A Low-Complexity Quasi-Z-Source Inverter Multi-Vector Model Predictive Control Method

Abstract

The present invention discloses a multi-vector model predictive control method for a low-complexity quasi-Z-source inverter, comprising the following steps: calculating the reference value i L1ref (k) of the inductor current by improving the sliding mode control method; obtaining the action time of the through vector by using the deadbeat control method; judging the sector where the optimal virtual switching vector is located according to the angle of the optimal virtual switching vector, and synthesizing the optimal virtual switching vector with the three vectors in this sector to obtain the action time of each vector; controlling each switching device of the quasi-Z-source inverter according to the found optimal switching state. The present invention enables closer approximation to the optimal virtual switching vector through multi-vector control, reduces harmonics, and improves the quality of the grid-connected current; greatly reduces the calculation amount through an efficient sector selection method; improves the sliding mode control to simultaneously control the capacitor voltage and the inductor current to the reference values, saves the cumbersome design of the weighting factor, and avoids the influence of the weighting factor on the control effect; the deadbeat control can reduce the inductor current ripple.

Description

Technical Field

[0001] The present invention relates to the technical field of PWM inversion, and particularly relates to a multi-vector model predictive control method for a quasi-Z-source inverter with low complexity. Background Art

[0002] With the rapid development of electric vehicles, large-scale charging piles have achieved large-scale development and application. At the same time, various large-scale energy storage devices have also developed rapidly. Among them, the inverter is an essential link. Traditional inverters need to add a DC-DC link at the front end to meet the grid connection requirements. This reduces efficiency, increases losses, and requires adding a dead zone link for control, reducing safety and the power quality, which is not beneficial to the power grid. The quasi-Z-source inverter can better solve the above problems.

[0003] There are various control strategies for quasi-Z-source inverters. Among them, the model predictive control strategy is widely used and has good control effects. The model predictive control strategy is a non-linear control strategy. This strategy determines the control quantity through a cost function based on the predicted value, and then controls the quasi-Z-source inverter to achieve the control effect of tracking the given value. The cost function in the traditional model predictive control strategy for quasi-Z-source inverters requires two weighting factors, which makes the design of the weighting factors difficult, and the control effect is also easily affected by the weighting factors. The ripple of the capacitor voltage and inductor current is also large, the harmonic content of the grid-connected current is high, and the current quality is low, which is very limited in application. Summary of the Invention

[0004] The purpose of the present invention is to solve the above-mentioned defects in the prior art and provide a multi-vector model predictive control method for a quasi-Z-source inverter with low complexity. This method first combines an improved sliding mode controller to improve the dynamic response of the quasi-Z-source inverter; then uses the improved sliding mode control method to control the capacitor voltage and inductor current simultaneously, so that there is no need to add weighting factors in the cost function, avoiding the influence of the weighting factors. At the same time, deadbeat control of the inductor current greatly reduces the inductor current ripple. Moreover, using multiple effective vectors in one control cycle reduces the harmonic content and greatly improves the power quality. By using a simple and efficient sector selection method, the calculation amount is greatly reduced.

[0005] The purpose of the present invention can be achieved by adopting the following technical solutions:

[0006] A multi-vector model predictive control method for a quasi-Z-source inverter with low complexity, the model predictive control method comprising the following steps:

[0007] S1. Collect the voltage V of the DC side at regular time intervals dc , and the voltages V of the two capacitors C1 and C2 on the DC side C1 (k), VC2 (k), the inductor current i L1 (k), the AC side voltage e a (k), e b (k), e c (k) and the current i a (k), i b (k), i c (k), and calculate the reference value of the inductor current i through improved sliding mode control L1ref (k);

[0008] S2. Calculate the action time t of the through vector according to the deadbeat control method d ;

[0009] S3. Judge the optimal sector where the optimal virtual switch vector is located according to the angle of the optimal virtual switch vector and the sector where the grid voltage vector is located, and synthesize the optimal virtual switch vector with the three vectors in this optimal sector to obtain the action time of each vector;

[0010] S4. Control each switching device of the Z-source inverter according to the found optimal switch vector sequence and the corresponding action time.

