Inverter single vector finite control set MPC optimization method and device based on suboptimal switch state switching

Abstract

The invention discloses an inverter single vector finite control set MPC optimization method and device based on suboptimal switch state switching, and the method comprises the steps: firstly collecting the three-phase current at a filter side and the current and voltage at an output side, predicting the output voltage at a next sampling moment, selecting a switch state corresponding to a minimum cost function as an optimal switch state, and carrying out the optimization of the optimal switch state; and recording the switching state corresponding to the second low-cost function as a sub-optimal switching state. Then judging whether the optimal switching state is the same as the current switching state or not, and if not, executing the optimal switching state; and if yes, executing sub-optimal switch state switching. And finally, calculating the switching time of the suboptimal switch state, and switching to the suboptimal switch state at the switching time to complete the control action. According to the method, the power quality of the output voltage of the inverter is effectively improved by selecting the suboptimal switching state for switching, meanwhile, the calculation amount about suboptimal switching time is small, implementation is easy, and sampling input does not need to be additionally added and a new predicted value does not need to be calculated.

Description

Technical Field

[0001] The present invention relates to the field of power electronics technology, and particularly to an optimization method and device for a single-vector finite control set MPC of an inverter based on sub-optimal switching state switching. Background Art

[0002] Since the 1980s, model predictive control (MPC) has been applied in many industrial process controls. Due to its advantages such as simple modeling, convenient parameter tuning, and strong robustness, it has been gradually extended to fields such as power electronics control and power system dispatching in recent years. In the field of power electronics, as a common control object, the inverter has the characteristics of a limited and discrete number of switching states, and is suitable for finite control set MPC (FCS-MPC). All switching states are substituted into the inverter prediction model. The switching state with the minimum cost function is regarded as the optimal switching state, and the switching action is executed. FCS-MPC is currently the most widely used inverter MPC method because it does not require PWM modulation and has relatively small computational complexity.

[0003] FCS-MPC traverses all possible switching states of the inverter through the prediction model, calculates the cost function, selects the switching state with the minimum cost function as the optimal switching state, and switches to the newly selected switching state when the inverter sampling time arrives. Since multiple consecutive switching states may be selected for the same switching, the switching states are the same. When the inverter switch remains on or off for a relatively long time, only a single switching state is adopted within the same sampling period, which not only reduces the service life of the switching device, but also reduces the power quality of the output voltage, with a large total harmonic distortion rate and an unfixed switching frequency. Therefore, improving the power quality of FCS-MPC is a research hotspot in the field of power electronics control. Summary of the Invention

[0004] In view of the above problems, the present invention proposes an optimization method and device for a single-vector finite control set MPC of an inverter based on sub-optimal switching state switching, which is used to control the output voltage of a voltage-source three-phase inverter. The present invention not only stabilizes the switching frequency, but also maximally utilizes each control period to improve the power quality of the inverter output voltage.

[0005] To achieve the above object, the present invention provides an optimization method for a single-vector finite control set MPC of an inverter based on sub-optimal switching state switching, including the following steps:

[0006] Step 1: Obtain the filter current, output-side current and voltage, and establish a discretized prediction model of the inverter;

[0007] Step 2: Traverse all inverter switching states, and obtain the optimal switching state and the sub-optimal switching state based on the discretized prediction model of the inverter;

[0008] Step 3: If the current switching state is the sub-optimal switching state, directly and immediately perform the optimal switching state transition; if the current switching state is the optimal switching state, perform the sub-optimal switching state transition and select the sub-optimal switching state as the switching state of the inverter for the next cycle.

[0009] Step 4: Calculate the sub-optimal switching time T based on the predicted output voltage and the reference voltage. com ;

[0010] Step 5: Switch to the sub-optimal switching state at the sub-optimal switching time T com to complete the control action.

[0011] Furthermore, in Step 1, collect the three-phase current and the output current and voltage on the filter side, transform them into the two-phase stationary alpha-beta coordinate system to obtain the filter current and the output current and voltage respectively, establish a discrete prediction model of the inverter in the alpha-beta coordinate system, and calculate the predicted output voltage.

[0012] Furthermore, in Step 2, calculate the cost function of each switching state based on the discrete prediction model of the inverter, select the switching state corresponding to the minimum cost function as the optimal switching state, and record the switching state corresponding to the second smallest cost function as the sub-optimal switching state.

[0013] Furthermore, in Step 4, calculate the sub-optimal switching time according to the predicted output voltage corresponding to the sub-optimal switching state and the cost function, the predicted output voltage corresponding to the optimal switching state and the cost function, and the reference voltage.

