Harmonic suppression and optimal SVPWM (Space Vector Pulse Width Modulation) method of single-phase three-level PWM rectifier
By adopting a singular regulator and optimized SVPWM algorithm in a single-phase three-level PWM rectifier, combining three-stage and five-stage modulation, the problems of midpoint potential fluctuation and harmonic suppression are solved, and low harmonic characteristics and high power are achieved The factor reduces switching losses and improves the stability and efficiency of the system.
Patent Information
- Application Number
- CN202410957364.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-15
AI Technical Summary
Traditional single-phase three-level PWM rectifiers have limitations in harmonic suppression and switching loss control, with midpoint potential fluctuations and high-frequency switching operation leading to problems such as unstable output voltage, electromagnetic interference and reduced power factor.
Using singular regulator and optimized SVPWM algorithm, the basic vector and sector division are designed, combined with three-stage and five-stage SVPWM modulation, the vector action time and order are adjusted, and the midpoint potential balance and harmonic suppression are achieved, and dynamic adjustment is achieved. Switching frequency to optimize switching losses.
Effectively control the midpoint potential fluctuation range, significantly reduce the current harmonic content on the grid side, improve the system power factor and overall efficiency, reduce switching losses, and improve system reliability.
Smart Images

Figure SMS_3 
Figure SMS_5 
Figure SMS_6
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of power electronics, and in particular to a harmonic suppression and optimal SVPWM modulation method of a single-phase three-level PWM rectifier. Background Art
[0002] Single-phase three-level PWM rectifiers have been widely used in the field of power electronics due to their advantages such as high efficiency, low harmonics and high power factor, and have become the mainstream technology, especially in high-voltage and high-power applications such as power transmission, renewable energy and industrial drives. However, with the expansion of application areas and the improvement of technical requirements, traditional modulation algorithms have gradually shown their limitations in harmonic suppression and switching loss control.
[0003] The traditional SVPWM algorithm mainly achieves the modulation of the rectifier by adjusting the switching state and controlling the voltage vector, but there are several significant problems in practical applications. First, the problem of midpoint potential fluctuation is particularly prominent in single-phase three-level rectifiers, which can easily lead to unstable output voltage, thus affecting system performance. Secondly, high-frequency switching operations not only increase the loss of switching devices, but may also cause electromagnetic interference and affect the reliability of the system. In addition, the increase in the harmonic content of the grid-side current will reduce the power factor of the system and have an adverse effect on the power grid.
[0004] In view of the above problems, some improvement methods have been proposed in the prior art. For example, by optimizing the switching sequence and adjusting the switching frequency, the midpoint potential can be controlled and harmonics can be reduced to a certain extent. However, these methods have limited effects in practical applications, and it is difficult to take into account the midpoint potential balance, harmonic suppression and switching loss optimization at the same time. Summary of the invention
[0005] The technical problem to be solved by the present invention is to provide a harmonic suppression and optimal SVPWM modulation method for a single-phase three-level PWM rectifier, which effectively suppresses harmonics and optimizes switching losses.
[0006] The present invention is achieved in this way.
[0007] A harmonic suppression and optimal SVPWM modulation method for a single-phase three-level PWM rectifier, comprising:
[0008] The basic vector and sector division of the single-phase three-level rectifier are designed. The vector allocation of the SVPWM algorithm is optimized through a singular regulator and harmonic suppression technology is adopted. The optimized SVPWM algorithm is used for regulation. When the control input is close to the boundary, a three-segment SVPWM is used; in other cases, a five-segment SVPWM is used.
[0009] Furthermore, the basic vectors and sector division of the single-phase three-level rectifier are designed including:
[0010] The single-phase three-level NPC pulse rectifier has 9 switching states, corresponding to 9 voltage vectors. According to the input voltage u under different switching states ab The size of V can be divided into three categories: large vector, small vector and zero vector. pn : Contains the positive vector V pn (1,-1) and negative large vector V pn (-1,1) two vectors, when this type of vector acts, the input voltage u ab The amplitudes are u dc * and -u dc * ; Small vector: There are 4 small vectors in total, which are divided into two groups of redundant vectors, among which V po (1,0) and V on (0,-1) are mutually redundant positive small vectors V po ; V no (-1,0) and V op (0,1) are mutually redundant negative small vectors. When this type of vector acts, the input voltage u ab The amplitudes are and Zero vector: The single-phase three-level NPC rectifier contains three zero vectors, namely V pp (1,1), V oo (0,0) and V nn (-1,-1), when this type of vector acts, u ab The amplitudes are all 0;
[0011] The 9 voltage vectors divide the αβ plane into 4 sectors with equal spacing. ab is the modulation signal, V ref is the reference voltage vector, the reference voltage vector V ref It rotates counterclockwise in the αβ plane at an angular velocity ω, with an initial phase angle of θ, where the reference voltage vector V ref The component on the α axis is u ab , define V ref The amplitude is V, then the input voltage u ab The expression is as follows:
[0012] u ab =Vcos(ωt-θ)
[0013] According to u ab The size of V ref It is divided into 4 sectors, and the division rules are as follows:
[0014] (1) Sector I: 0.5u dc <u ab <udc ;
[0015] (2) Sector II: 0<u ab <0.5u dc ;
[0016] (3) Sector III: -0.5u dc <u ab <0;
[0017] (4) Sector IV: -u dc <u ab <-0.5u dc .
