Fractional-order predictive control method for multi-level power factor correction circuits

Abstract

The invention discloses a fractional-order predictive control method applied to a multi-level power factor correction circuit. A multi-level power factor correction converter is subjected to fractional-order PI control by a PI control unit; a predictive control unit establishes a fractional-order model based on inductor current and capacitor voltage for the multi-level power factor correction converter; a fractional-order predictive control model is constructed according to the fractional-order model, the power quality on the input side is improved by adjusting the order of the inductor current, and the voltage balance between two capacitors on the output side is accelerated by adjusting the order of the capacitor voltage; a multi-objective fitness function reflecting the input side power quality, output side dynamic response, and capacitor voltage balance of the multi-level power factor correction converter is established, and the order of the inductor current, the order of the capacitor voltage, the fractional order of the fractional-order PI control, and the parameters of the fractional-order PI control are optimized by a gray wolf optimization unit to obtain an optimal parameter combination.

Description

Technical Field

[0001] The present invention relates to the technical field of power electronic control, and in particular to a fractional order predictive control method applied to a multi-level power factor correction circuit. Background Art

[0002] In order to reduce power transmission losses and improve system stability, more and more power electronic products include power factor correction functions. Generally speaking, the power factor correction circuit has the function of correcting the input current waveform on the AC side and regulating the output voltage on the DC side. Due to the characteristics of continuous current, power factor correction converters derived from boost circuits are widely used to achieve the required power factor correction function. Active front-end power factor correction converters play a key role in ac-dc power conversion in various applications to achieve high power quality power supply. Among them, high power factor (PF) close to unity, low total harmonic distortion (THD) of input current and DC output voltage regulation are the main advantages of power factor correction circuits.

[0003] For a conventional boost dc / dc converter, a single switch needs to withstand the DC output voltage when the single switch is blocked. In a multilevel boost dc / dc converter, multiple cascade switches and multiple cascade capacitors are connected together. When one switch is turned on and the other is blocked, if the multiple capacitor voltages are balanced, the blocked switch only needs to withstand a portion of the DC output voltage. It is worth noting that the inductor voltage in the multilevel boost dc / dc converter has multiple levels, which makes the multilevel boost dc / dc converter have smaller inductor current ripple than the boost converter at the same switching frequency. Therefore, multilevel boost converters are often used in high voltage ratio applications. In addition, high voltage semiconductor switches usually have higher cost and larger drain-source resistance compared to low voltage semiconductor switches. Therefore, multilevel boost converters have the additional advantages of low switching losses and high efficiency.

[0004] The main feature of predictive control is that it uses system models to predict future changes in control variables. Based on predefined optimization criteria, the controller will use this information to determine the optimal mode of operation. Compared with traditional PI control, hysteresis control, etc., predictive control has the advantages of simple and intuitive concept, dead time compensation, and easy nonlinear model presentation. However, since the power electronic devices in the converter all have fractional-order characteristics, the traditional integer-order model cannot well express the characteristics of some actual systems. With the continuous development of industry, researchers' requirements for the system are constantly increasing. In the case of time-varying and nonlinear, the traditional integer-order model predictive control may have the problem of insufficient control accuracy. Summary of the invention

[0005] The present invention provides a fractional-order predictive control method applied to a multi-level power factor correction circuit to solve the problem that the traditional model predictive control based on integer order may have insufficient control accuracy in the case of time-varying and nonlinear conditions.

[0006] The present invention provides a fractional-order predictive control method applied to a multi-level power factor correction circuit, comprising:

[0007] The PI control unit performs fractional-order PI control on the multi-level power factor correction converter; wherein the output end of the PI control unit is connected to the input end of the prediction control unit, and the output end of the prediction control unit is connected to the driving interface of the switch tube of the multi-level power factor correction converter;

[0008] The prediction control unit establishes a fractional-order model based on the inductor current and the capacitor voltage for the multi-level power factor correction converter; according to the fractional-order model, a fractional-order prediction control model is constructed, and the cost function of the fractional-order prediction control model includes a fractional-order inductor current model and a fractional-order capacitor voltage model; the power quality on the input side is improved by adjusting the order of the inductor current, and the voltage balance between the two capacitors on the output side is accelerated by adjusting the order of the capacitor voltage;

[0009] A multi-objective fitness function reflecting the input side power quality, output side dynamic response and capacitor voltage balance of the multi-level power factor correction converter is established. The order of the inductor current, the order of the capacitor voltage, the fractional order of the fractional-order PI control and the parameters of the fractional-order PI control are optimized through the Gray Wolf optimization unit to obtain the optimal parameter combination.

