An active filter control method of a grid-connected inverter

Abstract

The application discloses an active filter control method of a grid-connected inverter, and belongs to the field of power electronics technology and power quality control.The method solves the problems of the prior art, such as complex active filter control of the grid-connected inverter, low processor execution efficiency and poor control effect.The method comprises the following steps: predicting state variables at k+1 and k+2 time domains, wherein the filter inductance current at the k+2 time domain is predicted from the voltage vector at the k time; based on historical data of grid voltage, a virtual harmonic current of multiple harmonics is constructed by using a sliding update method, and the virtual harmonic current of multiple harmonics at the k+2 time domain is predicted; the filter inductance current at the k+2 time domain and the virtual harmonic current of multiple harmonics are evaluated according to the current reference signal at the k+2 time domain; an evaluation function is taken as an optimization target, the optimal voltage vector at the k time is obtained, and the grid-connected inverter is controlled according to a switching sequence signal corresponding to the optimal voltage vector.

Description

TECHNICAL FIELD

[0001] The present application relates to an active filter control method of a grid-connected inverter, and belongs to the field of power electronics technology and power quality control. BACKGROUND

[0002] Nonlinear loads will generate harmonic currents after being connected to the power grid, which will cause voltage and current distortion under the action of line impedance. When the power is large, the power quality of the power grid will be seriously affected, which will threaten the reliability of power supply. With the promotion of new power system construction and the increase of new energy generation penetration, there are more and more power electronic equipment in the power grid, especially inverters. Compared with traditional generators, inverter equipment has unique advantages in operational flexibility. Through the improvement of control strategy, it can achieve the additional function of active filtering on the basis of connecting new energy to the power grid, and realize the harmonic control and power quality improvement of the power grid. Therefore, using the grid-connected inverter equipment in the power grid to improve the power quality has good application prospect.

[0003] In the process of new energy conversion and power generation, in order to meet the requirement of unity power factor grid connection, the inverter usually only controls the fundamental wave of voltage and current. The commonly used methods are linear methods such as proportional integral decoupling control in dq domain and proportional resonant control in Domain. Linear control method has good steady-state and dynamic characteristics when implementing fundamental wave control, but its implementation process depends on the frequency of control variable. When applying proportional integral control, three-phase alternating current quantity needs to be converted into two-phase direct current component in dq domain, and the conversion process and decoupling link need to be based on the fundamental frequency. In the proportional resonant control method, the construction of the resonant controller also needs the frequency of the controlled variable. When the inverter realizes the function of active filtering, it is necessary to add control of harmonic components on the basis of ensuring the control of fundamental wave. In this case, if proportional integral control is used, parallel coordinate transformation and controller links need to be configured for fundamental wave and multiple harmonic components. When proportional resonant control is used, multiple harmonic controllers also need to be set in parallel. This means that when the two types of methods are used for active filtering control of the inverter, the calculation burden of the processor will be greatly increased, resulting in the influence of operation efficiency and control effect.

[0004] The existing linear control method of grid-connected inverter contains complex coordinate transformation, decoupling, resonant controller and other links, and its implementation strictly depends on the frequency of control variable. When the inverter realizes the function of active filtering, multiple parallel controllers need to be set to process fundamental wave and multiple harmonic components after harmonic extraction, which increases the complexity of the system, reduces the execution efficiency of the processor, and makes the control effect unsatisfactory. In addition, the characteristics of linear controller are greatly affected by parameters, and when multiple groups of components are processed in parallel, the complex structure leads to a large number of control system parameters and complex coupling mechanism, which brings difficulty to the optimization design of the control system. Summary of the Invention

[0005] To address the problems of complex active filter control, low processor execution efficiency, and poor control effect in existing technologies for grid-connected inverters, this invention provides an active filter control method for grid-connected inverters.

[0006] An active filter control method for a grid-connected inverter according to the present invention includes:

[0007] Obtain time k The state variables of the domain include the filter inductor current, the current reference signal, and the grid voltage.

