Improved deadbeat current predictive control method based on jury criterion

Abstract

The application discloses an improved zero-error current predictive control method based on a Jury criterion, and specifically comprises the following steps: step 1, collecting the grid-side inductance current of a T-type three-level inverter and the voltage at a grid connection point, and constructing a mathematical model in an alpha-beta coordinate system; step 2, performing forward Euler discretization on the model constructed in step 1 to obtain a reduced zero-error current predictive control model; step 3, performing prediction on the control model obtained in step 2 by using Newton interpolation to obtain predicted current and voltage; step 4, converting the predicted results obtained in step 3 to obtain the output voltage of the inverter; step 5, solving the value range of lambda by using the Jury criterion; and step 6, according to the solving result of step 5, using a pulse signal to drive the on-off of the switching tube in the inverter. The application solves the problem that the existing zero-error current predictive control has low precision and dynamic performance.

Description

Technical Field

[0001] The invention belongs to the technical field of power electronic control and stability analysis, and relates to an improved deadbeat current prediction control method based on Jury criterion. Background Art

[0002] With the rapid development of "dual-high" (high proportion of renewable energy and high proportion of power electronics) power grids, grid-connected inverters, as the interface devices for renewable energy grid connection, are often required to connect to weak grids. Due to the characteristics of weak grids, such as low voltage support capability and high grid impedance, they are susceptible to fluctuations and disturbances, placing higher demands on the control strategies of grid-connected devices. Predictive control, with its strong robustness and fast dynamic response, can effectively cope with grid fluctuations in weak grid environments, improving system stability and control accuracy, making it an ideal choice for weak grid control strategies for grid-connected inverters. To this end, an improved deadbeat current predictive control based on the Jury criterion is proposed. Compared to traditional deadbeat control, which is highly dependent on hardware parameters, the quality and accuracy of the output current waveform are lower, and the control of high-frequency harmonics is relatively limited. The proposed scheme reduces the order of the deadbeat current predictive control model by introducing a coupling coefficient λ. This simplifies the controller design while reducing dependence on hardware parameters, improving the quality and accuracy of the output current waveform, and enhancing the suppression of high-frequency harmonics. The appropriate parameter λ is selected in combination with the Jury criterion, which effectively improves the stability and robustness of the system, overcomes the shortcomings of traditional deadbeat control, and retains the advantages of fast response and high-precision tracking of deadbeat control. Summary of the Invention

[0003] The purpose of the present invention is to provide an improved deadbeat current predictive control method based on Jury criterion, which solves the problem of low accuracy and dynamic performance of existing deadbeat current predictive control.

[0004] The technical solution adopted by the present invention is an improved deadbeat current prediction control method based on Jury criterion, which specifically includes the following steps:

[0005] Step 1: Collect the grid-side inductor current and grid-connection point voltage of the T-type three-level inverter and construct a mathematical model in the αβ coordinate system;

[0006] Step 2: Perform forward Euler discretization on the model constructed in step 1 and calculate the inductor current i L and the grid-side inductor current i g The coupling relationship between them is used to obtain the reduced-order deadbeat current predictive control model;

[0007] Step 3: Use Newton interpolation to predict the control model obtained in step 2 to obtain predicted current and voltage;

[0008] Step 4: convert the prediction result obtained in step 3 to obtain the output voltage of the inverter;

[0009] Step 5, solve the value range of λ through Jury criterion;

[0010] Step 6: Based on the solution of step 5, a pulse signal is used to drive the switching tube in the converter to turn on and off.

[0011] The present invention is also characterized in that:

[0012] The specific process of step 1 is:

[0013] Step 1.1: Collect the grid-side inductor current and grid-connected point voltage of the T-type three-level inverter. The mathematical model of the grid-connected inverter in the three-phase stationary abc coordinate system is shown as follows:

[0014]

[0015]

[0016] Where, L is the filter inductor on the inverter side; C f is the filter capacitor of the LCL filter; L g The grid side is the filter inductor; i La 、i Lb 、i Lc are the currents flowing through the filter inductor L in the abc coordinate system respectively; i ga 、i gb 、i gc is the current flowing through the filter inductor L in the abc coordinate system g Current; u inva 、u invb 、u invc is the phase voltage on the inverter side in the abc coordinate system; u Ca 、u Cb 、u Cc is the filter capacitor voltage in the abc coordinate system; u ga 、u gb 、u gc is the grid-side phase voltage in the abc coordinate system;

[0017] Step 1.2, sort out formulas (1) to (3) to get the formula containing i L and i g The equation is shown in formula (4):

[0018]

[0019] Step 1.3, perform abc-αβ coordinate transformation on equation (4) to obtain equation (5):

[0020]

[0021] Where i Lα 、i Lβ is the current flowing through the filter inductor L in the αβ coordinate system, i gα 、i gβ is the current flowing through the filter inductor L in the αβ coordinate system g Current, u invα 、u invβ is the phase voltage on the inverter side in the αβ coordinate system; u gα 、u gβ is the grid-side voltage in the αβ coordinate system.

