CN110011359B - Grid-connected inverter parameter identification method under finite set model prediction control - Google Patents

Abstract

The invention discloses a grid-connected inverter parameter identification method under finite set model prediction control, which comprises the following steps: under a simplified model prediction control method, a dual-voltage prediction error equation is constructed, the relation between the actual value of the filter inductor and the equivalent resistor thereof and the parameter nominal value is obtained, the actual inductor and the actual resistor are identified in real time, and an accurate prediction model is constructed, so that the quality of grid-connected current is improved, and the robustness of finite set model prediction control is improved.

Description

Grid-connected inverter parameter identification method under finite set model prediction control

Technical Field

The invention belongs to the technical field of grid-connected inverter control, and particularly relates to a parameter identification method under simplified model prediction control, which is used for improving the grid-connected current waveform quality of an inverter and improving the parameter robustness of a grid-connected inverter under finite set model prediction control.

Background

Model Predictive Control (MPC) is a computer Control algorithm generated in the late 70 th of the 20 th century, has intuitive concept, is easy to Model, does not need precise Model and complex Control parameter design, has good effect on overcoming the problems of nonlinearity, uncertainty and the like in the industrial Control process, is easy to increase constraint, and has fast dynamic response and strong robustness.

In view of the great advantages, in the last decade, one of the MPCs is widely applied and developed in the related fields of power electronic converters, motor drives, power systems and the like, and it is inevitable that the FCS-MPC faces numerous challenges in practical application, such as a large amount of on-line calculation, if the calculation is overtime, the switching action is not the optimal switching action at that moment when applied, the effect of predictive control is reduced, and when a modeling error exists, the optimal switching function combination selected according to the traditional FCS-MPC algorithm loses optimality when applied to a three-phase inverter, so that the deviation occurs between the actual response curve of the controlled quantity of the system and the optimal response curve determined by the FCS-MPC algorithm, and the control performance of the system is finally affected. However, the existing identification method for the mismatch of the model prediction control parameters of the grid-connected inverter is based on the traditional model prediction current control, a large amount of calculation of an identification algorithm is superposed on the basis of a large amount of original calculation, the risk of calculation overtime is increased, most of the existing parameter identification schemes only identify the inductance, the resistance mismatch of the filter inductance is ignored, and when the resistance fluctuates in a large range, the steady-state error of the grid-connected current is increased, and the current quality is reduced.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provides a grid-connected inverter parameter identification method under the finite set model predictive control, so that an accurate model is obtained and the calculated amount of predictive control is reduced by introducing two predictive voltage error equations and identifying the inductance and the parasitic resistance thereof in real time, thereby improving the parameter robustness of the predictive control of the inverter model and improving the grid-connected current waveform quality of the inverter.

The invention adopts the following technical scheme for solving the technical problems:

the grid-connected inverter parameter identification method under the finite set model predictive control is characterized by comprising the following steps of:

step 1: when the k-time parameter mismatch is constructed by using the formula (1), a simplified nominal discrete model of the grid-connected inverter under the prediction control of a finite set model is as follows:

in the formula (1), V_α,β ^k*Is the optimal output voltage e of the grid-connected inverter under the stationary coordinate system of two phases αβ at the moment k_α,β ^kThe grid voltage under a stationary coordinate system of two phases αβ at the moment k, r 'is a nominal resistance, L' is a nominal inductance, i_{α,β_ref} ^k+1Is a reference current i in a stationary coordinate system of two phases αβ at the time of k +1_α，β ^kGrid-connected current under a stationary coordinate system of two phases αβ at the moment k, wherein Ts is a sampling time interval;

when the actual parameters at the moment k run, a simplified real discrete model of the grid-connected inverter under the prediction control of a finite set model is constructed by using the formula (2):

in the formula (2), V_α,β ^kIs the actual output voltage i of the grid-connected inverter under the stationary coordinate system of two phases αβ at the moment k_α,β ^k+1The actual current of the two-phase αβ stationary coordinate system at the moment of k +1, r is the actual resistance, and L is the actual inductance;

obtaining the optimal output voltage V of the grid-connected inverter under the static coordinate system of two phases αβ at the moment of k-1 by the formula (1) and the formula (2) respectively_α,β ^k-1*And the actual output voltage V of the grid-connected inverter under the actual two-phase αβ static coordinate system at the moment k-1_α,β ^k-1；