[0008] Combined with time k The state variables of the domain are predicted to be available at time k+1. The state variables of the domain are then used to predict the state at time k+2. The state variables of the domain, where, at time k+1 The filter inductor current in the domain, at time k+1 The grid voltage of the domain and the voltage vector at time k are predicted to obtain the voltage at time k+2. The filter inductor current in the domain;

[0009] Based on historical grid voltage data, a sliding update method is used to construct the k-time point. The virtual harmonic current of a given order in the domain is used to predict the time at k+2. Virtual harmonic current of a set order in the domain:

[0010]

[0011] In the formula, At time k+2 The grid voltage of the region, for time The grid voltage of the region, For harmonic order, Virtual resistor for harmonics of a set order. For the number of points in the sliding window, The damping coefficient is... It is a natural constant. represents an imaginary number;

[0012] Based on time k+2 The current reference signal of the domain at time k+2 The filter inductor current of the domain and the selected virtual harmonic currents of multiple set orders are evaluated to obtain the evaluation function at time k+2. The evaluation function at time k+2 is used as the optimization target to obtain the optimal voltage vector at time k. The grid-connected inverter is controlled according to the switching sequence signal corresponding to the optimal voltage vector.

[0013] As a preferred option, when the inverter is operating under complex grid conditions, reducing the damping coefficient r can improve the stability of the grid-connected inverter.

[0014] As a preferred option, the evaluation function at time k+2 is:

[0015]

[0016] in, , , They are at time k+2 respectively. The current reference signal, filter inductor current, and virtual harmonic current in the domain;

[0017] The voltage vector at time k corresponding to the minimum is the optimal voltage vector.

[0018] The beneficial effects of this invention are as follows: This invention utilizes a sliding update method to construct virtual harmonic currents. The computational load of each iteration is relatively small and does not increase with the sample size. Embedding the iterative process of constructing virtual harmonic currents into the prediction process of the FCS-MPC controller simplifies control compared to other methods that extract harmonics and then input them into the controller. The reduction in computational load and the simplification of the structure significantly reduce the processor's execution burden. This invention utilizes the FCS-MPC method to construct an active filter controller for grid-connected inverters, enabling simultaneous control of the fundamental and multiple harmonics with a single controller. Compared to linear control methods such as proportional-integral and proportional-resonant, it eliminates the need for parallel controllers, resulting in a simpler structure and higher control efficiency. Furthermore, it offers advantages such as convenient parameter tuning, high control bandwidth, fast dynamic response, and strong robustness. Additionally, when the inverter operates under complex grid conditions such as weak grids, this invention can improve the stability of the grid-connected inverter by adjusting the damping factor r in the sliding update process of the virtual harmonic current. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the principle of the present invention;

[0020] Figure 2 This is a voltage vector diagram of the three-phase two-level inverter of the present invention;

[0021] Figure 3 The waveforms of the grid voltage and grid current when the active filter control method of this application is not activated;

[0022] Figure 4 The grid voltage and grid current harmonic waveforms are shown when the active filter control method of this application is not activated;

[0023] Figure 5 The waveforms of the grid voltage and grid current after activating the active filter control method of this application;

[0024] Figure 6 The grid voltage and grid current harmonic waveforms after the active filter control method of this application is activated. Detailed Implementation

[0025] 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.

[0026] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0027] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0028] Whether it's PI control in the dq domain, or Proportional resonant control in the frequency domain essentially involves designing an independent controller for each frequency component of interest. This results in a controller number proportional to the harmonic order, complex system structure, enormous computational burden, and coupling between controllers, making parameter tuning difficult. This application abandons the design of independent controllers for each harmonic in the frequency domain and instead adopts the finite control set model predictive control method FCS-MPC. The FCS-MPC method directly... Operating in the time domain in a stationary coordinate system, this application unifies the handling of fundamental and harmonic control objectives through predictive models, rolling optimization, and feedback correction. In each control cycle, this application proposes predicting the system state at multiple future moments and directly evaluating the impact of all possible inverter switching states at the current moment on the control objective, selecting the optimal output. This avoids complex coordinate transformations and resonator arrays. The active filter control method for grid-connected inverters in this application includes:

[0029] Step 1: Obtain time k The state variables of the domain include the filter inductor current, the current reference signal, and the grid voltage; combined with time k... The state variables of the domain are predicted to be available at time k+1. The state variables of the domain are then used to predict the state at time k+2. The state variables of the domain, where, at time k+1 The filter inductor current in the domain, at time k+1 The grid voltage of the domain and the voltage vector at time k are predicted to obtain the voltage at time k+2. The filter inductor current in the domain, and the voltage vector at time k are unknowns, including multiple possibilities;

[0030] In step 1 of this application, the state variables at time k and before time k, and the optimal voltage vector determined at time k-1, are already determined quantities. Combining these with all currently applied possible voltage vectors, and based on the discrete mathematical model of the grid-connected inverter, the states at times k+1 and k+2 are predicted. This takes into account the dynamic characteristics of the system and provides a basis for making optimal control decisions in advance.