[0022] The specific process of step 2 is:

[0023] Step 2.1: Perform forward Euler discretization on Equation (5), as shown in Equation (6):

[0024]

[0025] Where u invα (k),u invβ (k) is the grid-side voltage at time k in the αβ coordinate system; i gα (k+1), i gβ (k+1) is the current flowing through the filter inductor L at time k+1 in the αβ coordinate system g Current; i gα (k), i gβ (k) is the current flowing through the filter inductor L at time k in the αβ coordinate system g Current; u gα (k),u gβ (k) is the grid-side voltage at time k in the αβ coordinate system; T s is the switching period; i Lα (k+1), i Lβ (k+1) is the current flowing through the filter inductor L at time k+1 in the αβ coordinate system; i Lα (k), i Lβ (k) is the current flowing through the filter inductor L at time k in the αβ coordinate system;

[0026] Step 2.2: Use the following formula (7) to convert the k+1 moment into the k moment i L The difference between time k+1 and time k is g The λ-fold difference is expressed as , and the reduced-order deadbeat current predictive control model is obtained, and equation (8) is obtained:

[0027]

[0028]

[0029] Step 2.3: According to formula (8), the inverter output voltages uinvα(k+1) and uinvβ at time k+1 are obtained: β (k+1), as shown in formula (9):

[0030]

[0031] The specific process of step 3 is:

[0032] Step 3.1, let (f(x i )-f(x j )) / (x i -x j ) is called the function f(x) at point x i 、x j The first-order difference quotient of , and recorded as f[x i , x j ],(f[x1,x2,...,x k ]-f[x0, x1, ..., x k -1]) / (x k -x0) is called the sum of f(x) at x0, x1, x2, ..., x k The k-order difference quotient of:

[0033] f(x)=f(x0)+f[x,x0](x-x0)

[0034] f(x)=f(x0)+f[x0,x1](x-x0)+f[x,x0,x1](x-x0)(x-x1)

[0035] f(x)=f(x0)+f[x0,x1](x-x0)+…+f[x0,x1,…,x n ](x-x0)(x-x1)…(xx n-1 )+f[x,x0…,x n ](x-x0)(x-x1)…(xx n-1 ) (10)

[0037] Step 3.2: Use Newton interpolation to predict the grid-connected current i at time k+1 gα (k+1), i gβ (k+1), as shown in the following formula (11):

[0038]

[0039] Where i gα * 、i gβ * is the expected value of the grid-connected current, i.e. the current value at time k+2, igα (k), i gβ (k) is the current value at time k, i gα (k-1), i gβ (k-1) is the current value at time k-1, i gα (k-2), i gβ (k-2) is the current value at time k-2;

[0040] Step 3.3, predict the grid-side voltage u gα (k+1), u gβ (k+1) using Newton interpolation, as shown in the following equation (12):

[0041]

[0042] wherein u gα (k), u gβ (k) is the voltage value at time k, u gα (k-1), u gβ (k-1) is the voltage value at time k-1, u gα (k-2), u gβ (k-2) is the voltage value at time k-2.

[0043] The specific process of step 4 is as follows:

[0044] Step 4.1, obtain the grid-side current i gα (k+1), i gβ (k+1) at time k+1 from step 3: gα (k+1), u gβ (k+1) into equation (9) to obtain the inverter output voltage u invα (k+1), u invβ (k+1) at time k+1, as shown in equation (13):

[0045]

[0046] The specific process of step 5 is as follows:

[0047] Step 5.1, the transfer function G ig (s) from inverter voltage to grid current is:

[0048]

[0049] wherein,

[0050] Step 5.3, z-transform G ig (s) by the residue theorem to obtain G ig (z):

[0051]

[0052] Wherein, Ts is the switching period;

[0053] Step 5.4: According to equation (13), the discrete transfer function of the improved deadbeat predictive controller is:

[0054]

[0055] In step 5.5, combining equations (15) and (16), we can obtain the closed-loop transfer function Φ(z) of the deadbeat control system as equation (17):

[0056]

[0057] Step 5.6: According to equation (17), the closed-loop characteristic equation of the discrete system is obtained as:

[0058] D(z)=1+G pc (z)K PWM (z)G ig (z) (18)

[0059] Step 5.7: Apply the Jury criterion to equation (18) and express the closed-loop characteristic equation using a polynomial expansion:

[0060] D(z)=a n z n +…+a2z 2 +a1z 1 +a0z 0 (19)

[0061] The characteristic equation of the deadbeat control system is expressed by polynomial expansion:

[0062] D(z)=a3z 3 +a2z 2 +a1z 1 +a0z 0 (20)

[0064] in:

[0065]

[0066] Step 5.8, solve the value range of λ according to the Jury criterion.

