Two-step voltage model prediction control method for three-level grid-connected inverter

Abstract

The invention discloses a two-step voltage model prediction control method for a three-level grid-connected inverter, and relates to the technical field of power electronics, and the method comprises the steps: building a topological structure of an NPC three-level inverter; establishing a mathematical model of the NPC three-level inverter based on the alpha-beta coordinate system; discretizing the mathematical model of the LC output filter by using a first-order Euler equation, predicting the inductive current and the output voltage of the (k + 1) th sampling period, and completing single-step prediction; the load side current value of the (k + 1) th sampling period is predicted, and k represents the index of the current sampling period; two-step prediction is completed based on the single-step prediction; and completing two-step voltage model prediction control of the current sampling period, entering the next sampling period, and repeating the process. According to the method, through two-step prediction control, the control precision is improved for the optimal switching sequence selection delay problem, and quick response and stable operation of the system are ensured.

Description

Technical Field

[0001] The present invention relates to the field of power electronics technology, and specifically to a two-step voltage model predictive control method for a three-level grid-connected inverter. Background Art

[0002] With the growth of global energy demand, countries attach importance to the development and utilization of new energy, which can reduce greenhouse gas emissions. When the proportion of new energy power generation is low, a synchronous generator can be used to connect to a stable power system; however, as the penetration rate expands, it is difficult for traditional synchronous generators to compensate for random fluctuations, posing challenges to the stable operation of the power system.

[0003] New energy power generation systems usually adopt the form of DC-AC or AC-DC power conversion. The grid-connected converter connects the power generation unit to the grid, and its structure and control strategy will directly affect the power generation efficiency and the power quality of grid connection. Although the grid-connected converter has a fast response speed, it has no moment of inertia and is difficult to participate in grid regulation.

[0004] The proposed virtual synchronous generator technology (VSG) enables the inverter to simulate the damping and inertia characteristics of a synchronous generator, and responds to the fluctuations of the grid frequency and voltage through two parts: active frequency control and reactive voltage control, making up for the lack of inertia in droop control and further improving the stability of the grid. However, the underlying control of general VSG technology still uses a voltage-current double closed-loop system as the control output, which has limitations in parameter tuning and dynamic response, and is costly. Therefore, a control method that combines model predictive control technology (MPC) with VSG technology is proposed, using MPC technology to replace the voltage-current double closed-loop to achieve the purpose of cost reduction and efficiency improvement.

[0005] Theoretically, single-step prediction can make the switching combination reach the ideal value within a certain control period. However, the response time consumed in this process will cause a delay phenomenon in the selection of the optimal switching sequence for a control system with a high sampling frequency. Summary of the Invention

[0006] The purpose of the present invention is to provide a two-step voltage model predictive control method for a three-level grid-connected inverter to solve the problems raised in the prior art.

[0007] To achieve the above purpose, the present invention provides the following technical solution: A two-step voltage model predictive control method for a three-level grid-connected inverter, the method includes:

[0008] Build the topological structure of the NPC three-level inverter;

[0009] Based on the αβ coordinate system, establish the mathematical model of the NPC three-level inverter, define the switching state combinations of the NPC three-level inverter, and generate voltage vectors;

[0010] The mathematical model of the LC output filter is discretized using the first-order Euler equation to predict the inductor current and output voltage at the (k + 1)-th sampling period, completing single-step prediction, where k represents the index of the current sampling period, i.e., the sampling point at the current moment;

[0011] Predict the load-side current value at the (k + 1)-th sampling period, and in combination with the single-step prediction, use the mathematical model to further predict the inductor current and output voltage at the (k + 2)-th sampling period;

[0012] Based on the comparison between the output voltage at the (k + 2)-th sampling period and the reference voltage, obtain the cost function, select the optimal voltage vector that minimizes the cost function, and complete two-step prediction;

[0013] Input the optimal voltage vector signal into the gates of the NPC three-level inverter to control the output of the desired voltage and current;

[0014] Complete the two-step voltage model predictive control for the current sampling period, enter the next sampling period, and repeat the above process.

[0015] According to the above scheme, the topological structure of the NPC three-level inverter includes the DC-side voltage, DC-side capacitor, filter capacitor, filter inductor, and line resistance; the NPC three-level inverter includes three bridge arms, each bridge arm consists of four switching tubes and two clamping diodes, and can output three levels.

[0016] According to the above scheme, the definition of the switching state combinations of the NPC three-level inverter and the generation of voltage vectors include:

[0017] The switching state is defined as S i ={1, 0, -1}; where S i represents the switching state of the bridge arm, i ∈ {a, b, c}, where i represents the bridge arm, and a, b, and c respectively represent the three bridge arms of the NPC three-level inverter;

[0018] When S i =1, the bridge arm outputs a high level, indicating a positive voltage output. When S i =0, the bridge arm outputs a zero level, indicating a medium voltage output. When S i =-1, the bridge arm outputs a low level, indicating a negative voltage output;

[0019] The three switching state combinations of the three arm bridges of the NPC three-level inverter are combined to define a total of 27 switching state combinations, and each switching state combination corresponds to a specific voltage vector. The formula is as follows:

[0020]

[0021] Among them, Uα Denoted as the α-axis voltage component; U β Denoted as the β-axis voltage component; U dc Denoted as the DC-side voltage; S a , S b and S c Denoted as the switching states of the three-phase bridge arms.