[0011] Furthermore, in the step S1, in order to control the inductor current and the capacitor voltage simultaneously, the reference value of the inductor current is calculated through improved sliding mode control, so that while the inductor current tracks its reference value, the capacitor voltage also tracks its reference value. Among them, the reference value of the inductor current is calculated as follows:

[0012]

[0013] In the formula, e(k) = V C1ref -V C1 (k), e(k - 1) = V C1ref -V C1 (k - 1), A is the sliding mode coefficient, C is the capacitance value of capacitors C1 and C2, the capacitance values of capacitors C1 and C2 are equal, D is the average duty ratio of the through vector, V C1ref is the reference value of the voltage of capacitor C1, V C1 (k) is the voltage of capacitor C1 at time k, V C1 (k - 1) is the voltage of capacitor C1 at time k - 1, P ref is the reference value of the active power, r s = r1 / T s , r1, r2 are the first and second correction coefficients to make the capacitor voltage accurately track its reference value, T s is the sampling period, y2(k - 1) = r2(V C1ref -V C1 (k - 1))t, t is the time point.

[0014] Further, in step S2, in order to enable the inductor current to track its reference value while having a small inductor current ripple, the duty ratio of the direct vector is calculated using the deadbeat control method, and then the control effect is achieved. The calculation process of the duty ratio d of the direct vector is as follows:

[0015]

[0016] In the formula, d is the duty ratio of the direct vector, i L1ref (k) is the reference value of the inductor current, i L1 (k) is the inductor current, l nst is the change rate of the inductor current when not in direct connection, l st is the change rate of the inductor current when in direct connection;

[0017] Then, the action time t of the direct vector is calculated d = d·T s , T s is the sampling period.

[0018] Further, the control effect is achieved on the DC side through the direct vector. On the AC side, the optimal sector where the optimal virtual switching vector is located needs to be judged by using the angle of the optimal virtual switching vector and the sector where the grid voltage vector is located. Then, the action times of the three vectors in the optimal sector are calculated as follows:

[0019]

[0020]

[0021] t3 = T s1 - t1 - t2

[0022] In the formula, t1, t2, and t3 are the action times of the three vectors respectively, Q1′, Q2′, and Q3′ are the change rates of the reactive power of the three vectors respectively, P1′, P2′, and P3′ are the change rates of the active power of the three vectors respectively, ΔP = P ref - P(k), ΔQ = Q ref - Q(k), T s1 = (1 - d)T s , P(k) and Q(k) are the actual values of the active power and reactive power at time k respectively, T s1 is the time when not in direct connection, T s is the sampling time, P ref , Q ref are the reference values of the active power and reactive power respectively.

[0023] The tangent value of the angle of the optimal virtual switching vector is as follows:

[0024]

[0025] In the formula, θ u is the angle of the optimal virtual switching vector, and θ e , θ are respectively the angle of the grid voltage vector and the included angle between the grid voltage vector and the optimal virtual switching vector.

[0026] Furthermore, according to the value calculated by the above formula, quickly determine the optimal sector where the optimal virtual switching vector is located, and then select the optimal switching vector sequence and action time, and then control the quasi-Z-source inverter to realize the function of the quasi-Z-source inverter.

[0027] The present invention has the following advantages and effects compared with the prior art:

[0028] (1) The present invention uses multiple effective vectors within one control period, making the optimal virtual switching vector closer to the actual situation, thereby reducing the harmonic content and greatly improving the power quality. At the same time, a simple and efficient sector selection method is proposed. Instead of optimizing one by one to determine the optimal sector, the optimal sector where it is located can be directly obtained by looking up the table according to the angle of the optimal virtual switching vector, greatly reducing the calculation amount;

[0029] (2) The present invention uses an improved sliding mode control method to control the inductor current and capacitor voltage simultaneously. The relationship between the inductor current and capacitor voltage is established by using power balance, and a sliding mode is added. By calculating the reference value of the inductor current and combining with the deadbeat control method, the inductor current at the next moment is made equal to the calculated reference value of the inductor current, and then the action time of the through vector is obtained, and the through vector is applied to the switching tube, so that the inductor current and capacitor voltage can respectively track their reference values simultaneously. As long as the inductor current tracks the reference value in this control method, the capacitor voltage will also track the reference value, which makes the cost function omit the weight factor, avoiding the design trouble of the weight factor and its influence on the control effect;