Step 2, obtaining the difference delta V between the optimal output voltage and the actual voltage under the stationary coordinate system of the two phases αβ at the moment of k-1 by using the formula (3)_α,β ^k-1：

In the formula (3), i_{α,β_ref} ^kIs a reference current i in a stationary coordinate system of two phases αβ at the time k_α,β ^k-1The grid-connected current is the grid-connected current under a stationary coordinate system of two phases αβ at the moment of k-1;

the difference delta between the optimal voltage estimation errors in the stationary coordinate system of the two-phase αβ at the time k-1 and the time k-2 is obtained by using the formula (4)_α,β ^k-1：

In the formula (4), i_{α,β_ref} ^k-1Is a reference current i in a stationary coordinate system of two phases αβ at the time of k-1_α,β ^k-2The grid-connected current is the grid-connected current under a stationary coordinate system of two phases αβ at the moment of k-2;

an identification expression L of the actual inductance L under the stationary coordinate system of the two-phase αβ at the moment k is constructed by using the formula (5)_{α,β_est} ^k：

Step 3, obtaining an estimated formula R of the actual resistance R under the stationary coordinate system of the two phases αβ at the moment k by using the formula (6)_{α,β_est} ^k：

Step 4, estimating the actual inductance L by using an α -axis inductance estimation formula shown in formula (7) to obtain an estimated value L of α -axis inductance at the time k_{α_est} ^k：

In the formula (7), L_{α_est} ^k-1Is an estimate of the α axis inductance, Δ, at time k-1_α ^k-1Is the difference between the α axis optimum voltage estimation errors at time k-1 and time k-2, i_αref ^k，i_αref ^k-1α -axis reference current, i, at time k, time k-1, respectively_α ^k，i_α ^k-1，i_α ^k-2α axis grid-connected current V at the time k, the time k-1 and the time k-2 respectively_α ^k-2，V_α ^k-1The output voltages of the α -axis grid-connected inverter at the time k-2 and the time k-1 respectively;

and 5: the actual resistance R is estimated by the equation (8) to obtain the resistance estimation value R at the time k_est ^k：

In formula (8), θ is the grid phase angle at the time of k-1, the variable a ═ 1 pi/4, 3 pi/4 ═ u [5 pi/4, 7 pi/4 ], the variable B ═ 0, 1 pi/4 ], [3 pi/4, 5 pi/4 ], [7 pi/4, 2 pi ];

step 6, estimating the α axis inductance value L at the k moment_{α_est} ^kTaking the resistance estimated value R at the k moment as a nominal inductance L_est ^kThe nominal resistance r' is used and substituted into the formula (1) to obtain a nominal discrete model at the moment k;

and 7: replacing k with k +1 to obtain a two-step prediction model in formula (1);

and 8: obtaining the optimal output voltage of the grid-connected inverter at the moment k +1 according to the two-step prediction model;

and step 9: obtaining a switching tube action signal S of the grid-connected inverter at the moment of k +1 by utilizing a value function optimization method according to the optimal output voltage_a ^k+1,S_b ^k+1,S_c ^k+1And outputs a switching tube action signal S of the grid-connected inverter after delaying for one period_a ^k+1,S_b ^k+1,S_c ^k+1The switching action of the grid-connected inverter at the moment of k +1 is realized;

step 10: and at the moment of k +1, assigning k +1 to k, and returning to the step 1 for execution.

Compared with the prior art, the invention has the beneficial effects that:

1. on the basis of simplified finite set model prediction control, the double-voltage error equation is constructed, so that the parasitic resistance of the inductor is identified while the inductor is identified, the parameter mismatch is completely eliminated, and the quality of grid-connected current is improved; the overall calculation amount of the finite set model predictive control based on parameter identification is reduced, and the risk of calculation timeout in one period is reduced.

