Three-level non-reverse buck-boost finite set model prediction control method

Abstract

The invention discloses a three-level non-reverse buck-boost finite set model prediction control method, which comprises the following steps of: recording a switching state of a previous period, and initializing an initial value of a cost function; sampling an input side voltage, an inductive current and an output side voltage of the buck-boost converter in the previous period; inputting the input side voltage, the inductive current and the output side voltage into a prediction model to obtain a predicted value of the inductive current, a predicted value of the input side capacitor voltage and a predicted value of the output side capacitor voltage; switching state optimization is carried out to determine a finite set, candidate switching modes in the period are determined, and the number of the candidate switching modes is recorded as N; and respectively calculating cost function values under 1-N candidate modes through a cost function, selecting a mode which enables the cost function value to be minimum, and enabling the switching pulse corresponding to the mode to act on the next control period. According to the three-level buck-boost converter, the switching loss of the converter is reduced while switching state optimization is realized, and the efficiency of the three-level buck-boost converter is effectively improved.

Description

Technical Field

[0001] The invention belongs to the field of converters, and relates to a control technology of a three-level non-inverting buck-boost converter, and in particular to a three-level non-inverting buck-boost finite set model predictive control method. Background Art

[0002] Driven by both battery swap stations and fast charging technology, not only has the "mileage anxiety" of electric vehicle users been greatly alleviated, but also the technological innovation of the entire industry chain has been promoted. As an important infrastructure for electric vehicle energy replenishment, the research on the topology and control method of battery swap stations has become a hot topic. However, the demand for charging voltage and power of electric vehicles is growing, from the traditional low voltage level to the high voltage range (such as 400V~800V), which puts more stringent requirements on charging facilities. Although the traditional two-level buck-boost converter has a simple structure and is easy to control, its disadvantages are becoming increasingly significant when facing such a wide voltage range. The switching devices need to withstand higher voltage stress, and the cost of the converter is also increasing. Therefore, the advantages of the three-level non-reversing buck-boost DC converter are gradually revealed.

[0003] Different from the two-level converter, the three-level structure effectively broadens the output voltage range and reduces the voltage stress of the switching device by increasing the number of levels, thereby reducing the requirements for the withstand voltage level of the switching device. However, the three-level non-reversing buck-boost converter has many switching modes, up to 16. With the increase of the voltage level, the switch state combination of the converter increases exponentially. How to select a limited number of modes from a large number of switching modes? The current research does not give a standard for state selection, which makes the control of the three-level buck-boost converter more difficult. In addition, although the global switch state polling can obtain the best output effect, this situation does not take into account the loss of the switching device of the converter. There is a possibility that multiple switching devices may act simultaneously in adjacent control cycles. In some cases, the action of the switching device in adjacent control cycles even causes 8 switching devices to act simultaneously. This leads to a significant increase in the switching loss of the switching device of the converter, severe heating of the converter, and serious impact on the reliable operation of the converter. Summary of the invention

[0004] Purpose of the invention: In order to overcome the deficiencies in the prior art, a three-level non-inverting buck-boost finite set model predictive control method is provided to achieve optimal switching state while reducing the switching loss of the converter, thereby effectively improving the efficiency of the three-level buck-boost converter.

[0005] Technical solution: To achieve the above object, the present invention provides a three-level non-inverting buck-boost finite set model predictive control method, comprising the following steps:

[0006] S1: Record the switch state Mi(k) used in the previous cycle, initialize the initial value of the cost function, and let J opt =∞, where J opt is the minimum value of the cost function in the current control cycle;

[0007] S2: Sample the input side voltage, inductor current and output side voltage of the buck-boost converter in the previous cycle;

[0008] S3: Inputting the input side voltage, the inductor current and the output side voltage into the constructed prediction model to obtain the predicted value of the inductor current, the predicted value of the input side capacitor voltage and the predicted value of the output side capacitor voltage;

[0009] S4: Perform switch state optimization to determine a finite set, determine the candidate switch modes of this cycle, and record the number of candidate switch modes as N. The serial numbers of all candidate switch modes are stored in the array Order[N].

