Five-level inverter model prediction control method capable of enhancing bus utilization rate

Abstract

A five-level inverter model prediction control method capable of enhancing the bus utilization rate comprises the steps that at the sampling k moment, the output voltage, the output current and the load voltage of an inverter are measured, and physical quantities under a dq coordinate system are obtained through the coordinate transformation principle; constructing a prediction model, and calculating the predicted current of the inverter under the dq coordinate system; according to the photovoltaic MPPT algorithm, the photovoltaic voltage vpv and the photovoltaic current ipv are used as input quantities, the deviation value of the output of the DC voltage reference # imgabs0 # photovoltaic MPPT algorithm and the actual DC side voltage is used as the input of a PI controller, and therefore a d-axis current reference value is obtained; on-line calculation and evaluation of a cost function J are carried out in combination with six switching states of the inverter, the switching state enabling the cost function value to be minimum is found, and the cost function J acts on the inverter in the next switching period. According to the invention, fast dynamic response can be realized, and robustness to power grid fluctuation and load fluctuation is stronger; the structure is simpler, the bus voltage utilization rate is maximized, and the voltage stress between the direct current side capacitor and the switching device is lower.

Description

Technical Field

[0001] The present invention relates to the field of power electronics technology, and particularly to a model predictive control method for a five-level inverter with enhanced bus utilization rate. Background Art

[0002] Compared with two-level inverters, five-level inverters can provide better power quality, higher efficiency, and significantly reduce electromagnetic interference. Therefore, they are widely used in high-voltage power transmission systems and renewable energy systems, etc.

[0003] In the industrial field, there are various five-level inverter topologies widely used, including Neutral Point Clamped (NPC), Flying Capacitor (FC), and Cascade H-bridge (CHB), etc. However, in these topologies, the number of components used is relatively large, and the utilization rate of the DC bus voltage is relatively low. Due to the addition of more clamping diodes and capacitors, the losses of NPC-type five-level inverters will also increase, and at the same time, a complex control scheme needs to be designed to achieve the balance of the DC link capacitor voltages. For CHB-type five-level inverters, the use of multiple phase-shifting transformers or isolated DC power supplies results in a larger volume and higher cost. FC-type five-level inverters require a complex control strategy to make the voltage levels between capacitors consistent.

[0004] PR (Proportional-Resonant) control is a method used to regulate the output voltage and frequency of an inverter. PR control combines proportional control and resonant control, where the proportional part reduces the difference between the actual output and the reference value; the resonant part introduces harmonics consistent with the input signal frequency to compensate for the error, sets a high gain at the target frequency, ensures no steady-state error control for the signal at this frequency, and effectively reduces the error of frequency matching. However, the dynamic response of PR control is limited by the system bandwidth. Under the conditions of load mutation or grid fluctuation, there may be a certain overshoot, which affects the grid quality. In addition, a single PR controller can only perform zero-static-error tracking for the set resonant frequency, and adding multiple resonant controllers to compensate for other harmonic frequencies will increase the complexity and computational amount of the system. MPC has obvious advantages in inverter control, such as excellent anti-interference ability, fast dynamic response, and flexible ability to handle multi-objective optimization.

[0005] The FCS-MPC first establishes a discrete mathematical model of the system, predicts the future state at a fixed step size, calculates the cost function of the controlled object and the reference object under finite switching states, and then determines the optimal switching state applied to the inverter in the next control period through online optimization. MPC performs optimization calculations on all switching states in each sampling period and predicts the current changes at multiple future moments in advance. Therefore, under conditions such as load mutation and grid disturbance, it can achieve a faster dynamic response. In addition, MPC fully considers the discrete characteristics and system constraints of the inverter, and can directly add a harmonic suppression term to the optimization objective to dynamically compensate for the harmonics generated by non-linear loads, effectively suppressing multiple harmonic components.

[0006] Disadvantages of the prior art: The complexity of the PR control strategy is relatively high, requiring precise parameter adjustment and may cause a large overshoot under rapidly changing grid conditions; the PR control introduces a resonant term to provide zero-steady-state error tracking at a specific frequency, but when the grid frequency deviates, the control performance of the system will decline and the steady-state error may increase. At the same time, the traditional five-level inverter topology uses a large number of components, has a low utilization rate of the DC bus voltage, and the increased switching devices will cause higher switching losses. Summary of the Invention

[0007] The object of the present invention is achieved through the following technical solutions.