[0018] Furthermore, the three-stage SVPWM achieves the balance of the midpoint potential by adjusting the vector action time, and the vector action order is V a →V b →V a or V b →V a →V b , in one switching cycle, V a Select V pn , V b Selected as redundant vector V po (1,0) and V op (0,1).
[0019] Furthermore, the vector action timing of the five-segment SVPWM midpoint potential control algorithm is: T b+ / 2→T a / 2→T b / 2→T a / 2→T b+ / 2 or T b / 2→T a / 2→T b+ / 2→T a / 2→T b / 2, where T a and T b They represent the action time of the large vector and the small vector in a switching cycle, respectively. Tb+ and Tb- are the redundant small vector V b+ and V b- The action time in one switching cycle, where V b+ Select V in sectors Ⅰ and Ⅱ po (1,0), select V in sectors III and IV no (-1,0); V b- Select V in sectors Ⅰ and Ⅱ on (0,-1), select V in sectors III and IV op (0,1).
[0020] Furthermore, the midpoint potential control function of the three-stage SVPWM is as follows:
[0021] Δ=sign(u1-u2)*sign(i s )
[0022] According to the reference voltage vector V ref The sector is different. When Δ=1, select the small vector V on (0,-1) or V op (0,1); when Δ=-1, select the small vector V po (1,0) or V no (-1,0).
[0023] Furthermore, the action time of the small vector is written as:
[0024]
[0025] When five-segment SVPWM modulation is adopted, k Δ ∈(-1,1); when three-stage SVPWM modulation is adopted, k Δ = ±1, the DC side voltage state equation is as follows:
[0026] The above equation can be written as the following state equation:
[0027]
[0028] The performance indicators are selected as follows:
[0029]
[0030] When J is the minimum value, the optimal control is achieved. The original singular regulator is converted into an equivalent standard regulator through linear transformation. After solving the standard regulator, it is converted into the solution of the original singular regulator.
[0031] x1(t)=x(t)-B(t)u1(t)
[0032]
[0033] Taking the derivative of both sides, we get: in
[0034] By linearly transforming the original performance index, we can get:
[0035]
[0036] H(t)=Q(t)B(t); R(t)=B T (t)Q(t)B(t)
[0037] right The integrand is formulated as x1 T (t)Q(t)x1(t)+2x1 T (t)H(t)u1(t)+u1 T (t)R(t)u1(t)
[0038] =x1 T [Q(t)+ΔQ(t)]x1(t)+[u1(t)+Δu1(t)] T R(t)[u1(t)+Δu1(t)]
[0039] Assuming R(t), multiply the right side of the above equation to get:
[0040] Δu1(t)=R -1 (t)H(t)x1(t)
[0041] ΔQ(t)=-H(t)R -1 (t)H T (t)
[0042] The performance index J is written as:
[0043]
[0044] Where: Q1(t) = Q(t) - H T (t)R -1 (t)H(t);Q1(t) is non-negative and definite,
[0045] u2(t)=u1(t)+R -1 (t)H(t)x1(t),
[0046] R(t) and Q1(t) are symmetrical positive definite matrices and symmetric non-negative definite matrices respectively. Assuming that R(t)>0, the non-negative definiteness of Q1(t) is proved.
