CN110932309A

CN110932309A - Inverter control system and method based on ACSF-MPC and PI dual-mode switching

Info

Publication number: CN110932309A
Application number: CN201911044351.1A
Authority: CN
Inventors: 王安娜; 张佳豪; 林森; 程科
Original assignee: Northeastern University China
Current assignee: Northeastern University China
Priority date: 2019-10-30
Filing date: 2019-10-30
Publication date: 2020-03-27

Abstract

The invention discloses an inverter control system and method based on ACSF-MPC and PI dual-mode switching, belonging to the technical field of current control of inverters, wherein the system control method adopts dual-mode switching, can realize no-static-error regulation on system current in a steady state and adopts simple and effective dual-PI control; meanwhile, a fixed frequency model is adopted for prediction during dynamic regulation, the optimal switching sequence control is predicted by the grid-connected inverter, a traversal optimization mode is adopted for improvement, and an ideal output voltage vector is obtained

And comparing the difference with a model of the difference between the central vectors of the 6 sectors to obtain the minimum deviation, so that the calculated amount during traversal is reduced, and the quick response is realized. In addition, in the controlIn the process, the virtual reference current error compensation value is obtained by adopting improved fuzzy compensation and is used as the compensation quantity of the current reference value output by the outer ring PI, so that the problems of periodic disturbance and current ripple are solved.

Description

Inverter control system and method based on ACSF-MPC and PI dual-mode switching

Technical Field

The invention relates to the technical field of current control of inverters, in particular to an inverter control system and method based on ACSF-MPC and PI dual-mode switching.

Background

With the mass access of new energy power generation systems and flexible direct current transmission (FACTS) devices to the power grid, the grid-connected inverter is widely applied to alternating current driving, active filters, static reactive compensators, power grid interconnection and the like as a bridge device. The grid-connected inverter is an important medium for transmitting the distributed electric energy to a power grid, realizes the function of converting direct current into alternating current, is the basis for realizing the effective utilization of renewable energy, and has the performance directly influencing the reliability and stability of a distributed power generation system.

Common control schemes for inverters include voltage mode control and current mode control. Due to the defects that voltage type control contains a large amount of low-order harmonics, dynamic performance is poor and the like, most grid-connected inverters adopt a current type control method, namely, a power grid is regarded as an infinite power supply, and the voltage and the frequency of the grid-connected inverter do not change along with the change of grid-connected current of the grid-connected inverter. At present, the mainstream current tracking control algorithm mainly comprises an artificial intelligence control algorithm represented by current hysteresis control, sliding mode control, linear control, predictive control and fuzzy neural network control. The MPC has the main idea of predicting all possible future outputs of the system by using all possible future actions of the system, and then selecting the optimal future action through constraint conditions, performance optimization indexes and the like, and is an optimal solution method in general. In the field of power electronics, Control actions, i.e., the on/off of a power device, and Model Predictive Control applied to the field of power electronics can be divided into two major categories, namely, Finite Control Set Model Predictive Control (FCS-MPC for short) and Continuous Control Set Model Predictive Control (CCS-MPC for short) according to different combinations of switching actions. The FCS-MPC solves the optimal control problem by utilizing the characteristic of a limited number of switching states of the power converter, each switching state is not changed in each control period, so the total amount of control actions is limited, the obtained optimal solution is not necessarily obtained due to the limited number of the controls in each control period, only one group of optimal switching states is selected in one switching period, and the duty ratio of the switch drive is nonzero and is one, so that the output current brings ripples. The CCS-MPC is an element that adds time to the control set of the FCS-MPC so that the switching action is variable every cycle. And controlling the action time of each of the switch states such that the controlled variable converges to the reference value along a fixed switching period. However, neither of the above two control methods solves the optimization problem in the inaccurate case of the system model.

When the system model is accurate, the model predictive control can realize the rapid and accurate tracking effect, but the double frequency fluctuation can not be eliminated, at the moment, the proportional-integral control can realize the non-static tracking of the current, and the adoption of the double PI control under the steady state condition becomes simple and effective. In an actual process, unknown deviations which are difficult to measure directly are introduced into system parameters due to uncertain factors such as device process, temperature, abnormal operation state and faults, the deviations bring periodic disturbance to a system, and when the disturbance is too large, the control effect is influenced, and even the system is crashed. How to control the inverter to realize grid connection under different working conditions through a set of effective and reasonable control scheme is an urgent problem to be solved.

Disclosure of Invention

In view of the above-mentioned deficiencies of the prior art, the present invention provides an inverter control system and method based on ACSF-MPC and PI dual-mode switching. The method is improved on the basis of CCS-MPC (continuous control set model predictive control), and provides Adaptive constant switching frequency model predictive control (ACSF-MPC).