[0008] Specifically, the present invention provides a five-level inverter model predictive control method with enhanced bus utilization rate, including:

[0009] At sampling time k, measure the output voltage, output current, and load voltage of the five-level inverter, and obtain the physical quantities in the dq coordinate system by using the coordinate transformation principle;

[0010] Construct a prediction model and calculate the predicted current of the inverter in the dq coordinate system;

[0011] Through the photovoltaic MPPT algorithm, take the photovoltaic voltage v pv and photovoltaic current i pv as input quantities, and the output is the DC voltage reference Take the deviation value between the DC voltage reference and the actual DC side voltage as the input of the PI controller to obtain the d-axis current reference value;

[0012] Combine the 6 switching states of the inverter to perform online calculation and evaluation of the cost function J, find the switching state that minimizes the cost function value, and apply it to the inverter in the next switching period.

[0013] Further, after calculating the predicted current, it further includes:

[0014] Calculate the difference in bus capacitor voltages to balance the neutral point voltage.

[0015] Furthermore, the topology of the five-level inverter includes six power switches (S1 - S6), two power diodes D1 and D2, and a floating capacitor C. F Among them, the switches (S1 - S4) form two half - bridges, the two power diodes are used to block reverse voltages, and the floating capacitor is placed between the two half - bridges.

[0016] Furthermore, the inverter is powered by a single DC power supply V. dc and two capacitors C1 and C2 are connected in series on the DC side; where V. dc is the DC - side voltage; V. p and V. n are the voltages of the upper capacitor and the lower capacitor respectively; i. c1 , i. c2 and i. o are the currents flowing through the upper capacitor, the lower capacitor, and the neutral - point current respectively; v. a , v. b and v. c are the output voltages of the inverter in phases a, b, and c respectively; L. f and R. f constitute the output LC filter; i. a , i. b and i. c are the output currents of the inverter in phases a, b, and c respectively; v. ga , v. gb and v. gc are the load voltages.

[0017] Furthermore, the drive signals of power switches S1 and S2 are always complementary, the drive signals of S3 and S4 are always complementary, and the drive signals of S5 and S6 are always the same.

[0018] Furthermore, obtaining the physical quantities in the dq coordinate system by using the coordinate transformation principle includes:

[0019] According to Kirchhoff's law, the state equations of the inverter in the dq rotating coordinate system are expressed as:

[0020]

[0021] Among them, i. d and i. q respectively represent the components of the inverter output current on the d - axis and q - axis; ω represents the angular frequency; R and L respectively represent the resistive load and inductive load; v. gd and v. gq respectively represent the components of the load voltage on the d - axis and q - axis; v. d and v. qrespectively represent the components of the inverter output voltage on the dq axes.

[0022] Further, the constructing the prediction model and calculating the predicted current of the inverter in the dq coordinate system includes:

[0023] At the k + 1 moment, the predicted currents of the dq axes of the inverter are expressed as:

[0024]

[0025] where, i d (k + 1) and i q (k + 1) respectively represent the predicted currents of the dq axes of the inverter at the k + 1 moment; T s represents the sampling period; i d (k) and i q (k) respectively represent the predicted currents of the dq axes of the inverter at the k moment; v d (k) and v q (k) respectively represent the output voltages of the dq axes of the inverter at the k moment; v gd (k) and v gq (k) respectively represent the load voltages of the dq axes of the inverter at the k moment.

[0026] Further, the voltages of the upper capacitor and the lower capacitor at the k + 1 moment are expressed as:

[0027]

[0028] where, V p (k + 1) and V n (k + 1) respectively represent the voltages of the upper and lower capacitors on the bus at the k + 1 moment; V p (k) and V n (k) respectively represent the voltages of the upper and lower capacitors on the bus at the k moment; i c1 (k) and i c2 (k) respectively represent the currents of the upper and lower capacitors on the bus at the k moment; C1 and C2 are respectively the capacitor on the bus and the capacitor under the bus.

[0029] Further, the cost function J is:

[0030]

[0031] where, and respectively represent the reference currents of the dq axes of the inverter; ΔV NP represents the difference between the voltages of the upper and lower capacitors on the bus.

[0032] Further, the difference between the voltages of the upper and lower capacitors on the bus is expressed as:

[0033]

[0034] The advantages of the present invention are as follows:

[0035] 1. In the control of grid-connected inverters, FCS-MPC performs optimized calculations on a limited number of switching states within each control cycle, predicts the current changes at multiple moments in advance, and can achieve a relatively fast dynamic response, especially suitable for systems with transient load disturbances.

[0036] 2. Compared with traditional PR control, FCS-MPC is easy to add system constraints, such as current constraints, voltage constraints, etc., so it has stronger robustness to grid fluctuations and load fluctuations.