[0047] x T (t)Q(t)x(t)=x1 T (t)Q(t)x1(t)+u1(t)+R -1 (t)H(t)x1(t)
[0048] +[u1(t)+R -1 (t)H(t)x1(t)] T R(t)[u1(t)+R -1 (t)H(t)x1(t)]≥0
[0049] Selection: u1(t) = -R -1 (t)H(t)x(t)
[0050] Get x1 T (t)Q(t)x1(t)≥0
[0051] Therefore Q1(t)≥0;
[0052] This modified state equation is further transformed into u2(t)=u1(t)+R -1 (t)H(t)1x(t) is substituted into the formula In the result:
[0053]
[0054] Where: A1(t) = A(t) - B1(t)R -1 (t)H T (t)
[0055] Without considering the control constraints |u (t) The modified singular regulator when |≤1 is:
[0056] System state equation:
[0057]
[0058] Performance indicators:
[0059]
[0060] When the above formula takes the minimum value, the optimal control u2 is obtained * (t) , u2 * (t) The expression is as follows:
[0061] u2 * (t)=-K(t)x1(t)
[0062] Where K(t) = R -1 (t)B1 T (t)P(t)
[0063] where P(t) is a symmetric non-negative definite matrix that satisfies the following Riccati differential equation
[0064]
[0065] Its boundary conditions are:
[0066] P(t)=0
[0067] Will H(t)=Q(t)B(t); R(t)=B T (t)Q(t)B(t) and u2(t)=u1(t)+R-1 Substitute (t)H(t)x1(t) into u2 * (t)=-K(t)x1(t) to obtain u1 * (t)=-K1(t)x1(t)
[0068] Where K1(t) = R -1 (t)[B1 T (t)P(t)+H T (t)]
[0069] K1(t) in the above formula is the feedback gain matrix of the modified singular regulator;
[0070]
[0071] because:
[0072] K1 T (t)B T (t)Q(t)K1(t)=[P(t)B1(t)+H(t)]R -1 (t)B T (t)Q(t)B(t)R -1 (t)[B T (t)P(t)+H T (t)]
[0073] =P(t)B1(t)R -1 (t)H T (t)+H T (t)R -1 (t)B T (t)P(t)+P(t)B1(t)R -1 (t)B T (t)P(t)+H(t)R -1 (t)H T (t)
[0074] The Riccati equation can be written as:
[0075]
[0076] The optimal performance index is:
[0077] J * =x1 T (0)P(0)x1(0)
[0078] For the modified singular regulator, its variables should satisfy the following relationship:
[0079] P(t)B(t)=0
[0080] K1(t)B(t)=I
[0081] J * =x1 T (0)P(0)x1(0)
[0082] K1(t)x(t)=0
[0083] After obtaining the solution of the modified singular regulator, it is necessary to convert it to the solution of the original singular regulator when it is not constrained;
[0084] Since x1(t)=x(t)-B(t)u1(t) holds, the following boundary constraints need to be met during the conversion:
[0085] x(0)=x1(0)+B(0)u1 * (0), x(0) is the initial state of the original singular optimal regulator;
[0086] u1 * (t) = -K1(t)x1(t) Substituting into the above formula, the boundary constraint condition is expressed as:
[0087] x(0)=x1(0)-B(0)K1(0)x1(0)
[0088] The solution of K1(t)x(t)=0 is proved to be a modified singular optimal regulator, and this solution can be transformed into the solution of the original singular optimal regulator;
[0089] The optimal control is:
[0090]
[0091] u(t): control input, A(t): system matrix, B(t): control matrix, x(t): system state vector, P(t): Riccati matrix, Q(t): weight matrix, H(t): matrix, defined as H(t) = Q(t)B(t), R(t): weight matrix, defined as R(t) = B T (t)Q(t)B(t), Δu(t): control input correction, ΔQ(t): matrix, defined as ΔQ(t) = -H T (t)R -1 (t)H(t), ΔR(t): matrix, defined as ΔR(t)=R(t)>0, Q1(t): symmetric non-negative definite matrix, defined as Q1(t)=Q(t)-H T (t)R -1 (t)H(t), u1(t): control input, defined as u1(t) = -R -1 (t)H(t)x(t), K(t): state feedback gain matrix, φ (t,0): state transfer matrix, P(t): boundary condition of the Riccati matrix, defined as P(t) = 0, The derivative of the state vector, x T (t): transpose of the state vector, u T (t): transpose of the control input, H T (t): transpose of matrix H(t), R -1 (t): The inverse of the matrix R(t).
[0092] Compared with the prior art, the present invention has the following beneficial effects:
[0093] The SVPWM modulation algorithm of the present invention adopts two modes: five-stage and three-stage. The five-stage SVPWM realizes balanced control of the DC side midpoint potential by adjusting the action time of the small vector; the three-stage SVPWM reduces the switching loss of the rectifier by reducing the average switching frequency of the system. Combined with the optimal SVPWM modulation algorithm of the singular regulator, the harmonic suppression effect is further optimized through real-time monitoring and dynamic adjustment, so that the system can maintain low harmonic characteristics and high power factor under different load conditions.
[0094] The method of the present invention can effectively control the midpoint potential balance, so that the midpoint voltage fluctuation range is kept within ±0.15V. The harmonic content of the grid-side current is significantly reduced, and the power factor of the system is improved. The switching frequency and switching loss of the rectifier are reduced, and the overall efficiency and reliability of the system are improved.
[0095] The invention has significant advantages in suppressing harmonics and optimizing switching losses, and provides technical support for high-efficiency and high-reliability power electronics applications. Simulation experiments have verified the effectiveness and superiority of the algorithm, which has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS
[0096] Figure 1 A schematic diagram of sector division of a single-phase three-level SVPWM provided by an embodiment of the present invention;
[0097] Figure 2 A five-segment vector action timing diagram of sector I provided by an embodiment of the present invention;
[0098] Figure 3 A three-stage vector action timing diagram of sector I provided by an embodiment of the present invention;
[0099] Figure 4 The optimal singular regulator structure diagram provided by the embodiment of the present invention;
[0100] Figure 5 A diagram showing the harmonic suppression effect of the method according to an embodiment of the present invention. DETAILED DESCRIPTION
[0101] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0102] A harmonic suppression and optimal SVPWM modulation method for a single-phase three-level PWM rectifier, comprising:
[0103] The basic vector and sector division of the single-phase three-level rectifier are designed. The vector allocation of the SVPWM algorithm is optimized through a singular regulator and harmonic suppression technology is adopted. The optimized SVPWM algorithm is used for regulation. When the control input is close to the boundary, a three-segment SVPWM is used; in other cases, a five-segment SVPWM is used.