[0022] According to the above scheme, the single-step prediction includes:

[0023] Adjust the active power and frequency through the rotor motion equation and the prime mover regulation equation to achieve the active frequency control of the virtual isomorphic generator;

[0024] Adjust the reactive power and voltage amplitude through the Q-U droop characteristic to achieve the reactive voltage control of the virtual isomorphic generator;

[0025] Based on the results of the active frequency control and the reactive voltage control, integrate the virtual power angle and the output voltage amplitude to generate the three-phase voltage value;

[0026] Predict the inductor current and output voltage at the (k + 1)-th sampling period to complete the single-step prediction.

[0027] According to the above scheme, the rotor motion equation is as follows:

[0028]

[0029] where J is denoted as the moment of inertia, D is denoted as the damping coefficient, P m is denoted as the mechanical power; P e is denoted as the electromagnetic power; ω is denoted as the actual angular velocity; ω0 is denoted as the rated angular velocity; θ is denoted as the virtual power angle;

[0030] The electromagnetic power P e is obtained through the instantaneous power theory, and the formula is as follows:

[0031]

[0032] where P e is denoted as the electromagnetic power; v α and v β are denoted as the components of the NPC three-level inverter output voltage in the αβ coordinate system, which are used to calculate the electromagnetic power; i α and i β are denoted as the components of the current in the αβ coordinate system; 3 / 2 is the proportionality coefficient, which is used to convert the two-phase power into three-phase power.

[0033] According to the above scheme, the prime mover regulation equation is obtained through the P-f droop characteristic, and the formula is as follows:

[0034] P m = P ref + m(ω0 - ω);

[0035] Among them, P m represents mechanical power; P ref represents the reference value of active power input; m represents the active droop control coefficient; ω represents the actual angular velocity; ω0 represents the rated angular velocity.

[0036] According to the above scheme, the reactive power and voltage amplitude are adjusted through the Q-U droop characteristic to achieve the reactive voltage control of the virtual isomorphic generator, including:

[0037] The Q-U droop characteristic has the following formula:

[0038] E = U N + n(Q ref - Q);

[0039] Among them, E represents the output voltage amplitude; U N represents the rated voltage; n represents the reactive voltage droop coefficient; Q ref represents the reference value of reactive power input; Q represents the reactive power output of the virtual isomorphic generator;

[0040] The reactive power Q output by the virtual isomorphic generator is obtained through the instantaneous power theory;

[0041] In the two-phase stationary αβ coordinate system, the reactive power Q output by the virtual isomorphic generator through the instantaneous power theory has the following formula:

[0042]

[0043] Among them, Q represents the reactive power output of the virtual isomorphic generator; u α and u β represent the components of the output voltage of the NPC three-level inverter in the αβ coordinate system and are used to calculate the reactive power; i α and i β represent the components of the current in the αβ coordinate system.

[0044] According to the above scheme, the integration of the virtual power angle and the output voltage amplitude to generate the three-phase voltage value includes:

[0045] Based on the active frequency control, the virtual power angle θ is obtained, and based on the reactive voltage control, the output voltage amplitude E is obtained;

[0046] The virtual power angle θ and the output voltage amplitude E are integrated into the three-phase voltage value, and the formula is as follows:

[0047]

[0048] Among them, v * is expressed as a three-phase voltage value.

[0049] Through Park transformation, the three-phase voltage value v * is converted into voltage components in the dq coordinate system, and a virtual impedance link is introduced. In the dq coordinate system, the formula is as follows:

[0050]

[0051] Among them, v * d is expressed as the d-axis voltage reference value; v * q is expressed as the q-axis voltage reference value; v * a , v * b and v * c are respectively expressed as the reference values of the three-phase voltages; γ is expressed as the rotation angle, which is the rotation angle of the dq coordinate system relative to the three-phase stationary coordinate system;

[0052] The three-phase voltage value is adjusted through a virtual resistor and a virtual inductor. The formula is as follows:

[0053]

[0054] Among them, v d_ref and v q_ref are respectively expressed as the components of the output voltage reference value of the virtual isomorphic generator in the dq coordinate axes; v * d is expressed as the d-axis voltage reference value; R is expressed as the virtual resistor; L vir is expressed as the virtual inductor; i d and i q are expressed as the current components in the dq coordinate system;

[0055] The virtual impedance link helps to achieve power decoupling, suppress circulating current, and improve the stability and dynamic response ability of the system.