[0010] Further, when a three-level power factor correction converter is used as an example to describe a multi-level power factor correction circuit, the switch tube includes a switch device T1 and a switch device T2, and the prediction control unit establishes a fractional-order model based on the inductor current and the capacitor voltage for the multi-level power factor correction converter, including:

[0011] In the nth cycle, if the switch device T1 and the switch device T2 are turned on at the same time, the fractional-order mathematical model is:

[0012]

[0013]

[0014] Where V in is the input voltage on the AC side, L is the filter inductance, i L is the inductor current, C 1 is the output capacitance, R is the output resistance, v C1 is the output capacitor C 1 Voltage, V 0is the output voltage, α is the order of the inductor current, and β is the order of the capacitor voltage.

[0015] Furthermore, in the nth cycle, if the switch device T1 is turned on and the switch device T2 is turned off, the fractional-order mathematical model is:

[0016]

[0017]

[0018] In the formula, v C2 Represents the output capacitance C 2 voltage.

[0019] Furthermore, in the nth cycle, if the switch device T1 is turned off and the switch device T2 is turned on, the obtained fractional-order mathematical model is:

[0020]

[0021]

[0022] Furthermore, in the nth cycle, if the switch device T1 and the switch device T2 are turned off at the same time, the fractional-order mathematical model is:

[0023]

[0024]

[0025] Further, according to the fractional-order model, a fractional-order predictive control model is constructed, wherein the cost function of the fractional-order predictive control model includes a fractional-order inductor current model and a fractional-order capacitor voltage model, and the steps include:

[0026] In the nth cycle, if the switch device T1 and the switch device T2 are turned on at the same time, the expressions of the inductor current and the capacitor voltage are:

[0027]

[0028]

[0029] In the formula, p 1-α is the discretized d after Oustaloup filtering algorithm 1-α / dt 1-α Fractional differential operator, i L_1 (n+1) represents the inductor current when the switch devices T1 and T2 are turned on at the same time, v C1_1 (n+1) represents the capacitance C when the switch devices T1 and T2 are turned on at the same time 1 Voltage, T Srepresents the switching cycle, V in (n) represents the rectified input voltage at the beginning of the nth cycle, i L (n) represents the rectified inductor current at the beginning of the nth cycle, v C1 (n) represents the rectified capacitance C at the beginning of the nth cycle 1 Voltage.

[0030] Furthermore, in the nth cycle, if the switch device T1 is turned on and the switch device T2 is turned off, the expressions of the inductor current and the capacitor voltage are:

[0031]

[0032]

[0033] In the formula, i L_2 (n+1) represents the inductor current when the switch device T1 is turned on and the switch device T2 is turned off; v C1_2 (n+1) represents the capacitor voltage when the switch device T1 is turned on and the switch device T2 is turned off; v C2 (n) is the capacitance C at the beginning of the nth cycle 2 voltage.

[0034] Furthermore, in the nth cycle, if the switch device T1 is turned off and the switch device T2 is turned on, the expressions of the inductor current and the capacitor voltage are:

[0035]

[0036]

[0037] In the formula, i L_3 (n+1) represents the inductor current when the switch device T1 is turned off and the switch device T2 is turned on; v C1_3 (n+1) represents the capacitor voltage when the switch device T1 is turned off and the switch device T2 is turned on.

[0038] Furthermore, in the nth cycle, if the switch device T1 and the switch device T2 are turned off at the same time, the expressions of the inductor current and the capacitor voltage are respectively:

[0039]

[0040]

[0041] In the formula, i L_4 (n+1) is the inductor current when the switch devices T1 and T2 are turned off at the same time; v C1_4 (n+1) is the capacitor voltage when the switch device T1 and the switch device T2 are turned off at the same time;

[0042] Based on the above four operating modes, the cost function of fractional-order predictive control is established as:

[0043]

[0044] In the formula, i ref (n) and V ref (n) represents the reference current signal and the reference voltage signal respectively, and λ is the weight coefficient of predictive control. By calculating the value of the cost function g, the switch state that minimizes the g value is selected as the state of the switch tube in the next cycle. The reference current i ref The value is obtained by the voltage outer loop, the reference voltage V ref The value is half of the output voltage.

