An improved droop control method and system based on adaptive virtual impedance

Abstract

The application relates to an improved droop control method and system based on adaptive virtual impedance, and the specific steps comprise the following: in a micro-grid system in which inverters are connected in parallel to supply power to an AC bus, a model of multiple inverters connected in parallel is established, and an inverter droop controller is built; reactive power output by each inverter is collected, a difference between the reactive power output by each inverter and that by an adjacent inverter in communication with the inverter is obtained, and a value of adaptive virtual impedance for adjusting line equivalent impedance is obtained through a radial basis function neural network; the value of the adaptive virtual impedance is input into the inverter droop controller, a voltage drop on the virtual impedance is calculated, the voltage drop is added to an original output voltage reference value, and an output voltage reference value considering the virtual impedance is obtained; in each sampling period, the value of the virtual impedance is updated according to the difference between the reactive power outputs of the inverters, and a switch tube driving signal is generated according to the output voltage reference value considering the virtual impedance.

Description

Technical Field

[0001] This invention relates to the field of power system technology, specifically to an improved droop control method and system based on adaptive virtual impedance. Background Technology

[0002] With the gradual advancement of the "dual carbon" goals, the "dual high" characteristics of my country's power system—high proportion of renewable energy and high proportion of power electronic equipment integration—are becoming increasingly prominent. AC microgrids are independent power generation and distribution systems composed of distributed power sources, power electronic equipment, and loads. They are widely used due to their ability to facilitate flexible and efficient utilization of distributed energy. Droop control simulates the droop characteristics of synchronous generators, achieving automatic power distribution among inverters through the linear relationship between active power output (frequency) and reactive power output (voltage) of distributed power sources. However, in low-voltage microgrids, the line impedance is not inductive, making it impossible to directly apply traditional droop control strategies through power decoupling. Furthermore, due to the mismatch between the line impedance from each inverter output port to the AC bus and the output reactive power, errors exist in the output voltage of each inverter, leading to unreasonable power distribution among inverters and the generation of circulating currents. Virtual impedance methods are used in droop control to improve these problems by changing the equivalent impedance of the line. However, existing virtual impedance methods suffer from complex parameter tuning and the control effect is affected by inverter output power oscillations.

[0003] The existing technology CN111130375B discloses a method for precisely adjusting virtual impedance and a Thevenin equivalent circuit based on the inverter in a single-unit inverter and an inverter parallel system. This method obtains the no-load voltage gain at the fundamental frequency f0 and the equivalent output impedance at the fundamental frequency f0, and proposes an improved virtual impedance value Z for addition. v0 It can compensate for the effect of the inverter's no-load voltage gain not being 1, and accurately adjust the equivalent output impedance.

[0004] In the existing technology, the equivalent output impedance of the inverter is precisely tuned based on the Thevenin equivalent circuit of the inverter, but the parameter tuning process is relatively complex and inefficient. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention proposes an improved droop control method and system based on adaptive virtual impedance.

[0006] The technical solution of the present invention is as follows:

[0007] On one hand, the present invention provides an improved droop control method based on adaptive virtual impedance, the specific steps of which include:

[0008] In a microgrid system where inverters operate in parallel and are connected to the AC bus to supply power to the load, a model of multiple inverters operating in parallel is established, and an inverter droop controller is built.

[0009] The reactive power output of each inverter is collected. Based on the difference in reactive power output between each inverter and its communicating neighboring inverter, the adaptive virtual impedance value for adjusting the equivalent impedance of the line is obtained through a radial basis function neural network.

[0010] Input the adaptive virtual impedance value into the inverter droop controller, calculate the voltage drop across the virtual impedance, and add the voltage drop to the original output voltage reference value to obtain the output voltage reference value considering the virtual impedance.

[0011] In each sampling period, the virtual impedance value is updated according to the difference in reactive power output of each inverter, and the switching transistor drive signal is generated according to the output voltage reference value considering the virtual impedance to adjust the reactive power distribution of each inverter so that the output port voltage of each inverter tends to be equal.

[0012] The specific steps for obtaining the adaptive virtual impedance value for adjusting the equivalent impedance of the line using a radial basis function neural network are as follows:

[0013] Estimating line resistance by line length Perform static virtual impedance numerical tuning, where the static virtual resistance... The formula for reducing line resistance is as follows:

[0014]

[0015]

[0016] in, This is an estimated value of the line resistance; For line length, The resistance parameter per unit length of the line;

[0017] Static virtual inductance To increase the inductance of the line, the specific calculation formula is as follows:

[0018]

[0019]

[0020] in, The angular frequency of the power system. For power system frequency;

[0021] Then, the radial basis function neural network module is used to calculate the first... i Taiwan inverter dynamic virtual inductor and dynamic virtual resistance To reduce reactive power distribution error and the derivative of reactive power distribution error As input information to the neural network module, the radial basis function neural network module is used to fit this nonlinear relationship. The specific calculation formula is as follows:

[0022] No. i Taiwan inverter dynamic virtual inductor :

[0023]

[0024] No. i Taiwan inverter dynamic virtual resistance :

[0025]

[0026] In the formula, This refers to the dynamic virtual resistance gain coefficient.

[0027] Using the first i Taiwan inverter dynamic virtual inductor and dynamic virtual resistance Update the adaptive virtual inductance used to adjust the line's equivalent impedance. and adaptive virtual resistance The specific calculation formula is as follows:

[0028]

[0029]

[0030] Adaptive virtual inductor and adaptive virtual resistance The values ​​that constitute the adaptive virtual impedance used to adjust the equivalent impedance of the line.

