexible distribution transformer (PV-SVG-MAB), the multi-active bridge converter acts as an energy router, facilitating the exchange of active power between DC ports of the grid-type SVG. Furthermore, the MAB not only provides more power electronic interfaces for renewable energy sources but also replaces the traditional bulky low-frequency transformer with a small high-frequency transformer. Therefore, compared with traditional grid-type SVG, the proposed PV-SVG-MAB structure of the flexible distribution transformer offers higher power density and greater flexibility.

[0051] Furthermore, based on the high-frequency isolation transformer, the three converters 202-s1, 202-s2, and 202-s3 on the power supply side of the multi-active bridge converter (MAB) and the one converter 202-e1 on the grid side form three DC links. At the same time, by replacing the traditional low-frequency isolation transformer with a high-frequency isolation transformer, the size and weight of the equipment are significantly reduced, installation is simpler, and spatial adaptability is enhanced. Meanwhile, the multi-interface characteristics of the MAB converter facilitate the access of renewable energy, greatly improving the system's flexibility and scalability. This makes it perform exceptionally well in distributed energy application scenarios and better adapt to complex and ever-changing energy supply and electricity demand environments.

[0052] In the embodiments, the flexible interconnection device of the grid-type SVG in the transformer area supports direct energy exchange between photovoltaic and energy storage through multiple active bridge converters, and can also realize energy interaction with the grid and load through SVG.

[0053] This invention also proposes a control method for a flexible transformer substation based on a grid-type SVG, such as... Figure 2 As shown, it includes:

[0054] Step 1: Based on the topology of the flexible transformer substation, obtain the DC bus capacitance and DC bus voltage of each converter on the power supply side, the equivalent connection inductance and phase shift ratio between the two DC ports on the power supply side, the switching frequency of the power supply side converter, the gain constant of the grid side converter constituting the DC link, and the phase shift ratio of each DC port on the power supply side.

[0055] The gyration gain between DC ports of a converter describes the degree to which a voltage or current disturbance at one DC port affects the other DC port, reflecting the dynamic coupling characteristics between the DC ports. With the DC ports of the grid-side converter connected to a grid-connected SVG as references, for any two DC ports sx and sy on the power supply side, where x, y = 1, 2, ..., n-1, n, and x ≠ y, the gyration gain between DC ports sx and sy satisfies the following expression:

[0056]

[0057] In the formula, G sxsy f is the gyration gain between DC ports sx and sy. s K is the switching frequency of the power supply converter. e1 K is the gain constant of the grid-side converter that constitutes the DC link. sx and K sy These are the phase shift ratios of DC ports sx and sy, respectively. The phase shift ratio is the normalized value of the phase offset of the control waveform per unit period, characterizing the efficiency and stability of energy transfer. sxsy The equivalent connection inductance (ECI) is the inductance between DC ports SX and SY. ECI describes the energy transfer characteristics and coupling degree between different DC ports. This inductance value affects the system's stability and energy exchange efficiency. Leakage inductance refers to the loss generated during current flow due to parasitic inductance in the circuit. DC port leakage inductance is mainly the impedance generated when current flows through the DC port due to the unavoidable parasitic inductance of the line. DC port leakage inductance affects the equivalent connection inductance because leakage inductance is an important factor in inductive coupling. sxsy The phase shift ratio between DC ports sx and sy represents the difference between their phase shift ratios, determining the energy exchange between them. The phase shift ratio reflects the dynamic coupling characteristics between the DC ports and is a result of phase shift ratio coupling modeling.

[0058] Step 2: Establish DC port voltage models for the three power-side converters and one grid-side converter that constitute the three DC links; establish DC port voltage models for each power-side converter that does not constitute a DC link.

[0059] In this embodiment, the power supply-side converters 202-s1, 202-s2, and 202-s3 and the grid-side converter 202-e1 constitute three DC links. Taking the DC port voltage of the grid-side converter constituting the three DC links as a reference, the DC port voltage models of the four converters constituting the three DC links satisfy the following relationship:

[0060]

[0061] In the formula, C τx v τx Let τx be the capacitance value of the DC bus capacitor and the DC bus voltage of the converter τx. The values ​​of τx and τ′x are both one of s1, s2, s3, and e1, and the values ​​of τx and τ′x are not the same. τxτ′x L is the phase shift ratio between DC ports τx and τ′x. τxτ′x i′ is the equivalent connection inductance between DC ports τx and τ′x. τx Let i be the DC current output by the converter at port τx. pvx Let x be the output current of the x-th distributed photovoltaic unit;

[0062] In the above formula, Equivalent to a controlled current source, the gyration gain is defined in this invention as the admittance between the DC input port and the DC output port.