2. The method carries out parameter identification on the basis of the simplified finite set model predictive control model in the step 1, avoids the huge total calculated amount caused by parameter identification on the basis of the traditional predictive current control, greatly reduces the total calculated amount and lightens the calculation burden of the controller;

3. in the invention, the double-voltage error equation is constructed in the step 2 and the step 3, the parasitic resistance of the inductor is identified while the inductor is identified, the mismatch of parameters is completely eliminated, and the robustness of the prediction control of the finite set model is improved.

4. In the invention, the singularity problem of the estimated inductance and estimated resistance expressions in the steps 2 and 3 is analyzed in detail in the steps 4 and 5, a simple and effective solution is provided, and the complexity of parameter estimation singularity problem analysis in the previous research is avoided.

Drawings

FIG. 1 is a block diagram of a grid-connected inverter system according to the present invention;

FIG. 2 is a schematic diagram of singularity processing in inductance estimation according to the present invention;

FIG. 3 is a schematic diagram of singularity processing in resistance estimation according to the present invention;

FIG. 4 is a diagram illustrating the effect of parameter estimation before and after adding parameter identification according to the present invention;

FIG. 5 is a graph of the effect of grid-connected current before and after parameter estimation is added in the present invention;

FIG. 6 is an FFT analysis chart of grid-connected current before parameter estimation is added in the present invention;

fig. 7 is an FFT analysis diagram of the grid-connected current after parameter estimation is added in the present invention.

Detailed Description

Referring to fig. 1, in this embodiment, a method for identifying parameters of a grid-connected inverter under simplified finite set model prediction control includes: the method comprises the steps of constructing a simplified model prediction control mismatch parameter prediction model and an actual model based on accurate parameters, constructing a double-voltage error equation according to the model, identifying real inductance and resistance in real time, constructing an accurate prediction model, calculating optimal output voltage, performing two-step prediction, and finally optimizing through a cost function to obtain optimal inverter switch tube action. Specifically, the method comprises the following steps:

step 1: when the k-time parameter mismatch is constructed by using the formula (1), a nominal discrete model based on the calculated optimal voltage of the grid-connected inverter under the prediction control of a finite set model is as follows:

in the formula (1), V_α,β ^k*Is the optimal output power of the grid-connected inverter under the stationary coordinate system of two phases αβ at the moment of kPressure, e_α,β ^kThe grid voltage under a stationary coordinate system of two phases αβ at the moment k, r 'is a nominal resistance, L' is a nominal inductance, i_{α,β_ref} ^k+1Is a reference current i in a stationary coordinate system of two phases αβ at the time of k +1_α,β ^kGrid-connected current under a stationary coordinate system of two phases αβ at the moment k, wherein Ts is a sampling time interval;

when the actual parameters at the moment k run, a grid-connected inverter real discrete model under the prediction control of a finite set model is constructed by using the formula (2) based on the calculation of the optimal voltage:

in the formula (2), V_α,β ^kIs the actual output voltage i of the grid-connected inverter under the stationary coordinate system of two phases αβ at the moment k_α，β ^k+1The actual current of the two-phase αβ stationary coordinate system at the moment of k +1, r is the actual resistance, and L is the actual inductance;

the optimal output voltage V of the grid-connected inverter under the static coordinate system of two phases αβ calculated at the moment of k-1 is obtained by the formula (1) and the formula (2) respectively_α,β ^k-1*And the actual output voltage V of the grid-connected inverter under the actual two-phase αβ static coordinate system at the moment k-1_α，β ^k-1But actually outputs a voltage V_α，β ^k-1The direct current voltage of the inverter and the switching action of the inverter at the moment of k-1 can be directly obtained; the prediction model based on the optimal voltage calculation only needs 1 time of calculation of the optimal voltage, and compared with 8 times of calculation of the predicted current of the traditional predicted current model, the calculation amount is greatly reduced, so that the overall calculation amount can be greatly reduced by parameter identification based on simplified model prediction control for calculating the optimal voltage.