[0031] In a preferred embodiment, the step of collecting the reactive power output of each inverter and calculating the difference in reactive power output between each inverter and its communicating neighboring inverter specifically comprises:

[0032] Treating an inverter as a node forms i Nodes communicate with each other j Node communication adjacency matrix :

[0033]

[0034] Where n is the number of inverters, and the elements in the matrix are... The value can be 0 or 1;

[0035] compute nodesi Its neighboring nodes j reactive power distribution error

[0036]

[0037] in, For nodes i The set of connected nodes For the first i Taiwan inverter reactive power-voltage droop factor, For the first i The reactive power output of the inverter; n j For the first i Adjacent inverters of the Taiwan inverter j The reactive power-voltage droop factor; Q j For the first i Adjacent inverters of the Taiwan inverter j The output reactive power.

[0038] In a preferred embodiment, the dynamic virtual inductance of the i-th inverter is calculated using a radial basis function neural network module. The specific steps are as follows:

[0039] Construct a radial basis function neural network, including an input layer, hidden layers, and an output layer; the input layer takes the first hidden layer as input. i Taiwan inverter reactive power distribution error Passing the exam i Taiwan inverter reactive power distribution error derivative The output of the input layer is , ;

[0040] The input to the hidden layer is:

[0041]

[0042] The output of the hidden layer is:

[0043]

[0044] in, Let n be the center vector of the n1-th neuron in the hidden layer. The width of the n1-th neuron in the hidden layer;

[0045] Input of the output layer for:

[0046]

[0047] in, The weight of the n1-th neuron in the hidden layer;

[0048] The output of the output layer is the first i Taiwan inverter dynamic virtual impedance The formula is as follows:

[0049]

[0050] in, This represents the upper limit of the dynamic virtual impedance.

[0051] Since the inverter output power will experience transient oscillations after the change in the equivalent impedance of the line, and then gradually adjust to a new steady state, in order to reduce the impact of the oscillation process on the fitting effect of the neural network, the performance index function of the neural network is selected as follows:

[0052]

[0053] in: For the first i Taiwan inverter reactive power distribution error The moving average, the first i Taiwan inverter t Moving average of reactive power distribution error at time step The specific calculation formula is as follows:

[0054]

[0055] in, t Indicates the current sampling time. For the first i The reactive power distribution error of the inverter at time t. For moving average weights, The range of values ​​is .

[0056] In a preferred embodiment, the step of inputting an adaptive virtual impedance value into the inverter droop controller, calculating the voltage drop across the virtual impedance, and adding the voltage drop to the original output voltage reference value to obtain an output voltage reference value considering the virtual impedance specifically comprises:

[0057] First, calculate the voltage drop across the adaptive virtual impedance. , :

[0058]

[0059] in, For the first i The voltage drop across the d-axis of the virtual impedance of the inverter. For the first i The voltage drop across the q-axis of the virtual impedance of the inverter. For the firsti The d-axis components of the ABC phase currents output from the inverter are obtained after Parker transformation. For the first i The q-axis components of the ABC phase currents output from the inverter are obtained after Parker transformation; R vi For adaptive virtual resistance; L vi For adaptive virtual inductance;

[0060] Further update the output voltage reference value considering the voltage drop across the adaptive virtual impedance:

[0061] ,

[0062] in, and The first i The d-axis and q-axis voltages input to the outer voltage loop of the voltage-current dual closed-loop control system of the inverter; and The first i The power calculation in the droop controller of the inverter yields the d-axis and q-axis components of the output voltage reference value when virtual impedance is neglected.

[0063] On the other hand, this invention proposes an improved droop control system based on adaptive virtual impedance, comprising:

[0064] The droop controller construction module establishes a model of multiple inverters operating in parallel in a microgrid system where inverters are connected to the AC bus to supply power to the load, and builds an inverter droop controller.

[0065] The adaptive virtual impedance calculation module collects the reactive power output of each inverter and calculates the adaptive virtual impedance value for adjusting the equivalent impedance of the line through a radial basis function neural network based on the difference in reactive power output between each inverter and its adjacent communicating inverter.

[0066] The output voltage reference value calculation module inputs the adaptive virtual impedance value into the inverter droop controller, calculates the voltage drop across the virtual impedance, and adds the voltage drop to the original output voltage reference value to obtain the output voltage reference value considering the virtual impedance.

[0067] The drive signal adjustment module updates the virtual impedance value based on the reactive power output difference of each inverter in each sampling period, and generates a switching transistor drive signal based on the output voltage reference value considering the virtual impedance to adjust the reactive power distribution of each inverter so that the output port voltage of each inverter tends to be equal.

[0068] The specific steps for obtaining the adaptive virtual impedance value for adjusting the equivalent impedance of the line using a radial basis function neural network are as follows:

[0069] Estimating line resistance by line length Perform static virtual impedance numerical tuning, where the static virtual resistance... The formula for reducing line resistance is as follows:

[0070]

[0071]

[0072] in, This is an estimated value of the line resistance; For line length, The resistance parameter per unit length of the line;

[0073] Static virtual inductance To increase the inductance of the line, the specific calculation formula is as follows:

[0074]

[0075]

[0076] in, The angular frequency of the power system. For power system frequency;

[0077] Then, the radial basis function neural network module is used to calculate the first... i Taiwan inverter dynamic virtual inductor and dynamic virtual resistance To reduce reactive power distribution error and the derivative of reactive power distribution error As input information to the neural network module, the radial basis function neural network module is used to fit this nonlinear relationship. The specific calculation formula is as follows:

[0078] No. i Taiwan inverter dynamic virtual inductor :

[0079]

[0080] No. i Taiwan inverter dynamic virtual resistance :

[0081]

[0082] In the formula, This refers to the dynamic virtual resistance gain coefficient.