[0067] The specific process of step 5.8 is:

[0068] Step 5.8.1, construct the Jury array;

[0069] Step 5.8.2: The highest power of the polynomial expansion of the characteristic equation of the deadbeat control system is cubic, so only the third row element b needs to be solved. k Then, as shown in the following formula (25):

[0070]

[0071] In step 5.8.3, use the Jury array to determine the necessary and sufficient conditions for all roots of the characteristic equation D(z) = 0 to lie within the unit circle on the z plane. This condition is used to determine the range of the λ parameter. The necessary and sufficient conditions are as follows:

[0072] aD(1)>0;

[0073] b.(-1) n D(-1)>0;

[0074] c.|a0|>a n ;

[0075] d.|b0|>|b n-1 |,|c0|>|c n-2 |, ..., |m0|>|m2|;

[0076] In step 5.8.4, based on the grid-connected inverter system parameters, the value range of λ is obtained as shown in formula (26):

[0077] λ∈(0, 0.8)

[0078] (26).

[0079] The specific process of step 6 is:

[0080] Substitute the value of λ obtained in step 5 into equation (13), and the inverter output voltage u at time k+1 is obtained. invα (k+1) and u invβ (k+1), after αβ-abc transformation, it is sent to SPWM modulation and converted into a corresponding pulse signal, which is used to drive the on and off of the switch tube in the converter.

[0081] The beneficial effect of the present invention is that the present invention first proposes a method for simplifying the control model based on the coupling relationship between the inductance and capacitance of the LCL type three-phase grid-connected inverter, by establishing a control model including the k+1 moment and the k moment i L and i gThe functional relationship of the original third-order control equation is reduced to a lower order. The complexity of the control strategy is effectively simplified, the main dynamic characteristics of the system are retained, and the accuracy and stability of the system in rapid response are ensured; secondly, by introducing the Jury criterion to determine the appropriate value range of the selection coefficient λ, the robustness and stability of the system are further enhanced, so that the system can maintain reliable control performance under various operating conditions and disturbances; in addition, the Newton interpolation algorithm is used to predict and control the voltage and current, which greatly improves the control accuracy and the dynamic response of the system, making the system more accurate in tracking the reference signal, reducing the prediction error, and improving the power quality and control effect. These improvements enable the present invention to exhibit excellent performance in a complex grid-connected environment, and have good practicality and application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 This is a deadbeat current prediction control block diagram in the improved deadbeat current prediction control method based on Jury criterion of the present invention;

[0083] Figure 2 This is a block diagram of closed-loop feedback predictive current control in the discrete domain in the improved deadbeat current predictive control method based on Jury criterion of the present invention;

[0084] Figure 3 It is the zero-pole diagram in the improved deadbeat current predictive control method based on Jury criterion of the present invention;

[0085] Figure 4 This is a grid-connected current simulation waveform diagram obtained by using improved deadbeat control in the improved deadbeat current prediction control method based on Jury criterion of the present invention;

[0086] Figure 5 This is an FFT analysis diagram of the grid-connected current obtained by using the improved deadbeat control in the improved deadbeat current predictive control method based on Jury criterion of the present invention;

[0087] Figure 6 This is the grid-connected current simulation waveform obtained using traditional deadbeat control;

[0088] Figure 7 This is the FFT analysis diagram of the grid-connected current obtained using traditional deadbeat control;

[0089] Figure 8 This is a simulated waveform diagram of the phase relationship between the output current and voltage of the grid-side phase A obtained by the improved deadbeat control in the improved deadbeat current prediction control method based on the Jury criterion of the present invention;

[0090] Figure 9This is a simulated waveform diagram of the phase relationship between the grid-side A-phase output current and voltage obtained by conventional deadbeat control in the improved deadbeat current predictive control method based on Jury criterion of the present invention;

[0091] Figure 10 The simulated waveform of the load current when the current jumps from 10A to 15A obtained by using the improved deadbeat control in the improved deadbeat current predictive control method based on Jury criterion of the present invention;

[0092] Figure 11 The simulated waveform of the load current when the current jumps from 10A to 15A obtained by using traditional deadbeat control in the improved deadbeat current predictive control method based on Jury criterion of the present invention;

[0093] Figure 12 Experimental waveform of grid-connected current obtained by using improved deadbeat control in the improved deadbeat current predictive control method based on Jury criterion of the present invention;

[0094] Figure 13 Experimental waveform diagram of the phase relationship between the output current and voltage of phase A on the grid side obtained by the improved deadbeat control in the improved deadbeat current predictive control method based on Jury criterion of the present invention;

[0095] Figure 14 The improved deadbeat current predictive control method based on Jury criterion of the present invention uses the improved deadbeat control to obtain the experimental waveform of the load current when the current jumps from 10A to 15A. DETAILED DESCRIPTION

[0096] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0097] Example 1

[0098] The improved deadbeat current prediction control method based on Jury criterion of the present invention refers to Figure 1 , including the following steps:

[0099] Step 1: Collect the grid-side inductor current and grid-connection point voltage of the T-type three-level inverter, write the KVL and KCL equations, and then transform them into abc-αβ to obtain the mathematical model in the αβ coordinate system. When modeling, all switching devices are assumed to be ideal devices, and the turn-on and turn-off time delays of the devices are not considered.