[0010] S5: According to the predicted value obtained in step S3 and the inductor current of this cycle, the cost function values ​​under 1 to N candidate modes are calculated respectively through the cost function, the mode with the smallest cost function value is selected, and the switching pulse corresponding to the mode is applied to the next control cycle.

[0011] Furthermore, the prediction model in step S3 is a discretized state space model, and the establishment method is:

[0012] The eight switching devices S1 to S8 of the three-level buck-boost converter are divided into four half-bridges, and the state of each half-bridge is represented by H1 to H4. i for:

[0013]

[0014] According to Kirchhoff's voltage law, the voltage at the midpoint of the left bridge arm is v ab And the right bridge arm midpoint voltage v cd The relationship is:

[0015]

[0016] Among them, R eq is the equivalent resistance, L is the inductance, i L is the inductor current; similarly, the relationship between the bridge arm voltage and the half-bridge state is:

[0017]

[0018] Among them, V i1 is the capacitor voltage on the input side, V i2 is the capacitor voltage below the input side, V o1 is the capacitor voltage on the output side, V o2 is the capacitor voltage below the output side;

[0019] According to Kirchhoff's current law, the input current and output current equations are written as:

[0020]

[0021] Among them, C i1 , C i2 is the value of the two capacitors on the input side, C o1 , C o2 are the two capacitor values ​​on the output side, R is the output load resistance, i o is the output current, i i is the input current;

[0022] The inductor current i L 、Input voltage V i1 、V i2 And the output voltage V o1 、V o2 As the state variable, the state equation in the continuous domain is obtained; then, the continuous domain state equation is discretized and the inductor current i L Discretized into:

[0023]

[0024] Among them, T s is the control period, k is the current moment, k+1 is the next moment, and the superscript p represents the predicted value at the next moment;

[0025] Similarly, the input capacitor voltage V i1 Discretized into:

[0026]

[0027] The input capacitor voltage V i2 Discretized into:

[0028]

[0029] The output capacitor voltage V o1 Discretized into:

[0030]

[0031] The output capacitor V o2 Discretized into:

[0032]

[0033] Furthermore, the method for optimizing the switch state in step S4 includes:

[0034] A1: Buck mode switch state optimization;

[0035] Definition M i =H1H2H3H4, 4 are selected from the 16 modes of the converter to make it work in buck mode. In buck mode, the left half-bridge states H1 and H2 are in dynamic switching mode while the right half-bridge state remains in fixed mode and H3=1, H4=0, and the modes M1, M2, M3, and M4 are selected;

[0036] A2: Optimizing the switch state of boost mode;

[0037] Select 4 modes from the 16 modes of the converter to make it work in boost mode. In boost mode, the left half-bridge states H1 and H2 maintain fixed states H1=1 and H2=0, while the right half-bridge states H3 and H4 are in dynamic switching mode, and the modes M1, M6, M7, and M8 are selected;

[0038] A3: Using the common mode M1 in step A1 and step A2, the converter can be switched from buck mode to boost mode and from boost mode to buck mode;

[0039] A4: Optimizing the state of capacitor voltage balance at input and output sides;

[0040] In order to ensure the voltage balance of the capacitors on the input and output sides, four additional modes Me1=1111, Me2=1100, Me3=0011 and Me4=0000 are introduced from the remaining modes. The introduction of additional modes not only satisfies the voltage balance of the input and output sides in any mode, but also minimizes the number of switching actions of adjacent mode devices.

[0041] A5: Determine a finite set;

[0042] The set consisting of M1~M7 and Me1~Me4 is determined as the finite set to be optimized.