In order to solve the technical problems, the technical scheme adopted by the invention is as follows: an inverter control system based on ACSF-MPC and PI dual-mode switching, the system structure is shown in FIG. 1, and the system comprises:

the data acquisition module is used for acquiring the voltage and the current value of the AC side and the voltage value of the DC side of the inverter at the current moment;

the outer ring control module compares the sampled direct-current side voltage with a given voltage value and obtains a reference current value through a PI (proportional integral) controller;

the mode switching module compares the reference current value with an alternating current side current value obtained by sampling at the power grid side, so as to determine and select a proper control mode;

a data control mode 1, wherein a current inner ring adopts a PI control mode;

a data control mode 2, wherein a constant frequency model is adopted by a current inner loop to predict a control mode;

and the function optimization module adopts improved fuzzy compensation on the current reference value in the fixed frequency model prediction to obtain the compensation quantity of the virtual reference current value, feeds the compensation quantity back to the outer ring control module, and finally takes the optimal switching sequence and the action time thereof as the control signal of the inverter control system.

The control diagram of the control system is shown in fig. 2, and a method for controlling an inverter by using an inverter control system based on ACSF-MPC and PI dual-mode switching is shown in fig. 3, and the method comprises the following steps:

step 1: collecting voltage, current value and direct current side voltage value of an alternating current side of the inverter, and obtaining alternating current side current quantity and direct current side voltage quantity through a mean value filtering module of the DSP;

step 2: voltage outer loop control is adopted;

comparing the voltage value sampled at the DC side with a given voltage value, and obtaining a current reference value i under dq coordinate system through a PI (proportional-integral) controller^* _drefAnd i^* _qref；

Making q-axis given current zero, comparing the voltage value obtained by sampling at the DC side with the given value and carrying out PI control to obtain d-axis given current reference value i of the current inner ring^* _dref；

Wherein, V^* _dcrefThe reference value of the voltage on the direct current side,

setting current parameter for q axisExamination value, V_dcDC-side voltage sampling value, K_p、K_iRespectively, a proportional coefficient and an integral coefficient of the proportional PI controller.

And step 3: for the current reference value i obtained in the step 2^* _dref、i^* _qrefComparing the current value obtained by sampling at the power grid side, if the deviation is less than 0.5A, adopting a PI control mode for the current inner loop, executing the step 4 and ending; if the deviation is larger than 0.5A, the current inner loop adopts a fixed frequency model prediction control mode, and step 5 is executed;

and 4, step 4: making a difference between the current reference value and the sampled power grid side current value, performing PI regulation to obtain an inverter output voltage vector, obtaining a PWM output wave through an SVPWM (space vector pulse width modulation) unit, and turning to the step 9;

the inner ring adopts PI control as follows:

wherein k is_p、k_iProportional and integral coefficients, v, of the inner loop PI controller_d、v_qFor the inverter output voltage, i, in dq coordinate system_d、i_qSampling the current, e, for the inverter side in dq coordinate system_d、e_qIs grid-connected voltage under dq coordinate system, L is filter inductance, omega is grid frequency, and s is complex frequency.

And 5: introducing the system parameters obtained in the previous control period of the current amount and the voltage amount obtained in the step (1) into a constant-frequency current prediction model, and calculating the predicted current value of the next control period;

for the convenience of design, the current and the voltage of the three-phase grid-connected inverter under a two-phase static coordinate system are converted into αβ coordinate system through Clark conversion;

calculating the predicted current value of the next period:

wherein i_α(k) And i_β(k) Is αβThree-phase sampling current on the grid side at the beginning of the kth (k is 1,2,3 …) control period under the coordinate system; f. of_αiAnd f_βi(i ═ 1,2,3) is the increment of the instantaneous current under Clark transformation, t_i(i is 1,2,3) is the corresponding inverter output voltage vector v_iThe action time of (c);

if it is assumed that during the sampling interval f_αiAnd f_βiIs constant, then f_αi＝f_αi,kAnd f_βi＝f_βi,kWherein f is_αi,kAnd f_βi,kIs a vector v_iThe instantaneous k values of (a) are as follows:

wherein R is equivalent resistance at the power grid side, L is filter inductance, i_α、i_β，v_α、v_β，e_α、e_βRespectively, grid-connected current, inverter output voltage and grid-connected side voltage under an αβ coordinate system.