[0104] The basic vectors and sector divisions for designing a single-phase three-level rectifier include:
[0105] The single-phase three-level NPC pulse rectifier has 9 switching states, corresponding to 9 voltage vectors. According to the input voltage u under different switching states ab The size of V can be divided into three categories: large vector, small vector and zero vector. pn : Contains the positive vector V pn (1,-1) and negative large vector V pn (-1,1) two vectors, when this type of vector acts, the input voltage u ab The amplitudes are u dc * and -u dc * ; Small vector: There are 4 small vectors in total, which are divided into two groups of redundant vectors, among which V po (1,0) and V on (0,-1) are mutually redundant positive small vectors V po ; V no (-1,0) and V op (0,1) are mutually redundant negative small vectors. When this type of vector acts, the input voltage u ab The amplitudes are and Zero vector: The single-phase three-level NPC rectifier contains three zero vectors, namely V pp (1,1), V oo (0,0) and V nn (-1,-1), when this type of vector acts, u ab The amplitudes are all 0;
[0106] Figure 1 A schematic diagram of sector division is given.
[0107] The 9 voltage vectors divide the αβ plane into 4 sectors with equal spacing. ab is the modulation signal, V ref is the reference voltage vector, the reference voltage vector V ref It rotates counterclockwise in the αβ plane at an angular velocity ω, with an initial phase angle of θ, where the reference voltage vector V ref The component on the α axis is u ab , define V ref The amplitude is V, then the input voltage u ab The expression is as follows:
[0108] u ab =Vcos(ωt-θ)
[0109] According to u ab The size of V ref It is divided into 4 sectors, and the division rules are as follows:
[0110] (1) Sector I: 0.5u dc <u ab <u dc ;
[0111] (2) Sector II: 0<u ab <0.5u dc ;
[0112] (3) Sector III: -0.5u dc <u ab <0;
[0113] (4) Sector IV: -u dc <u ab <-0.5u dc .
[0114] The vector action timing of the five-segment SVPWM midpoint potential control algorithm is: T b+ / 2→T a / 2→T b / 2→T a / 2→T b+ / 2 or T b / 2→T a / 2→T b+ / 2→T a / 2→T b / 2, where T a and T b They represent the action time of the large vector and the small vector in a switching cycle, T b+ and T b- They are respectively redundant small vectors V b+ and V b- The action time in one switching cycle, where V b+Select V in sectors Ⅰ and Ⅱ po (1,0), select V in sectors III and IV no (-1,0); V b- Select V in sectors Ⅰ and Ⅱ on (0,-1), select V in sectors III and IV op (0,1).
[0115] Where T b+ and T b- are the redundant small vectors V b+ and V b- The action time in one switching cycle and meets the following requirements:
[0116] T b+ +T b- =T b
[0117] Figure 2 The output vector timing diagram of sector I in a switching cycle is given. It can be seen from the figure that the small vector V b+ and V b- Select V po and V on , large vector V a Select V pn , and it can be seen that the PWM output pulse is symmetrical, so at the beginning of the next cycle, no switching will occur in the two bridge arms, thereby reducing switching losses.
[0118] Table 1 shows the vector selection for the four sectors.
[0119]
[0120] The five-segment SVPWM selects a pair of redundant small vectors to adjust the midpoint potential within a switching cycle, so the action time of each small vector needs to be calculated separately. Assume that after a switching cycle (t, t+Ts), the midpoint voltage can satisfy u dc (t+T s )=u dc * (t), then
[0121]
[0123] The above formulas respectively represent the action time of the redundant small vectors. These formulas are used to adjust the DC side midpoint potential balance in the five-segment SVPWM modulation algorithm to ensure the balance of the midpoint potential and reduce harmonic interference.
[0124] To ensure the stability of the system, the midpoint voltage change uΔ(t) must satisfy the following control law:
[0125]
[0126] Since the five-segment SVPWM modulation algorithm selects a pair of redundant small vectors to work together in one switching cycle, the switching state changes frequently, and the grid-side current also changes four times. Therefore, the algorithm has higher control accuracy and less low-order harmonic content in the grid-side current. However, the use of a pair of redundant vectors to control the midpoint potential balance also leads to a higher average switching frequency, and the high-order harmonics of the grid-side current are mainly distributed around twice the switching frequency.
[0127] Three-stage SVPWM midpoint potential control algorithm: The three-stage SVPWM also achieves the balance of midpoint potential by adjusting the vector action time, but its switching frequency is lower than that of the five-stage type, which is suitable for occasions with low switching losses. The vector action order is V a →V b →V a or V b →V a →V b , in one switching cycle, V a Select V pn , V b Selected as redundant vector V po (1,0) and V op (0,1). The difference between the three-stage SVPWM and the five-stage SVPWM lies in the different definitions of the basic vectors Va and Vb and the different action sequences of the output vectors. Since the design of the midpoint potential controller is closely related to the selection of the basic vectors, the control principle of the midpoint potential needs to be given priority during the design.