[0043] Furthermore, in the buck mode in step A1, when the mode is M1=1010, the load power is provided by the input power supply and the inductor; when the mode is M2=1110, the power required by the load is provided by the input side capacitor C i1 and the inductor together provide the capacitor C i1 Discharge; When the mode is M3 = 0110, the power required by the load is provided by the inductor and the supporting capacitor. The capacitor C i1 and Ci2 Bypassed; when the mode is M4 = 0010, the power required by the load is provided by the input side capacitor C i2 and the inductor together provide the capacitor C i2 The preferred converter modes M1 and M2, M2 and M3, M3 and M4, and M4 and M1 have a switching state change of 1 between adjacent modes.

[0044] Furthermore, in the boost mode in step A2, when the mode is M1=1010, the load power is provided by the input power supply and the inductor; when the mode is M5=1011, the load power is provided by the supporting capacitor C o1 and C o2 Provided, in addition, that capacitor C o1 Charging C o2 Discharge; When the mode is M6=1001, the power required by the load is supplied by the support capacitor, and the two inductors store energy during this period; When the mode is M7=1000, the power required by the load is provided by the support capacitor, and the capacitor C o2 Charging C o1 The preferred converter modes M1 and M5, M5 and M6, M6 and M7, and M7 and M1 have a switching state variation of 1 between adjacent modes.

[0045] Furthermore, the cost function in step S5 is formulated as follows:

[0046] The output voltage of the sampling converter is compared with the reference voltage value, and the difference is passed through the PI controller to obtain the reference value i of the inner loop current. L ref ; In order to make the inner loop current track quickly, the inductor current is introduced into the cost function; In order to ensure the balance of the upper and lower capacitor voltages on the input side and the output side, the capacitor voltage is introduced into the cost function; In order to protect the device, the inductor current limit λ iL_limit Introduced into the cost function;

[0047] The cost function is determined as:

[0048]

[0049] Inductor current limit λ iL_limit for:

[0050]

[0051] Furthermore, the method of selecting the mode that minimizes the cost function value in step S5 is:

[0052] If the candidate cost function value J j <J opt , then J j Assign to J opt, and record j at this time. When the optimization of N modes is completed, the jth mode is selected as the best mode for the next control cycle, and the corresponding 8 switch pulses are output; finally, the underlying switch device is driven by the driving circuit. The above is the entire process of optimizing a control cycle. The next cycle process is similar to this, rolling optimization, so as to select the optimal mode of each cycle and determine the switch state S1~S8.

[0053] The method of the present invention is based on the state space model of the three-level buck-boost converter, and selects finite modes from numerous switching modes. It not only achieves the requirements of input and output side capacitor voltage balance and smooth output in a wide voltage range, but also ensures that the number of switching actions in adjacent control cycles is minimal, greatly reducing the loss of the device. The present invention can achieve smooth transition between boost, buck, and buck-boost modes without monitoring the working state of the converter. In addition, its output voltage gain is high, bidirectional power flow can be achieved, and the output voltage polarity is the same as the input voltage polarity, which can meet the charging needs of loads with different voltage levels in battery swap stations.

[0054] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0055] 1. By combing the modes of buck mode, boost mode, switching mode and capacitor voltage balancing mode, the method of the present invention selects a set of finite modes from numerous modes, which provides a new idea for the control of the three-level buck-boost converter, which not only ensures the smooth transition of the converter between various modes, but also can meet the requirements of a wide range of output voltages; in addition, the dynamic response of the method of the present invention is faster.

[0056] 2. Since the method of the present invention considers the switching between adjacent modes in the process of modal optimization, the number of changes between the switch states of adjacent control cycles is always 1, which is the minimum. The reduction in the number of switching actions means that the loss generated by the switching device is lower. Therefore, the method of the present invention can effectively reduce the loss of the switching device, simplify the structural design of the heat dissipation system, and the efficiency of the converter system using the method of the present invention will be further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of a three-level non-inverting buck-boost topology;

[0058] Figure 2 It is a schematic diagram of the state transition of a three-level non-inverting buck-boost finite set switch;

[0059] Figure 3 It is a control diagram of a three-level non-inverting buck-boost model prediction system;

[0060] Figure 4 It is a schematic diagram of the predictive control process of the three-level non-inverting buck-boost model;