Step 2: the optimal output voltage V of the inverter calculated in the step 1 is calculated_α,β ^k-1*With the actual output voltage V_α,β ^k-1Making a difference: Δ V_α,β ^k-1＝V_α,β ^k-1*-V_α,β ^k-1Obtaining the optimal output voltage and the optimal output voltage of the two-phase αβ stationary coordinate system at the moment of k-1 by using the formula (3)Difference between the voltages:

in the formula (3), i_{α,β_ref} ^kIs a reference current i in a stationary coordinate system of two phases αβ at the time k_α,β ^k-1The grid-connected current is the grid-connected current under a stationary coordinate system of two phases αβ at the moment of k-1;

further constructing the difference between the optimal voltage estimation errors at the front moment and the rear moment: delta_α,β ^k-1＝ΔV_α,β ^k-1-ΔV_α,β ^k-2And obtaining the difference of the optimal voltage estimation errors under the static coordinate system of the two phases αβ at the time k-1 and the time k-2 by using the formula (4):

in the formula (4), i_{α,β_ref} ^k-1Is a reference current i in a stationary coordinate system of two phases αβ at the time of k-1_α,β ^k-2The grid-connected current is the grid-connected current under a stationary coordinate system of two phases αβ at the moment of k-2;

in formula (4), there are, in general:

therefore, the resistance term in the equation (4) is omitted, and the identification expression L of the actual inductance L in the stationary coordinate system of the two phases αβ at the time k shown in the equation (6) is obtained_{α,β_est} ^k：

And 3, substituting the formula (6) in the step 2 into a voltage error expression (3) to obtain a relational expression J related to the actual resistance r and the nominal resistance r' under the stationary coordinate system of the two phases αβ at the moment of k-1 as shown in the formula (7)_α,βAs shown in formula (7):

obtaining an estimated expression R of the actual resistance R in the stationary coordinate system of the two-phase αβ at the time k from the expression (7)_{α,β_est} ^k：

And 4, step 4: when the inductance calculation formula (6) calculated in the step 2 is used for calculation, if the molecular term is 0, the peak value of the estimated value appears, and the estimated value is regarded as a singularity, and the singularity needs to be avoided to obtain a stable estimated value. The direct analysis of the singularity problem of the inductor is difficult, and according to the inverter switching vector relation diagram 2, a voltage vector relation shown as a formula (9) can be obtained:

in the formula (9), V_α,β ^k-2，V_α,β ^k-1The inverter output voltage V under the stationary coordinate system of two phases αβ at the time of k-2 and k-1 respectively_{grid_α,β} ^k-2，V_{grid_α,β} ^k-1The grid voltage under the stationary coordinate system of the two phases αβ at the moment k-2 and k-1 respectively.

In the case that the sampling frequency is sufficiently large, the amplitudes of the grid voltages at two adjacent time instants are approximately equal, and equation (9) can be converted into:

the form of the singularity in the formula (6) in the step 2 can be converted into the difference between the two front and rear switching actions by the formula (10), so that the singularity judgment process is greatly simplified. As can be seen from the inverter switching vector diagram 2, the singularity of the α -axis inductance estimation formula exists only when the inverter switching actions are identical at two adjacent times, while the singularity of the β -axis inductance estimation formula exists not only in the above case, but also when the inverter switching actions are switched between 000 and 100, between 001 and 101, between 010 and 110, and between 111 and 011, and the number of the singularities is more than that of the α -axis inductance estimation formula, so the response speed of the α -axis inductance estimation is faster than that of the β -axis inductance estimation formula, and the α -axis inductance estimation formula is used to estimate the actual inductance.