[0128] First, define the midpoint potential control function as follows:
[0129] Δ=sign(u1-u2)*sign(i s )
[0130] According to the different sectors where the reference voltage vector Vref is located, when Δ=1, the small vector V on or V on ; When Δ=-1, select the small vector V po or V no .
[0131] Figure 3The following is a timing diagram of selecting different small vectors to act in one switching cycle, taking sector I as an example, where the large vector Va is selected as Vpn, and the small vector Vb is selected as Vpo and Von respectively. As can be seen from the figure, the same as the five-segment SVPWM, when the switch state is switched, only the switching function of one bridge arm changes. Since the three-segment SVPWM modulation algorithm selects one of the members of a pair of redundant small vectors to act in one switching cycle, the switch state will only change twice, and the grid-side current will only change twice, which leads to lower control accuracy and higher low-order harmonic content in the grid-side current, but its average switching frequency is lower, and the high-order harmonics of the grid-side current are also mainly distributed near the odd multiples of the switching frequency.
[0132] The five-segment SVPWM uses a pair of redundant small vectors to work together in one switching cycle to adjust the midpoint potential. This method causes frequent changes in the switching state and four changes in the grid-side current, thereby improving control accuracy and reducing low-order harmonic content.
[0133] Advantages: Due to high control accuracy, the low-order harmonic content in the grid-side current is less.
[0134] Disadvantages: The average switching frequency is high, and the high-order harmonics of the grid-side current are mainly distributed around twice the switching frequency, which increases the switching loss.
[0135] The three-stage SVPWM only selects one member of a pair of redundant small vectors to act in one switching cycle, resulting in only two changes in the switching state. This method has a low switching frequency, but the control accuracy is relatively low and the midpoint potential fluctuates greatly.
[0136] Advantages: Low average switching frequency, reducing switching losses.
[0137] Disadvantages: Low control accuracy and higher content of low-order harmonics in the grid-side current.
[0138] Conditions for using five-stage and three-stage SVPWM
[0139] When the control input is close to the boundary, three-stage SVPWM is used; in other cases, five-stage SVPWM is used. The specific conditions are as follows:
[0140] The three-stage SVPWM is suitable for situations where the switching loss is low but the harmonic control requirements are not high.
[0141] The five-stage SVPWM is suitable for situations that require higher control accuracy and lower harmonic content, but allows higher switching losses.
[0142] The five-stage SVPWM has more advantages in control accuracy and low-order harmonic control, but at the cost of higher switching losses. The three-stage SVPWM performs better in reducing switching losses, but has weaker control accuracy and harmonic control capabilities.
[0143] Based on the above two modes, the optimal SVPWM modulation algorithm is designed in combination with the singular regulator. This algorithm can significantly reduce the average switching frequency of the system while achieving midpoint potential balance, thereby reducing switching losses.
[0144] Since the five-segment SVPWM and the three-segment SVPWM respectively use a pair of redundant small vectors and one member of a pair of redundant small vectors to control the midpoint potential balance, the action time of their small vectors can be similarly written as:
[0145]
[0146]
[0147] When five-segment SVPWM modulation is adopted, k Δ ∈(-1,1); when three-stage SVPWM modulation is adopted, k Δ = ±1, the DC side voltage state equation is as follows:
[0148] The above equation can be written as the following state equation:
[0149]
[0150] The performance indicators are selected as follows:
[0151]
[0152] When J is the minimum value, the optimal control is achieved. The original singular regulator is converted into an equivalent standard regulator through linear transformation. After solving the standard regulator, it is converted into the solution of the original singular regulator.
[0153] x1(t)=x(t)-B(t)u1(t)
[0154]
[0155] Taking the derivative of both sides, we get: in
[0156] By linearly transforming the original performance index, we can get:
[0157]
[0158] H(t)=Q(t)B(t); R(t)=B T(t)Q(t)B(t)
[0159] The integrand is formulated as x1 T (t)Q(t)x1(t)+2x1 T (t)H(t)u1(t)+u1 T (t)R(t)u1(t)
[0160] =x1 T [Q(t)+ΔQ(t)]x1(t)+[u1(t)+Δu1(t)] T R(t)[u1(t)+Δu1(t)]
[0161] Assuming R(t), multiply the right side of the above equation to get:
[0162] Δu1(t)=R -1 (t)H(t)x1(t)
[0163] ΔQ(t)=-H(t)R -1 (t)H T (t)
[0164] The performance index J is written as:
[0165]
[0166] Where: Q1(t) = Q(t) - H T (t)R -1 (t)H(t);Q1(t) is non-negative and definite,
[0167] u2(t)=u1(t)+R -1 (t)H(t)x1(t),
[0168] R(t) and Q1(t) are symmetrical positive definite matrices and symmetric non-negative definite matrices respectively. Assuming that R(t)>0, the non-negative definiteness of Q1(t) is proved.