[0031] Step 2: Based on historical grid voltage data, construct time step k using the sliding update method. The virtual harmonic current of a given order in the domain is used to predict the time at k+2. Virtual harmonic current of a set order in the domain:

[0032] (1)

[0033] In the formula, At time k+2 The grid voltage of the region, for time The grid voltage of the region, For harmonic orders, multiple values ​​can be selected simultaneously. Virtual resistor for harmonics of a set order. For the number of points in the sliding window, The damping factor, It is a natural constant. Represents an imaginary number; i vh (k+1) is the virtual harmonic current at time k+1, which contains the accumulated information of the power grid voltage historical samples from k+1-N to k+1 for a total of N sampling periods; As a rotation factor, the virtual harmonic current i at time k+1 is... vh (k+1) Rotate by a fixed phase so that the recursive process can track the phase change of the harmonic component. r is the damping factor used to ensure system stability, and its range is (0~1). This represents the sliding update process of sample data, that is, adding new sample data of grid voltage at time k+2. and remove the previous Old sample data at time Multiplied by This indicates that the removed sample was attenuated. The sliding update construction method for virtual harmonic currents can be organically integrated with the prediction process of FCS-MPC. Compared with the scheme of first extracting harmonics and then handing them over to the controller for control, it can greatly reduce the complexity of the inverter active filter system. The computational cost of each iteration process is small and does not change with the increase of the sample size N.

[0034] Step 2 of this application calculates the virtual harmonic current generated when a virtual path with low impedance to a specific harmonic exists, based on the predicted grid voltage at time k+2. This virtual current essentially defines the harmonic current filtering target that the controller needs to track. Combining the virtual impedance characteristics of specific harmonics (such as the 5th, 7th, 11th, and 13th harmonics) is equivalent to setting the filtering requirements for these harmonics in the control target all at once. Step 2 of this application embeds the sliding filter process into the prediction optimization process of FCS-MPC, realizing efficient parallel calculation of virtual harmonic currents of each order. By independently setting the virtual harmonic resistance values, flexible control of multiple harmonic currents can be achieved simultaneously.

[0035] Step 3: Based on time k+2 The current reference signal of the domain at time k+2 The filter inductor current of the domain and the selected virtual harmonic currents of multiple set orders are evaluated to obtain the evaluation function at time k+2. The evaluation function at time k+2 is used as the optimization target to obtain the optimal voltage vector at time k. The grid-connected inverter is controlled according to the switching sequence signal corresponding to the optimal voltage vector.

[0036] Step 3 of this application iterates through the switching states of the grid-connected inverter and calculates the difference between the predicted future filter inductor current, the virtual harmonic current, and the current reference signal after each voltage vector is applied. The voltage vector that minimizes this difference is selected as the optimal output. This application achieves both accurate tracking of the current reference signal and effective suppression of the set subharmonic through a single optimization calculation, perfectly integrating the two control objectives within a single framework.

[0037] This application utilizes the time-domain optimization capability of model predictive control, combined with composite harmonic virtual impedance, to set a unified harmonic filtering target, thereby replacing the traditional complex parallel multi-frequency domain controller with a simple framework.