[0056] According to the above scheme, the predicted load-side current value at the (k + 1)-th sampling period is calculated by the following formula:

[0057]

[0058] Among them, i αβ (k + 1) is expressed as the load-side current at the (k + 1)-th sampling period; i fαβ (k + 1) is expressed as the inductor current at the (k + 1)-th sampling period; C is expressed as the filter capacitor; T sis denoted as the sampling period; v αβ (k + 1) is denoted as the load - side voltage at the (k + 1)-th sampling period.

[0059] According to the above - mentioned scheme, using the mathematical model, further predict the inductor current and output voltage at the (k + 2)-th sampling period, including:

[0060] Based on the load - side current value at the (k + 1)-th sampling period and the single - step prediction, obtain the optimal prediction value at the current sampling period, and the optimal prediction value at the current sampling period includes the optimal load - side current prediction value, inductor current prediction value, and output voltage prediction value at the (k + 1)-th sampling period;

[0061] Use the optimal prediction value at the current sampling period to predict the inductor current prediction value at the (k + 2)-th sampling period, and the formula is as follows:

[0062]

[0063] where, i fαβ (k + 2) is denoted as the predicted inductor current value at the (k + 2)-th sampling period; T s is denoted as the sampling period; L fil is denoted as the inductance value of the filter inductor; U αβ (k + 2) is denoted as the predicted voltage vector at the (k + 2)-th sampling period; R is denoted as the line resistance; i fαβ (k + 1) is denoted as the inductor current at the (k + 1)-th sampling period; v αβ (k + 1) is denoted as the load - side voltage at the (k + 1)-th sampling period;

[0064] Based on the predicted inductor current value, predict the output voltage prediction value at the (k + 2)-th sampling period, and the formula is as follows:

[0065]

[0066] where, v αβ (k + 2) is denoted as the predicted output voltage value at the (k + 2)-th sampling period; C is denoted as the filter capacitor; i fαβ (k + 1) is denoted as the inductor current at the (k + 1)-th sampling period; i αβ (k + 1) is denoted as the load - side current at the (k + 1)-th sampling period; v αβ (k + 1) is denoted as the load - side voltage at the (k + 1)-th sampling period.

[0067] According to the above - mentioned scheme, compare the output voltage at the (k + 2)-th sampling period with the reference voltage to obtain the value function, including:

[0068] Substitute the reference voltage output by the virtual isomorphic generator into the value function. The formula is as follows:

[0069] g = |v αβ_ref - v αβ (k + 2)|;

[0070] where g represents the value function, and v αβ-ref represents the reference voltage; v αβ (k + 2) represents the predicted value of the output voltage at the (k + 2)-th sampling period.

[0071] According to the above scheme, select the optimal voltage vector that minimizes the value function to complete the two-step prediction, including:

[0072] Traverse all 27 voltage vectors, calculate the value function of each voltage vector; select the voltage vector that minimizes the value function as the optimal vector; input the optimal voltage vector signal into the gate of the NPC three-level inverter to control the output of the desired voltage and current, and complete the two-step voltage model predictive control.

[0073] Compared with the prior art, the beneficial effects of the present invention are:

[0074] 1. The present invention effectively addresses the problems of grid frequency and voltage fluctuations through active frequency control and reactive voltage control, makes up for the deficiencies of traditional droop control in inertia support, and significantly improves the stability and anti-interference ability of the power grid;

[0075] 2. The present invention combines model predictive control technology with virtual isomorphic generator technology, overcomes the limitations of traditional PI control in parameter tuning complexity and dynamic response speed, and realizes more efficient and flexible system control;

[0076] 3. Through two-step predictive control, the present invention improves the control accuracy for the problem of delay in selecting the optimal switching sequence in high-sampling-frequency systems, and ensures the fast response and stable operation of the system under complex working conditions;

[0077] 4. Through two-step predictive control, the present invention effectively reduces the total harmonic distortion rate of the output current. Compared with the traditional single-step predictive control algorithm, the current quality is significantly improved, and the power output performance of the system is further optimized. Brief Description of the Drawings

[0078] Figure 1 It is a schematic diagram of the topological structure of the NPC three-level inverter for the two-step voltage model predictive control method of a three-level grid-connected inverter according to the present invention;

[0079] Figure 2Schematic diagram of 27 switching state combinations of a two-step voltage model predictive control method for a three-level grid-connected inverter of the present invention;

[0080] Figure 3 Flow chart of the two-step voltage model predictive control steps of a two-step voltage model predictive control method for a three-level grid-connected inverter of the present invention;

[0081] Figure 4 Waveform diagram of load-side voltage and current of traditional single-step prediction;

[0082] Figure 5 Waveform diagram of load-side voltage and current of two-step prediction of a two-step voltage model predictive control method for a three-level grid-connected inverter of the present invention;