[0169] x T (t)Q(t)x(t)=x1 T (t)Q(t)x1(t)+u1(t)+R -1 (t)H(t)x1(t)
[0170] +[u1(t)+R -1 (t)H(t)x1(t)] T R(t)[u1(t)+R -1 (t)H(t)x1(t)]≥0
[0171] Selection: u1(t) = -R -1(t)H(t)x(t)
[0172] Get x1 T (t)Q(t)x1(t)≥0
[0173] Therefore Q1(t)≥0;
[0174] This modified state equation is further transformed into u2(t)=u1(t)+R -1 (t)H(t)1x(t) is substituted into the formula In the result:
[0175]
[0176] Where: A1(t) = A(t) - B1(t)R -1 (t)H T (t)
[0177] Without considering the control constraints |u (t) The modified singular regulator when |≤1 is:
[0178] System state equation:
[0179]
[0180] Performance indicators:
[0181]
[0182] When the above formula takes the minimum value, the optimal control u2 is obtained * (t) , u2 * (t) The expression is as follows:
[0183] u2 * (t)=-K(t)x1(t)
[0184] Where K(t) = R -1 (t)B1 T (t)P(t)
[0185] where P(t) is a symmetric non-negative definite matrix that satisfies the following Riccati differential equation
[0186]
[0187] Its boundary conditions are:
[0188] P(t)=0
[0189] Will H(t)=Q(t)B(t); R(t)=B T(t)Q(t)B(t) and u2(t)=u1(t)+R -1 Substitute (t)H(t)x1(t) into u2 * (t)=-K(t)x1(t) to obtain u1 * (t)=-K1(t)x1(t)
[0190] Where K1(t) = R -1 (t)[B1 T (t)P(t)+H T (t)]
[0191] K1(t) in the above formula is the feedback gain matrix of the modified singular regulator;
[0192]
[0193] because:
[0194] K1 T (t)B T (t)Q(t)K1(t)=[P(t)B1(t)+H(t)]R -1 (t)B T (t)Q(t)B(t)R -1 (t)[B T (t)P(t)+H T (t)]
[0195] =P(t)B1(t)R -1 (t)H T (t)+H T (t)R -1 (t)B T (t)P(t)+P(t)B1(t)R -1 (t)B T (t)P(t)+H(t)R -1 (t)H T (t) The Riccati equation can be written as:
[0196]
[0197] The optimal performance index is:
[0198] J * =x1 T (0)P(0)x1(0)
[0199] For the modified singular regulator, its variables should satisfy the following relationship:
[0200] P(t)B(t)=0
[0201] K1(t)B(t)=I
[0202] J * =x1 T (0)P(0)x1(0)
[0203] K1(t)x(t)=0
[0204] After obtaining the solution of the modified singular regulator, it is necessary to convert it to the solution of the original singular regulator when it is not constrained;
[0205] Since x1(t)=x(t)-B(t)u1(t) holds, the following boundary constraints need to be met during the conversion:
[0206] x(0)=x1(0)+B(0)u1 * (0), x(0) is the initial state of the original singular optimal regulator;
[0207] u1 * (t) = -K1(t)x1(t) Substituting into the above formula, the boundary constraint condition is expressed as:
[0208] x(0)=x1(0)-B(0)K1(0)x1(0)
[0209] The solution of K1(t)x(t)=0 is proved to be a modified singular optimal regulator, and this solution can be transformed into the solution of the original singular optimal regulator;
[0210] The optimal control is:
[0211]
[0212] u(t): control input, A(t): system matrix, B(t): control matrix, x(t): system state vector, P(t): Riccati matrix, Q(t): weight matrix, H(t): matrix, defined as H(t) = Q(t)B(t), R(t): weight matrix, defined as R(t) = B T (t)Q(t)B(t), Δu(t): control input correction, ΔQ(t): matrix, defined as ΔQ(t) = -H T (t)R -1 (t)H(t), ΔR(t): matrix, defined as ΔR(t)=R(t)>0, Q1(t): symmetric non-negative definite matrix, defined as Q1(t)=Q(t)-H T (t)R -1 (t)H(t), u1(t): control input, defined as u1(t) = -R -1 (t)H(t)x(t), K(t): state feedback gain matrix, φ (t,0): state transfer matrix, P(t): boundary condition of the Riccati matrix, defined as P(t) = 0, The derivative of the state vector, x T (t): transpose of the state vector, u T (t): transpose of the control input, H T (t): transpose of matrix H(t), R -1 (t): the inverse of the matrix R(t). The modified optimal singular regulator is Figure 4 shown.
[0213] The present invention adopts advanced harmonic suppression technology to reduce the harmonic content of grid-side current by optimizing the vector distribution of SVPWM algorithm. Figure 5 shown.
[0214] The optimal SVPWM modulation algorithm based on the singular regulator can not only effectively control the balance of the midpoint potential, but also make the midpoint potential fluctuate within a smaller range. Compared with the five-segment SVPWM, the grid-side current harmonic content increases less and the switching times of the device are reduced, which effectively reduces the switching loss of the device. Through real-time monitoring and dynamic adjustment, the harmonic suppression effect is further optimized, so that the system can maintain low harmonic characteristics under different load conditions.