Step 6: defining a target function by adopting a least square optimization method, obtaining the action time of 3 output voltage vectors when the target function is at the minimum value by utilizing an extremum solving method, and solving the voltage vector value of a sector by adopting an improved optimization strategy;

the cost function of the finite set model predictive control is the square of the difference between the current predicted value and the current reference value, for the fixed frequency model predictive control, similar performance optimization indexes are adopted, and the performance optimization evaluation function is defined as follows:

J＝(E'_α)²+(E'_β)²

wherein, E'_αAnd E'_βRespectively representing the difference between the predicted current value and the reference value at the time k +1,

and

for a given reference current value at the end of the control period, i_α(k +1) and i_β(k +1) is a predicted current value at the moment of k + 1;

the smaller the value of the evaluation function J is, the smaller the deviation between the output current at the moment of k +1 and the reference current is, and the action time of 3 output voltage vectors when the target function takes the minimum value can be obtained by using the method of obtaining the extreme value:

in addition

Finally, calculating the output voltage vector v of the corresponding inverter in the step 5_iTime of action t_iComprises the following steps:

wherein the content of the first and second substances,

t is the switching period, f_αiAnd f_βi(i ═ 1,2,3) is the increment of the instantaneous current under Clark transformation;

by using the corresponding central vector of each sector, the voltage vector of ideal output can be obtained

In the sector, the objective function is:

wherein the content of the first and second substances,

the values of the current increment corresponding to the central vector αβ of each sector in a coordinate system are respectively obtained by calculating the average value of the calculated numbers of the corresponding three vectors of the central vector of the sector in a αβ coordinate system and are converted into the following values by Clark:

wherein the content of the first and second substances,

respectively, the value of the center vector of the sector in the αβ coordinate system, v_α0、v_β0Is a zero vector v₀Voltage vector corresponding to value in αβ coordinate system v_αi、v_βiAnd v_α(i+1)、v_β(i+1)Are each v_iAnd v_(i+1)Voltage vectors corresponding to the values of the voltage vectors adjacent to the two sectors in the αβ coordinate system;

the ideal output voltage vector can be obtained by calculating the minimum value of J

And then the acting time of the three point pressure vectors can be obtained by utilizing SVPWM (space vector pulse width modulation).

And 7: adopting a virtual reference current value obtained by the improved fuzzy compensation as a compensation quantity of a current reference value output by the outer ring PI;

step 7.1: adopting a two-dimensional fuzzy controller to calculate the inductive current error w at the time k_kAnd rate of change of error w_k-w_k-1The output quantity delta w of the fuzzy control is obtained as the input of a two-dimensional fuzzy controller; squaring the error to prevent error inundation, and then storing the past 2n output quantities in the DSP as a database for obtaining a virtual reference current compensation quantity and a weight coefficient M;

firstly, historical data of fuzzy control output quantity is stored in the DSP, and the data obtained for 2n times is as follows:

Δw²＝[Δw²(k-2n) Δw²(k-2n+1) ... Δw²(k-1)]

the change of the virtual reference current of the last n times of sampling is as follows:

Δi_ref＝[Δi_ref(k-n) Δi_ref(k-n+1) ... Δi_ref(k-1)]

step 7.2: modeling the system by adopting a linear fitting method, wherein the formula for obtaining the virtual reference is as follows:

wherein M is^T＝[M₁M₂… M_n]A weighting factor representing an error of the virtual reference current change;

then, one can obtain:

step 7.3: using the latest set of error data [ Δ w ]²(k-n) w²(k-n+1) ... Δw²(k-1)]^TAnd a virtual reference current variation Δ i_ref(k) Error replaces [ Δ w ] in the weighting factor M equation described in step 7.2²(k-2n) w²(k-2n+1)... Δw²(k-n-1)]^TAnd Δ i_ref(k-n), resulting in a weight factor M for the latest iteration;

step 7.4: by multiplying the weighting factors M and [ Δ w ]²(k-n) w²(k-n+1) ... Δw²(k-1)]^TMultiplying to obtain a virtual reference Δ i_ref(k) Expressed as:

Δi_ref(k)＝M[Δw²(k-n) Δw²(k-n+1) ... Δw²(k-1)]^T

step 7.5: error Δ i of changing virtual reference current_ref(k) Compensating for reference current on α axis

The method comprises the following steps:

wherein i^* _αref(k) The reference value of the inductive current sampled at the kth time is obtained by obtaining an output value by a PI controller of a voltage outer ring and performing Clark inverse transformation,

representing the inductor current reference value sampled at the (k +1) th time on the power grid side;

is subjected to an inverse Clark transformation of the output reference value in dq coordinates:

and (6) substituting the inductor current reference value of the (k +1) th sampling of the power grid side under the αβ coordinate system into the step 6 for updating the current reference value, namely performing model prediction control.