[0045] Furthermore, in the step of obtaining the optimal parameter combination by optimizing the order of the inductor current, the order of the capacitor voltage, the fractional order of the fractional-order PI control, and the parameters of the fractional-order PI control through the gray wolf optimization unit, the expression of the gray wolf's prey search process is:

[0046]

[0047]

[0048]

[0049] In the formula, X α , Y γ , X β are the current positions of α wolf, γ wolf and β wolf respectively; C1, C2 and C3 are random perturbations of α wolf, γ wolf and β wolf respectively, and X(t+1) is the best position of the tth generation;

[0050] The calculation formula for vectors A and C is:

[0051]

[0052] In the formula, r 1 、r 2 is a random number in [0,1], and the convergence factor a decreases linearly from 2 to 0 as the number of iterations increases;

[0053] The performance index of the intelligent optimization algorithm is evaluated by establishing a multi-objective fitness function. The objective function used is the time integral of absolute value (ITAE). The optimal parameter combination is obtained by minimizing ITAE.

[0054] The expression of ITAE is:

[0055] ITAE=∫t|e(t)|dt;

[0056] Among them, the expression of e(t) is:

[0057]

[0058] In the formula, I h is the effective value of the total harmonic current on the input side; I 1 is the effective value of the fundamental current on the input side; v C1 is the capacitance C 1 The voltage value, V 0 is the output voltage value, V ref is the reference value of the output voltage; μ and λ represent the inertia weight values ​​in the fitness function, which can reflect the power quality at the input side of the PFC circuit, the balance of the capacitor voltage and the dynamic response of the output voltage.

[0059] The beneficial effects of the present invention are as follows: a fractional-order predictive control method applied to a multi-level power factor correction circuit provided by the present invention performs fractional-order PI control on a multi-level power factor correction converter through a PI control unit; wherein the output end of the PI control unit is connected to the input end of the predictive control unit, and the output end of the predictive control unit is connected to the drive interface of the switch tube of the multi-level power factor correction converter, and the switch tube includes a switch device T1 and a switch device T2; the predictive control unit establishes a fractional-order model based on the inductor current and the capacitor voltage for the multi-level power factor correction converter; according to the fractional-order model, a fractional-order predictive control model is constructed, and the cost function of the fractional-order predictive control model The multi-level power factor correction converter is designed to improve the power quality of the input side, the dynamic response of the output side, and the capacitor voltage balance. The multi-objective fitness function is established to reflect the power quality of the input side, the dynamic response of the output side, and the capacitor voltage balance of the multi-level power factor correction converter. The order of the inductor current, the order of the capacitor voltage, the fractional order of the fractional-order PI control, and the parameters of the fractional-order PI control are optimized through the Gray Wolf optimization unit to obtain the optimal parameter combination, which solves the problem that the traditional model predictive control based on integer order may have insufficient control accuracy in time-varying and nonlinear conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0061] Figure 1 1 is a circuit diagram of a three-level power factor correction converter according to an embodiment of the present invention.

[0062] Figure 2This is a circuit control principle diagram of a three-level power factor correction converter according to an embodiment of the present invention.

[0063] Figure 3 This is a level diagram of the Grey Wolf Optimization Algorithm according to an embodiment of the present invention.

[0064] Figure 4 This is a control block diagram of the grey wolf optimization algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. The technical solutions provided by the embodiments of the present invention are described in detail below in conjunction with the drawings.