[0014] Furthermore, the specific process of calculating the sub-optimal switching time T com in Step 4 is as follows:

[0015] When the current switching state is the optimal switching state, the sub-optimal switching state transition time T com is:

[0016]

[0017] where μ is the combination coefficient, represents the sampling period.

[0018] Furthermore, compare the predicted output voltage after the sub-optimal switching time T com with the reference output voltage, and represent it by the square of the spatial distance between the two voltage vectors obtained by decomposition on the alpha-axis and beta-axis, and then calculate the combination coefficient when the distance is the smallest.

[0019] Furthermore, the combination coefficient μ is calculated as follows:

[0020]

[0021] Among them, the optimal switching state is n opt The predicted output voltage v o opt (k + 1) on the alpha-axis and beta-axis are respectively denoted as v oα opt (k + 1) and v oβ opt (k + 1), the sub-optimal switching state is n sub The predicted output voltage v o sub (k + 1) on the alpha-axis and beta-axis are respectively denoted as v oα sub (k + 1) and v oβ sub (k + 1), g subα and g subβ Are respectively the components of the cost function of the sub-optimal switching state on the alpha-axis and beta-axis.

[0022] On the other hand, the present invention provides an inverter single-vector finite control set MPC optimization device based on sub-optimal switching state switching, and the device includes an inverter discretized prediction model construction module, a switching state judgment module, a sub-optimal switching time calculation module, and a switching state switching module;

[0023] The inverter discretized prediction model construction module is used to obtain filter current, output-side current and voltage, and establish an inverter discretized prediction model;

[0024] The switching state judgment module is used to traverse all inverter switching states, and obtain the optimal switching state and sub-optimal switching state based on the inverter discretized prediction model; if the current switching state is the sub-optimal switching state, directly perform the optimal switching state switching immediately; if the current switching state is the optimal switching state, perform the sub-optimal switching state switching, and select the sub-optimal switching state as the switching state of the inverter in the next cycle;

[0025] The sub-optimal switching time calculation module is used to calculate the sub-optimal switching time T based on the predicted output voltage and the reference voltage com ;

[0026] The switching state switching module is used to switch to the sub-optimal switching state at the switching time T com To complete the control action.

[0027] Advantages of the present invention:

[0028] (1) Compared with the traditional single-vector finite control set MPC that only adopts the optimal switching state, the present invention maximizes the control efficiency per cycle while maintaining a fixed switching frequency by adding the option of sub-optimal switching states, and effectively improves the power quality of the inverter output voltage by dynamically adjusting the switching duration.

[0029] (2) In the present invention, the calculation amount for the sub-optimal switching time is small, the implementation is simple, and there is no need to additionally increase the sampling input and calculate new predicted values. Description of the Drawings

[0030] Figure 1 It is a flowchart for implementing the sub-optimal switching optimization method based on the single-vector finite control set MPC of the inverter and the model of the inverter connected by an LC inverter.

[0031] Figure 2 It is an analysis diagram of the inverter output voltage under the traditional finite control set MPC.

[0032] Figure 3 It is an analysis diagram of the inverter output voltage under the control of the present invention.

[0033] Figure 4 It is a schematic diagram of a sub-optimal switching optimization device based on the single-vector finite control set MPC of the inverter according to the present invention. Detailed Embodiment

[0034] The following further elaborates on the specific embodiments of the present invention in conjunction with the drawings.

[0035] As Figure 1 shown, the present invention provides an MPC optimization method for a single-vector finite control set of an inverter based on sub-optimal switching states, and the specific implementation scheme is as follows:

[0036] Step 1: Collect the three-phase current on the filter side and the output current and voltage when the LC inverter is connected to the AC bus model as Figure 1 shown, and transform them to the two-phase stationary alpha-beta coordinate system to obtain the filter current and the output current and voltage respectively, and establish a discrete prediction model of the inverter in the alpha-beta coordinate system:

[0037]

[0038]

[0039]

[0040] where Ts represents the sampling period, k represents the current sampling moment, k + 1 represents the next sampling moment, and i f(k) is the filter current at the current sampling moment, v o (k) is the output voltage at the current sampling moment, i o (k) is the output current at the current sampling moment, v o (k + 1), i o (k + 1) are the predicted output voltage and current at the next sampling moment respectively, v n (k) is the inverter output voltage vector, n takes values from 0 to 7, corresponding to 8 groups of inverter switch states, g n represents the cost function for model predictive control MPC under different switch states, v oα (k + 1), v oβ (k + 1) are respectively v o (k + 1) components on the alpha-axis and beta-axis, v α * (k + 1), v β * (k + 1) is the inverter reference output voltage v * (k + 1) components on the alpha-axis and beta-axis respectively, Cf is the capacitance value of the LC filter, Lf is the inductance value of the LC filter.