The actual inductance L was estimated using the α -axis inductance estimation formula shown in formula (11), and the estimated value L of α -axis inductance at time k was obtained_{α_est} ^k：

In the formula (11), L_{α_est} ^k-1Is an estimate of the α axis inductance, Δ, at time k-1_α ^k-1Is the difference between the α axis optimum voltage estimation errors at time k-1 and time k-2, i_αref ^k，i_αref ^k-1α -axis reference current, i, at time k, time k-1, respectively_α ^k，i_α ^k-1，i_α ^k-2α axis grid-connected current V at the time k, the time k-1 and the time k-2 respectively_α ^k-2，V_α ^k-1The output voltages of the α -axis grid-connected inverter at the time k-2 and the time k-1 respectively;

and 5: step 3, the singularity also exists in the resistance estimation formula (8), but the singularity is simple in form, namely the zero crossing point of the two-axis current, the phase difference of the currents of the two axes under the two-phase alpha beta static coordinate system is 90 degrees, and the phase difference of the singularity of the two-axis resistance estimation formula is also 90 degrees. By switching the estimation axes, the singularity of the estimation formula of each axis is easily avoided, namely, the singularity of the estimation formula of the alpha axis is estimated by using the resistance estimation formula of the beta axis within a certain range, the alpha axis stops estimation, and when the singularity of the beta axis reaches within a certain range, the resistance estimation formula of the alpha axis is used for estimation, and the beta axis stops estimation. However, the selection of the range near the two axes singularity affects the estimation effect, and theoretically, the farther away from the singularity, the smaller the oscillation of the estimation value is, and the better the estimation effect is.

In step 3, in equation (7), r' is the actual application value of the resistor in the prediction model at the time k-1, and the estimated value of r is the application value at the time k, and in the case that the sampling frequency is high enough, the amplitudes of the current before and after the current are considered to be equal, and equation (7) can be approximated by subtracting the voltage of the resistor at the time k from the voltage of the resistor at the time k-1 to obtain (12):

J_α,β≈V_r ^k-1-V_r ^k(12)

in the formula (14), V_r ^k-1，V_r ^kThe voltage across the resistor r at time k-1 and k, respectively, is divided by the sampling period Ts to obtain equation (13):

J_α,β/Ts≈(V_r ^k-1-V_r ^k)/Ts＝-ΔV_r/Ts (13)

in the formula (13), Δ V_rThe difference between the voltages across resistor r at time k, k-1.

Known that J is_α,βHysteresis V_r ^k-190 deg. and the voltage over resistor r is in phase with the current, so J_α,βAlso lags behind i_α,β ^k-190 deg. then J_α/i_α ^k-1Can be represented as (14):

J_α/i_α ^k-1＝J_{_α}sin(θ-90)/I_{_α}×sin(θ)＝J_{_α}/I_{_α}tan(θ-90) (14)

in formula (14), J_αIs the α -axis component, i, of the voltage difference at adjacent times across the resistor r_α ^k-1α -axis component of grid-connected current at time k-1, J_{_α}For the voltage difference amplitude α axis component across the resistor at adjacent times, I_{_α}And theta is the amplitude of the α axis component of the grid-connected current, and is the grid phase angle at the moment k-1.

In the same way, J_β/i_β ^k-1Can be represented by formula (15):

J_β/i_β ^k-1＝＝J_{_β}/I_{_β}tanθ (15)

in formula (15), J_βIs the β -axis component, i, of the voltage difference at adjacent times across the resistor r_β ^k-1β -axis component of grid-connected current at time k-1, J_{_β}For the amplitude of the β axial component of the voltage difference across the resistor at adjacent times, I_{_β}Is the magnitude of the β axis component of the grid-tie current.

As can be seen from the graph 3 of the equations (14) and (15), the oscillation of the α -axis resistance estimated value is small near pi/2 and 3 pi/2, and the oscillation of the β -axis resistance estimated value is small near pi and 2 pi, so that the constant amplitude oscillation point is taken as the critical point of the two-axis identification conversion, and the estimation of the inductance is stable overall.