[0100] Step 2: Perform forward Euler discretization on the model obtained in step 1 and calculate the inductor current i L and the grid-side inductor current i g The coupling relationship between them is obtained, and the reduced-order deadbeat current predictive control model is obtained.

[0101] Step 3, the grid-connected current i gα (k+1) is predicted by Newton interpolation respectively at k+1 moment in formula (2) gβ (k+1) and the grid-side voltage u gα (k+1) at k+1 moment gβ (k+1) is predicted by Newton interpolation. Wherein, the grid-connected current i gα (k+2) is predicted by Newton interpolation respectively at k+2 moment in formula (2) gβ (k+2) is the given current i gα * (k+2) obtained by voltage outer loop control gβ * .

[0102] To obtain the inverter output voltage u invα (k+1) at k+1 moment, it is necessary to know the grid-connected current i invβ (k+2) at k+2 moment gα (k+2) and the grid-connected current i gβ (k+1) at k+1 moment, and the grid-side voltage u gα (k+1) at k+1 moment gβ (k+1) is predicted by Newton interpolation. Wherein, the grid-connected current i gα (k+2) is predicted by Newton interpolation respectively at k+2 moment in formula (2) gβ (k+2) is the given current i gα gβ (k+2) obtained by voltage outer loop control gα * (k+1) at k+1 moment gβ * (k+1) at k+1 moment gα (k+1) is predicted by Newton interpolation. Wherein, the grid-side voltage u gβ (k+1) at k+1 moment gα (k+1) needs to be predicted by Newton interpolation. gβ

[0103] Step 4, i gα (k+1) obtained by Newton interpolation prediction in step 3, i gβ (k+1) is obtained, and the inverter output voltage u invα (k+1) at k+1 moment invβ (k+1) is obtained

[0104] Step 5, in order to ensure the stability of the control system, optimize the dynamic response, improve the harmonic suppression ability and control accuracy, the value range of λ is solved by Jury criterion;

[0105] Step 6, according to the solving result of step 5, the on-off of the switching tube in the converter is driven by pulse signal.​​

[0106] Example 2

[0107] The specific process of step 1 is:

[0108] In step 1.1, collect the grid-side inductor current and grid-connection point voltage of the T-type three-level inverter and write the KVL and KCL equations. When modeling, all switching devices are assumed to be ideal, and device turn-on and turn-off time delays are not considered. The mathematical model of the grid-connected inverter in the three-phase stationary abc coordinate system is as follows:

[0109]

[0110]

[0111] Where, L is the filter inductor on the inverter side; C f is the filter capacitor of the LCL filter; L g The grid side is the filter inductor; i La ,i Lb ,i Lc is the current flowing through the filter inductor L in the abc coordinate system; i ga ,i gb ,i gc is the current flowing through the filter inductor L in the abc coordinate system g Current, also the grid current; u inva ,u invb ,u invc is the phase voltage on the inverter side in the abc coordinate system; u Ca ,u Cb ,u Cc is the filter capacitor voltage in the abc coordinate system; u ga ,u gb ,u gc is the grid-side phase voltage in the abc coordinate system.

[0112] Step 1.2, sort out equations (1) to (3) to get the equation containing i L and i g The equation is shown in formula (4):

[0113]

[0114] Step 1.3, perform abc-αβ coordinate transformation on equation (4) to obtain equation (5):

[0115]

[0116] Where i Lα ,i Lβ is the current flowing through the filter inductor L in the αβ coordinate system, which is also the inverter side current; igα ,i gβ is the current flowing through the filter inductor L in the αβ coordinate system g Current, also the grid current; u invα ,u invβ is the phase voltage on the inverter side in the αβ coordinate system; u gα ,u gβ is the grid-side voltage in the αβ coordinate system.