[0038] Step 1 of this application specifically includes:

[0039] Step 11: Obtain time k The state variables of the domain include the filter inductor current, the current reference signal, and the mains voltage; such as Figure 1 As shown, U dc L is the DC input voltage of the inverter.f For filter inductance, R lf C is the parasitic resistance of the filter inductor. f For the filter capacitor, R f Z is the damping resistor. g Z is the power grid impedance. vh This represents the virtual harmonic impedance, and PCC is the inverter's grid connection point. fabc i represents the three-phase inductor current of the inverter. gabc For the three-phase power grid current, u gabc The three-phase grid voltage and the inverter three-phase inductor current i are given. fabc After Clarke transform, time k is obtained. Filter inductor current in the domain Three-phase grid voltage u gabc After Clarke transform, time k is obtained. Grid voltage in the region ;

[0040] k moment The grid voltage in the region is used to calculate the grid angle via a phase-locked loop (PLL). The angle of the power grid With current setpoint Simultaneously, the input is fed into the inverse Park unit, which calculates time k. Current reference signal of the domain ;

[0041] Step 12: To reduce the impact of system delay on control performance, a two-step prediction method is used for delay compensation. First, based on time k... State variables of the domain, predicting time k+1 The state variables of the domain specifically include:

[0042] according to Figure 1 The continuous state equation of the main circuit of the grid-connected inverter can be established as follows:

[0043] (2)

[0044] Represents a voltage vector;

[0045] The discrete state equations of the inverter can be derived using the forward Euler method.

[0046] (3)

[0047] in, Let k represent the sampling time, and let equation (3) be used as the prediction model for the inverter's filter inductor current. That is: from time k... Grid voltage in the region The optimal voltage vector determined at time k-1 Predicting time k+1 Filter inductor current in the domain ;

[0048] Predicting time k+1 using the extension method The grid voltage of the region is:

[0049] (4)

[0050] in, , They are respectively , time The grid voltage of the region;

[0051] The k+1 time step can also be predicted using the extension method. The current reference signal for the domain is:

[0052] (5)

[0053] in, , They are respectively , time The current reference signal of the domain;

[0054] Based on time k+1 Predicting the state variables of the domain at time k+2 The state variables of the domain specifically include:

[0055] Based on time k+1 Filter inductor current in the domain k+1 time Grid voltage in the region and the voltage vector at time k Predict time k+2 The filter inductor current of the domain is:

[0056] (6)

[0057] in, This represents the voltage vector at time k, including all possible states of the inverter switches. It is a three-phase two-level inverter voltage vector diagram, as shown below. Figure 2 As shown, there are a total of 8 voltage vectors: , , , , , , , This corresponds to 8 switch states: , , , , , , , 1 represents that the upper arm of each phase is on, and 0 represents that the lower arm is on. It represents the imaginary unit.

[0058] Time k+2 The current reference signal for the domain is:

[0059] (7)

[0060] in, They are respectively time The current reference signal of the domain.

[0061] At time k+2 The grid voltage of the region is:

[0062] (8)

[0063] in, for time The current reference signal of the domain.

[0064] Step 2 of this application uses historical grid voltage data and a sliding update method to construct virtual harmonic currents of a specific order, as shown in equation (1). In actual operation, the operating stability of the grid-connected inverter can be improved by adjusting the damping coefficient r. Specifically, when the inverter operates under complex grid conditions such as weak grids, the damping coefficient is reduced. Until the grid voltage and grid current stop oscillating.

[0065] Step 3 of this application specifically includes:

[0066] At time k+2 Current reference signal of the domain Using k+2 time as a baseline, the evaluation function at time k+2 is used to evaluate the predicted time k+2. Domain virtual harmonic current and k+2 time Filter inductor current in the domain An evaluation is performed to determine the optimal voltage vector. The evaluation function at time k+2 can be expressed as:

[0067] (10)

[0068] In the formula, The voltage vector that achieves the minimum value is called the optimal voltage vector. Applying the corresponding switching sequence to the switching transistors in the inverter's main circuit can achieve closed-loop control of the current, thereby achieving harmonic filtering.

[0069] This application implements the active filtering function of a grid-connected inverter, enabling simultaneous control of the fundamental frequency and multiple harmonics using a single controller. The controller does not need to distinguish between harmonic orders, simplifying the structure of the active filtering control system for the grid-connected inverter while maintaining control performance. Furthermore, adjusting the damping coefficient *r* during the virtual harmonic current sliding update process can improve the stable operation of the grid-connected inverter under complex grid conditions.

[0070] To verify the active filter control method for a grid-connected inverter proposed in this invention, reference is made to... Figure 1 A grid-connected inverter simulation model was built, and a diode rectifier bridge with resistors and inductors as loads was installed at the PCC as a nonlinear load. The system parameters are shown in Table 1.