[0083] Using the first i Taiwan inverter dynamic virtual inductor and dynamic virtual resistance Update the adaptive virtual inductance used to adjust the line's equivalent impedance. and adaptive virtual resistance The specific calculation formula is as follows:

[0084]

[0085]

[0086] Adaptive virtual inductor and adaptive virtual resistance The values ​​that constitute the adaptive virtual impedance used to adjust the equivalent impedance of the line.

[0087] In a preferred embodiment, the step of collecting the reactive power output of each inverter and calculating the difference in reactive power output between each inverter and its communicating neighboring inverter specifically comprises:

[0088] Treating an inverter as a node forms i Nodes communicate with each other j Node communication adjacency matrix :

[0089]

[0090] Where n is the number of inverters, and the elements in the matrix are... The value can be 0 or 1;

[0091] compute nodes i Its neighboring nodes j reactive power distribution error

[0092]

[0093] in, For nodes i The set of connected nodes For the first i Taiwan inverter reactive power-voltage droop factor, For the first i Taiwan inverter output reactive power; n j For the first i Adjacent inverters of the Taiwan inverter j The reactive power-voltage droop factor; Q j For the first i Adjacent inverters of the Taiwan inverter j The output reactive power.

[0094] In a preferred embodiment, the dynamic virtual inductance of the i-th inverter is calculated using a radial basis function neural network module. The specific steps are as follows:

[0095] Construct a radial basis function neural network, including an input layer, hidden layers, and an output layer; the input layer takes the first hidden layer as input. i Taiwan inverter reactive power distribution error Passing the exam i Taiwan inverter reactive power distribution error derivative The output of the input layer is , ;

[0096] The input to the hidden layer is:

[0097]

[0098] The output of the hidden layer is:

[0099]

[0100] in, Let n be the center vector of the n1-th neuron in the hidden layer. The width of the n1-th neuron in the hidden layer;

[0101] Input of the output layer for:

[0102]

[0103] in, The weight of the n1-th neuron in the hidden layer;

[0104] The output of the output layer is the first i Taiwan inverter dynamic virtual impedance The formula is as follows:

[0105]

[0106] in, This represents the upper limit of the dynamic virtual impedance.

[0107] Since the inverter output power will experience transient oscillations after the change in the equivalent impedance of the line, and then gradually adjust to a new steady state, in order to reduce the impact of the oscillation process on the fitting effect of the neural network, the performance index function of the neural network is selected as follows:

[0108]

[0109] in: For the first i Taiwan inverter reactive power distribution error The moving average, the first i Taiwan inverter t Moving average of reactive power distribution error at time step The specific calculation formula is as follows:

[0110]

[0111] in, t Indicates the current sampling time. For the first i The reactive power distribution error of the inverter at time t. For moving average weights, The range of values ​​is .

[0112] In a preferred embodiment, the step of inputting an adaptive virtual impedance value into the inverter droop controller, calculating the voltage drop across the virtual impedance, and adding the voltage drop to the original output voltage reference value to obtain an output voltage reference value considering the virtual impedance specifically comprises:

[0113] First, calculate the voltage drop across the adaptive virtual impedance. , :

[0114]

[0115] in, For the first i The voltage drop across the d-axis of the virtual impedance of the inverter. For the first i The voltage drop across the q-axis of the virtual impedance of the inverter. For the first i The d-axis components of the ABC phase currents output from the inverter are obtained after Parker transformation. For the first i The q-axis components of the ABC phase currents output from the inverter are obtained after Parker transformation; R vi For adaptive virtual resistance; L vi For adaptive virtual inductance;

[0116] Further update the output voltage reference value considering the voltage drop across the adaptive virtual impedance:

[0117] ,

[0118] in, and The first i The d-axis and q-axis voltages input to the outer voltage loop of the voltage-current dual closed-loop control system of the inverter; and The first i The power calculation in the droop controller of the inverter yields the d-axis and q-axis components of the output voltage reference value when virtual impedance is neglected.

[0119] The present invention has the following beneficial effects:

[0120] 1. This invention improves the droop control power distribution problem and circulating current problem caused by the non-inductive line impedance and the mismatch between the line impedance and the rated reactive power of the inverter in the microgrid system by using adaptive virtual impedance.

[0121] 2. By using the moving average term of reactive power distribution error as a performance index function, this invention can reduce the impact of transient oscillations in inverter output power due to changes in the equivalent inductance of the line on the fitting effect of the neural network.

[0122] 3. This invention utilizes a radial basis function neural network to generate adaptive virtual impedance, which has a faster convergence speed and avoids complex parameter tuning; at the same time, it reduces the impact of transient oscillations in inverter output power due to changes in the equivalent inductance of the line on the neural network fitting effect, effectively improving the power distribution effect. Attached Figure Description

[0123] Figure 1 This is a schematic diagram of the steps of the present invention;

[0124] Figure 2 A circuit diagram of a microgrid with two inverters operating in parallel;

[0125] Figure 3 The output reactive power waveforms of the two inverters are shown.

[0126] Figure 4 The output voltage amplitude waveforms of the two inverters are shown.

[0127] Figure 5 This is a block diagram illustrating the control principle of the circuit. Detailed Implementation

[0128] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0129] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0130] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0131] The terms “comprising” and “including” indicate the presence of the described feature, whole, step, operation, element and / or component, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0132] The term “and / or” refers to any combination of one or more of the associated listed items, as well as all possible combinations, and includes these combinations.

[0133] Example 1:

[0134] See Figure 1 An improved droop control method based on adaptive virtual impedance, the specific steps of which include:

[0135] In a microgrid system where inverters operate in parallel and are connected to the AC bus to supply power to the load, a model of multiple inverters operating in parallel is established, and an inverter droop controller is built.