[0030] (3) The present invention uses the deadbeat control method to make the inductor current track the reference value of the inductor current within one control period, while the traditional method requires multiple periods to complete the tracking, significantly improving the inductor current ripple;

[0031] (4) The present invention combines an improved sliding mode controller, and by designing a corresponding sliding mode surface, the dynamic response of the quasi-Z-source inverter can be improved. Description of the Drawings

[0032] The drawings described herein are used to provide a further understanding of the present invention, and constitute a part of this application. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0033] Figure 1 It is the control block diagram of a low - complexity quasi - Z - source inverter multi - vector model predictive control method proposed in the present invention;

[0034] Figure 2 It is the circuit topology diagram of the quasi - Z - source inverter in the present invention;

[0035] Figure 3 It is the equivalent circuit diagram of the quasi - Z - source inverter in the through - state and non - through - state; where, Figure 3 (a) is the equivalent circuit diagram of the quasi - Z - source inverter in the non - through - state; Figure 3 (b) is the equivalent circuit diagram of the quasi - Z - source inverter in the through - state;

[0036] Figure 4 It is the grid - side current waveform diagram of a low - complexity quasi - Z - source inverter multi - vector model predictive control method proposed in the present invention;

[0037] Figure 5 It is the schematic diagram of the output active power and reactive power of a low - complexity quasi - Z - source inverter multi - vector model predictive control method proposed in the present invention;

[0038] Figure 6 It is the inductor current waveform diagram of the DC side of a low - complexity quasi - Z - source inverter multi - vector model predictive control method proposed in the present invention;

[0039] Figure 7 It is the schematic diagram of the DC - side capacitor voltage of a low - complexity quasi - Z - source inverter multi - vector model predictive control method proposed in the present invention;

[0040] Figure 8 It is the optimal sector judgment diagram of a low - complexity quasi - Z - source inverter multi - vector model predictive control method proposed in the present invention; where, Figure 8 (a) is the schematic diagram when the grid voltage vector is in the shaded area; Figure 8 (b) is the schematic diagram when the grid voltage vector is in another shaded area. Detailed implementation manners

[0041] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0042] Embodiment 1

[0043] Figure 2It is a topological diagram of a quasi-Z-source inverter circuit disclosed in this embodiment. The quasi-Z-source inverter is composed of a quasi-Z-source and an inverter. The quasi-Z-source part is composed of inductors L1 and L2, capacitors C1 and C2, and diode D. The structure of the inverter is a basic three-phase bridge inverter. Inductors L1 and L2 are connected in series through diode D, and inductor L1 is connected to the DC-side power supply V dc , capacitor C2 is connected in parallel with diode D and inductor L2, capacitor C1 is connected in parallel with inductor L1, diode D, and the DC-side power supply. Inductor L2 and capacitor C1 are connected to the PN terminals of the inverter. The AC side of the inverter is connected to the power grid through filter inductor L f . The model predictive control method includes the following steps:

[0044] S1. Collect the voltage V of the DC side, the two capacitor voltages V dc of the DC side, C1 (k), V C2 (k), the inductor current i L1 (k), the AC-side voltages e a (k), e b (k), e c (k), and the currents i a (k), i b (k), i c (k) at regular time intervals, and calculate the reference value i L1ref (k) of the inductor current through improved sliding mode control;

[0045] S2. Calculate the action time t of the through vector according to the deadbeat control method d ;

[0046] S3. Judge the optimal sector where the optimal virtual switch vector is located according to the angle of the optimal virtual switch vector and the sector where the grid voltage vector is located, and then synthesize the optimal virtual switch vector with the three vectors in this sector to obtain the action time of each vector;

[0047] S4. Control each switching device of the quasi-Z-source inverter according to the found optimal switch vector sequence and the corresponding action time.

[0048] Specifically, it is described as follows:

[0049] As Figure 2 shown, the AC side of the quasi-Z-source inverter is directly connected to the power grid through filter inductor L f . The mathematical model of the AC side in the two-phase stationary αβ coordinate system is as follows:

[0050]

[0051] Among them, e α , e β , iα and i β and u α and u β represent the grid phase voltage, line current, and inverter switch voltage in the stationary two-phase αβ coordinate system. L f is the value of the grid-side filter inductor.

[0052] In a three-phase balanced system, the following relationship holds:

[0053]

[0054] where ω is the grid-side angular frequency.