[0071] Table 1 Parameters of Grid-Connected Inverter Active Filter System

[0072]

[0073] Figure 3 The waveforms of the grid voltage and grid current are shown when active filter control is not enabled. It can be seen that the grid voltage and current waveforms are severely distorted due to the influence of nonlinear load. Figure 4 The corresponding harmonic spectra show that the voltage and current harmonic amplitudes are relatively high at frequencies of 250Hz, 350Hz, 550Hz, and 650Hz, corresponding to harmonic orders of 5, 7, 11, and 13, respectively. Calculations show that the total harmonic distortion (THD) of the grid voltage and current is 9.2% and 5.6%, respectively, indicating a high harmonic content.

[0074] After enabling active filter control, the waveforms of the grid voltage and grid current are as follows: Figure 5 As shown, with Figure 3 Compared to the previous method, the waveform quality is significantly improved. The corresponding harmonic spectrum is as follows: Figure 6 As shown, with Figure 4 In comparison, the harmonic amplitudes of grid voltage and current have decreased significantly, with THD dropping to 2.5% and 3.2%, respectively.

[0075] Simulation results show that the proposed active filter control method can effectively reduce the distortion of grid voltage and current when nonlinear loads are connected, and significantly improve power quality.

[0076] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. An active filter control method for a grid-connected inverter, characterized in that, include: Obtain time k The state variables of the domain include the filter inductor current, the current reference signal, and the grid voltage. Combined with time k The state variables of the domain are predicted to be available at time k+1. The state variables of the domain are then used to predict the state at time k+2. The state variables of the domain, where, at time k+1 The filter inductor current in the domain, at time k+1 The grid voltage of the domain and the voltage vector at time k are predicted to obtain the voltage at time k+2. The filter inductor current in the domain; Based on historical grid voltage data, a sliding update method is used to construct the k-time point. The virtual harmonic current of a given order in the domain is used to predict the time at k+2. Virtual harmonic current of a set order in the domain: In the formula, At time k+2 The grid voltage of the region, for time The grid voltage of the region, For harmonic order, Virtual resistor for harmonics of a set order. The number of points in the sliding window. The damping coefficient is... It is a natural constant. represents an imaginary number; Based on time k+2 The current reference signal of the domain at time k+2 The filter inductor current of the domain and the selected virtual harmonic currents of multiple set orders are evaluated to obtain the evaluation function at time k+2. The evaluation function at time k+2 is used as the optimization target to obtain the optimal voltage vector at time k. The grid-connected inverter is controlled according to the switching sequence signal corresponding to the optimal voltage vector.

2. The active filter control method for a grid-connected inverter according to claim 1, characterized in that, When the inverter is operating under complex grid conditions, reducing the damping coefficient r improves the stability of the grid-connected inverter.

3. The active filter control method for a grid-connected inverter according to claim 1, characterized in that, The evaluation function at time k+2 is: in, , , They are at time k+2 respectively. The current reference signal, filter inductor current, and virtual harmonic current in the domain; The voltage vector at time k corresponding to the minimum is the optimal voltage vector.

4. The active filter control method for a grid-connected inverter according to claim 1 or 3, characterized in that, Time k+2 The filter inductor current of the domain is: in, Represents time k+1 The filter inductor current in the domain; Indicates the filter inductance; Represents the voltage vector at time k; This represents the parasitic resistance of the filter inductor; Indicates the sampling time; Represents time k+1 The grid voltage of the region.

5. The active filter control method for a grid-connected inverter according to claim 1 or 3, characterized in that, Time k+2 The current reference signal for the domain is: in, , , They are respectively , , time The current reference signal of the domain.

6. The active filter control method for a grid-connected inverter according to claim 5, characterized in that, Methods for obtaining the current reference signal at time k include: k moment The grid voltage in the region is used to calculate the grid angle via a phase-locked loop (PLL). The angle of the power grid With current setpoint Simultaneously, the input is fed into the inverse Park unit, which calculates time k. Current reference signal of the domain .

7. The active filter control method for a grid-connected inverter according to claim 1, characterized in that, Time k+2 The grid voltage of the region is: in, , , They are respectively , , time The grid voltage of the region.

Images