Therefore, formula (14) is equal to formula (15), resulting in formula (16):

|J_{_α}/I_{_α}tan(θ-90)|＝|J_{_β}/I_{_β}tanθ| (16)

solving the critical angles of identification and conversion of the two-axis resistance under the two-phase αβ static coordinate system to pi/4, 3 pi/4, 5 pi/4 and 7 pi/4, estimating the actual resistance R by using a formula (17) to obtain a resistance estimation value R at the moment k_est ^k：

In formula (17), θ is a grid phase angle at the time of k-1, and the variable a ═ 1 pi/4, 3 pi/4 ═ u [5 pi/4, 7 pi/4 ], and the variable B ═ 0, 1 pi/4 ], [3 pi/4, 5 pi/4 ], [7 pi/4, 2 pi ],; the resistance estimation value obtained by the method has the minimum total oscillation amplitude and is in periodic oscillation, so that a smooth resistance estimation curve can be obtained by adopting the sliding average filtering.

Step 6, estimating the α axis inductance value L at the moment k_{α_est} ^kAs the nominal inductance L', the resistance estimate R at time k is taken_est ^kThe nominal resistance r' is used and substituted into the formula (1) to obtain a nominal discrete model at the moment k;

and 7: replacing k with k +1 to obtain a two-step prediction model in formula (1);

and 8: obtaining the optimal output voltage of the grid-connected inverter at the moment k +1 according to the two-step prediction model;

and step 9: obtaining a switching tube action signal S of the grid-connected inverter at the moment k +1 by using a value function optimization method according to the optimal output voltage_a ^k+1,S_b ^k+1,S_c ^k+1And outputs a switching tube action signal S of the grid-connected inverter after delaying for one period_a ^k+1,S_b ^k+1,S_c ^k+1The switching action of the grid-connected inverter at the moment of k +1 is realized;

step 10: and at the moment of k +1, assigning k +1 to k, and returning to the step 1 for execution.

In order to verify the effectiveness of the parameter identification method provided by the invention, a three-phase grid-connected inverter model with the rated capacity of 18kw is built in Matlab/simulink, the nominal values of the inductor and the resistor are respectively 1mH and 0.5 omega, and the actual inductor and the resistor are respectively 5mH and 0.3 omega. Fig. 4 is a graph showing the effect of identifying the inductance and the resistance before and after adding the parameter identification, and a parameter identification algorithm is added in 0.5s, wherein the nominal value is before 0.5s, and the identified actual inductance value is after 0.5 s. As can be seen from fig. 4, identifying the inductance and the resistance can quickly and accurately track the actual value of the inductance resistance. Fig. 5 shows the grid-connected current waveforms before and after parameter identification is added, and it can be seen that the three-phase grid-connected current distortion is serious before 0.1s and the current waveform is greatly improved after 0.1s by adding the parameter identification algorithm in 0.1 s. Fig. 6 is a graph of the FFT analysis result of the grid-connected current before the parameter identification is added, in which the grid-connected current amplitude significantly deviates from the current command value 40A, the current waveform distortion is severe, and the THD value is too high. Fig. 7 is a graph of the FFT analysis result of the grid-connected current after the parameter identification is added, the deviation of the amplitude of the grid-connected current is significantly improved, the current waveform is sinusoidal, and the THD value is greatly reduced. The parameter identification based on the simplified model prediction control can well improve grid-connected current amplitude deviation and current distortion caused by parameter mismatch, and greatly improves the robustness of the finite set model prediction control.

Claims (1)

1. A grid-connected inverter parameter identification method under the prediction control of a finite set model is characterized by comprising the following steps:

step 1: when the k-time parameter mismatch is constructed by using the formula (1), a simplified nominal discrete model of the grid-connected inverter under the prediction control of a finite set model is as follows:

in the formula (1), V_α,β ^k*Is the optimal output voltage e of the grid-connected inverter under the stationary coordinate system of two phases αβ at the moment k_α,β ^kThe grid voltage under a stationary coordinate system of two phases αβ at the moment k, r 'is a nominal resistance, L' is a nominal inductance, i_{α,β_ref} ^k+1Is a reference current i in a stationary coordinate system of two phases αβ at the time of k +1_α,β ^kGrid-connected current under a stationary coordinate system of two phases αβ at the moment k, wherein Ts is a sampling time interval;

when the actual parameters at the moment k run, a simplified real discrete model of the grid-connected inverter under the prediction control of a finite set model is constructed by using the formula (2):