[0063] In this embodiment, the converters 202-s4, ..., 202-sn on the power supply side do not form a DC link, and the DC port voltages satisfy the following relationship:

[0064]

[0065] In the formula, C sx v sx Let D be the capacitance value of the DC bus capacitor and the DC bus voltage of the converter sx on the power supply side. sxsy L is the phase shift ratio between DC ports sx and sy. sxsy i is the equivalent connection inductance between DC ports sx and sy. pvx Let x be the output current of the x-th distributed photovoltaic unit;

[0066] The output current of each distributed photovoltaic system satisfies the following expression:

[0067]

[0068] In the formula, i pvx Let P be the output current of the x-th distributed photovoltaic unit, where x = 1, 2, ..., n-1. pvx Let U be the output power of the x-th distributed photovoltaic power generation. pvx Let x be the output voltage of the x-th distributed photovoltaic unit;

[0069] This invention incorporates the calculated output current of each distributed photovoltaic system into the overall system model, enhancing the physical realism of the overall model. At the same time, the established DC port voltage model adopts phase shift ratio coupling modeling, replacing the traditional approximate model that ignores the phase shift ratio.

[0070] Step 3: Based on the DC port voltage model, establish the decoupling control model of the device;

[0071] Specifically, step 3 includes:

[0072] Step 3.1: Based on the DC port voltage model, introduce the phase shift parameter of the DC port, establish a coupling term between the DC port voltage and the phase shift ratio, and use the coupling term as the output of the decoupling control model;

[0073] Based on the DC port voltage model, by introducing the phase offset parameter of the DC port, the following relationship is obtained for the n DC port voltages on the power supply side:

[0074]

[0075] In the formula, δ s1 δ s2 , ..., δ sn Let A be the phase offset parameters of the n DC ports on the power supply side, and let A be the first coupling matrix. s1 v s2 ... v s(n-1) v sn Let C be the DC bus voltage of the n converters on the power supply side. s1 C s2 ... C s(n-1) C sn For the DC bus capacitors of the n converters on the power supply side;

[0076] The first coupling matrix is ​​as follows:

[0077]

[0078] in, x, y = 1, 2, 3, ..., n;

[0079] Step 3.2, using current as the input to the decoupling control model, satisfies the following expression:

[0080]

[0081] In the formula, i s1 i s2 ... s(n-1) i sn Let B be the input current of the n converters on the power supply side, and B be the second coupling matrix.

[0082] The second coupling matrix is ​​as follows:

[0083]

[0084] Among them, b xy =a xy·v sy x, y = 1, 2, 3, ..., n;

[0085] Step 3.3: Based on the input and output of the decoupled control model, establish the decoupled control model of the device, satisfying the following expression:

[0086]

[0087] The decoupling control model of the device is the result of a deformation of the DC port voltage model. In the model, the right side of the equal sign is the input, and the left side of the equal sign is the decoupled DC port voltage and phase shift ratio. Furthermore, by introducing a phase offset parameter, the coupling terms are decoupled to modulate the voltage phase. The decoupled DC port voltage and phase shift ratio can both be correlated and controlled by controlling the first coupling matrix, the second coupling matrix, and the input current, thereby achieving the control of the device.

[0088] Step 4: Using the improved extreme learning machine method, the decoupling control model of the device is iteratively solved to obtain the phase shift ratio of each DC port; based on the phase shift ratio of each DC port, the modulation voltage phase of each converter is determined to control the multi-active bridge converter.

[0089] Specifically, step 4 includes:

[0090] Step 4.1: Using the extreme learning machine method, an iterative model of the inverse function of the output term of the decoupled control model is established to improve control accuracy and solve the problem that traditional models are difficult to model online.