[0037] 3. Compared with the traditional five-level topology, the topology proposed by the present invention requires fewer switching devices, has a simpler structure, maximizes the utilization rate of the bus voltage, has lower voltage stress between the DC-side capacitor and the switching devices, and is suitable for photovoltaic inverter applications. Description of the Drawings

[0038] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0039] Figure 1 A five-level inverter circuit diagram according to an embodiment of the present invention is shown.

[0040] Figure 2 A block diagram of a five-level inverter model predictive control strategy according to an embodiment of the present invention is shown.

[0041] Figure 3 A steady-state experimental waveform diagram under a resistive-inductive load according to an embodiment of the present invention is shown. Detailed Embodiments

[0042] Hereinafter, the exemplary embodiments of the present disclosure will be described in more detail with reference to the drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0043] Term Explanation:

[0044] Multilevel Inverters (MLIs): Compared with traditional two-level inverters, multilevel inverters have more significant advantages. Multilevel inverters can output voltages of more levels, can generate output voltages closer to sine waves, effectively reducing harmonic distortion and electromagnetic interference. In addition, multilevel inverters can output higher voltages through combinations of switching devices with lower voltages, improving the voltage and power capacity of the system. Therefore, multilevel inverters are widely used in high-voltage power transmission systems and renewable energy systems.

[0045] Grid-connected inverter: A grid-connected inverter is a device that converts DC electrical energy into AC electrical energy and connects to the power grid. Grid-connected inverters are widely used in photovoltaic power generation systems, wind power generation systems, and energy storage systems, converting the DC electricity generated by renewable energy into the AC electricity required by the power grid, showing a high energy utilization rate.

[0046] Finite Control Set Model Predictive Control (FCS-MPC): FCS-MPC first establishes a discrete mathematical model of the system, then constructs a suitable cost function according to the control instructions and constraint conditions, calculates the cost function values under finite switching states within a control period, and then determines the optimal switching state applied to the inverter at the next moment through iterative optimization. FCS-MPC fully considers the nonlinearity of the system and has great advantages in the process of multi-objective optimization and multi-objective constraints.

[0047] The specific implementation method of the present invention is as follows:

[0048] A. Model construction

[0049] The topological structure of the five-level inverter proposed by the present invention is as Figure 1 shown. The topology consists of six power switching tubes (S1 - S6), two power diodes D1 and D2, and a floating capacitor C F constituting it. Among them, the switching tubes (S1 - S4) form two half-bridges, the diodes are used to block reverse voltages, and the floating capacitor is placed between the two half-bridges. The inverter is powered by a single DC power supply V dc and two capacitors C1 and C2 are connected in series on the DC side.

[0050] Among them, V dc is the DC side voltage. V p and V n are the voltages of the upper capacitor and the lower capacitor respectively. i c1 , i c2 and i o are the currents flowing through the upper capacitor, the lower capacitor, and the neutral point current respectively. va , v b and v c are the output voltages of the inverter in phases a, b, and c respectively; L f and R f constitute the output LC filter; i a , i b and i c are the output currents of the inverter in phases a, b, and c respectively; v ga , v gb and v gc are the load voltages.

[0051] For the power switching devices of each leg of the five-level inverter, "1" represents that the power switching device is on, and "0" represents that the power switching device is off. Among them, the driving signals of switching devices S1 and S2 are always complementary, the driving signals of S3 and S4 are always complementary, and the driving signals of S5 and S6 are always the same. For the convenience of modeling, it is assumed that all switching components are ideal devices, the dead time is ignored, and the output end of the inverter is connected to the power grid through a filter. Then the proposed five-level inverter circuit can be equivalent to the corresponding switching model.

[0052] Table 1 Different Switching Device Combinations and Corresponding Output Voltages

[0053]

[0054]

[0055] Table 1 lists all the switching states of the proposed five-level inverter topology. G represents the charging state of capacitor C F , D represents the discharging state, CH represents the charging state, and NE represents neither charging nor discharging. v out represents the output voltage of one phase of the inverter. It can be seen from Table 1 that the inverter can generate 6 different switching states, and the peak value of its output voltage is equal to the input voltage; capacitor C F is continuously charged and discharged within one control period. Since power transistors S5 and S6 carry the charging current and the load current, their rated current values should be higher than those of other switching devices.

[0056] According to Kirchhoff's law, the state equation of the inverter in the dq rotating coordinate system can be expressed as:

[0057]

[0058] Among them, i d and i q respectively represent the components of the inverter output current on the dq axes; ω represents the angular frequency; R and L respectively represent the resistive load and the inductive load; v gd and v gqrespectively represent the components of the load voltage on the dq axes; v d and v q respectively represent the components of the inverter output voltage on the dq axes.