[0066] The embodiment of the present invention provides a fractional-order predictive control method applied to a multi-level power factor correction circuit, and the scheme for implementing the method includes a PI control unit, a predictive control unit, and a gray wolf optimization unit. The output end of the PI control unit is connected to the input end of the predictive control unit, and the output end of the predictive control unit is connected to the driving interface of the switch tube of the multi-level power factor correction converter, and the switch tube includes a switch device T1 and a switch device T2. The method specifically includes the following steps:

[0067] The PI control unit performs fractional-order PI control on the multi-level power factor correction converter. The prediction control unit establishes a fractional-order model based on the inductor current and the capacitor voltage for the multi-level power factor correction converter; according to the fractional-order model, a fractional-order prediction control model is constructed, and the cost function of the fractional-order prediction control model includes a fractional-order inductor current model and a fractional-order capacitor voltage model; the power quality on the input side is improved by adjusting the order of the inductor current, and the voltage balance between the two capacitors on the output side is accelerated by adjusting the order of the capacitor voltage. A multi-objective fitness function reflecting the input-side power quality, output-side dynamic response, and capacitor voltage balance of the multi-level power factor correction converter is established, and the order of the inductor current, the order of the capacitor voltage, the fractional order of the fractional-order PI control, and the parameters of the fractional-order PI control are optimized by the gray wolf optimization algorithm of the gray wolf optimization unit to obtain the optimal parameter combination.

[0068] The present invention is based on a fractional-order model of a multi-level power factor correction circuit, and a fractional-order model predictive control is built. Based on the problem of insufficient control accuracy of traditional control, the present invention improves the control accuracy through fractional-order predictive control. In the PI control loop, fractional-order PI control is adopted. In the predictive control loop, a fractional-order predictive control model is built based on the fractional-order model of the inductor current and the capacitor voltage. By adjusting the order, the improvement of the power quality on the input side and the voltage balance between the two capacitors on the output side are achieved. In addition, the gray wolf optimization algorithm is adopted to optimize the parameters such as PI parameters, fractional order, and predictive control weight coefficient, so as to further improve the system performance.

[0069] In this embodiment, the multi-level power factor correction converter is described by taking a three-level power factor correction converter as an example. The prediction control unit establishes a fractional order model based on the inductor current and the capacitor voltage for the multi-level power factor correction converter, including:

[0070] In the nth cycle, if the switch device T1 and the switch device T2 are turned on at the same time, the fractional-order mathematical model is:

[0071]

[0072]

[0073] Where V in is the input voltage on the AC side, L is the filter inductance, i L is the inductor current, C 1 is the output capacitance, R is the output resistance, v C1 is the output capacitor C 1 Voltage, V 0 is the output voltage, α is the order of the inductor current, and β is the order of the capacitor voltage.

[0074] In the nth cycle, if the switch device T1 is turned on and the switch device T2 is turned off, the fractional-order mathematical model is:

[0075]

[0076]

[0077] In the formula, v C2 Represents the output capacitance C 2 voltage.

[0078] In the nth cycle, if the switch device T1 is turned off and the switch device T2 is turned on, the obtained fractional-order mathematical model is:

[0079]

[0080]

[0081] In the nth cycle, if the switch device T1 and the switch device T2 are disconnected at the same time, the obtained fractional-order mathematical model is:

[0082]

[0083]

[0084] According to the fractional-order model, a fractional-order predictive control model is constructed, wherein the cost function of the fractional-order predictive control model includes a fractional-order inductor current model and a fractional-order capacitor voltage model. The steps include:

[0085] In the nth cycle, if the switch device T1 and the switch device T2 are turned on at the same time, the expressions of the inductor current and the capacitor voltage are:

[0086]

[0087]

[0088] In the formula, p 1-α is the discretized d after Oustaloup filtering algorithm 1-α / dt 1-α Fractional differential operator, i L_1 (n+1) represents the inductor current when the switch devices T1 and T2 are turned on at the same time, v C1_1 (n+1) represents the capacitance C when the switch devices T1 and T2 are turned on at the same time 1 Voltage, T S represents the switching cycle, V in (n) represents the rectified input voltage at the beginning of the nth cycle, i L (n) represents the rectified inductor current at the beginning of the nth cycle, v C1 (n) represents the rectified capacitance C at the beginning of the nth cycle 1 Voltage.

[0089] In the nth cycle, if the switch device T1 is turned on and the switch device T2 is turned off, the expressions of the inductor current and capacitor voltage are:

[0090]

[0091]

[0092] In the formula, i L_2 (n+1) represents the inductor current when the switch device T1 is turned on and the switch device T2 is turned off; v C1_2 (n+1) represents the capacitor voltage when the switch device T1 is turned on and the switch device T2 is turned off; vC2 (n) is the capacitance C at the beginning of the nth cycle 2 voltage.