[0136] The reactive power output of each inverter is collected. Based on the difference in reactive power output between each inverter and its communicating neighboring inverter, the adaptive virtual impedance value for adjusting the equivalent impedance of the line is obtained through a radial basis function neural network.

[0137] Input the adaptive virtual impedance value into the inverter droop controller, calculate the voltage drop across the virtual impedance, and add the voltage drop to the original output voltage reference value to obtain the output voltage reference value considering the virtual impedance.

[0138] Within each sampling period, the virtual impedance value is updated based on the difference in reactive power output of each inverter. A switching transistor drive signal is generated based on the output voltage reference value considering the virtual impedance to adjust the reactive power distribution of each inverter, so that the output port voltage of each inverter tends to be equal.

[0139] The specific steps for obtaining the adaptive virtual impedance value for adjusting the equivalent impedance of the line using a radial basis function neural network are as follows:

[0140] Estimating line resistance by line length Perform static virtual impedance numerical tuning, where the static virtual resistance... The formula for reducing line resistance is as follows:

[0141]

[0142]

[0143] in, This is an estimated value of the line resistance; For line length, The resistance parameter per unit length of the line;

[0144] In this embodiment, since the load and operating status of the low-voltage microgrid are not fixed, and the actual line impedance is affected by environmental factors and line aging, the line impedance is difficult to calculate accurately; therefore, the line resistance is estimated by estimating the line length. Perform static virtual impedance numerical tuning.

[0145] Static virtual inductance To increase the inductance of the line, the specific calculation formula is as follows:

[0146]

[0147]

[0148] in, The angular frequency of the power system. This refers to the frequency of the power system.

[0149] In this embodiment, the static virtual inductor A positive value is used to increase the inductive component of the circuit, representing a static virtual resistance. Taking a negative value reduces the resistive component of the line. Since the microgrid is a low-voltage line, the line is not inductive. By setting the static virtual impedance, the line impedance can be made approximately inductive, thus achieving decoupling of active and reactive power control.

[0150] Then, the radial basis function neural network module is used to calculate the first... i Taiwan inverter dynamic virtual inductor and dynamic virtual resistance To reduce reactive power distribution error and the derivative of reactive power distribution error As input information to the neural network module, the radial basis function neural network module is used to fit this nonlinear relationship. The specific calculation formula is as follows:

[0151] No. i Taiwan inverter dynamic virtual inductor :

[0152]

[0153] No. i Taiwan inverter dynamic virtual resistance :

[0154]

[0155] In the formula, This refers to the dynamic virtual resistance gain coefficient.

[0156] In this embodiment, dynamic virtual inductance It is used to adjust line impedance to reduce reactive power distribution error, and is a function related to reactive power distribution error. Nonlinear functions.

[0157] Using the first i Taiwan inverter dynamic virtual inductor and dynamic virtual resistance Update the adaptive virtual inductance used to adjust the line's equivalent impedance. and adaptive virtual resistance The specific calculation formula is as follows:

[0158]

[0159]

[0160] Adaptive virtual inductor and adaptive virtual resistance The values ​​that constitute the adaptive virtual impedance used to adjust the equivalent impedance of the line.

[0161] In a preferred embodiment of this invention, the step of collecting the reactive power output of each inverter and calculating the difference in reactive power output between each inverter and its communicating neighboring inverters specifically comprises:

[0162] Treating an inverter as a node forms i Nodes communicate with each other j Node communication adjacency matrix :

[0163]

[0164] Where n is the number of inverters, and the elements in the matrix are... The value can be 0 or 1;

[0165] compute nodes i Its neighboring nodes j reactive power distribution error

[0166]

[0167] in, For nodes i The set of connected nodes For the first i Taiwan inverter reactive power-voltage droop factor, For the first i Taiwan inverter output reactive power; n j For the first i Adjacent inverters of the Taiwan inverter j The reactive power-voltage droop factor; Q jFor the first i Adjacent inverters of the Taiwan inverter j The output reactive power.

[0168] In this embodiment, the elements in the matrix The value of is determined by whether there is communication between node i and node j. If node j i With nodes j To communicate between them, If node i With nodes j If there is no communication between them, then .

[0169] In a preferred embodiment of this example, the calculation of the dynamic virtual inductance of the i-th inverter using a radial basis function neural network module is described. The specific steps are as follows:

[0170] Construct a radial basis function neural network, including an input layer, hidden layers, and an output layer; the input layer takes the first hidden layer as input. i Taiwan inverter reactive power distribution error Passing the exam i Taiwan inverter reactive power distribution error derivative The output of the input layer is , ;

[0171] The input to the hidden layer is:

[0172]

[0173] The output of the hidden layer is:

[0174]

[0175] in, Let n be the center vector of the n1-th neuron in the hidden layer. The width of the n1-th neuron in the hidden layer;

[0176] Input of the output layer for:

[0177]

[0178] in, The weight of the n1-th neuron in the hidden layer;

[0179] The output of the output layer is the first i Taiwan inverter dynamic virtual impedance The formula is as follows:

[0180]

[0181] in, This represents the upper limit of the dynamic virtual impedance.

[0182] Since the inverter output power will experience transient oscillations after the change in the equivalent impedance of the line, and then gradually adjust to a new steady state, in order to reduce the impact of the oscillation process on the fitting effect of the neural network, the performance index function of the neural network is selected as follows:

[0183]

[0184] in: For the first i Taiwan inverter reactive power distribution error The moving average, the first i Taiwan inverter t Moving average of reactive power distribution error at time step The specific calculation formula is as follows:

[0185]

[0186] in, t Indicates the current sampling time. For the first i The reactive power distribution error of the inverter at time t. For moving average weights, The range of values ​​is .