[0061] Figure 5 It is a schematic diagram for comparing the efficiency of the control method of the present invention and the traditional control method. DETAILED DESCRIPTION

[0062] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0063] The three-level non-inverting buck-boost topology in this embodiment is as follows: Figure 1 As shown, where V in is the total voltage on the input side, V i1 is the capacitor voltage on the upper left side, V i2 is the capacitor voltage at the bottom left; V o is the total voltage on the output side, V o1 is the capacitor voltage on the upper right side, V o2 is the capacitor voltage at the lower right side; C i1 , C i2 is the value of the two capacitors on the input side, C o1 , C o2 are the values ​​of the two capacitors on the output side; R eq is the equivalent resistance, L is the inductance, i L is the inductor current; R is the output load resistance, i o is the output current, i i is the input current.

[0064] For the above three-level non-inverting buck-boost, this embodiment provides a three-level non-inverting buck-boost finite set model predictive control method, such as Figure 3 and Figure 4 As shown, it includes the following steps:

[0065] S1: Record the switch state Mi(k) used in the previous cycle, initialize the initial value of the cost function, and let J opt =∞, where J opt is the minimum value of the cost function in the current control cycle;

[0066] S2: Sample the input side voltage, inductor current and output side voltage of the buck-boost converter in the previous cycle;

[0067] S3: Inputting the input side voltage, the inductor current and the output side voltage into the constructed prediction model to obtain the predicted value of the inductor current, the predicted value of the input side capacitor voltage and the predicted value of the output side capacitor voltage;

[0068] The prediction model is a discretized state space model, and the establishment method is:

[0069] The eight switching devices S1 to S8 of the three-level buck-boost converter are divided into four half-bridges, and the state of each half-bridge is represented by H1 to H4. i for:

[0070]

[0071] According to Kirchhoff's voltage law, the voltage at the midpoint of the left bridge arm is v ab And the right bridge arm midpoint voltage v cd The relationship is:

[0072]

[0073] Among them, R eq is the equivalent resistance, L is the inductance, i L is the inductor current; similarly, the relationship between the bridge arm voltage and the half-bridge state is:

[0074]

[0075] Among them, V i1 is the capacitor voltage on the input side, V i2 is the capacitor voltage below the input side, V o1 is the capacitor voltage on the output side, V o2 is the capacitor voltage below the output side;

[0076] According to Kirchhoff's current law, the input current and output current equations are written as:

[0077]

[0078] Among them, C i1 , C i2 is the value of the two capacitors on the input side, C o1 , C o2 are the two capacitor values ​​on the output side, R is the output load resistance, i o is the output current, i i is the input current;

[0079] The inductor current i L 、Input voltage V i1 、V i2 And the output voltage V o1 、V o2As the state variable, the state equation in the continuous domain is obtained; then, the continuous domain state equation is discretized and the inductor current i L Discretized into:

[0080]

[0081] Among them, T s is the control period, k is the current moment, k+1 is the next moment, and the superscript p represents the predicted value at the next moment;

[0082] Similarly, the input capacitor voltage V i1 Discretized into:

[0083]

[0084] The input capacitor voltage V i2 Discretized into:

[0085]

[0086] The output capacitor voltage V o1 Discretized into:

[0087]

[0088] The output capacitor V o2 Discretized into:

[0089]

[0090] S4: Perform switch state optimization to determine a finite set, determine the candidate switch modes of this cycle, and record the number of candidate switch modes as N. The serial numbers of all candidate switch modes are stored in the array Order[N].