[0117] Example 3

[0118] The specific process of step 2 is as follows:

[0119] Step 2.1: Perform forward Euler discretization on Equation (5), as shown in Equation (6):

[0120]

[0121] Where u invα (k)u invβ (k) is the grid-side voltage at time k in the αβ coordinate system; i gα (k+1), i gβ (k+1) is the current flowing through the filter inductor L at time k+1 in the αβ coordinate system g Current; i gα (k), i gβ (k) is the current flowing through the filter inductor L at time k in the αβ coordinate system g Current; u gα (k),u gβ (k) is the grid-side voltage at time k in the αβ coordinate system; T s is the switching period; i Lα (k+1), i Lβ (k+1) is the current flowing through the filter inductor L at time k+1 in the αβ coordinate system; i Lα (k), i Lβ (k) is the current flowing through the filter inductor L at time k in the αβ coordinate system.

[0122] Step 2.2, since the current is the moment value, and the inductor current i L and the grid-side inductor current i g There is a certain coupling relationship. Therefore, in formula (6), we can use (7) to combine the k+1 moment with the k moment i L The difference between time k+1 and time k is g The λ-times difference is expressed as λ, and the reduced-order deadbeat current prediction control model is obtained, where λ is a variable constant coefficient representing the coupling coefficient of the filter, and equation (8) is obtained:

[0123]

[0124] Step 2.3: According to formula (8), the inverter output voltage u at time k+1 can be obtained: invα (k+1),u invβ (k+1), as shown in formula (9):

[0125]

[0126] Example 4

[0127] The specific process of step 3 is as follows:

[0128] Step 3.1, write out the Newton interpolation polynomial, (f(x i )-f(x j )) / (x i -x j )(i≠j) is called the function f(x) at point x i 、x j The first-order difference quotient of , and recorded as f[x i , x j ],(f[x1,x2,...,x k ]-f[x0, x1, ..., x k -1]) / (x k -x0) is called f(x) in x0, x1, x2, ..., x k The k-order difference quotient of:

[0129] f(x)=f(x0)+f[x,x0](x-x0)

[0130] f(x)=f(x0)+f[x0,x1](x-x0)+f[x,x0,x1](x-x0)(x-x1)

[0131] f(x)=f(x0)+f[x0,x1](x-x0)+…+f[x0,x1,…,x n ](x-x0)(x-x1)…(xx n-1 )+f[x,x0…,x n ](x-x0)(x-x1)…(xx n-1 ) (10)

[0133] Step 3.2: Use Newton interpolation to predict the grid-connected current i at time k+1 gα (k+1), i gβ (k+1):

[0134]

[0135] Where i gα * 、igβ * is the expected value of the grid-connected current, i.e. the current value at time k+2, i gα (k), i gβ (k) is the current value at time k, i gα (k-1), i gβ (k-1) is the current value at time k-1, i gα (k-2), i gβ (k-2) is the current value at time k-2.

[0136] Step 3.3: Use Newton interpolation to predict the grid-side voltage u at time k+1 gα (k+1),u gβ (k+1):

[0137]

[0138] Where u gα (k),u gβ (k) is the voltage value at time k, u gα (k-1),u gβ (k-1) is the voltage value at time k-1, u gα (k-2),u gβ (k-2) is the voltage value at time k-2.

[0139] Example 5

[0140] The specific process of step 4 is as follows:

[0141] The grid-connected current i at time k+1 obtained in step 3 is gα (k+1), i gβ (k+1) and grid-side voltage u gα (k+1),u gβ Substituting (k+1) into equation (9) further yields the inverter output voltage u at time k+1: invα (k+1),u invβ (k+1), as shown in formula (13):

[0142]

[0143] Example 6

[0144] The specific process of step 5 is as follows:

[0145] Step 5.1: To ensure system stability, optimize dynamic response, improve harmonic suppression capability and control accuracy, it is necessary to determine the value range of λ. Figure 2As shown in the closed-loop feedback predictive current control block diagram, it is necessary to solve the discrete transfer function Gpc(z) of the improved deadbeat predictive controller and the discrete transfer function G of the LCL filter. ig (z), and finally the closed-loop transfer function Φ(z) of the deadbeat control system is obtained. PWM (z) is the equivalent proportional link of PWM modulation. The grid-connected current feedback is unit negative feedback, so H is set to 1.

[0146] Step 5.2, write the transfer function G from the inverter voltage to the grid current ig (s) is:

[0147]

[0148] in,

[0149] Step 5.3, use the residue theorem to calculate G ig (s) performs z-transformation to obtain G ig (z), where Ts is the switching period:

[0150]

[0151] Step 5.4: According to equation (13), the discrete transfer function of the improved deadbeat predictive controller is:

[0152]

[0153] In step 5.5, combining equations (15) and (16), we can obtain the closed-loop transfer function Φ(z) of the deadbeat control system as equation (17):

[0154]

[0155] Step 5.6: According to equation (17), the closed-loop characteristic equation of the discrete system is obtained as:

[0156] D(z)=1+G pc (z)K PWM (z)G ig (z) (18)

[0157] Step 5.7: Apply the Jury criterion to equation (18) and express the closed-loop characteristic equation using a polynomial expansion:

[0158] D(z)=a n z n +…+a2z 2 +a1z 1 +a0z 0 (19)

[0159] The characteristic equation of the deadbeat control system is expressed by polynomial expansion:

[0160] D(z)=a3z 3 +a2z 2 +a1z 1 +a0z 0 (20)

[0162] in:

[0163]

[0164] Step 5.8, solve the value range of λ according to the Jury criterion:

[0165] 1) First, the Jury array needs to be constructed according to certain rules, as shown in Table 1:

[0166] Table 1 Jury array

[0167]

[0168]

[0169] 2) According to the polynomial expansion of the characteristic equation of the deadbeat control system, the highest power is 3, so we only need to solve the element b in the third row. k (k=0,1,2) then:

[0170]

[0171] 3) Using the Jury array, determine the necessary and sufficient conditions for all roots of the characteristic equation D(z) = 0 to lie within the unit circle on the z plane, and solve for the range of the λ parameter. The necessary and sufficient conditions are as follows:

[0172] eD(1)>0;

[0173] f.(-1) n D(-1)>0;

[0174] g.|a0|>a n ;

[0175] h.|b0|>|b n-1 |,|c0|>|c n-2 |,…,|m0|>|m2|.

[0176] In step 5.9, combined with the following system parameter table, the value range of λ is obtained as shown in formula (26), and then substituted into formula (13):

[0177] Table 2 Grid-connected inverter system parameters

[0178] parameter Numerical Inverter side inductance L / mH 5 Filter capacitor C f / F]]> <![CDATA[5*10 -5 ]]> <![CDATA[网侧电感L g / mH]]> 0.6 <![CDATA[开关频率f s / Hz]]> 20k

[0179] λ∈(0, 0.8)

[0180] (26).

[0181] Step 6: Substitute the value of λ obtained in step 5 into equation (13), and finally obtain the inverter output voltage u at time k+1: invα (k+1) and u invβ (k+1), after αβ-abc transformation, it is sent to SPWM modulation and converted into a corresponding pulse signal, which is used to drive the on and off of the switch tube in the converter.

[0182] Example 7

[0183] Figure 3 The zero-pole diagram is drawn using the improved deadbeat current predictive control method based on the Jury criterion of the present invention. The value range of λ is solved by the Jury criterion, and λ equal to 0.7 is substituted into the control system to draw the closed-loop zero-pole diagram of the system. It can be seen that the pole distribution of the response is 0.9286+0.3688i, 0.9286-0.3688i, 0.625+0.1329i, 0.625-0.1329i, and -0.1158+0i, all of which are within the unit circle. Therefore, the system is stable, which proves the effectiveness of solving λ based on the Jury criterion.

[0184] Figure 4 and Figure 5 They are respectively a waveform diagram of the grid-connected current obtained by using the improved deadbeat control in the improved deadbeat current prediction control method based on Jury criterion of the present invention, and an FFT analysis diagram. Figure 4 Given the expected current value is 10A, the actual output value is 10.02A, which shows that the prediction accuracy is relatively high.

[0185] Figure 5 The horizontal axis is the harmonic order n, and the vertical axis is the content of each harmonic. From this figure, it can be seen that the THD content of the current calculated using the improved deadbeat control is 0.48%, and the waveform quality is high.

[0186] Figure 6 and Figure 7 They are respectively the waveform diagram of the grid-connected current obtained by using traditional deadbeat control and the FFT analysis diagram. Figure 6 Given the expected current value is 10A, the actual output value is 9.903A. Figure 4 It can be seen that its prediction accuracy is low. Figure 7 The horizontal axis is the harmonic number n, and the vertical axis is the content of each harmonic. Figure 5, using traditional deadbeat control to calculate the THD content of its current is 1.03%, and the waveform quality is poor.

[0187] Figure 8 and Figure 9 They are respectively a simulation waveform diagram of the grid-side A-phase output current-voltage phase relationship obtained by the improved deadbeat control in the improved deadbeat current prediction control method based on Jury criterion of the present invention, and a simulation waveform diagram of the grid-side A-phase output current-voltage phase relationship obtained by the traditional deadbeat control. It can be found that the grid-connected current and voltage of the improved deadbeat current prediction control and the traditional deadbeat control are in phase, which ensures the effective transmission and utilization of electric energy and can maintain the stability and efficiency of the power grid.

[0188] Figure 10 In the improved deadbeat current predictive control method based on the Jury criterion of the present invention, the current waveform at the grid-side inductor of the inverter is obtained by using the improved deadbeat control at 1.04s when the current increases from 10A to 15A. The response time is short, only 721μs, and the dynamic response is fast.