[0215] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A harmonic suppression and optimal SVPWM modulation method for a single-phase three-level PWM rectifier, characterized in that: include: The basic vector and sector division of the single-phase three-level rectifier are designed. The vector allocation of the SVPWM algorithm is optimized through a singular regulator and harmonic suppression technology is adopted. The optimized SVPWM algorithm is used for regulation. When the control input is close to the boundary, a three-segment SVPWM is used; in other cases, a five-segment SVPWM is used.
2. The harmonic suppression and optimal SVPWM modulation method of a single-phase three-level PWM rectifier according to claim 1, characterized in that: The basic vectors and sector divisions for designing a single-phase three-level rectifier include: The single-phase three-level NPC pulse rectifier has 9 switching states, corresponding to 9 voltage vectors. According to the input voltage u under different switching states ab The size of V can be divided into three categories: large vector, small vector and zero vector. pn : Contains the positive vector V pn (1,-1) and negative large vector V pn (-1,1) two vectors, when this type of vector acts, the input voltage u ab The amplitudes are u dc * and -u dc * ; Small vector: There are 4 small vectors in total, which are divided into two groups of redundant vectors, among which V po (1,0) and V on (0,-1) are mutually redundant positive small vectors V po ; V no (-1,0) and V op (0,1) are mutually redundant negative small vectors. When this type of vector acts, the input voltage u ab The amplitudes are and Zero vector: The single-phase three-level NPC rectifier contains three zero vectors, namely V pp (1,1), V oo (0,0) and V nn (-1,-1), when this type of vector acts, u ab The amplitudes are all 0; The 9 voltage vectors divide the αβ plane into 4 sectors with equal spacing. ab is the modulation signal, V ref is the reference voltage vector, the reference voltage vector V ref It rotates counterclockwise in the αβ plane at an angular velocity ω, with an initial phase angle of θ, where the reference voltage vector V ref The component on the α axis is u ab , define V ref The amplitude is V, then the input voltage u ab The expression is as follows: you ab =Vcos(ωt-θ) According to u ab The size of V ref It is divided into 4 sectors, and the division rules are as follows: (1) Ogi Ward I: 0.5u dc <u ab <u dc ; (2) Sector II: 0<u ab <0.5u dc ; (3) Sector III: -0.5u dc <u ab <0; (4) Ogi Ward IV: -u dc <u ab <-0.5u dc .
3. The harmonic suppression and optimal SVPWM modulation method of a single-phase three-level PWM rectifier according to claim 2, characterized in that: The three-stage SVPWM achieves the balance of midpoint potential by adjusting the vector action time. The vector action order is V a →V b →V a or V b →V a →V b , in one switching cycle, V a Select V pn , V b Selected as redundant vector V po (1,0) and V op (0,1).
4. The harmonic suppression and optimal SVPWM modulation method of a single-phase three-level PWM rectifier according to claim 2, characterized in that: The vector action timing of the five-segment SVPWM midpoint potential control algorithm is: T b+ / 2→T a / 2→T b / 2→T a / 2→T b+ / 2 or T b / 2→T a / 2→T b+ / 2→T a / 2→T b / 2, where T a and T b They represent the action time of the large vector and the small vector in a switching cycle, respectively. Tb+ and Tb- are the redundant small vector V b+ and V b- The action time in one switching cycle, where V b+ Select V in sectors Ⅰ and Ⅱ po (1,0), select V in sectors III and IV no (-1,0); V b- Select V in sectors Ⅰ and Ⅱ on (0,-1), select V in sectors III and IV op (0,1).
5. The harmonic suppression and optimal SVPWM modulation method of a single-phase three-level PWM rectifier according to claim 3, characterized in that: The midpoint potential control function of the three-stage SVPWM is as follows: Δ=sign(u1-u2)*sign(i s ) According to the reference voltage vector V ref The sector is different. When Δ=1, select the small vector V on (0,-1) or V op (0,1); when Δ=-1, select the small vector V po (1,0) or V no (-1,0).