[0091] The iterative model satisfies the following relationship:

[0092]

[0093] In the formula, Let i be the predicted value for time point (t+1). t-d This indicates that the input combination for step d is a multidimensional vector, β j Let g be the connection weight from the j-th hidden layer node to the output, g() be the activation function, and a be the connection weight from the j-th hidden layer node to the output. j Let b be the input weight of the j-th hidden layer node. j For bias, L is the total number of different labeled samples.

[0094] In the Extreme Learning Machine model, the connection weight β from the last hidden node to the output is... j+1 Satisfy the following expression:

[0095]

[0096] In the formula, i n H is the input current, and H is the output matrix of the hidden layer. THere, H is the transpose of H, T is the output of the neural network, and C is the regularization coefficient, which is usually greater than 0 and is used to balance the relationship between empirical error and regularization term.

[0097] Step 4.2: Use swarm intelligence optimization to optimize the initial values ​​of the connection weights from the hidden layer nodes to the output to avoid getting trapped in local optima.

[0098] The swarm intelligence optimization method simulates the region boundary selection strategy during oyster mantis oviposition, satisfying the following expression:

[0099]

[0100] In the formula, T * Da represents the current local optimum. * and Ga * Let represent the lower and upper bounds of the spawning region, respectively. S is the radius of the dashed circle. S = 1 - t / Hmax, where t represents the current iteration number and Hmax represents the maximum iteration number. Da and Ga represent the lower and upper bounds of the optimization problem, respectively. mix() is used to linearly combine the two vectors.

[0101] The algorithm described above enhances its global search capability by introducing various biologically inspired mechanisms (such as foraging and egg-laying simulation).

[0102] During the iteration process, the position of the egg is dynamic, satisfying the following expression:

[0103] U i (t+1)=T * +a1×(U i (t)-Da * )+a2×(U i (t)-Ga * )

[0104] In the t-th iteration, U i (t) represents the position information of the i-th egg, and a1 and a2 are two independent random vectors of size 1×Q, where Q represents the dimension of the search space.

[0105] The boundary of the optimal foraging region satisfies the following expression:

[0106]

[0107] In the formula, T a Da represents the globally optimal foraging location. a and Ga a Let represent the lower and upper limits of the oviposition area, respectively. The location update (data prediction update) of the oyster cockroach satisfies the following expression:

[0108] f n(t+1)=f n (t)+k1×(f n (t)-Da q )+k2×(f n (t)-Ga q )

[0109] Among them, f n (t) represents the location information of the nth oyster cockroach in the t-th iteration, k1 is a random number following a normal distribution, k2 is a random vector belonging to (0, 1), and Da q and Ga q Let the lower and upper limits of the oviposition region in the q-th dimension be defined respectively.

[0110] The DBO algorithm consists of a population of Q target individuals, each representing a candidate solution. The position vector of the nth individual satisfies the following expression:

[0111] f i (t)=(f i1 (t),f i2 (t),......f in (t))

[0112] In the original DELM, DBO is used to optimize the initial weights of the DELM, and the objective function satisfies the following expression:

[0113] fitness=MSE(train)+MSE(test)

[0114] In the formula, the mean square error MSE satisfies the following expression:

[0115]

[0116] In the formula, k is the number of data points, and x i It is the actual target value (true value) of the i-th data point. It is the model's prediction for the i-th data point.

[0117] The iterative steps satisfy:

[0118] Initialize the particle swarm. For the i-th dung beetle rolling a dung ball, update its position using the boundary of the oyster mantis's egg-laying region or the boundary of the optimal foraging region. For the i-th small oyster cockroach searching for food, update its position using the small oyster cockroach's position update expression or the dynamic position expression of the egg ball.

[0119] The iteration termination condition is satisfied:

[0120] When t ≥ tmax, the iteration terminates, and the optimal position Ta, fitness fa, weight β, and output are given. And calculate:

[0121] fitness=MSE(train)+MSE(test)

[0122] The PV-SVG-MAB system of the present invention brings many significant advantages.

[0123] An improved Extreme Learning Machine (ELM) algorithm is used to fit the inverse function. The inverse function model from step three is used as the basis to train an improved ELM iterative model. Initial weight optimization of the improved ELM algorithm based on swarm intelligence is employed, ultimately yielding an improved ELM iterative model of the proposed inverse function, achieving the final decoupling control objective.