[0093] In the nth cycle, if the switch device T1 is turned off and the switch device T2 is turned on, the expressions of the inductor current and capacitor voltage are:

[0094]

[0095]

[0096] In the formula, i L_3 (n+1) represents the inductor current when the switch device T1 is turned off and the switch device T2 is turned on; v C1_3 (n+1) represents the capacitor voltage when the switch device T1 is turned off and the switch device T2 is turned on.

[0097] In the nth cycle, if the switch device T1 and the switch device T2 are turned off at the same time, the expressions of the inductor current and the capacitor voltage are:

[0098]

[0099]

[0100] In the formula, i L_4 (n+1) is the inductor current when the switch devices T1 and T2 are turned off at the same time; v C1_4 (n+1) is the capacitor voltage when the switch device T1 and the switch device T2 are turned off at the same time;

[0101] Based on the above four operating modes, the cost function of fractional-order predictive control is established as:

[0102]

[0103] In the formula, i ref (n) and v ref (n) represents the reference current signal and the reference voltage signal respectively, and λ is the weight coefficient of predictive control. By calculating the value of the cost function g, the switch state that minimizes the g value is selected as the state of the switch tube in the next cycle. The reference current i ref The value is obtained by the voltage outer loop, the reference voltage v ref The value is half of the output voltage.

[0104] Among them, the outer loop PI control adopts fractional-order PI control to improve the dynamic response and voltage accuracy of the system output side. By establishing a multi-objective fitness function that reflects the power quality on the input side, dynamic response on the output side, and capacitor voltage balance of the three-level power factor correction converter, the inductor current, capacitor voltage, and fractional order of the outer loop PI control and the parameters of PI control are optimized through the gray wolf algorithm to obtain the optimal parameter combination and achieve further performance improvement.

[0105] The gray wolf optimization algorithm is inspired by the gray wolf hierarchy and hunting process. It achieves optimization by simulating the chasing, encircling, harassing and attacking behaviors of different levels of wolves in the total wolf pack. The gray wolf pack is divided into four levels, such as Figure 3 As shown, they are:

[0106] 1) Leadership: The leader of a gray wolf pack is usually composed of a female gray wolf and a male gray wolf. They make decisions on the hunting, daily routines and other behaviors of the entire group. This class is called the α class.

[0107] 2) Second level: The second level of gray wolves assist the leader in decision-making or other wolf activities, and are subordinates of the leader, known as the β level. They are the best candidates for the leader level, and play the role of advisors and team trainers for the leader level.

[0108] 3) Subordinate class: The gray wolf packs in this class must obey the orders of the wolf packs in the first two classes, and are called the δ class. Some of them come from the first two classes and rule the gray wolf packs in the next class; their duties include monitoring the border, ensuring the safety of the wolf pack, assisting in hunting, and taking care of the old, weak, sick, etc.

[0109] 4) The lowest level: This level of gray wolf packs has the lowest status and is called the ω level. The ω level gray wolf packs usually eat last and are unimportant groups.

[0110] The various steps of the Grey Wolf Algorithm are as follows: Figure 4 As shown, the description is as follows:

[0111] Step 1: Surround the prey. The gray wolf's position is updated as follows:

[0112] D=|C·X p (t)-X(t)|;

[0113] X(t+1)=X p (t)-A·D;

[0114] In the formula, A and C are coefficient vectors; Xp is the position of the prey; X is the current position of the gray wolf individual; and t is the number of iterations. The calculation formulas for system vectors A and C are:

[0115] A=2a·r 1-a;

[0116] C=2r 2 ;

[0117] Where r1 and r2 are random numbers in the range [0,1]. The convergence factor a decreases linearly from 2 to 0 as the number of iterations increases. Step 2: Searching for prey. The expression of the gray wolf's search process is:

[0118] D α =|C 1 ·X α (t)-X(t)|;

[0119] D β =|C2·X β (t)-X(t)|;

[0120] D δ =|C 3 ·X δ (t)-X(t)|;

[0121] Where, X α , X β , X δ are the current positions of α wolf, γ wolf and β wolf respectively. C1, C2 and C3 are random perturbations of α wolf, γ wolf and β wolf respectively.