[0083] Figure 6 Schematic diagram of output voltage after 0.5 s of startup of traditional single-step prediction;

[0084] Figure 7 Schematic diagram of output voltage after 0.5 s of startup of two-step prediction of a two-step voltage model predictive control method for a three-level grid-connected inverter of the present invention;

[0085] Figure 8 Schematic diagram of output voltage when the power grid frequency fluctuates in traditional single-step prediction;

[0086] Figure 9 Schematic diagram of output voltage when the power grid frequency fluctuates in two-step prediction of a two-step voltage model predictive control method for a three-level grid-connected inverter of the present invention;

[0087] Figure 10 Schematic diagram of output voltage after returning to stability in traditional single-step prediction;

[0088] Figure 11 Schematic diagram of output voltage after the power grid returns to stability in two-step prediction of a two-step voltage model predictive control method for a three-level grid-connected inverter of the present invention;

[0089] Figure 12 Schematic diagram of active and reactive power waveforms output by a virtual isomorphic generator of a two-step voltage model predictive control method for a three-level grid-connected inverter of the present invention. Detailed implementation manners

[0090] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0091] Embodiment: AsFigures 1 - 12 As shown in the figure, the present invention provides a technical solution, a two-step voltage model predictive control method for a three-level grid-connected inverter, and the method includes the steps of:

[0092] S1. Build the topological structure of the NPC three-level inverter;

[0093] Specifically, the topological structure of the NPC three-level inverter is as Figure 1 shown; the topological structure of the NPC three-level inverter includes a DC-side voltage U dc , a DC-side capacitor C1, a DC-side capacitor C2, a filter capacitor C, a filter inductor L, and a line resistance R; the NPC three-level inverter includes three bridge arms, each bridge arm is composed of four switching tubes and two clamping diodes, and outputs three levels;

[0094] Further, use MATLAB / Simulink to establish a simulation model, and the main parameters of the simulation experiment are shown in Table 1:

[0095] Table 1: Main parameters of the simulation experiment

[0096] Symbol Parameter Value Udc DC voltage 1000V L Filter inductor 3mH R Line resistance 0.05Ω C Filter capacitor 35μF C1, C2 DC - side capacitor 5600μF Ts Sampling period 50μs J Virtual inertia coefficient 3.36 D Virtual damping coefficient 103.13

[0097] S2. Establish a mathematical model of the NPC three-level inverter based on the αβ coordinate system, define the switching state combinations of the NPC three-level inverter, and generate voltage vectors;

[0098] Specifically, the switching state is defined as S i ={1, 0, -1}; where S i represents the switching state of the bridge arm, i ∈ {a, b, c}, where i represents the bridge arm, and a, b, and c respectively represent the three bridge arms of the NPC three-level inverter; when S i =1, the bridge arm outputs a high level, indicating a positive voltage output, when S i =0, the bridge arm outputs a zero level, indicating a medium voltage output, when S i =-1, the bridge arm outputs a low level, indicating a negative voltage output, and the switching state of each bridge arm is defined as:

[0099]

[0100] where S i represents the switching state of the bridge arm; S i1 , S i2 , S i3 , and S i4 respectively represent the four switches on the bridge arm;

[0101] The three switching state combinations of the three arm bridges of the NPC three-level inverter are combined to define a total of 27 switching state combinations. As Figure 2 shown, each switching state combination corresponds to a specific voltage vector for model predictive control. The formula is as follows:

[0102]

[0103] Among them, U α represents the voltage component on the α-axis; U β represents the voltage component on the β-axis; U dc represents the DC-side voltage; S a , S b and S c represent the switching states of the three-phase bridge arms.

[0104] S3. Discretize the mathematical model of the LC output filter using the first-order Euler equation to predict the inductor current and output voltage at the (k + 1)-th sampling period and complete a single-step prediction;

[0105] Specifically, the single-step prediction includes: adjusting the active power and frequency through the rotor motion equation and the prime mover regulation equation to achieve the active frequency control of the virtual synchronous generator; adjusting the reactive power and voltage amplitude through the Q-U droop characteristic to achieve the reactive voltage control of the virtual synchronous generator; integrating the virtual power angle and output voltage amplitude based on the results of the active frequency control and the reactive voltage control to generate three-phase voltage values; predicting the inductor current and output voltage at the (k + 1)-th sampling period to complete a single-step prediction.

[0106] Furthermore, the rotor motion equation has the following formula:

[0107]

[0108] Among them, J represents the moment of inertia, D represents the damping coefficient, P m represents the mechanical power; P e represents the electromagnetic power; ω represents the actual angular velocity; ω0 represents the rated angular velocity; θ represents the virtual power angle;

[0109] The electromagnetic power P e is obtained through the instantaneous power theory. The formula is as follows:

[0110]

[0111] Among them, P e represents the electromagnetic power; v α and v β represent the components of the NPC three-level inverter output voltage in the αβ coordinate system and are used to calculate the electromagnetic power; i αand i β are expressed as the components of the current in the αβ coordinate system; 3 / 2 is the proportionality coefficient used to convert two-phase power to three-phase power.