[0041] Step 2: According to the inverter discretized prediction model in Step 1, traverse the 8 groups of switch states of the inverter, select the switch state corresponding to the minimum cost function as the optimal switch state denoted as n opt , record the switch state corresponding to the second smallest cost function as the sub-optimal switch state n sub , judge whether the optimal switch state n opt is the same as the current switch state n, if different, execute the optimal switch state n opt ; if the same, execute the sub-optimal switch state n sub switching, that is, select the sub-optimal switch state as the switch state of the inverter in the next cycle.

[0042] Step 3: The predicted output voltage v opt under the optimal switch state n o opt (k + 1) components on the alpha-axis and beta-axis are respectively denoted as v oα opt (k + 1) and v oβ opt (k + 1), the sub-optimal switch state n sub under the predicted output voltage v o sub (k + 1) components on the alpha-axis and beta-axis are respectively denoted as v oα sub (k + 1) and v oβsub (k + 1), the components of the reference output voltage v*(k + 1) on the alpha-axis and beta-axis are denoted as v α *(k + 1) and v β *(k + 1); when the current switching state is the same as the optimal switching state, the sub-optimal switching time T com is as follows:

[0043]

[0044] where,

[0045]

[0046] where, µ is the combination coefficient, g subα and g subβ are respectively the components of the cost function of the sub-optimal switching state on the alpha-axis and beta-axis.

[0047] Furthermore, the calculation of the delayed switching time in step 3 is further described as follows:

[0048] When the optimal switching state n opt is the same as the current switching state n, the predicted output voltage v com after the sub-optimal switching time T o com (k + 1) is expressed as:

[0049]

[0050] At this time, the v o com (k + 1) combined by the combination coefficient µ has the minimum error with v*(k + 1), and the corresponding sub-optimal switching time is µTs.

[0051] Step 4: Switch to the sub-optimal switching state at the switching time T com to complete the control action.

[0052] To reflect the control effect of the present invention, an inverter using the traditional FCS-MPC is compared with an inverter using the control method of the present invention. The sampling frequency is 10k for both, the amplitude of the reference output voltage is 311V, and the frequency is 50Hz. As Figure 2 shown is the output voltage situation of the inverter under the traditional finite control set MPC. The FFT fast Fourier transform analysis function of MATLAB / SIMULINK is used to analyze the a-phase output voltage of the inverter in the steady state, and it is obtained that its amplitude is 308.8V and the total harmonic distortion rate is 1.23%; as Figure 3The figure shows the output voltage of the inverter under the control of the present invention. Using the same spectrum analysis function to analyze the output voltage of phase a of the inverter in the steady state, the amplitude is obtained as 310.3 V, and the total harmonic distortion rate is 1.07%. It can be seen that for the inverter controlled by the present invention, the amplitude of its output voltage is closer to the reference value, and its total harmonic distortion rate is also smaller, improving the power quality of the output voltage of the inverter.

[0053] Corresponding to the foregoing embodiment of an optimization method for a single-vector finite control set MPC of an inverter based on sub-optimal switching state switching, the present invention also provides an embodiment of an optimization device for a single-vector finite control set MPC of an inverter based on sub-optimal switching state switching.

[0054] As Figure 4 shown, the present invention also provides an optimization device for a single-vector finite control set MPC of an inverter based on sub-optimal switching state switching. The device includes an inverter discretized prediction model construction module, a switching state judgment module, a sub-optimal switching time calculation module, and a switching state switching module;

[0055] The inverter discretized prediction model construction module is used to collect the three-phase current and output-side current and voltage on the filter side, transform them to the two-phase stationary alpha-beta coordinate system to obtain the filter-side current and output-side current and voltage respectively, establish an inverter discretized prediction model in the alpha-beta coordinate system, and calculate the predicted output voltage;

[0056] The switching state judgment module is used to traverse the switching states of the inverter. The switching state judgment module is used to traverse the switching states of the inverter. According to the calculated cost function, select the switching state corresponding to the minimum cost function as the optimal switching state, record the switching state corresponding to the second minimum cost function as the sub-optimal switching state, and judge whether the optimal switching state is the same as the current switching state. If they are different, execute the optimal switching state; if they are the same, execute the sub-optimal switching state switching, that is, select the sub-optimal switching state as the switching state of the inverter in the next cycle.

[0057] The sub-optimal switching time calculation module is used to calculate the sub-optimal switching time T according to the predicted output voltage corresponding to the sub-optimal switching state and the cost function, the predicted output voltage corresponding to the optimal switching state and the cost function, and the reference voltage com ;

[0058] The switching state switching module is used to switch to the sub-optimal switching state at the switching time T com to complete the control action.