[0122] X 1 =X α (n)-A 1 D α ;

[0123] X 2 =X β (n)-A 2 D β ;

[0124] X 3 =X δ (n)-A 3 D δ ;

[0125]

[0126] Where X(t+1) is the best position of the tth generation.

[0127] Step 3: Attack the prey. If the value of a is smaller, the gray wolf will get closer to the prey. The mathematical model of the gray wolf approaching the prey is:

[0128]

[0129] Where t is the current iteration number, and Tmax is the set maximum iteration number.

[0130] The establishment of a multi-objective fitness function can effectively evaluate the performance indicators of the intelligent optimization algorithm. The objective function used here is the time integral of absolute value (ITAE), and the optimal parameter combination is obtained by minimizing ITAE. The expression of ITAE is:

[0131] ITAE=∫t|e(t)|dt;

[0132] Among them, the expression of e(t) is:

[0133]

[0134] In the formula, I h is the effective value of the total harmonic current on the input side; I 1 is the effective value of the fundamental current on the input side; v C1 is the capacitance C 1 The voltage value, V 0 is the output voltage value, V ref is the reference value of the output voltage; μ and λ represent the inertia weight values ​​in the fitness function, which can reflect the power quality at the input side of the PFC circuit, the balance of the capacitor voltage and the dynamic response of the output voltage.

[0135] Through the real-time optimization of parameter combinations by the Grey Wolf algorithm, the power quality on the input side of the three-level PFC converter, the dynamic response on the output side and the voltage balance between the two output capacitors can be effectively improved.

[0136] The above-described embodiments of the present invention do not limit the protection scope of the present invention.

Claims

1. A fractional-order predictive control method for multi-level power factor correction circuits. It is characterized in that include: The PI control unit performs fractional-order PI control on the multi-level power factor correction converter; wherein the output end of the PI control unit is connected to the input end of the prediction control unit, and the output end of the prediction control unit is connected to the driving interface of the switch tube of the multi-level power factor correction converter; The prediction control unit establishes a fractional-order model based on the inductor current and the capacitor voltage for the multi-level power factor correction converter; according to the fractional-order model, a fractional-order prediction control model is constructed, and the cost function of the fractional-order prediction control model includes a fractional-order inductor current model and a fractional-order capacitor voltage model; the power quality on the input side is improved by adjusting the order of the inductor current, and the voltage balance between the two capacitors on the output side is accelerated by adjusting the order of the capacitor voltage; A multi-objective fitness function is established to reflect the input-side power quality, output-side dynamic response, and capacitor voltage balance of the multi-level power factor correction converter. The order of the inductor current, the order of the capacitor voltage, the fractional order of the fractional-order PI control, and the parameters of the fractional-order PI control are optimized by the Gray Wolf optimization unit to obtain the optimal parameter combination. The multi-level power factor correction converter is a three-level power factor correction converter, the switch tube includes a switch device T1 and a switch device T2, and the prediction control unit establishes a fractional order model based on the inductor current and the capacitor voltage for the multi-level power factor correction converter, including: In the nth cycle, if the switch device T1 and the switch device T2 are turned on at the same time, the fractional-order mathematical model is: Where V in is the input voltage on the AC side, L is the filter inductance, i L is the inductor current, C 1 is the output capacitance, R is the output resistance, v C1 is the output capacitor C 1 Voltage, V 0 is the output voltage, α is the order of the inductor current, and β is the order of the capacitor voltage; In the nth cycle, if the switch device T1 is turned on and the switch device T2 is turned off, the fractional-order mathematical model is: Where V C2 Represents the output capacitance C 2 Voltage; In the nth cycle, if the switch device T1 is turned off and the switch device T2 is turned on, the obtained fractional-order mathematical model is: In the nth cycle, if the switch device T1 and the switch device T2 are disconnected at the same time, the fractional-order mathematical model is: According to the fractional-order model, a fractional-order predictive control model is constructed, wherein the cost function of the fractional-order predictive control model includes a fractional-order inductor current model and a fractional-order capacitor voltage model. The steps include: In the nth cycle, if the switch device T1 and the switch device T2 are turned on at the same time, the expressions of the inductor current and the capacitor voltage are: In the formula, p 1-α is the discretized d after Oustaloup filtering algorithm 1-α / dt 1-α Fractional differential operator, i L_1 (n+1) represents the inductor current when the switch devices T1 and T2 are turned on at the same time, V C1_1 (n+1) represents the capacitance C when the switch devices T1 and T2 are turned on at the same time 1 Voltage, T S represents the switching cycle, V in (n) represents the rectified input voltage at the beginning of the nth cycle, i L (n) represents the rectified inductor current at the beginning of the nth cycle, v C1 (n) represents the rectified capacitance C at the beginning of the nth cycle 1 Voltage.