[0187] In a preferred embodiment of this invention, the steps of inputting an adaptive virtual impedance value into the inverter droop controller, calculating the voltage drop across the virtual impedance, and adding the voltage drop to the original output voltage reference value to obtain an output voltage reference value considering the virtual impedance are as follows:

[0188] First, calculate the voltage drop across the adaptive virtual impedance. , :

[0189]

[0190] in, For the first i The voltage drop across the d-axis of the virtual impedance of the inverter. For the first i The voltage drop across the q-axis of the virtual impedance of the inverter. For the first i The d-axis components of the ABC phase currents output from the inverter are obtained after Parker transformation. For the first i The q-axis components of the ABC phase currents output from the inverter are obtained after Parker transformation; R vi For adaptive virtual resistance; L viFor adaptive virtual inductance;

[0191] Further update the output voltage reference value considering the voltage drop across the adaptive virtual impedance:

[0192] ,

[0193] in, and The first i The d-axis and q-axis voltages input to the outer voltage loop of the voltage-current dual closed-loop control system of the inverter; and The first i The power calculation in the droop controller of the inverter yields the d-axis and q-axis components of the output voltage reference value when virtual impedance is neglected.

[0194] Example 2:

[0195] A simulation experiment was conducted based on the improved droop control method based on adaptive virtual impedance described above:

[0196] First, a power transfer model based on droop control for a microgrid composed of multiple inverters is constructed. For example... Figure 2 Taking the parallel operation model of two inverters shown as an example, the inverters are equivalent to DC voltage sources. U dci Indicates the first i The output port voltage of the inverter. L fi , C fi They represent the first i The inductor and capacitor of the LC filter in the Taiwanese inverter. U oi , I oi They represent the first i Output voltage and output current at the port of the inverter's LC filter Z li Indicates the first i The line impedance of the inverter connected to the AC bus. U pcc This indicates the voltage of the AC bus. i Taiwan inverter output active power reactive power for:

[0197]

[0198] In the formula, For the first i The voltage phase output by the inverter. For the first iThe line impedance angle of the inverter connected to the AC bus.

[0199] Because this method introduces static virtual impedance , This effectively reduces line resistance and increases line inductance, making the line impedance in a low-voltage microgrid approximately inductive. ; and because It is relatively small and can be approximated as , Then the first i Taiwan inverter output active power reactive power It can be simplified to:

[0200]

[0201] From the simplified output power expression, it can be seen that at this time... and , and The relationship is first-order linear, and there is approximately no coupling between active and reactive power. A droop control equation can be established:

[0202]

[0203] in , The first i The Taiwanese inverter's output angular frequency and output voltage reference value are neglected due to virtual impedance. , The first i The rated reference values ​​for the output angular frequency and output voltage of the inverter. , The first i Active power droop factor and reactive power droop factor of Taiwan inverter , The first i The rated active and reactive power of the inverter. Therefore, to ensure that inverters of different capacities in a microgrid share the power load proportionally according to their capacity, the droop factor must be inversely proportional to the power:

[0204]

[0205] Based on the droop control equation, the expression for output reactive power can be rewritten as follows:

[0206]

[0207] As can be seen, it is only necessary to make the line inductance proportional to the droop factor, for example:

[0208]

[0209] At this point, the condition that inverters of different capacities share reactive load according to their capacity ratio can be met.

[0210] This method introduces adaptive virtual impedance, which can effectively change the line impedance value by adding a virtual voltage drop to the control loop, and the impedance adjustment satisfies the following:

[0211]

[0212] To verify the effectiveness of the improved droop control strategy based on adaptive virtual impedance, a system was built in the MATLAB / Simulink platform as follows: Figure 2 The simulation model of two inverters operating in parallel is shown in Table 1.

[0213] Table 1 Main parameters of the simulation model

[0214]

[0215] Inverter 1 and Inverter 2 have the same rated output reactive power. A reactive load of 10.5 MVar is connected to the AC bus. Before 0.15s, the adaptive virtual impedance proposed in this invention is not introduced; only conventional droop control is used. After 0.15s, the adaptive virtual impedance proposed in this invention is added. The output reactive power waveforms and inverter output voltage amplitudes of the two inverters are as follows: Figure 3 , Figure 4 As shown, , The reactive power outputs of inverter 1 and inverter 2 are respectively... , These are the output voltage amplitudes of inverter 1 and inverter 2, respectively.

[0216] Depend on Figure 3 , Figure 4 Simulation results show that before 0.15s, due to the non-inductive line impedance and mismatch between the line impedance and the inverter reactive power capacity in the low-voltage microgrid system, although the two inverters have the same rated output reactive power and the same reactive power droop factor, the actual output reactive power cannot be distributed according to the rated output reactive power ratio, and there is a difference in the output voltage. After 0.15s, by adding an adaptive virtual impedance using the method proposed in this paper, after a transient adjustment process, the output reactive power of the two inverters is distributed in a 1:1 ratio after 0.35s, and the actual output voltage of the two inverters... , Approximately equal. The droop control method based on adaptive virtual impedance proposed in this paper can effectively improve power distribution, reduce the voltage difference between inverters, and suppress circulating current.

[0217] The control principle block diagram based on this embodiment is as follows: Figure 5 As shown.

[0218] Example 3:

[0219] An improved droop control system based on adaptive virtual impedance includes:

[0220] The droop controller construction module establishes a model of multiple inverters operating in parallel in a microgrid system where inverters are connected to the AC bus to supply power to the load, and builds an inverter droop controller.

[0221] The adaptive virtual impedance calculation module collects the reactive power output of each inverter and calculates the adaptive virtual impedance value for adjusting the equivalent impedance of the line through a radial basis function neural network based on the difference in reactive power output between each inverter and its adjacent communicating inverter.