[0059] When the sampling period T s is small, the formula (1) is discretized using the forward Euler formula to obtain the recurrence formula for the state variables at the k-th and k+1-th moments:

[0060]

[0061] Then at the k+1-th moment, the predicted dq-axis currents of the inverter can be expressed as:

[0062]

[0063] where, i d (k + 1) and i q (k + 1) respectively represent the predicted dq-axis currents of the inverter at the k+1-th moment;

[0064] i d (k) and i q (k) respectively represent the predicted dq-axis currents of the inverter at the k-th moment; v d (k) and v q (k) respectively represent the dq-axis output voltages of the inverter at the k-th moment; v gd (k) and v gq (k) respectively represent the dq-axis load voltages of the inverter at the k-th moment.

[0065] B. DC-side model

[0066] According to Figure 1 , the relationship between the bus capacitor current and the neutral point current is:

[0067] i o =i c1 -i c2 (4)

[0068] The DC bus capacitor voltage is:

[0069]

[0070] Using the forward Euler method to discretize the formula (5), the voltages of the upper and lower capacitors at the k+1-th moment can be expressed as:

[0071]

[0072] where, V p (k + 1) and V n(k + 1) represents the upper and lower capacitor voltages of the bus at the (k + 1)-th moment respectively;

[0073] V p (k) and V n (k) represent the upper and lower capacitor voltages of the bus at the k-th moment respectively; i c1 (k) and i c2 (k) represent the upper and lower capacitor currents of the bus at the k-th moment respectively; C1 and C2 are the upper and lower capacitors of the bus respectively, and their values are equal, C1 = C2 = C.

[0074] C. Construct the cost function

[0075] To achieve accurate reference current tracking and NP voltage balance, the cost function J of FCS-MPC is:

[0076]

[0077] where and represent the reference currents of the dq axes of the inverter respectively; ΔV NP represents the difference between the upper and lower capacitor voltages.

[0078] The given q-axis current of the five-level inverter is 0; the given d-axis current is calculated by the maximum power point tracking (MPPT) algorithm.

[0079] As can be seen from Equation (6), the difference between the upper and lower capacitor voltages of the bus is expressed as:

[0080]

[0081] The finite switching state model predictive control strategy of the five-level inverter is as Figure 2 shown. At the sampling k-th moment, measure the output voltage, output current, and load voltage of the inverter, and obtain the physical quantities in the dq coordinate system using the coordinate transformation principle; construct a prediction model to calculate the predicted current of the inverter in the dq coordinate system; the photovoltaic MPPT algorithm takes the photovoltaic voltage v pv and the photovoltaic current i pv as input quantities, and the output is the DC voltage reference The deviation value between the output of the photovoltaic MPPT algorithm and the actual DC side voltage is used as the input of the PI controller, and thus the reference value of the d-axis current is obtained; combined with the 6 switching states of the inverter, the cost function J is calculated and evaluated online, and the switching state that minimizes the cost function value is found and applied to the inverter in the next switching period.

[0082] Different from traditional PR control, this strategy does not require dq-axis current PI controllers. By constructing a prediction model of the inverter and combining the six switching states of the inverter to estimate the predicted current at the next moment, a faster dynamic response can be achieved, which is easy to debug and is especially suitable for systems with transient load disturbances. In addition, this strategy uses a cost function to evaluate the optimal switching vector, with voltage constraints and current constraints, and has stronger adaptability in the case of grid fluctuations and load fluctuations.

[0083] D. Steady-state performance experiment

[0084] To verify the steady-state performance of the proposed five-level inverter topology, an inductive load of 10Ω + 35mH is connected to the output terminal of the inverter, and the DC voltage V dc is 110V. Figure 3 (a) of shows the single-phase output voltage v o of the inverter, the output current i o , and the voltage v F across the capacitor C CF . It can be seen from the figure that the output voltage waveform includes five different levels: 0V, ±55V, and ±110V, with a bus voltage utilization rate of 100%. The maximum value of v o reaches 110V; the output current is approximately a sine wave, and the peak value reaches 14.7A; Figure 3 (b) of shows the DC voltage V dc , the DC-side capacitor voltages v c1 and v c2 . It can be seen from Figure 3 that the voltages v c1 and v c2 can be well balanced and have a peak value of 55V.

[0085] The advantages of the present invention are as follows:

[0086] 1. In the control of grid-connected inverters, FCS-MPC performs optimization calculations on a limited number of switching states in each control cycle, predicts the current changes at multiple moments in advance, and can achieve a faster dynamic response, especially suitable for systems with transient load disturbances.