2. The fractional-order predictive control method for a multi-level power factor correction circuit according to claim 1, It is characterized in that In the nth cycle, if the switch device T1 is turned on and the switch device T2 is turned off, the expressions of the inductor current and capacitor voltage are: In the formula, i L_2 (n+1) represents the inductor current when the switch device T1 is turned on and the switch device T2 is turned off; v C1_2 (n+1) represents the capacitor voltage when the switch device T1 is turned on and the switch device T2 is turned off; v C2 (n) is the capacitance C at the beginning of the nth cycle 2 voltage.

3. The fractional-order predictive control method for a multi-level power factor correction circuit according to claim 2, It is characterized in that In the nth cycle, if the switch device T1 is turned off and the switch device T2 is turned on, the expressions of the inductor current and capacitor voltage are: In the formula, i L_3 (n+1) represents the inductor current when the switch device T1 is turned off and the switch device T2 is turned on; v C1_3 (n+1) represents the capacitor voltage when the switch device T1 is turned off and the switch device T2 is turned on.

4. The fractional-order predictive control method for a multi-level power factor correction circuit according to claim 3, It is characterized in that In the nth cycle, if the switch device T1 and the switch device T2 are turned off at the same time, the expressions of the inductor current and the capacitor voltage are: In the formula, i L4 (n+1) is the inductor current when the switch devices T1 and T2 are turned off at the same time; v C1_4 (n+1) is the capacitor voltage when the switch device T1 and the switch device T2 are turned off at the same time; Based on the above four operating modes, the cost function of fractional-order predictive control is established as: In the formula, i ref (n) and V ref (n) represents the reference current signal and the reference voltage signal respectively, λ is the weight coefficient of predictive control; By calculating the value of the cost function g, the switch state that minimizes the g value is selected as the state of the switch tube in the next cycle, and the reference current i ref The value is obtained by the voltage outer loop, the reference voltage v ref The value is half of the output voltage.

5. The fractional-order predictive control method for a multi-level power factor correction circuit according to claim 1, It is characterized in that In the step of obtaining the optimal parameter combination by optimizing the order of the inductor current, the order of the capacitor voltage, the fractional order of the fractional-order PI control, and the parameters of the fractional-order PI control through the gray wolf optimization unit, the expression of the gray wolf's prey search process is: Where, X α , X γ , X β are the current positions of α wolf, γ wolf and β wolf respectively; C1, C2 and C3 are random perturbations of α wolf, γ wolf and β wolf respectively, and X(t+1) is the best position of the tth generation; The calculation formula for vectors A and C is: In the formula, r 1 、r 2 is a random number in [0,1], and the convergence factor a decreases linearly from 2 to 0 as the number of iterations increases; The performance index of the intelligent optimization algorithm is evaluated by establishing a multi-objective fitness function. The objective function used is the time integral of absolute value (ITAE). The optimal parameter combination is obtained by minimizing ITAE. The expression of ITAE is: ITAE=∫t|e(t)|dt; Among them, the expression of e(t) is: In the formula, I h is the effective value of the total harmonic current on the input side; I 1 is the effective value of the fundamental current on the input side; v C1 is the capacitance C 1 The voltage value, V 0 is the output voltage value, V ref is the reference value of the output voltage; μ and λ represent the inertia weight values ​​in the fitness function, which can reflect the power quality at the input side of the PFC circuit, the balance of the capacitor voltage and the dynamic response of the output voltage.

Images