[0091] See also Figure 2 , the methods for optimizing the switch state include:

[0092] A1: Buck mode switch state optimization;

[0093] Definition M i =H1H2H3H4, 4 are selected from the 16 modes of the converter to make it work in buck mode. In buck mode, the left half-bridge states H1 and H2 are in dynamic switching mode while the right half-bridge state remains in fixed mode and H3=1, H4=0, and the modes M1, M2, M3, and M4 are selected;

[0094] In buck mode, when the mode is M1=1010, the load power is provided by the input power supply and the inductor; when the mode is M2=1110, the power required by the load is provided by the input side capacitor C i1 and the inductor together provide the capacitor Ci1 Discharge; When the mode is M3 = 0110, the power required by the load is provided by the inductor and the supporting capacitor. The capacitor C i1 and C i2 Bypassed; when the mode is M4 = 0010, the power required by the load is provided by the input side capacitor C i2 and the inductor together provide the capacitor C i2 The preferred converter modes M1 and M2, M2 and M3, M3 and M4, and M4 and M1 have a switching state change of 1 between adjacent modes.

[0095] A2: Optimizing the switch state of boost mode;

[0096] Select 4 modes from the 16 modes of the converter to make it work in boost mode. In boost mode, the left half-bridge states H1 and H2 maintain fixed states H1=1 and H2=0, while the right half-bridge states H3 and H4 are in dynamic switching mode, and the modes M1, M6, M7, and M8 are selected;

[0097] In boost mode, when the mode is M1=1010, the load power is provided by the input power supply and the inductor; when the mode is M5=1011, the load power is provided by the support capacitor C o1 and C o2 Provided, in addition, that capacitor C o1 Charging C o2 Discharge; When the mode is M6=1001, the power required by the load is supplied by the support capacitor, and the two inductors store energy during this period; When the mode is M7=1000, the power required by the load is provided by the support capacitor, and the capacitor C o2 Charging C o1 The preferred converter modes M1 and M5, M5 and M6, M6 and M7, and M7 and M1 have a switching state variation of 1 between adjacent modes.

[0098] A3: Using the common mode M1 in step A1 and step A2, the converter can be switched from buck mode to boost mode and from boost mode to buck mode;

[0099] A4: Optimizing the state of capacitor voltage balance at input and output sides;

[0100] In order to ensure the voltage balance of the capacitors on the input and output sides, four additional modes Me1=1111, Me2=1100, Me3=0011 and Me4=0000 are introduced from the remaining modes. The introduction of additional modes not only satisfies the voltage balance of the input and output sides in any mode, but also minimizes the number of switching actions of adjacent mode devices.

[0101] A5: Determine a finite set;

[0102] The set consisting of M1~M7 and Me1~Me4 is determined as the finite set to be optimized.

[0103] S5: According to the predicted value obtained in step S3 and the inductor current of this cycle, the cost function values ​​under 1 to N candidate modes are calculated respectively by using the cost function, the mode with the smallest cost function value is selected, and the switch pulse corresponding to the mode is applied to the next control cycle;

[0104] The cost function is formulated as:

[0105] The output voltage of the sampling converter is compared with the reference voltage value, and the difference is passed through the PI controller to obtain the reference value i of the inner loop current. L ref ; In order to make the inner loop current track quickly, the inductor current is introduced into the cost function; In order to ensure the balance of the upper and lower capacitor voltages on the input side and the output side, the capacitor voltage is introduced into the cost function; In order to protect the device, the inductor current limit λ iL_limit Introduced into the cost function;

[0106] The cost function is determined as:

[0107] J(k)=|i L ref -i L p (k+1)|+λ1(|V o1 p (k+1)-V o2 p (k+1)|)+λ2(|V i1 p (k+1)-V i2 p (k+1)|)+λ iL_limit (10)

[0108] Among them, λ1 and λ2 are weight coefficients;

[0109] Inductor current limit λ iL_limit for:

[0110]

[0111] Subsequently, the cost functions under different modes are calculated, and the state that minimizes the cost function is selected as the state output of the next control cycle, and the pulse signal is sent to the switch devices S1 to S8; the method for selecting the mode that minimizes the cost function value is:

[0112] If the candidate cost function value J j <J opt , then J jAssign to J opt , and record j at this time. When the optimization of N modes is completed, the jth mode is selected as the best mode for the next control cycle, and the corresponding 8 switch pulses are output; finally, the underlying switch device is driven by the driving circuit. The above is the entire process of optimizing a control cycle. The next cycle process is similar to this, rolling optimization, so as to select the optimal mode of each cycle and determine the switch state S1~S8.