[0087] 2. Compared with traditional PR control, FCS-MPC is easy to add system constraints such as current constraints and voltage constraints, so it has stronger robustness to grid fluctuations and load fluctuations.

[0088] 3. Compared with traditional five-level topologies, the topology proposed in the present invention requires fewer switching devices, has a simpler structure, maximizes the bus voltage utilization rate, has lower voltage stress between the DC-side capacitor and the switching devices, and is suitable for photovoltaic inverter applications.

[0089] As described above, it is only the preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims described above.

Claims

1. A model predictive control method for a five-level inverter with enhanced bus utilization rate, characterized in that, Including: At the sampling time k, measure the output voltage, output current, and load voltage of the five-level inverter, and obtain the physical quantities in the dq coordinate system by using the coordinate transformation principle; Construct a prediction model to calculate the predicted current of the inverter in the dq coordinate system; The photovoltaic voltage v is processed through a photovoltaic MPPT algorithm pv and the photovoltaic current i pv are used as input quantities, and the output is the DC voltage reference The deviation value between the DC voltage reference and the actual DC side voltage is used as the input of the PI controller to obtain the d-axis current reference value; Combine the six switching states of the inverter to perform online calculation and evaluation of the cost function J, find the switching state that minimizes the cost function value, and apply it to the inverter in the next switching period.

2. The method according to claim 1, wherein After calculating the predicted current, it further includes: Calculate the difference in bus capacitor voltages to balance the neutral point voltage.

3. The method according to claim 1, wherein The topology of the five-level inverter includes six power switches (S1 - S6), two power diodes D1 and D2, and a floating capacitor C F ; among them, the switches (S1 - S4) form two half-bridges, the two power diodes are used to block the reverse voltage, and the floating capacitor is placed between the two half-bridges.

4. The method according to claim 3, wherein The inverter is powered by a single DC power supply V dc and two capacitors C1 and C2 are connected in series on the DC side; Among them, V dc is the DC-side voltage; V p and V n are the voltages of the upper capacitor and the lower capacitor respectively; i c1 , i c2 and i o are the currents flowing through the upper capacitor, the lower capacitor and the neutral-point current respectively; v a , v b and v c are the output voltages of the inverter in phases a, b, and c respectively; L f and R f constitute the output LC filter; i a , i b and i c are the output currents of the inverter in phases a, b, and c respectively; v ga , v gb and v gc are the load voltages.

5. The method according to claim 4, wherein The drive signals of power switch tubes S1 and S2 are always complementary, the drive signals of S3 and S4 are always complementary, and the drive signals of S5 and S6 are always the same.

6. The method according to claim 1 or 5, wherein The obtaining of the physical quantities in the dq coordinate system by using the coordinate transformation principle includes: According to Kirchhoff's law, the state equation of the inverter in the dq rotating coordinate system is expressed as: wherein, i d and i q respectively represent the components of the inverter output current on the dq axes; ω represents the angular frequency; R and L respectively represent the resistive load and the inductive load; v gd and v gq respectively represent the components of the load voltage on the dq axes; v d and v q respectively represent the components of the inverter output voltage on the dq axes.

7. The method according to claim 6, wherein The constructing of the prediction model to calculate the predicted current of the inverter in the dq coordinate system includes: At the time k + 1, the dq-axis predicted current of the inverter is expressed as: where i d (k + 1) and i q (k + 1) respectively represent the dq-axis predicted currents of the inverter at the (k + 1)-th moment; T s represents the sampling period; i d (k) and i q (k) respectively represent the dq-axis predicted currents of the inverter at the k-th moment; v d (k) and v q (k) respectively represent the dq-axis output voltages of the inverter at the k-th moment; v gd (k) and v gq (k) respectively represent the dq-axis load voltages of the inverter at the k-th moment.

8. The method according to claim 7, wherein The voltages of the upper capacitor and the lower capacitor at the time k + 1 are expressed as: Among them, V p (k + 1) and V n (k + 1) represent the upper and lower capacitor voltages on the bus at the (k + 1)-th moment respectively; V p (k) and V n (k) represent the upper and lower capacitor voltages on the bus at the k-th moment respectively; i c1 (k) and i c2 (k) represent the upper and lower capacitor currents on the bus at the k-th moment respectively; C1 and C2 are the capacitor on the bus and the capacitor under the bus respectively.

9. The method according to claim 8, wherein The cost function J is: Among them, and respectively represent the reference currents of the dq axes of the inverter; ΔV NP represents the difference in the capacitor voltages above and below the bus.

10. The method according to claim 9, wherein The difference in the upper and lower bus capacitor voltages is expressed as:

Images