[0189] Figure 11 The current waveform at the grid-side inductor of the inverter is obtained by using traditional deadbeat control. The current changes from 10A to 15A in 1.04s. The response time is short, only 500μs, and the dynamic response is fast. Figure 10 , which is 221μs faster, but sacrifices some filtering characteristics and may be insufficient in harmonic suppression and current quality.

[0190] Figure 12 This is an experimental waveform diagram of the grid-connected current obtained by using improved deadbeat control in the improved deadbeat current predictive control method based on the Jury criterion of the present invention. The expected current value is 10A, and the actual output value is 9.86A. The waveform quality is high, verifying the feasibility of the proposed scheme.

[0191] Figure 13 The experimental waveform diagram of the phase relationship between the grid-side A-phase output current and voltage obtained by the improved deadbeat current predictive control method based on the Jury criterion of the present invention shows that the grid-connected current and voltage are in phase with each other, ensuring the effective transmission and utilization of electric energy, maintaining the stability and efficiency of the power grid, and verifying the feasibility of the proposed scheme.

[0192] Figure 14The experimental waveform of the load current jump from 10A to 15A obtained by using the improved deadbeat control in the improved deadbeat current predictive control method based on the Jury criterion of the present invention. The pink-purple vertical line in the figure is the time interval measurement line, which is used to observe the time interval of the dynamic response of the load current jump from 10A to 15A. It can be seen that the response time is short, only 1.03ms, which is close to the simulation result, verifying the feasibility of the proposed scheme.

[0193] The improved deadbeat current prediction control based on Jury criterion is designed for the deadbeat control of traditional LCL grid-connected inverters. It aims to improve the accuracy and response speed of the system while maintaining the original third-order system structure and not neglecting the filter capacitor. It uses Newton interpolation prediction to compensate for time delay. In order to reduce the complexity of digital control when predicting the grid current at time k+3, the i in discrete state is established. L and i g The corresponding functional relationship is reduced in order. At the same time, in order to ensure the stability of the system, the value range of λ is obtained based on the Jury criterion to ensure the stable operation of the system.

Claims

1. An improved deadbeat current predictive control method based on Jury criterion, characterized by: The specific steps include: Step 1: Collect the grid-side inductor current and grid-connection point voltage of the T-type three-level inverter and construct a mathematical model in the αβ coordinate system; Step 2: Perform forward Euler discretization on the model constructed in step 1 and calculate the inductor current on the inverter side. i L and grid-side inductor current i g The coupling relationship between them is used to obtain the reduced-order deadbeat current predictive control model; Step 3: Use Newton interpolation to predict the control model obtained in step 2 to obtain predicted current and voltage; Step 4: Convert the prediction result obtained in step 3 to obtain the output voltage of the inverter. The specific process of step 4 is as follows: The grid-connected current at time k+1 obtained in step 3 i gα (k+1), i gβ (k+1) and grid-side voltage u gα (k+1), u gβ Substitute (k+1) into the inverter output voltage output in step 2 to obtain the inverter output voltage at time k+1 u invα (k+1), u invβ (k+1), as shown in formula (1): （1）； Where, i gα * 、 i gβ * is the expected value of the grid-connected current, that is k +2 moment current value, i gα (k) i gβ (k) is k The current value at the moment, i gα (k-1), i gβ (k-1) is k -1 moment current value, i gα (k-2), i gβ (k-2) is k -2 current value at moment; u gα (k) u gβ (k) is k The voltage value at the moment, u gα (k-1), u gβ (k-1) is k The voltage value at the moment -1, u gα (k-2), u gβ (k-2) is k -2 moment voltage value, L is the filter inductor on the inverter side, L g The grid side is the filter inductor, T s is the switching cycle; Step 5, solve the value range of λ through Jury criterion; Substitute the value of λ solved in step 5 into equation (1), and the inverter output voltage at time k+1 is finally obtained. uinvα (k+1) and u invβ(k+1), after αβ-abc transformation, is sent to SPWM modulation and converted into a corresponding pulse signal, which is used to drive the switching tube in the converter on and off; Step 6: Based on the solution of step 5, a pulse signal is used to drive the switching tube in the converter to turn on and off.

2. The improved deadbeat current predictive control method based on Jury criterion according to claim 1, characterized in that: The specific process of step 1 is: Step 1.1: Collect the grid-side inductor current and grid-connected point voltage of the T-type three-level inverter. The mathematical model of the grid-connected inverter in the three-phase static abc coordinate system is shown in the following equations (2) to (4): （2） （3） （4） Where, L is the filter inductor on the inverter side; C f is the filter capacitor of the LCL filter; L g The grid side is the filter inductor; i La 、 i Lb 、 i Lc They are the currents flowing through the filter inductor L in the abc coordinate system respectively; i ga 、 i gb 、 i gc is the current flowing through the filter inductor L in the abc coordinate system g current; u inva 、 u invb 、 u invc is the phase voltage on the inverter side in the abc coordinate system; u Ca 、 u Cb 、 u Cc is the filter capacitor voltage in the abc coordinate system; u ga 、 u gb 、 u gc is the grid-side phase voltage in the abc coordinate system; Step 1.2, sort out formulas (2) to (4) to obtain i L and i g The equation is shown in formula (5): （5） Step 1.3, perform abc-αβ coordinate transformation on equation (5) to obtain equation (6): （6） Where, i Lα 、 i Lβ is the current flowing through the filter inductor L in the αβ coordinate system, i gα 、 i gβ is the current flowing through the filter inductor L in the αβ coordinate system g Current, u invα 、 u invβ is the phase voltage on the inverter side in the αβ coordinate system; u gα 、 u gβ is the grid-side voltage in the αβ coordinate system.