[0113] In order to verify the effectiveness and effect of the method of the present invention, the following experiments and data analysis were performed in this example:

[0114] A SiC Mosfet with a voltage level of 650V / 80A was selected, and the model number was C3M0025065J1. According to its device manual, its turn-on loss, turn-off loss, and conduction loss were imported into PLECS. Through the electric-thermal joint model, the losses of each semiconductor device of the converter were statistically analyzed, and the efficiency curves under the control strategy of the present invention and the traditional conventional optimization control were plotted. The details are shown in the attached figure. Figure 5 As shown. The loss analysis shows that the efficiency of the system under the finite control set model predictive control provided by the present invention is always higher than the efficiency under the traditional global optimization control mode. And the efficiency of the control method proposed by the present invention is basically maintained at more than 98%, indicating that the control method of the present invention can meet the requirements of efficient operation of the converter.

Claims

1. A three-level non-inverting buck-boost finite set model predictive control method, characterized in that: The steps include: S1: Record the switch state Mi(k) used in the previous cycle, initialize the initial value of the cost function, and let J opt =∞, where J opt is the minimum value of the cost function in the current control cycle; S2: Sample the input side voltage, inductor current and output side voltage of the buck-boost converter in the previous cycle; S3: Inputting the input side voltage, the inductor current and the output side voltage into the constructed prediction model to obtain the predicted value of the inductor current, the predicted value of the input side capacitor voltage and the predicted value of the output side capacitor voltage; S4: Perform switch state optimization to determine a finite set, determine the candidate switch modes of this cycle, and record the number of candidate switch modes as N. The serial numbers of all candidate switch modes are stored in the array Order[N]. S5: According to the predicted value obtained in step S3 and the inductor current of this cycle, the cost function values ​​under 1 to N candidate modes are calculated respectively through the cost function, the mode with the smallest cost function value is selected, and the switching pulse corresponding to the mode is applied to the next control cycle.

2. A three-level non-inverting buck-boost finite set model predictive control method according to claim 1, characterized in that: The prediction model in step S3 is a discretized state space model, and the establishment method is: The eight switching devices S1 to S8 of the three-level buck-boost converter are divided into four half-bridges, and the state of each half-bridge is represented by H1 to H4. i for: According to Kirchhoff's voltage law, the voltage at the midpoint of the left bridge arm is v ab And the right bridge arm midpoint voltage v cd The relationship is: Among them, R eq is the equivalent resistance, L is the inductance, i L is the inductor current; similarly, the relationship between the bridge arm voltage and the half-bridge state is: Among them, V i1 is the capacitor voltage on the input side, V i2 is the capacitor voltage below the input side, V o1 is the capacitor voltage on the output side, V o2 is the capacitor voltage below the output side; According to Kirchhoff's current law, the input current and output current equations are written as: and Among them, C i1 , C i2 is the value of the two capacitors on the input side, C o1 , C o2 are the two capacitor values ​​on the output side, R is the output load resistance, i o is the output current, i i is the input current; The inductor current i L 、Input voltage V i1 、V i2 And the output voltage V o1 、V o2 As the state variable, the state equation in the continuous domain is obtained; then, the continuous domain state equation is discretized and the inductor current i L Discretized into: Among them, T s is the control period, k is the current moment, k+1 is the next moment, and the superscript p represents the predicted value at the next moment; Similarly, the input capacitor voltage V i1 Discretized into: The input capacitor voltage V i2 Discretized into: The output capacitor voltage V o1 Discretized into: The output capacitor V o2 Discretized into:

3. A three-level non-inverting buck-boost finite set model predictive control method according to claim 1, characterized in that: The method for optimizing the switch state in step S4 includes: A1: Buck mode switch state optimization; Definition M i =H1H2H3H4, 4 are selected from the 16 modes of the converter to make it work in buck mode. In buck mode, the left half-bridge states H1 and H2 are in dynamic switching mode while the right half-bridge state remains in fixed mode and H3=1, H4=0, and the modes M1, M2, M3, and M4 are selected; A2: Optimizing the switch state of boost mode; Select 4 modes from the 16 modes of the converter to make it work in boost mode. In boost mode, the left half-bridge states H1 and H2 maintain fixed states H1=1 and H2=0, while the right half-bridge states H3 and H4 are in dynamic switching mode, and the modes M1, M6, M7, and M8 are selected; A3: The preferred switching state change between adjacent modes of the converter modes M1 and M5, M5 and M6, M6 and M7, and M7 and M1 is 1, and the preferred switching state change between adjacent modes of the converter modes M1 and M2, M2 and M3, M3 and M4, and M4 and M1 is 1; using the common mode M1 in the two modes of step A1 and step A2, the converter can be switched from buck mode to boost mode, and from boost mode to buck mode; A4: Optimizing the state of capacitor voltage balance at input and output sides; To ensure the voltage balance of the input and output capacitors, four additional modes Me1 = 1111, Me2 = 1100, Me3 = 0011 and Me4 = 0000 are introduced from the remaining modes; A5: Determine a finite set; The set consisting of M1~M7 and Me1~Me4 is determined as the finite set to be optimized.

4. A three-level non-inverting buck-boost finite set model predictive control method according to claim 3, characterized in that: In the buck mode in step A1, when the mode is M1=1010, the load power is provided by the input power supply and the inductor; when the mode is M2=1110, the power required by the load is provided by the input side capacitor C i1 and the inductor together provide the capacitor C i1 Discharge; When the mode is M3 = 0110, the power required by the load is provided by the inductor and the supporting capacitor. The capacitor C i1 and C i2 Bypassed; when the mode is M4 = 0010, the power required by the load is provided by the input side capacitor C i2 and the inductor together provide the capacitor C i2 Discharge.

5. A three-level non-inverting buck-boost finite set model predictive control method according to claim 3, characterized in that: In the boost mode in step A2, when the mode is M1=1010, the load power is provided by the input power supply and the inductor; when the mode is M5=1011, the load power is provided by the support capacitor C o1 and C o2 Provided, in addition, that capacitor C o1 Charging C o2 Discharge; When the mode is M6=1001, the power required by the load is supplied by the support capacitor, and the two inductors store energy during this period; When the mode is M7=1000, the power required by the load is provided by the support capacitor, and the capacitor C o2 Charging C o1 Discharge.

6. A three-level non-inverting buck-boost finite set model predictive control method according to claim 2, characterized in that: The cost function in step S5 is formulated as follows: The output voltage of the sampling converter is compared with the reference voltage value, and the difference is passed through the PI controller to obtain the reference value i of the inner loop current. L ref ; In order to make the inner loop current track quickly, the inductor current is introduced into the cost function; In order to ensure the balance of the upper and lower capacitor voltages on the input side and the output side, the capacitor voltage is introduced into the cost function; In order to protect the device, the inductor current limit λ iL_limit Introduced into the cost function; The cost function is determined as: J(k)=|i L ref -i L p (k+1)|+λ1(|V o1 p (k+1)-V o2 p (k+1)|)+λ2(|V i1 p (k+1)-V i2 p (k+1)|)+λ iL_limit (10)。 7. A three-level non-inverting buck-boost finite set model predictive control method according to claim 6, characterized in that: The inductor current limit λ in the cost function iL_limit for:

8. A three-level non-inverting buck-boost finite set model predictive control method according to claim 1, characterized in that: The method for selecting the mode that minimizes the cost function value in step S5 is: If the candidate cost function value J j <J opt , then J j Assign to J opt , and record j at this time. After the optimization of N modes is completed, the jth mode is selected as the best mode for the next control cycle, and the corresponding 8 switch pulses are output; finally, the underlying switch devices are driven by the driving circuit, and rolling optimization is performed to select the optimal mode for each cycle and determine the switch states S1~S8.

Images