[0055] Based on the instantaneous power theory in the αβ coordinate system, the expressions for the active and reactive powers injected by the inverter into the grid are as follows:

[0056]

[0057] Combining the above formulas, in the stationary two-phase αβ coordinate system, the expressions for the derivatives of active and reactive powers are as follows:

[0058]

[0059] where P′ represents the derivative of active power and Q′ represents the derivative of reactive power.

[0060] Then, the power at time k + 1 is as follows:

[0061]

[0062] where T s is the sampling period, P(k) and Q(k) are the active and reactive powers at time k, and P(k + 1) and Q(k + 1) are the predicted values of the active and reactive powers at time k + 1.

[0063] The quasi-Z-source inverter has two states, one is the through state and the other is the non-through state. Their equivalent circuits are as Figure 3 shown, where L1 = L2 = L and C1 = C2 = C.

[0064] 1) Non-through state: According to Figure 3 the non-through state equivalent circuit shown in (a), the current through inductor L1 is

[0065]

[0066] where L is the inductance value of the inductor; V dc and V C1 and i L1 are the DC-side voltage, the voltage across capacitor C1, and the current through inductor L1, respectively.

[0067] 2) Through - conduction state: As shown in Figure 3 (b), the diode is turned off during the through - conduction state. Since the current of inductor L1 is the same as that of inductor L2, the current of inductor L1 can be expressed as

[0068]

[0069] Based on the state - space average model of the quasi - Z - source inverter, we will get

[0070]

[0071] where D represents the average through - conduction duty cycle.

[0072] Next, ignoring the inductor current ripple and capacitor voltage ripple, assuming that the inductor current is controlled at its reference value, the relationship between the inductor current and capacitor voltage is derived according to power balance as

[0073]

[0074] where I L1ref is the reference current of inductor L1; C is the capacitance of capacitors C1 and C2; P represents the active power injected into the grid; V C1 and V C2 represent the average voltages of capacitors C1 and C2 within one control period.

[0075] Assuming that P is equal to its reference value, substituting Equation (8) into Equation (9), Equation (9) can be rewritten as

[0076]

[0077] Based on the above formula, the sliding mode surface is designed as

[0078]

[0079] where A is the sliding - mode coefficient; V C1ref is the reference voltage of capacitor C1, which can be expressed as

[0080] V C1ref =0.5(V dc +V PNref ) (Equation 12)

[0081] where V PNref is the reference value of the peak voltage at the PN terminal.

[0082] By solving the equation s = 0, the reference value of the inductor current can be calculated as

[0083]

[0084] Among them, P ref is the reference value of the active power, and D can be calculated by the following formula

[0085]

[0086] To compensate for the errors caused by assumptions and parameter variations, a correction function is added to formula (13), and the correction function is expressed as

[0087]

[0088] Among them, y is the correction function; r1 and r2 are correction coefficients.

[0089] Discretize formula (13) and formula (15) to obtain the final reference value of the inductor current of inductor L1, expressed as

[0090]

[0091] In the formula, e(k) = V C1ref - V C1 (k), e(k - 1) = V C1ref - V C1 (k - 1), A is the sliding mode coefficient, C is the capacitance value of capacitor C1, D is the average duty cycle of the direct-through vector, V C1ref is the reference value of the voltage of capacitor C1, V C1 (k) is the voltage of capacitor C1 at time k, V C1 (k - 1) is the voltage of capacitor C1 at time k - 1, P ref is the reference value of the active power, r s = r1 / T s , r1, r2 are the first and second correction coefficients that enable the capacitor voltage to accurately track its reference value, T s is the sampling period, y2(k - 1) = r2(V C1ref - V C1 (k - 1))t, t is the time point.

[0092] Different from the weight factors in the cost function, these coefficients do not affect the steady-state performance. They only affect transient processes such as transient time and overshoot.

[0093] If the average inductor current I L1 can track the reference value of the inductor current calculated by formula (16), then the capacitor voltage can be controlled to V C1ref , and thus the weight factor is not required, avoiding the trouble of the weight factor.

[0094] Next, deadbeat control is used for the inductor current to obtain an optimized direct-conduction duty ratio. According to Equation (6) and Equation (7), the slopes of the inductor current in the non-direct-conduction state and the direct-conduction state can be expressed as

[0095]

[0096] where l nST and l ST are the slopes of the inductor current in the non-direct-conduction state and the direct-conduction state, respectively.