[0222] The output voltage reference value calculation module inputs the adaptive virtual impedance value into the inverter droop controller, calculates the voltage drop across the virtual impedance, and adds the voltage drop to the original output voltage reference value to obtain the output voltage reference value considering the virtual impedance.

[0223] The drive signal adjustment module updates the virtual impedance value based on the reactive power output difference of each inverter in each sampling period, and generates a switching transistor drive signal based on the output voltage reference value considering the virtual impedance to adjust the reactive power distribution of each inverter so that the output port voltage of each inverter tends to be equal.

[0224] The specific steps for obtaining the adaptive virtual impedance value for adjusting the equivalent impedance of the line using a radial basis function neural network are as follows:

[0225] Estimating line resistance by line length Perform static virtual impedance numerical tuning, where the static virtual resistance... The formula for reducing line resistance is as follows:

[0226]

[0227]

[0228] in, This is an estimated value of the line resistance; For line length, The resistance parameter per unit length of the line;

[0229] Static virtual inductance To increase the inductance of the line, the specific calculation formula is as follows:

[0230]

[0231]

[0232] in, The angular frequency of the power system. For power system frequency;

[0233] Then, the radial basis function neural network module is used to calculate the first... i Taiwan inverter dynamic virtual inductor and dynamic virtual resistance To reduce reactive power distribution error and the derivative of reactive power distribution error As input information to the neural network module, the radial basis function neural network module is used to fit this nonlinear relationship. The specific calculation formula is as follows:

[0234] No. i Taiwan inverter dynamic virtual inductor :

[0235]

[0236] No. i Taiwan inverter dynamic virtual resistance :

[0237]

[0238] In the formula, This refers to the dynamic virtual resistance gain coefficient.

[0239] Using the first i Taiwan inverter dynamic virtual inductor and dynamic virtual resistance Update the adaptive virtual inductance used to adjust the line's equivalent impedance. and adaptive virtual resistance The specific calculation formula is as follows:

[0240]

[0241]

[0242] Adaptive virtual inductor and adaptive virtual resistance The values ​​that constitute the adaptive virtual impedance used to adjust the equivalent impedance of the line.

[0243] In a preferred embodiment of this invention, the step of collecting the reactive power output of each inverter and calculating the difference in reactive power output between each inverter and its communicating neighboring inverters specifically comprises:

[0244] Treating an inverter as a node forms i Nodes communicate with each other j Node communication adjacency matrix :

[0245]

[0246] Where n is the number of inverters, and the elements in the matrix are... The value can be 0 or 1;

[0247] compute nodes i Its neighboring nodes j reactive power distribution error

[0248]

[0249] in, For nodes i The set of connected nodes For the first i Taiwan inverter reactive power-voltage droop factor, For the first i The reactive power output of the inverter; n j For the first i Adjacent inverters of the Taiwan inverter j The reactive power-voltage droop factor; Q j For the first i Adjacent inverters of the Taiwan inverter j The output reactive power.

[0250] In a preferred embodiment of this example, the calculation of the dynamic virtual inductance of the i-th inverter using a radial basis function neural network module is described. The specific steps are as follows:

[0251] Construct a radial basis function neural network, including an input layer, hidden layers, and an output layer; the input layer takes the first hidden layer as input. i Taiwan inverter reactive power distribution error Passing the exam i Taiwan inverter reactive power distribution error derivative The output of the input layer is , ;

[0252] The input to the hidden layer is:

[0253]

[0254] The output of the hidden layer is:

[0255]

[0256] in, Let n be the center vector of the n1-th neuron in the hidden layer. The width of the n1-th neuron in the hidden layer;

[0257] Input of the output layer for:

[0258]

[0259] in, The weight of the n1-th neuron in the hidden layer;

[0260] The output of the output layer is the first i Taiwan inverter dynamic virtual impedance The formula is as follows:

[0261]

[0262] in, This represents the upper limit of the dynamic virtual impedance.

[0263] Since the inverter output power will experience transient oscillations after the change in the equivalent impedance of the line, and then gradually adjust to a new steady state, in order to reduce the impact of the oscillation process on the fitting effect of the neural network, the performance index function of the neural network is selected as follows:

[0264]

[0265] in: For the first i Taiwan inverter reactive power distribution error The moving average, the first i Taiwan inverter t Moving average of reactive power distribution error at time step The specific calculation formula is as follows:

[0266]

[0267] in, t Indicates the current sampling time. For the first i The reactive power distribution error of the inverter at time t. For moving average weights, The range of values ​​is .

[0268] In a preferred embodiment of this invention, the steps of inputting an adaptive virtual impedance value into the inverter droop controller, calculating the voltage drop across the virtual impedance, and adding the voltage drop to the original output voltage reference value to obtain an output voltage reference value considering the virtual impedance are as follows:

[0269] First, calculate the voltage drop across the adaptive virtual impedance. , :

[0270]

[0271] in, For the first i The voltage drop across the d-axis of the virtual impedance of the inverter. For the first i The voltage drop across the q-axis of the virtual impedance of the inverter. For the first i The d-axis components of the ABC phase currents output from the inverter are obtained after Parker transformation. For the first i The q-axis components of the ABC phase currents output from the inverter are obtained after Parker transformation; R vi For adaptive virtual resistance; L vi For adaptive virtual inductance;

[0272] Further update the output voltage reference value considering the voltage drop across the adaptive virtual impedance:

[0273] ,

[0274] in, and The first i The d-axis and q-axis voltages input to the outer voltage loop of the voltage-current dual closed-loop control system of the inverter; and The first i The power calculation in the droop controller of the inverter yields the d-axis and q-axis components of the output voltage reference value when virtual impedance is neglected.