[0124] Based on the structure of a flexible interconnection device for distribution substations using a grid-type SVG, a model of a flexible interconnection system for distribution substations based on grid-type SVG, considering photovoltaic and energy storage access, is established. The impact of DC port power fluctuations of a multi-active bridge converter on the DC-side voltage stability is analyzed. An improved Extreme Learning Machine (ELM) fitting algorithm is used to approximate the nonlinear physical model of the multi-active bridge converter. Using the input current of each DC port as input and the converter's shift ratio as output, robust decoupling control of the multi-port DC voltage is achieved, thereby improving the system's operational stability and dynamic response performance.

[0125] The decoupling control proposed in this invention includes at least the following: analysis of the overall system mathematical model based on a multi-active-bridge converter; a series of mathematical processing and transformations based on the original multi-active-bridge converter mathematical model; and decoupling control based on an improved Extreme Learning Machine (ELM) method. The improved ELM method uses an ELM algorithm to fit an inverse function, using the inverse function model as the basis for training an improved ELM iterative model. The initial weights of the improved ELM algorithm are optimized using the DBO algorithm to obtain the improved ELM iterative model of the proposed inverse function, enabling independent control of the three DC bus voltages of a grid-connected SVG (Static Var Generator) system capable of integrating distributed photovoltaic (PV) power. By establishing an overall system mathematical model of the grid-connected SVG system based on a multi-active-bridge converter, the mathematical model of the original multi-active-bridge converter under the "Y" and "Δ" equivalent circuits is derived in detail, and the coupling relationship between the DC port voltages and shift ratios of each PV grid connection is analyzed. Based on the original multi-active-bridge converter mathematical model, a series of mathematical processing and transformations are performed, treating the overall coupling terms of each port voltage and shift ratio as control inputs, forming a new decoupling control system. The voltages and shift ratios at each port are decoupled from the overall coupling term. An improved Extreme Learning Machine (ELM) method is used to approximate the inverse function with the largest overall coupling term. The improved ELM algorithm is used to fit the inverse function, with the inverse function model from step three as the basis for training the improved ELM iterative model. Initial weight optimization using a swarm intelligence-based algorithm for the improved ELM algorithm is employed, ultimately yielding the improved ELM iterative model of the proposed inverse function, achieving the final decoupling control objective.

[0126] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0127] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0128] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0129] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions to implement various aspects of this disclosure.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A flexible transformer substation device based on a grid-type SVG, characterized in that, include: Distributed photovoltaic (PV) systems, multi-active bridge converters (MPB), grid-type SVG (Static Var Generator), parallel compensators, and energy storage devices; distributed PV systems and energy storage devices are connected to the power supply side of the MPB, the grid side of the MPB is connected to the grid-type SVG, and the grid-type SVG is connected to the grid via the parallel compensator. The multi-active bridge converter includes: a high-frequency isolation transformer, n converters on the power supply side, and m converters on the grid side. The DC ports of the first to (n-1)th converters on the power supply side of the multi-active bridge converter are all connected to distributed photovoltaics, the DC port of the nth converter is connected to an energy storage device, and the AC ports of the first to nth converters are all connected to the primary winding of the high-frequency isolation transformer. The DC ports of the first to mth converters on the grid side of the multi-active bridge converter are connected to a grid-type SVG, and the AC ports of the first to mth converters are all connected to the secondary winding of the high-frequency isolation transformer.

2. The flexible transformer station device based on a grid-type SVG according to claim 1, characterized in that, Based on a high-frequency isolation transformer, three converters on the power supply side of the multi-active bridge converter and one converter on the grid side form three DC links; the DC ports of the three power supply-side converters participating in forming the DC links are all connected to photovoltaics, and the DC port of the one grid-side converter participating in forming the DC links is connected to a grid-type SVG.