3. The improved deadbeat current predictive control method based on Jury criterion according to claim 2, characterized in that: The specific process of step 2 is: Step 2.1, perform forward Euler discretization on Equation (6), as shown in Equation (7): （7） Where, u invα (k) u invβ (k) is the phase voltage on the inverter side at time k in the αβ coordinate system; i gα (k+1), i gβ (k+1) is the current flowing through the filter inductor L at time k+1 in the αβ coordinate system g current; i gα (k) i gβ (k) is the current flowing through the filter inductor L at time k in the αβ coordinate system g current; u gα (k) u gβ (k) is the grid-side voltage at time k in the αβ coordinate system; T s is the switching cycle; i Lα (k+1), i Lβ (k+1) is the current flowing through the filter inductor L at time k+1 in the αβ coordinate system; i Lα (k) i Lβ (k) is the current flowing through the filter inductor L at time k in the αβ coordinate system; Step 2.2, use the following formula (8) to convert k +1 moment with k time i L The difference is used k +1 moment with k time i g The λ-times difference is expressed as λ, and the reduced-order deadbeat current predictive control model is obtained, and equation (9) is obtained: （8） （9） Step 2.3, according to formula (9), we can get k Inverter output voltage at time +1 u invα (k+1), u invβ (k+1), as shown in formula (10): （10） Where, L on the inverter side is the filter inductor; L g The grid side is the filter inductor.

4. The improved deadbeat current predictive control method based on Jury criterion according to claim 3, characterized in that: The specific process of step 3 is as follows: Step 3.1, let called a function At the point x i 、 x j The first-order difference quotient of , called exist x 0. x 1. x 2. ... x k of k Order quotient: （11） Step 3.2: Use Newton interpolation to predict the grid-connected current at time k+1 i gα (k+1), i gβ (k+1), as shown in the following formula (12): （12） Step 3.3: Use Newton interpolation to predict the grid voltage at time k+1 u gα (k+1), u gβ (k+1), as shown in the following formula (13): （13） Where, u gα (k) u gβ (k) is k The voltage value at the moment, u gα (k-1), u gβ (k-1) is k The voltage value at the moment -1, u gα (k-2), u gβ (k-2) is k -2 The voltage value at time.

5. The improved deadbeat current predictive control method based on Jury criterion according to claim 4, characterized in that: The specific process of step 5 is as follows: Step 5.1, transfer function G from inverter voltage to grid current ig ( s )for: （14） in, , ; Step 5.3, by the residue theorem Perform z-transform to get : (15) Wherein, Ts is the switching period; Step 5.4: According to equation (1), the discrete transfer function of the improved deadbeat predictive controller is: （16） Step 5.5: Combining Equations (15) and (16), we can obtain the closed-loop transfer function Φ(z) of the deadbeat control system as Equation (17): （17） Step 5.6: According to Equation (17), the closed-loop characteristic equation of the discrete system is obtained as: （18） Step 5.7: Apply the Jury criterion to equation (18) and express the closed-loop characteristic equation using a polynomial expansion: （19） The characteristic equation of the deadbeat control system is expressed by polynomial expansion: （20） in: （21） （22） （23） （24） Step 5.8, solve the value range of λ according to the Jury criterion.

6. The improved deadbeat current predictive control method based on Jury criterion according to claim 5, characterized in that: The specific process of step 5.8 is as follows: Step 5.8.1, construct the Jury array; Step 5.8.2: The highest power of the polynomial expansion of the characteristic equation of the deadbeat control system is cubic, so only the third row of elements needs to be solved. b k Then, as shown in the following formula (25): （25） In step 5.8.3, use the Jury array to determine the necessary and sufficient conditions for all roots of the characteristic equation D(z) = 0 to lie within the unit circle on the z plane. This condition is used to determine the range of the λ parameter. The necessary and sufficient conditions are as follows: In step 5.8.4, based on the grid-connected inverter system parameters, the value range of λ is obtained as shown in formula (26): （26） in, The variable constant coefficient represents the coupling coefficient of the filter.

Images