[0275] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. An improved droop control method based on adaptive virtual impedance, characterized in that, The specific steps include: In a micro-grid system in which inverters are connected in parallel to access an AC bus to supply power to loads, a model of parallel operation of multiple inverters is established, and a droop controller of the inverter is built; The reactive power output by each inverter is collected, and according to the difference between the reactive power output by each inverter and the adjacent inverter in communication therewith, an adaptive virtual impedance value for adjusting the equivalent impedance of the line is obtained through a radial basis function neural network; The adaptive virtual impedance value is input into the droop controller of the inverter, the voltage drop on the virtual impedance is calculated, the voltage drop is added to the original output voltage reference value, and an output voltage reference value considering the virtual impedance is obtained; In each sampling period, the virtual impedance value is updated according to the difference between the reactive power outputs of the inverters, and a switch tube driving signal is generated according to the output voltage reference value considering the virtual impedance, so as to adjust the reactive power distribution of the inverters and make the output port voltages of the inverters tend to be equal; The step of obtaining the adaptive virtual impedance value for adjusting the equivalent impedance of the line through the radial basis function neural network specifically includes: Estimating line resistance by line length Carrying out static virtual impedance numerical setting, wherein the static virtual resistance For reducing line resistance, the calculation formula is specifically: wherein, is a line resistance estimate; is a line length, is a line resistance per unit length parameter; Static virtual inductance The formula for increasing the line inductance is as follows: wherein is the power system angular frequency, is the power system frequency; The radial basis function neural network module is further used to calculate the first derivative of the reactive power distribution error i Dynamic virtual inductance of the inverter And dynamic virtual resistance The reactive power distribution error And the derivative of the reactive power distribution error As input information of the neural network module, the radial basis function neural network module is used to fit the nonlinear relationship, and the specific calculation formula is as follows: First i Dynamic virtual inductance of a power converter : First i Dynamic virtual resistance of a power converter : In the formula, is a dynamic virtual resistance gain coefficient; With the first i Dynamic virtual inductance of the inverter And dynamic virtual resistance Update the adaptive virtual inductance for adjusting the line equivalent impedance And adaptive virtual resistance The specific formula is as follows: The adaptive virtual inductance and the adaptive virtual resistance comprise a value of an adaptive virtual impedance for adjusting a line equivalent impedance.

2. The improved droop control method based on adaptive virtual impedance according to claim 1, wherein, The step of collecting the reactive power output by each inverter and according to the difference between the reactive power output by each inverter and the adjacent inverter in communication therewith specifically includes: Consider an inverter as a node, forming i Nodes and their communication j Communication adjacency matrix of nodes : where n is the number of inverters, the elements of the matrix takes the value 0 or 1 ; Computing node i reactive power distribution error j of its neighboring nodes wherein, is a set of nodes connected to the node i , is a reactive-voltage droop coefficient of the nth i inverter, is a reactive power output of the nth i inverter; n j is a reactive-voltage droop coefficient of a neighboring inverter i of the nth j inverter; Q j is a reactive power output of a neighboring inverter i of the nth j inverter.

3. The improved droop control method based on adaptive virtual impedance according to claim 1, wherein, The calculating the dynamic virtual inductance of the i-th inverter by the radial basis function neural network module The steps are specifically: A radial basis function neural network is established, including an input layer, a hidden layer and an output layer; the input of the input layer is the first i The error of reactive power distribution of the inverter And the i The error derivative of reactive power distribution of the inverter The output of the input layer is 、 ; The input of the hidden layer is: The output of the hidden layer is: wherein, is a center vector of the n1th neuron of the hidden layer, is a width of the n1th neuron of the hidden layer; Input to the output layer is: wherein, is the weight of the nth1 neuron of the hidden layer; The output of the output layer is the first i Dynamic virtual impedance for inverters The formula is specifically: wherein, is an upper limit value of the dynamic virtual impedance; Since the equivalent impedance of the line changes, the inverter output power will produce transient oscillation and gradually adjust to a new steady state, in order to reduce the influence of the oscillation process on the fitting effect of the neural network, the performance index function of the neural network is selected as: Wherein: is the first i Inverter reactive power distribution error of the moving average, the first i Inverter at the first t reactive power distribution error moving average at the first The calculation formula is: wherein, t denotes the current sampling time, is the t-th i is the t-th time error of reactive power distribution of the inverter, is the moving average weight, the value range is .

4. The improved droop control method based on adaptive virtual impedance according to claim 1, wherein, The step of inputting the adaptive virtual impedance value into the droop controller of the inverter, calculating the voltage drop on the virtual impedance, and adding the voltage drop to the original output voltage reference value to obtain an output voltage reference value considering the virtual impedance specifically includes: First, the voltage drop over the adaptive virtual impedance is calculated , : wherein, is the first i is the voltage drop across the virtual impedance of the inverter on the d-axis, is the first i is the voltage drop across the virtual impedance of the inverter on the q-axis, is the first i is the d-axis component of the ABC phase currents of the inverter output after Park transformation, is the first i is the q-axis component of the ABC phase currents of the inverter output after Park transformation; R vi is the adaptive virtual resistance; L vi is the adaptive virtual inductance; The output voltage reference value considering the voltage drop on the adaptive virtual impedance is updated: 、 wherein, and are the d-axis and q-axis voltage components, respectively, of the voltage reference input to the outer voltage loop of the voltage-current dual-loop control of the mth inverter; i and are the d-axis and q-axis voltage components, respectively, of the voltage reference input to the outer voltage loop of the voltage-current dual-loop control of the mth inverter; and i are the d-axis and q-axis voltage components, respectively, of the voltage reference input to the outer voltage loop of the voltage-current dual-loop control of the mth inverter.