[0112] Furthermore, the prime mover regulation equation is obtained through the P-f droop characteristic, and the formula is as follows:

[0113] P m = P ref + m(ω0 - ω);

[0114] where, P m is expressed as mechanical power; P ref is expressed as the reference value of the active power input; m is expressed as the active droop control coefficient; ω is expressed as the actual angular velocity; ω0 is expressed as the rated angular velocity.

[0115] Furthermore, the Q-U droop characteristic, the formula is as follows:

[0116] E = U N + n(Q ref - Q);

[0117] where, E is expressed as the amplitude of the output voltage; U N is expressed as the rated voltage; n is expressed as the reactive voltage droop coefficient; Q ref is expressed as the reference value of the reactive power input; Q is expressed as the reactive power output of the virtual isomorphic generator;

[0118] The reactive power Q output by the virtual isomorphic generator is obtained through the instantaneous power theory; in the two-phase stationary αβ coordinate system, the reactive power Q output by the virtual isomorphic generator through the instantaneous power theory, the formula is as follows:

[0119]

[0120] where, Q is expressed as the reactive power output of the virtual isomorphic generator; u α and u β are expressed as the components of the output voltage of the NPC three-level inverter in the αβ coordinate system, used to calculate the reactive power; i α and i β are expressed as the components of the current in the αβ coordinate system.

[0121] Furthermore, based on the active frequency control, the virtual power angle θ is obtained, and based on the reactive voltage control, the output voltage amplitude E is obtained; the virtual power angle θ and the output voltage amplitude E are integrated into a three-phase voltage value, the formula is as follows:

[0122]

[0123] where, v *Expressed as three-phase voltage values.

[0124] Furthermore, through Park transformation, the three-phase voltage values v * are converted into voltage components in the dq coordinate system, and a virtual impedance link is introduced. In the dq coordinate system, the formula is as follows:

[0125]

[0126] where v * d is expressed as the d-axis voltage reference value; v * q is expressed as the q-axis voltage reference value; v * a , v * b and v * c are respectively expressed as the reference values of the three-phase voltages; γ is expressed as the rotation angle, which is the rotation angle of the dq coordinate system relative to the three-phase stationary coordinate system;

[0127] The three-phase voltage values are adjusted through a virtual resistor and a virtual inductor. The formula is as follows:

[0128]

[0129] where v d_ref and v q_ref are respectively expressed as the components of the output voltage reference value of the virtual isomorphic generator in the dq coordinate axes; v * d is expressed as the d-axis voltage reference value; R is expressed as the virtual resistor; L vir is expressed as the virtual inductor; i d and i q are expressed as the current components in the dq coordinate system;

[0130] The virtual impedance link helps to achieve power decoupling, suppress circulating current, and improve the stability and dynamic response ability of the system.

[0131] S4. Predict the load-side current value at the (k + 1)-th sampling period, where k represents the index of the current sampling period; and in combination with the single-step prediction, use the mathematical model to further predict the inductor current and output voltage at the (k + 2)-th sampling period;

[0132] Specifically, the formula for predicting the load-side current value at the (k + 1)-th sampling period is as follows:

[0133]

[0134] where i αβ(k + 1) represents the load - side current at the (k + 1)-th sampling period; i fαβ (k + 1) represents the inductor current at the (k + 1)-th sampling period; C represents the filter capacitor; T s represents the sampling period; v αβ (k + 1) represents the load - side voltage at the (k + 1)-th sampling period;

[0135] Furthermore, based on the load - side current value at the (k + 1)-th sampling period and combining with the single - step prediction, the optimal prediction value of the current sampling period is obtained. The optimal prediction value of the current sampling period includes the optimal load - side current prediction value, inductor current prediction value, and output voltage prediction value at the (k + 1)-th sampling period; Using the optimal prediction value of the current sampling period, the inductor current prediction value at the (k + 2)-th sampling period is predicted. The formula is as follows:

[0136]

[0137] where, i fαβ (k + 2) represents the predicted inductor current value at the (k + 2)-th sampling period; T s represents the sampling period; L fil represents the inductance value of the filter inductor; U αβ (k + 2) represents the predicted voltage vector at the (k + 2)-th sampling period; R represents the line resistance; i fαβ (k + 1) represents the inductor current at the (k + 1)-th sampling period; v αβ (k + 1) represents the load - side voltage at the (k + 1)-th sampling period;

[0138] Based on the inductor current prediction value, the output voltage prediction value at the (k + 2)-th sampling period is predicted. The formula is as follows:

[0139]

[0140] where, v αβ (k + 2) represents the output voltage prediction value at the (k + 2)-th sampling period; C represents the filter capacitor; i fαβ (k + 1) represents the inductor current at the (k + 1)-th sampling period; i αβ (k + 1) represents the load - side current at the (k + 1)-th sampling period; v αβ (k + 1) represents the load - side voltage at the (k + 1)-th sampling period.