in the formula (2), V_α,β ^kIs the actual output voltage i of the grid-connected inverter under the stationary coordinate system of two phases αβ at the moment k_α,β ^k+1The actual current of the two-phase αβ stationary coordinate system at the moment of k +1, r is the actual resistance, and L is the actual inductance;

obtaining the optimal output voltage V of the grid-connected inverter under the static coordinate system of two phases αβ at the moment of k-1 by the formula (1) and the formula (2) respectively_α,β ^k-1*And the actual output voltage V of the grid-connected inverter under the actual two-phase αβ static coordinate system at the moment k-1_α,β ^k-1；

Step 2, obtaining the difference delta V between the optimal output voltage and the actual voltage under the stationary coordinate system of the two phases αβ at the moment of k-1 by using the formula (3)_α,β ^k-1：

In the formula (3), i_{α,β_ref} ^kIs a reference current i in a stationary coordinate system of two phases αβ at the time k_α,β ^k-1The grid-connected current is the grid-connected current under a stationary coordinate system of two phases αβ at the moment of k-1;

the equation (4) is used to obtain αβ stillness of two phases at the k-1 time and the k-2 timeDifference delta of optimum voltage estimation error in coordinate system_α,β ^k-1：

In the formula (4), i_{α,β_ref} ^k-1Is a reference current i in a stationary coordinate system of two phases αβ at the time of k-1_α,β ^k-2The grid-connected current is the grid-connected current under a stationary coordinate system of two phases αβ at the moment of k-2;

an identification expression L of the actual inductance L under the stationary coordinate system of the two-phase αβ at the moment k is constructed by using the formula (5)_{α,β_est} ^k：

Step 3, obtaining an estimated formula R of the actual resistance R under the stationary coordinate system of the two phases αβ at the moment k by using the formula (6)_{α,β_est} ^k：

Step 4, estimating the actual inductance L by using an α -axis inductance estimation formula shown in formula (7) to obtain an estimated value L of α -axis inductance at the time k_{α_est} ^k：

In the formula (7), L_{α_est} ^k-1Is an estimate of the α axis inductance, Δ, at time k-1_α ^k-1Is the difference between the α axis optimum voltage estimation errors at time k-1 and time k-2, i_αref ^k，i_αref ^k-1α -axis reference current, i, at time k, time k-1, respectively_α ^k，i_α ^k-1，i_α ^k-2α axis grid-connected current V at the time k, the time k-1 and the time k-2 respectively_α ^k-2，V_α ^k-1The output voltages of the α -axis grid-connected inverter at the time k-2 and the time k-1 respectively;

and 5: the actual resistance R is estimated by the equation (8) to obtain the resistance estimation value R at the time k_est ^k：

In formula (8), θ is the grid phase angle at the time of k-1, the variable a ═ 1 pi/4, 3 pi/4 ═ u [5 pi/4, 7 pi/4 ], the variable B ═ 0, 1 pi/4 ], [3 pi/4, 5 pi/4 ], [7 pi/4, 2 pi ];

step 6, estimating the α axis inductance value L at the k moment_{α_est} ^kTaking the resistance estimated value R at the k moment as a nominal inductance L_est ^kThe nominal resistance r' is used and substituted into the formula (1) to obtain a nominal discrete model at the moment k;

and 7: replacing k with k +1 to obtain a two-step prediction model in formula (1);

and 8: obtaining the optimal output voltage of the grid-connected inverter at the moment k +1 according to the two-step prediction model;

and step 9: obtaining a switching tube action signal S of the grid-connected inverter at the moment of k +1 by utilizing a value function optimization method according to the optimal output voltage_a ^k+1,S_b ^k+1,S_c ^k+1And outputs a switching tube action signal S of the grid-connected inverter after delaying for one period_a ^k+1,S_b ^k+1,S_c ^k+1The switching action of the grid-connected inverter at the moment of k +1 is realized;

step 10: and at the moment of k +1, assigning k +1 to k, and returning to the step 1 for execution.

Images