5. An improved droop control system based on adaptive virtual impedance, characterized in that, It includes: A droop controller building module, in a micro-grid system in which inverters are connected in parallel to access an AC bus to supply power to loads, a model of parallel operation of multiple inverters is established, and a droop controller of the inverter is built; An adaptive virtual impedance calculation module, the reactive power output by each inverter is collected, and according to the difference between the reactive power output by each inverter and the adjacent inverter in communication therewith, an adaptive virtual impedance value for adjusting the equivalent impedance of the line is obtained through a radial basis function neural network; An output voltage reference value calculation module, the adaptive virtual impedance value is input into the droop controller of the inverter, the voltage drop on the virtual impedance is calculated, the voltage drop is added to the original output voltage reference value, and an output voltage reference value considering the virtual impedance is obtained; A driving signal adjustment module, in each sampling period, the virtual impedance value is updated according to the difference between the reactive power outputs of the inverters, and a switch tube driving signal is generated according to the output voltage reference value considering the virtual impedance, so as to adjust the reactive power distribution of the inverters and make the output port voltages of the inverters tend to be equal; The step of obtaining the adaptive virtual impedance value for adjusting the equivalent impedance of the line through the radial basis function neural network specifically includes: Estimating line resistance by line length Carrying out static virtual impedance numerical setting, wherein the static virtual resistance For reducing line resistance, the calculation formula is specifically: wherein, is a line resistance estimate; is a line length, is a line resistance per unit length parameter; Static virtual inductance The formula for increasing the line inductance is given by wherein is the power system angular frequency, is the power system frequency; The reactive power distribution error and the derivative of the reactive power distribution error are calculated by the radial basis function neural network module i and the dynamic virtual resistance of the inverter , and the reactive power distribution error and the derivative of the reactive power distribution error are taken as input information of the neural network module, so that the radial basis function neural network module fits the nonlinear relationship, and the specific calculation formula is as follows: First i Dynamic virtual inductance of a power converter : First i Dynamic virtual resistance of a power converter : In the formula, is a dynamic virtual resistance gain factor; With the first i Dynamic virtual inductance of the inverter And dynamic virtual resistance Update the adaptive virtual inductance for adjusting the line equivalent impedance And adaptive virtual resistance The specific formula is as follows: The adaptive virtual inductance and the adaptive virtual resistance comprise a value of an adaptive virtual impedance for adjusting a line equivalent impedance.

6. The improved droop control system based on adaptive virtual impedance as claimed in claim 5 wherein, The reactive power outputted by each inverter is collected, and the difference between the reactive power outputted by each inverter and the adjacent inverter in communication with the inverter is calculated. Consider an inverter as a node, forming i Nodes and their communication j Communication adjacency matrix of nodes : where n is the number of inverters, the elements of the matrix takes the value 0 or 1 ; Computing node i reactive power distribution error j of its neighboring nodes wherein, is a set of nodes connected to the node i , is a reactive-voltage droop coefficient of the nth i inverter, is a reactive power output of the nth i inverter; n j is a reactive-voltage droop coefficient of a neighboring inverter i of the nth j inverter; Q j is a reactive power output of a neighboring inverter i of the nth j inverter.

7. The improved droop control system based on adaptive virtual impedance as claimed in claim 5 wherein, The calculating the dynamic virtual inductance of the i-th inverter by the radial basis function neural network module The steps are specifically: A radial basis function neural network is established, including an input layer, a hidden layer and an output layer; the input of the input layer is the first i Error of reactive power distribution of the inverter And the first i Error derivative of reactive power distribution of the inverter ; the output of the input layer is , ; The input of the hidden layer is: The output of the hidden layer is: wherein, is a center vector of the n1th neuron of the hidden layer, is a width of the n1th neuron of the hidden layer; Input to the output layer is: wherein, is the weight of the nth1 neuron of the hidden layer; The output of the output layer is the first i Dynamic virtual impedance for inverters The formula is specifically: wherein, is an upper limit value of the dynamic virtual impedance; Due to the change of the equivalent impedance of the line, the output power of the inverter will produce transient oscillation and gradually adjust to the new steady state, in order to reduce the influence of the oscillation process on the fitting effect of the neural network, the performance index function of the neural network is selected as: Wherein: is the first i Inverter reactive power distribution error of the moving average, the first i Inverter at the first t reactive power distribution error moving average at the first The calculation formula is: wherein, t denotes the current sampling time, is the t-th i is the t-th time error of reactive power distribution of the inverter, is the moving average weight, the value range is .

8. The improved droop control system based on adaptive virtual impedance as claimed in claim 5 wherein, The adaptive virtual impedance value is inputted into the droop controller of the inverter, the voltage drop on the virtual impedance is calculated, the voltage drop is added to the original output voltage reference value, and the output voltage reference value considering the virtual impedance is obtained. First, the voltage drop over the adaptive virtual impedance is calculated , : wherein, is the first i is the first is the first i is the first is the first i is the first is the first i is the first vi is the adaptive virtual resistance; L vi is the adaptive virtual inductance; The output voltage reference value considering the voltage drop on the adaptive virtual impedance is updated again. 、 wherein, and are the d-axis and q-axis voltage components, respectively, of the voltage reference input to the outer voltage loop of the voltage-current double-loop control of the mth inverter; i and are the d-axis and q-axis voltage components, respectively, of the voltage reference input to the outer voltage loop of the voltage-current double-loop control of the mth inverter; and i are the d-axis and q-axis voltage components, respectively, of the voltage reference input to the outer voltage loop of the voltage-current double-loop control of the mth inverter.

Images