6. The harmonic suppression and optimal SVPWM modulation method of a single-phase three-level PWM rectifier according to claim 4, characterized in that: The action time of the small vector is written as: When five-segment SVPWM modulation is adopted, k Δ ∈(-1,1); when three-stage SVPWM modulation is adopted, k Δ = ±1, the DC side voltage state equation is as follows: The above equation can be written as the following state equation: The performance indicators are selected as follows: When J is the minimum value, the optimal control is achieved. The original singular regulator is converted into an equivalent standard regulator through linear transformation. After solving the standard regulator, it is converted into the solution of the original singular regulator. x1(t)=x(t)-B(t)u1(t) Taking the derivative of both sides, we get: in By linearly transforming the original performance index, we can get: H(t)=Q(t)B(t);R(t)=B T (t)Q(t)B(t) right The integrand is formulated as x1 T (t)Q(t)x1(t)+2x1 T (t)H(t)u1(t)+u1 T (t)R(t)u1(t)=x1 T [Q(t)+ΔQ(t)]x1(t)+[u1(t)+Δu1(t)] T R(t)[u1(t)+Δu1(t)] Assuming R(t), multiply the right side of the above equation to get: Δu1(t)=R -1 (t)H(t)x1(t) ΔQ(t)=-H(t)R -1 (t)H T (t) The performance index J is written as: Where: Q1(t) = Q(t) - H T (t)R -1 (t)H(t);Q1(t) is non-negative and definite, u2(t)=u1(t)+R -1 (t)H(t)x1(t), R(t) and Q1(t) are symmetrical positive definite matrices and symmetric non-negative definite matrices respectively. Assuming that R(t)>0, the non-negative definiteness of Q1(t) is proved. x T (t)Q(t)x(t)=x1 T (t)Q(t)x1(t)+u1(t)+R -1 (t)H(t)x1(t)+[u1(t)+R -1 (t)H(t)x1(t)] T R(t)[u1(t)+R -1 (t)H(t)x1(t)]≥0 Selection: u1(t) = -R -1 (t)H(t)x(t) Get x1 T (t)Q(t)x1(t)≥0 Therefore Q1(t)≥0; This modified state equation is further transformed into u2(t)=u1(t)+R -1 (t)H(t)1x(t) is substituted into the formula In the result: Where: A1(t) = A(t) - B1(t)R -1 (t)H T (t) Without considering the control constraints |u (t) The modified singular regulator when |≤1 is: System state equation: Performance indicators: When the above formula takes the minimum value, the optimal control u2 is obtained * (t) , u2 * (t) The expression is as follows: u2 * (t)=-K(t)x1(t) Where K(t) = R -1 (t)B1 T (t)P(t) where P(t) is a symmetric non-negative definite matrix that satisfies the following Riccati differential equation Its boundary conditions are: P(t)=0 Will H(t)=Q(t)B(t); R(t)=B T (t)Q(t)B(t) and u2(t)=u1(t)+R -1 Substitute (t)H(t)x1(t) into u2 * (t)=-K(t)x1(t) to obtain u1 * (t)=-K1(t)x1(t) Where K1(t) = R -1 (t)[B1 T (t)P(t)+H T (t)] K1(t) in the above formula is the feedback gain matrix of the modified singular regulator; -P(t)B1(t)R -1 (t)B1 T (t)P(t)+Q(t)-H(t)R -1 (t)H T (t) because: K1 T (t)B T (t)Q(t)K1(t)=[P(t)B1(t)+H(t)]R -1 (t)B T (t)Q(t)B(t)R -1 (t)[B T (t)P(t)+H T (t)] =P(t)B1(t)R -1 (t)H T (t)+H T (t)R -1 (t)B T (t)P(t)+P(t)B1(t)R -1 (t)B T (t)P(t)+H(t)R -1 (t)H T (t) The Riccati equation can be written as: The optimal performance index is: J * =x1 T (0)P(0)x1(0) For the modified singular regulator, its variables should satisfy the following relationship: P(t)B(t)=0 K1(t)B(t)=I J * =x1 T (0)P(0)x1(0) K1(t)x(t)=0 After obtaining the solution of the modified singular regulator, it is necessary to convert it to the solution of the original singular regulator when it is not constrained; Since x1(t)=x(t)-B(t)u1(t) holds, the following boundary constraints need to be met during the conversion: x(0)=x1(0)+B(0)u1 * (0), x(0) is the initial state of the original singular optimal regulator; u1 * (t) = -K1(t)x1(t) Substituting into the above formula, the boundary constraint condition is expressed as: x(0)=x1(0)-B(0)K1(0)x1(0), the solution of K1(t)x(t)=0 is proved to be a modified singular optimal regulator, And this solution can be transformed into the solution of the original singular optimal regulator; The optimal control is: Among them, u(t): control input, A(t): system matrix, B(t): control matrix, x(t): system state vector, P(t): Riccati matrix, Q(t): weight matrix, H(t): matrix, defined as H(t) = Q(t)B(t), R(t): weight matrix, defined as R(t) = B T (t)Q(t)B(t), Δu(t): control input correction, ΔQ(t): matrix, defined as ΔQ(t) = -H T (t)R -1 (t)H(t), ΔR(t): matrix, defined as ΔR(t)=R(t)>0, Q1(t): symmetric non-negative definite matrix, defined as Q1(t)=Q(t)-H T (t)R -1 (t)H(t), u1(t): control input, defined as u1(t) = -R -1 (t)H(t)x(t), K(t): state feedback gain matrix, φ (t,0) : state transfer matrix, P(t): boundary condition of the Riccati matrix, defined as P(t) = 0, The derivative of the state vector, x T (t): transpose of the state vector, u T (t): transpose of the control input, H T (t): transpose of matrix H(t), R -1 (t): The inverse of the matrix R(t).
Citation Information
Cited By
Method for controlling Vienna rectifier under fault power grid based on frequency multiplication SVPWM (Space Vector Pulse Width Modulation) modulation
CN120638876A