[0097] Based on Equation (17), the predicted inductor current at the next moment can be calculated as

[0098] i L1 (k + 1) = i L1 (k) + l ST T s d + l nST T s (1 - d) (Equation 18)

[0099] where d is the expected direct-conduction duty ratio, i L1 (k + 1) is the inductor current at the (k + 1)-th moment, and i L1 (k) is the inductor current at the k-th moment.

[0100] The expected direct-conduction duty ratio can achieve deadbeat control of the inductor current, that is

[0101] i L1 (k + 1) = i L1ref (k) (Equation 19)

[0102] By solving Equation (19), the optimized direct-conduction duty ratio is obtained as

[0103]

[0104] According to the above formula, the action time of the direct-conduction vector can be calculated

[0105] t d = d·T s (Equation 21)

[0106] In this way, the inductor current reference value can be accurately tracked within one control period.

[0107] Next, a simple and efficient method is used to select the optimal sector to reduce the calculation amount. Assuming that the grid voltage in the αβ plane is

[0108] e α = Esinωt, e β = -Ecosωt (Equation 22)

[0109] Among them, E represents the peak value of the grid voltage.

[0110] To achieve unity power factor, the reference value of the output reactive power is set to 0. Therefore, when the optimal virtual switching vector is applied, the grid current is in phase with the grid voltage. Using the phasor method, the relationship between the grid-side current and the grid voltage can be expressed as

[0111] U∠θ=jωL f I∠0°+E∠0° (Equation 23)

[0112] In the formula, U is the peak value of the output voltage of the quasi-Z-source inverter; I is the peak value of the grid-side current; θ is the angle by which the output voltage of the quasi-Z-source inverter leads the grid voltage. From Equation (23), the tangent value of θ can be derived as

[0113]

[0114] In a balanced three-phase system, when unity power factor is satisfied, the output power of the quasi-Z-source inverter can be calculated as

[0115] P=1.5EI (Equation 25)

[0116] According to Equation (25), it can be obtained that

[0117]

[0118] First, it is necessary to find the sector of the grid voltage. The included angle of the grid voltage in the αβ coordinate system satisfies

[0119]

[0120] Among them, θ e is the angle of the grid voltage in the αβ plane.

[0121] The sector of the grid voltage can be found through Equation (27). For example, when the value of Equation (27) is between 0 and , the sector of the grid voltage is Sector I or Sector IV. If the α component of the grid voltage is greater than 0, the sector of the grid voltage is Sector I; if the α component of the grid voltage is less than 0, the sector of the grid voltage is Sector IV. By calculating the value of Equation (27) and e α , we can determine the sector of the grid voltage through Table 1.

[0122] Table 1. Sector Table of Grid Voltage

[0123]

[0124] Then, the angle of the output voltage of the quasi-Z-source inverter in the αβ plane is

[0125] θu = θ e + θ (Equation 28)

[0126] Then, its tangent value can be obtained

[0127]

[0128] Finally, the optimal sector can be determined from Equation (29) and the sector of the grid voltage. For example, when the value of Equation (29) is between 0 and and, the possible optimal sectors are as Figure 8 shown. In the phasor domain, for the inverter, θ is between 0° and 90°. Therefore, if the grid voltage vector is located in the Figure 8 shaded area of (a), the optimal sector is Sector I; if the grid voltage vector remains in the Figure 8 shaded area of (b), the optimal sector is Sector IV. Therefore, by calculating Equation (29), the optimal sector of the optimal virtual switching vector can be determined through Table 2

[0129] Table 2. Optimal Sector Table

[0130]

[0131] Next, select the corresponding 3 vectors in the optimal sector and insert the direct vector into the zero vector. These 4 vectors are combined into a vector called the optimal vector sequence. To minimize the switching losses during the entire control period, a single-phase direct vector is used, following the principle that only one switch is allowed to change at a time. All possible optimal vector sequences are listed in Table 3

[0132] Table 3. Optimal Vector Sequence Table

[0133] Sector Optimal vector sequence Ⅰ OOO-ZOO-POO-PPO II OOO-OZO-OPO-PPO III OOO-OZO-OPO-OPP IV OOO-OOZ-OOP-OPP V OOO-OOZ-OOP-POP VI OOO-ZOO-POO-POP