[0059] For the realization process of the functions and roles of each module in the above device, please refer to the realization process of the corresponding steps in the above method for details, and will not be elaborated here.

[0060] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the descriptions of the method embodiments. The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separated, and some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of the present invention. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0061] The above embodiments are used to explain the present invention rather than limit the present invention. Any modifications and changes made to the present invention within the spirit and scope of the claims of the present invention fall within the protection scope of the present invention.

Claims

1. An inverter single vector finite control set MPC optimization method based on suboptimal switch state switching, characterized in that: The steps include: Step 1: Obtain the filter current and output side current and voltage, and establish a discrete prediction model for the inverter; Step 2: Traverse all inverter switching states and obtain the optimal switching state and suboptimal switching state based on the inverter discretization prediction model; Step 3: If the current switch state is a suboptimal switch state, the optimal switch state is switched immediately; if the current switch state is an optimal switch state, the suboptimal switch state is switched, and the suboptimal switch state is selected as the switch state of the inverter in the next cycle; Step 4: Calculate the suboptimal switching time T based on the predicted output voltage and reference voltage com ; Step 5: At the suboptimal switching time T com Switch to the suboptimal switching state to complete the control action.

2. The inverter single vector finite control set MPC optimization method based on suboptimal switch state switching according to claim 1, characterized in that: In step 1, the three-phase current of the filter side and the output side current and voltage are collected and transformed into a two-phase stationary alpha-beta coordinate system to obtain the filter current and the output side current and voltage respectively, establish a discrete prediction model of the inverter in the alpha-beta coordinate system, and calculate the predicted output voltage.

3. The inverter single vector finite control set MPC optimization method based on suboptimal switch state switching according to claim 1, characterized in that: In step 2, the cost function of each switch state is calculated based on the inverter discretization prediction model, the switch state corresponding to the minimum cost function is selected as the optimal switch state, and the switch state corresponding to the second smallest cost function is recorded as the suboptimal switch state.

4. The inverter single vector finite control set MPC optimization method based on suboptimal switch state switching according to claim 1, characterized in that: In step 4, the suboptimal switching time is calculated according to the predicted output voltage corresponding to the suboptimal switching state and the cost function, the predicted output voltage corresponding to the optimal switching state and the cost function and the reference voltage.

5. The inverter single vector finite control set MPC optimization method based on suboptimal switch state switching according to claim 1, characterized in that: The suboptimal switching time T is calculated in step 4 com The specific process is: If the current switch state is the optimal switch state, the suboptimal switch state switching time T com for: ； Among them, µ is the combination coefficient, Indicates the sampling period.

6. The inverter single vector finite control set MPC optimization method based on suboptimal switch state switching according to claim 4, characterized in that: After the suboptimal switching time T com The predicted output voltage after the calculation is compared with the reference output voltage, and the square representation of the spatial distance between the two voltage vectors is obtained by decomposing the two voltage vectors on the alpha axis and the beta axis, and then the combination coefficient when the distance is minimum is calculated.

7. The inverter single vector finite control set MPC optimization method based on suboptimal switch state switching according to claim 4, characterized in that: The combination coefficient µ is calculated as follows: ； Among them, the optimal switching state n opt The predicted output voltage v o opt The components of (k+1) on the alpha and beta axes are denoted as v oα opt (k+1) and v oβ opt (k+1), suboptimal switching state n sub The predicted output voltage v o sub The components of (k+1) on the alpha and beta axes are denoted as v oα sub (k+1) and v oβ sub (k+1), g subα and g subβ are the components of the cost function of the suboptimal switching state on the alpha and beta axes, respectively.

8. An inverter single vector finite control set MPC optimization device based on suboptimal switch state switching for implementing the method according to any one of claims 1 to 7, characterized in that: The device includes an inverter discretization prediction model building module, a switch state judgment module, a suboptimal switching time calculation module and a switch state switching module; The inverter discretization prediction model building module is used to obtain the filter current and the output side current and voltage, and establish the inverter discretization prediction model; The switch state judgment module is used to traverse all the switch states of the inverters, and obtain the optimal switch state and the suboptimal switch state based on the inverter discretization prediction model; if the current switch state is the suboptimal switch state, the optimal switch state switching is directly and immediately executed; if the current switch state is the optimal switch state, the suboptimal switch state switching is executed, and the suboptimal switch state is selected as the switch state of the inverter in the next cycle; The suboptimal switching time calculation module is used to calculate the suboptimal switching time T based on the predicted output voltage and the reference voltage. com ; The switch state switching module is used to switch the state of the switch at a switching time T com Switch to the suboptimal switching state to complete the control action.

Images