[0141] S5. Compare the output voltage at the (k + 2)-th sampling period with the reference voltage to obtain the value function, and select the optimal voltage vector that minimizes the value function to complete the two - step prediction;

[0142] Specifically, substitute the reference voltage output by the virtual isomorphic generator into the value function, and the formula is as follows:

[0143] g = |v αβ_ref - v αβ (k + 2)|;

[0144] Among them, g represents the value function, and v αβ-ref represents the reference voltage; v αβ (k + 2) represents the predicted value of the output voltage at the (k + 2)-th sampling period;

[0145] Traverse all 27 voltage vectors, calculate the value function of each voltage vector; select the voltage vector that minimizes the value function as the optimal vector.

[0146] S6. Input the optimal voltage vector signal into the gates of the NPC three-level inverter to control the output of the desired voltage and current;

[0147] Specifically, input the optimal voltage vector signal into the gates of the NPC three-level inverter to control the output of the desired voltage and current, and complete the two-step voltage model predictive control.

[0148] S7. Complete the two-step voltage model predictive control of the current sampling period, enter the next sampling period, and repeat the above process.

[0149] The present invention provides another technical solution. As Figure 3 shown, the process flow of the two-step voltage model predictive control steps:

[0150] Specifically, initialize the parameters and the model, and first start the single-step prediction;

[0151] Specifically, determine the three bridge arms of the NPC three-level inverter. Each bridge arm consists of four switching tubes and two clamping diodes and can output three levels; the three bridge arms are combined to generate 27 switching states, and each state corresponds to a specific voltage vector; use the first-order Euler equation to discretize the mathematical model of the LC output filter. According to the inductor current and voltage vector at the k-th moment, predict the voltage vector at the (k + 1)-th moment; predict the output voltage at the (k + 1)-th moment, and substitute it into the value function for calculation to obtain the value g; select the optimal vector that minimizes the value function from the 27 voltage vectors, and output the current of the NPC three-level inverter measured under the predicted optimal vector at the (k + 1)-th moment and the load-side voltage.

[0152] Further, after the single-step prediction is completed, based on the single-step prediction result, perform the two-step prediction;

[0153] Specifically, based on the predicted value of the inductor current and the predicted value of the output voltage at the (k + 1)-th moment, predict the voltage vector at the (k + 2)-th moment; predict the output voltage at the (k + 2)-th moment, and substitute the output voltage into the value function for calculation to obtain the value g; select the optimal vector that minimizes the value function from 27 voltage vectors, and input the optimal vector signal to the gate of the NPC three-level inverter to control the NPC three-level inverter to output the desired voltage and current; end the two-step prediction.

[0154] After completing the two-step prediction, end the control of the current sampling period, enter the next sampling period, and repeat the above process.

[0155] The present invention provides another technical solution, which measures the load-side voltage and current of the traditional single-step prediction and the two-step prediction respectively, and conducts a comparative analysis.

[0156] Build a simulation model through MATLAB / Simulink, set the grid voltage to 380V, the initial frequency to 50Hz, and the frequency to step down by 0.2Hz at 1s and recover to 50Hz at 2s.

[0157] Conduct simulations through MATLAB / Simulink, and run the traditional single-step prediction and the two-step prediction method of the present invention respectively. Among them, the load-side voltage and current of the traditional single-step prediction are as Figure 4 shown, and the load-side voltage and current of the two-step prediction of the present invention are as Figure 5 shown. The simulation results show that both the traditional single-step prediction and the two-step prediction method of the present invention can effectively cope with the frequency fluctuation problem of the power grid, but the two-step prediction of the present invention can reach the desired stable load-side output voltage and current faster.

[0158] The present invention provides another technical solution, which measures the thd values of the traditional single-step prediction and the two-step prediction in different working time periods respectively, and conducts a comparative analysis, as Figures 6 - 10 shown;

[0159] Specifically, as Figure 6 and Figure 7 shown, within 25 cycles after the system starts for 0.5s, the thd value of the output voltage of the traditional single-step prediction is 0.31%, while the thd value of the output voltage of the two-step prediction is 0.08%. The simulation results show that the two-step prediction can reach the desired stable load-side output voltage faster. As Figure 8 and Figure 9 shown, when the power grid is subject to a frequency fluctuation of -0.2Hz, the thd value of the output voltage of the traditional single-step prediction is 2.26%, while the thd value of the output voltage of the two-step prediction is 2.18%; when there is a frequency fluctuation, the control of the virtual synchronous generator system plays an effective role, making the output voltage remain stable to a certain extent during the fluctuation, and among them, the two-step prediction has a better effect. As Figure 10and Figure 11 As shown, after the system runs stably, the THD value of the output voltage of the traditional single-step prediction is 0.03%, and the THD value of the output voltage of the two-step prediction is 0.02%. Compared with the traditional single-step prediction, the harmonic distortion rate of the output voltage of the two-step prediction is reduced by 0.01% compared with the traditional single-step prediction.