[0134] After selecting the vector sequence according to Table 3, the next step is to calculate their action time within one control period. The inductor current and capacitor voltage can be adjusted simultaneously by using deadbeat control for the inductor current, which is independent of the active vector and the zero vector. In this case, the cost function only includes the active and reactive powers injected into the grid, expressed as

[0135] g = [P ref - P(k + 1)] 2 + [Q ref - Q(k + 1)] 2 (Equation 30)

[0136] where g is the cost function for calculating the action time; Q refis the reactive power reference value, usually set to 0 to achieve unity power factor; P(k + 1) and Q(k + 1) are the predicted values of active and reactive power at the (k + 1)-th moment, and the expressions are as follows

[0137]

[0138] where P(k) and Q(k) are the active and reactive powers at the k-th moment; Pi′ and Qi′ are the derivatives of the active and reactive powers of the i-th vector respectively; t i is the action time of the i-th vector.

[0139] To obtain the optimal action time, the present invention uses an unconstrained optimization method, which can be expressed as

[0140]

[0141] where t1, t2, and t3 are the action times of the first, second, and third vectors respectively.

[0142] By solving formula (32), the optimal action time of each vector can be calculated as

[0143]

[0144] In the formula, t1, t2, and t3 are the action times of the three vectors respectively, Q1′, Q2′, and Q3′ are the change rates of the reactive powers of the three vectors respectively, P1′, P2′, and P3′ are the change rates of the active powers of the three vectors respectively, ΔP = P ref - P(k), ΔQ = Q ref - Q(k), T s1 = (1 - d)T s , P(k) and Q(k) are the actual values of the active and reactive powers at the k-th moment respectively, T s1 is the non-direct-through time, T s is the sampling time, P ref , Q ref are the reference values of the active and reactive powers respectively.

[0145] Finally, according to the obtained optimal switching vector sequence and their action times, each power switch tube is controlled by DSP to achieve the preset effect.

[0146] Embodiment 2

[0147] In Figure 1 a control block diagram of a low-complexity quasi-Z-source inverter multi-vector model predictive control method is given, and the control method is summarized as follows: First, an improved sliding mode controller is used to generate the inductor current reference value i L1ref(k). Secondly, the obtained inductor current reference value is used as the reference value of the deadbeat control method to calculate the action time of the direct-through vector, so that the inductor current can accurately track the reference value within one control period, achieving the effect of simultaneously controlling the inductor current and the capacitor voltage. Then, the tangent value of the angle of the optimal virtual switching vector and the sector where the grid voltage is located are used to determine the optimal sector where it is located. Furthermore, the optimal vector sequence of this sector is selected and its action time is calculated. Finally, the power switching devices of the Z-source inverter are controlled according to the optimal switching vector sequence and their action times, so that the active power and reactive power track the active power reference value and the reactive power reference value, and the power quality is very good. The ripples of the inductor current and capacitor voltage on the DC side are also greatly improved.

[0148] A low-complexity quasi-Z-source inverter multi-vector model predictive control system is constructed using the simulation software MATLAB / Simulink for simulation verification, and the parameters are shown in Table 4.

[0149] Table 4. Simulation parameter table

[0150] Parameter Value <![CDATA[Filter inductor L f > 6mH Quasi-Z-source capacitor C 1000uF Quasi-Z-source inductor L 4mH <![CDATA[Peak value e of the grid phase voltage max > 100V DC-side input given voltage value Vdc 150V DC-side voltage peak reference value Vpn 300V Active power reference value Pref 800W - 1200W Reactive power reference value Qref 0Var Grid frequency f 50Hz Sampling frequency fs 20kHz Sliding mode coefficient A 10 Correction coefficient r1 0.2 Correction coefficient r2 0.5