3. A control method for a flexible transformer substation based on a grid-type SVG, applicable to the flexible transformer substation based on a grid-type SVG as described in claim 1 or 2, characterized in that, include: Based on the topology of the flexible transformer substation, the DC bus capacitance and DC bus voltage of each converter on the power supply side, the equivalent connection inductance and phase shift ratio between the two DC ports on the power supply side, the switching frequency of the power supply side converter, the gain constant of the grid side converter constituting the DC link, and the phase shift ratio of each DC port on the power supply side are obtained. Establish DC port voltage models for the three power-side converters and one grid-side converter that constitute three DC links; establish DC port voltage models for each power-side converter that does not constitute a DC link. Based on the DC port voltage model, a decoupling control model for the device is established. An improved extreme learning machine method is used to iteratively solve the decoupling control model of the device to obtain the phase shift ratio of each DC port; based on the phase shift ratio of each DC port, the modulation voltage phase of each converter is determined to control the multi-active bridge converter.

4. The control method for a flexible transformer station based on a grid-type SVG according to claim 3, characterized in that, The DC port voltage models of the three power-side converters and one grid-side converter constituting the three DC links are as follows: In the formula, C τx v τx Let f be the capacitance value of the DC bus capacitor and the DC bus voltage of the converter τx. The values ​​of τx and τ′x are both one of s1, s2, s3, and e1, and the values ​​of τx and τ′x are not the same. s K is the switching frequency of the power supply converter. e1 D is the gain constant of the grid-side converter that constitutes the DC link. τxτ′x For the shift ratio between DC ports τx and τ′x, L τxτ′x K is the equivalent connection inductance between DC ports τx and τ′x. τx and K τ′x The phase shift ratios of DC ports sx and sy are respectively, i′ τx Let i be the DC current output by the converter at port τx. pvx Let x be the output current of the xth distributed photovoltaic system.

5. The control method for a flexible transformer station based on a grid-type SVG according to claim 4, characterized in that, The DC port voltage models of the individual power-side converters that do not form a DC link are as follows: In the formula, C sx v sx Let D be the capacitance value of the DC bus capacitor and the DC bus voltage of the converter sx on the power supply side. sxsy Compared to the shift between DC ports sx and sy, L sxsy i is the equivalent connection inductance between DC ports sx and sy. pvx Let x be the output current of the xth distributed photovoltaic system.

6. The control method for a flexible transformer station based on a grid-type SVG according to claim 5, characterized in that, Based on the DC port voltage model, a decoupling control model for the device is established, including: Based on the DC port voltage model, the phase shift parameter of the DC port is introduced to establish a coupling term between the DC port voltage and the phase shift ratio. The coupling term is used as the output of the decoupling control model, and the current is used as the input of the decoupling control model. Based on the input and output of the decoupling control model, the decoupling control model of the device is established.

7. The control method for a flexible transformer station based on a grid-type SVG according to claim 6, characterized in that, Based on the DC port voltage model, by introducing the phase offset parameter of the DC port, the following relationship is obtained for the n DC port voltages on the power supply side: In the formula, δ s1 δ s2 , ..., δ sn Let A be the phase offset parameters of the n DC ports on the power supply side, and let A be the first coupling matrix. s1 v s2 ... v s(n-1) v sn Let C be the DC bus voltage of the n converters on the power supply side. s1 C s2 ... C s(n-1) C sn For the DC bus capacitors of the n converters on the power supply side; The first coupling matrix is ​​as follows: in, x,y=1,2,3,…n.

8. The control method for a flexible transformer station based on a grid-type SVG according to claim 7, characterized in that, Using current as the input to the decoupled control model, the following expression is satisfied: In the formula, i s1 i s2 ... s(n-1) i sn Let B be the input current of the n converters on the power supply side, and B be the second coupling matrix. The second coupling matrix is ​​as follows: Among them, b xy =a xy ·v sy , x, y = 1, 2, 3, ... n.

9. The control method for a flexible transformer station based on a grid-type SVG according to claim 8, characterized in that, The decoupling control model of the device satisfies the following expression:

10. The control method for a flexible transformer station based on a grid-type SVG according to claim 2, characterized in that, An improved extreme learning machine method is used to iteratively solve the decoupling control model of the device, obtaining the phase shift ratio of each DC port, including: An iterative model of the inverse function of the output term of the decoupled control model is established using the extreme learning machine method; wherein, the initial values ​​of the connection weights from the hidden layer nodes to the output in the iterative model are optimized using the swarm intelligence optimization method.

Images