[0160] The present invention provides another technical solution. As Figure 12 shown, after simulation operation, the active power reaches the reference active power value stably within 0.5 s after startup, and responds to the grid frequency change at 1 s. The electromagnetic power P e gradually rises from 170 kW to about 210 kW in about 0.5 s; the reactive power Q fluctuates slightly around 0 Var. According to the above results, it shows that the two-step voltage model predictive control method performs excellently in both dynamic response and steady-state performance, and can effectively improve the control performance of the three-level grid-connected inverter. This is only for illustrative purposes and not for limitation.

[0161] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claims involved.

Claims

1. A two-step voltage model predictive control method for a three-level grid-connected inverter, characterized in that: The method includes: Constructing the topological structure of the NPC three-level inverter; Establishing the mathematical model of the NPC three-level inverter based on the αβ coordinate system, defining the switching state combinations of the NPC three-level inverter, and generating voltage vectors; Discretizing the mathematical model of the LC output filter using the first-order Euler equation, predicting the inductor current and output voltage at the (k + 1)-th sampling period, and completing single-step prediction, where k represents the index of the current sampling period; Predicting the load-side current value at the (k + 1)-th sampling period, and combining the single-step prediction, using the mathematical model to further predict the inductor current and output voltage at the (k + 2)-th sampling period; Comparing the output voltage at the (k + 2)-th sampling period with the reference voltage to obtain the cost function, selecting the optimal voltage vector that minimizes the cost function, and completing two-step prediction; Inputting the optimal voltage vector signal into the gates of the NPC three-level inverter to control the output of the desired voltage and current; Completing the two-step voltage model predictive control for the current sampling period, entering the next sampling period, and repeating the above process.

2. The two-step voltage model predictive control method for a three-level grid-connected inverter according to claim 1, wherein: The defining the switching state combinations of the NPC three-level inverter and generating voltage vectors includes: The switch state is defined as S i ={1, 0, -1}; where S i represents the switch state of the arm, i ∈ {a, b, c}, where i represents the arm, and a, b, and c respectively represent the three arms of the NPC three-level inverter; When S i = 1, the bridge arm outputs a high level, indicating a positive voltage output. When S i = 0, the bridge arm outputs a zero level, indicating a neutral voltage output. When S i = -1, the bridge arm outputs a low level, indicating a negative voltage output; The three switching states of the three arm bridges of the NPC three-level inverter are combined, and a total of 27 switching state combinations are defined. Each switching state combination corresponds to a specific voltage vector, and the formula is as follows: Among them, U α represents the α-axis voltage component; U β represents the β-axis voltage component; U dc represents the DC-side voltage; S a , S b and S c represent the switching states of the three-phase bridge arms.

3. A two-step voltage model predictive control method for a three-level grid-connected inverter according to claim 1, characterized in that: The single-step prediction includes: Adjusting the active power and frequency through the rotor motion equation and the prime mover regulation equation to achieve the active frequency control of the virtual synchronous generator; Adjusting the reactive power and voltage amplitude through the Q-U droop characteristic to achieve the reactive voltage control of the virtual synchronous generator; Based on the results of the active frequency control and the reactive voltage control, integrating the virtual power angle and the output voltage amplitude to generate three-phase voltage values; Predicting the inductor current and output voltage at the (k + 1)-th sampling period to complete single-step prediction.

4. A two-step voltage model predictive control method for a three-level grid-connected inverter according to claim 3, characterized in that: The rotor motion equation and the prime mover regulation equation include: The rotor motion equation, and the formula is as follows: Among them, J represents the moment of inertia, D represents the damping coefficient, and P m represents the mechanical power; P e represents the electromagnetic power; ω represents the actual angular velocity; ω0 represents the rated angular velocity; θ represents the virtual power angle; The electromagnetic power P e is obtained through the instantaneous power theory, and the formula is as follows: Among them, P e represents the electromagnetic power; v α and v β represent the components of the output voltage of the NPC three-level inverter in the αβ coordinate system, which are used to calculate the electromagnetic power; i α and i β represent the components of the current in the αβ coordinate system; 3 / 2 is a proportionality coefficient, which is used to convert the two-phase power into three-phase power; The prime mover regulation equation is obtained through the P-f droop characteristic, and the formula is as follows: P m = P ref + m(ω0 - ω); Among them, P m represents mechanical power; P ref represents the reference value of active power input; m represents the active droop control coefficient; ω represents the actual angular velocity; ω0 represents the rated angular velocity.