[0151] In Figure 4 , the grid-side current waveform diagram of a low-complexity quasi-Z-source inverter multi-vector model predictive control method is given; in Figure 5 , the output active power and reactive power of a low-complexity quasi-Z-source inverter multi-vector model predictive control method are given; in Figure 6 , the inductor current on the DC side of a low-complexity quasi-Z-source inverter multi-vector model predictive control method is given; in Figure 7 , the DC-side capacitor voltage of a low-complexity quasi-Z-source inverter multi-vector model predictive control method is given. In Figure 8 , the optimal sector judgment diagram of a low-complexity quasi-Z-source inverter multi-vector model predictive control method is given. It can be seen from the above simulation results that a low-complexity quasi-Z-source inverter multi-vector model predictive control method uses an improved sliding mode control method and omits the weight factor. It not only avoids the design difficulty of the weight factor, but also the control effect is not affected by the weight factor. At the same time, the ripples of the inductor current and capacitor voltage are greatly improved. Moreover, a simple and efficient sector selection method is used to greatly reduce the calculation amount. Using multiple vectors within one cycle reduces the harmonic content and greatly improves the power quality. It can completely track the active power reference value and the reactive power reference value and has a very fast dynamic response.

[0152] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be equivalent replacement methods and are all included in the protection scope of the present invention.

Claims

1. A low-complexity quasi-Z-source inverter multi-vector model predictive control method, characterized in that The model predictive control method includes the following steps: S1. Collect the DC-side voltage V at certain time intervals dc , the voltages V C1 (k) and V C2 (k) of the two DC-side capacitors C1 and C2, the inductor current i L1 (k), the AC-side voltages e a (k), e b (k), e c (k) and the currents i a (k), i b (k), i c (k), and calculate the reference value i L1ref (k) of the inductor current through improved sliding-mode control; S2. Calculate the action time t of the direct vector according to the deadbeat control method d In the step S2, the process of calculating the duty cycle d of the direct vector by the deadbeat control method is as follows: where d is the duty cycle of the direct vector, and i L1ref (k) is the reference value of the inductor current, and i L1 (k) is the inductor current, and l nst is the change rate of the inductor current when not in direct connection, and l st is the change rate of the inductor current when in direct connection; Then calculate the action time of the direct-through vector , T s is the sampling period; S3. Determine the optimal sector where the optimal virtual switching vector is located according to the angle of the optimal virtual switching vector and the sector where the grid voltage vector is located, and synthesize the optimal virtual switching vector with the three vectors in this optimal sector to obtain the action time of each vector. The method for determining the optimal sector where the optimal virtual switching vector is located by using the angle of the optimal virtual switching vector and the sector where the grid voltage vector is located in step S3 is as follows: In the formula, is the angle of the optimal virtual switching vector, , are the grid voltage vector angle and the included angle between the grid voltage vector and the optimal virtual switching vector, respectively; The action times of the three vectors in the optimal sector are as follows: Wherein, t1, t2, and t3 are the acting times of three vectors respectively, , , are the change rates of the reactive power of three vectors respectively, , , are the change rates of the active power of three vectors respectively, , , , P(k) and Q(k) are the actual values of the active power and the reactive power at the kth moment respectively, is the non-direct-through time, is the sampling time, P ref , Q ref are the reference values of the active power and the reactive power respectively; S4. Control each switching device of the Z-source inverter according to the found optimal switching vector sequence and the corresponding action time.

2. A low-complexity quasi-Z-source inverter multi-vector model predictive control method according to claim 1, characterized in that In the step S1, the reference value i L1ref (k) of the inductor current is calculated by improving the sliding mode control, where the reference value i L1ref (k) of the inductor current is calculated as follows: where \(e(k) = V C1ref -V C1 (k)\), \(e(k - 1)=V C1ref -V C1 (k - 1)\), \(A\) is the sliding mode coefficient, \(C\) is the capacitance value of capacitor \(C1\), \(D\) is the average duty cycle of the direct vector, \(V C1ref is the reference value of the voltage of capacitor \(C1\), \(V C1 (k)\) is the voltage of capacitor \(C1\) at time \(k\), \(V C1 (k - 1)\) is the voltage of capacitor \(C1\) at time \(k - 1\), \(P ref is the reference value of the active power, \(r s = r1 / T s \), \(r1\) and \(r2\) are the first and second correction coefficients for accurately tracking the corresponding reference values of the capacitor voltage, \(T s is the sampling period, \(y2(k - 1)=r2(V C1ref -V C1 (k - 1))t\), \(t\) is the time point.

3. A low-complexity quasi-Z-source inverter multi-vector model predictive control method according to claim 1, characterized in that, In step S4, control the Z-source inverter according to the optimal switching vector sequence and action time selected according to the optimal sector.

Images