5. A two-step voltage model predictive control method for a three-level grid-connected inverter according to claim 3, characterized in that: The through the Q-U droop characteristic includes: The Q-U droop characteristic, and the formula is as follows: E = U N + n(Q ref - Q); Among them, E represents the output voltage amplitude; U N represents the rated voltage; n represents the reactive voltage droop coefficient; Q ref represents the reactive power input reference value; Q represents the reactive power output of the virtual isomorphic generator; The reactive power Q output by the virtual synchronous generator is obtained through the instantaneous power theory; Under the two-phase stationary αβ coordinate system, the reactive power Q output by the virtual synchronous generator is obtained through the instantaneous power theory, and the formula is as follows: Among them, Q represents the reactive power output by the virtual isomorphic generator; u α and u β represent the components of the output voltage of the NPC three-level inverter in the αβ coordinate system, which are used to calculate the reactive power; i α and i β represent the components of the current in the αβ coordinate system.

6. A two-step voltage model predictive control method for a three-level grid-connected inverter according to claim 3, characterized in that: The integrating the virtual power angle and the output voltage amplitude to generate three-phase voltage values includes: Based on the active frequency control, obtaining the virtual power angle θ, and based on the reactive voltage control, obtaining the output voltage amplitude E; Integrating the virtual power angle θ and the output voltage amplitude E into three-phase voltage values, and the formula is as follows: Among them, v * represents three-phase voltage values; Through Park transformation, the three-phase voltage value v * is converted into voltage components in the dq coordinate system, and a virtual impedance link is introduced. In the dq coordinate system, the three-phase voltage value is adjusted by a virtual resistor and a virtual inductor. The formula is as follows: Among them, v d_ref and v q_ref respectively represent the components of the output voltage reference value of the virtual isomorphic generator in the dq coordinate axes; v * d represents the d-axis voltage reference value; R represents the virtual resistance; L vir represents the virtual inductance; i d and i q represent the current components in the dq coordinate system.

7. According to the two-step voltage model predictive control method for a three-level grid-connected inverter described in claim 1, wherein: The formula for predicting the load-side current value at the (k + 1)-th sampling period is as follows: where, i αβ (k + 1) represents the load - side current at the (k + 1)-th sampling period; i fαβ (k + 1) represents the inductor current at the (k + 1)-th sampling period; C represents the filter capacitor; T s represents the sampling period; v αβ (k + 1) represents the load - side voltage at the (k + 1)-th sampling period.

8. A two-step voltage model predictive control method for a three-level grid-connected inverter according to claim 1, characterized in that: The using the mathematical model to further predict the inductor current and output voltage at the (k + 2)-th sampling period includes: Based on the load - side current value in the (k + 1)-th sampling period and the single - step prediction, obtain the optimal prediction value in the current sampling period. The optimal prediction value in the current sampling period includes the optimal load - side current prediction value, inductor current prediction value, and output voltage prediction value for predicting the (k + 1)-th sampling period; Use the optimal prediction value in the current sampling period to predict the inductor current prediction value in the (k + 2)-th sampling period. The formula is as follows: where i fαβ (k + 2) represents the predicted value of the inductor current at the (k + 2)-th sampling period; T s represents the sampling period; L fil represents the inductance value of the filter inductor; U αβ (k + 2) represents the predicted voltage vector at the (k + 2)-th sampling period; R represents the line resistance; i fαβ (k + 1) represents the inductor current at the (k + 1)-th sampling period; v αβ (k + 1) represents the load-side voltage at the (k + 1)-th sampling period; Based on the inductor current prediction value, predict the output voltage prediction value in the (k + 2)-th sampling period. The formula is as follows: Among them, v αβ (k + 2) represents the predicted value of the output voltage at the (k + 2)-th sampling period; C represents the filter capacitor; i fαβ (k + 1) represents the inductor current at the (k + 1)-th sampling period; i αβ (k + 1) represents the load-side current at the (k + 1)-th sampling period; v αβ (k + 1) represents the load-side voltage at the (k + 1)-th sampling period.

9. A two-step voltage model predictive control method for a three-level grid-connected inverter according to claim 1, characterized in that: Compare the output voltage in the (k + 2)-th sampling period with the reference voltage to obtain the cost function, including: Substitute the reference voltage output by the virtual isomorphic generator into the cost function. The formula is as follows: g = |v αβ_ref - v αβ (k + 2)|; Among them, g represents the value function, and v αβ-ref represents the reference voltage; v αβ (k + 2) represents the predicted value of the output voltage at the (k + 2)-th sampling period.

10. A two-step voltage model predictive control method for a three-level grid-connected inverter according to claim 1, characterized in that: Select the optimal voltage vector that minimizes the cost function to complete the two - step prediction, including: Traverse all 27 voltage vectors, calculate the cost function of each voltage vector; select the voltage vector that minimizes the cost function as the optimal vector; input the optimal voltage vector signal into the gate of the NPC three - level inverter to control the output of the desired voltage and current, and complete the two - step voltage model predictive control.

Images