And 8: according to the minimum switching loss principle, a zero vector concentration implementation method is adopted, namely, a modulation signal is obtained according to the distribution principle of voltage vectors in 5 levels of SVPWM to control a switching tube of the inverter, and the step 9 is carried out;

and step 9: and applying the output result, namely the control signal to a switching tube of the inverter.

Adopt the produced beneficial effect of above-mentioned technical scheme to lie in:

1. the invention adopts dual-mode switching, can realize no-static-error adjustment on the system in a steady state, and simultaneously adopts fixed-frequency model prediction in dynamic adjustment to realize quick response;

2. aiming at the optimal switch sequence control of grid-connected inverter prediction, the invention adopts a traversal optimization mode for improvement to ideally output a voltage vector

Comparing with the mode of the difference between the central vectors of the 6 sectors to obtain the minimum deviation, reducing the calculation amount of the traversal type and saving the time;

3. in the control process, the virtual reference current error compensation value obtained by adopting the improved fuzzy compensation is used as the compensation quantity of the current reference value output by the outer ring PI, so that the problems of periodic disturbance and current ripple are solved, and meanwhile, an improved method of taking the square of the error and then weighting is adopted to realize the accurate compensation of the compensation value to the reference value aiming at the problem of error inundation in the calculation process.

Drawings

FIG. 1 is a structural diagram of an inverter control system based on ACSF-MPC and PI dual mode switching according to the present invention;

FIG. 2 is a system control diagram of an inverter control system based on ACSF-MPC and PI dual mode switching according to the present invention;

FIG. 3 is a flowchart of a method for controlling an inverter using an inverter control system based on ACSF-MPC and PI dual mode switching according to the present invention;

FIG. 4 is a diagram of a three-phase two-level grid-connected inverter with an inductance load according to the present invention;

fig. 5 is a space vector diagram of the three-phase two-level grid-connected inverter under αβ coordinate system corresponding to 8 switching states and voltages;

FIG. 6 is a three-phase waveform diagram corresponding to different switching sequences of switches at different positions of a reference voltage vector according to the present invention;

(a) the position of UREF is a zone I (theta is more than or equal to 0 degree and less than or equal to 60 degrees);

(b) the position of UREF is a zone II (theta is more than or equal to 60 degrees and less than or equal to 120 degrees);

(c) the position of UREF is a zone III (theta is more than or equal to 120 degrees and less than or equal to 180 degrees);

(d) the position of UREF is an IV zone (theta is more than or equal to 180 degrees and less than or equal to 240 degrees);

(e) the position of UREF is a V zone (theta is more than or equal to 240 degrees and less than or equal to 300 degrees);

(f) the position of the UREF is a VI area (theta is more than or equal to 300 degrees and less than or equal to 360 degrees);

FIG. 7 is a grid-connected graph with unknown inductance and resistance of the three-phase grid-connected inverter of the invention;

FIG. 8 is a graph of the current variation under the action of the voltage vector sequence of the present invention;

(a)i_α(ii) a change in (c);

(b)i_β(ii) a change in (c);

wherein-is the output current, -is the reference current.

Detailed Description

The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.

As shown in fig. 1, an ac sf-MPC and PI dual-mode switching based inverter control system includes:

the data control module 1 adopts a PI control mode for the current inner loop;

the data solving module 2 adopts a fixed frequency model prediction control mode for the current inner loop;

As shown in fig. 4, the topology of the three-phase grid-connected inverter is composed of four parts, namely an input direct current source, a three-phase inverter bridge, an output filter and an alternating current grid. V_dcIs a DC bus voltage, Q₁～Q₆Denotes 6 insulated gate transistors (IGBT) with antiparallel diodes i_a、i_b、i_cFor the grid-connected current of the inverter, v_a、v_b、v_cFor the three-phase inverter bridge output voltage of the inverter, e_a、e_a、e_aThe three-phase grid-connected voltage is obtained, L is a power grid side filter inductor, R is an equivalent series impedance of an inverter loop, and L, R is a prediction model parameter.

In this embodiment, an L-type filter is adopted, and other types of filters are also suitable, and the basic voltage equation of the three-phase grid-connected inverter obtained from fig. 4 is as follows:

through Clark transformation with unchanged voltage amplitude, a mathematical model of the three-phase grid-connected inverter under a two-phase static coordinate system is obtained as follows:

wherein v is an inverter output voltage vector; e is a grid voltage vector; i is the output current vector.

The method for controlling the inverter by adopting the inverter control system based on ACSF-MPC and PI dual-mode switching has a flow chart shown in FIG. 2, and comprises the following steps:

step 2: voltage outer loop control is adopted;

given a current reference value, V, for the q-axis_dcDC-side voltage sampling value, K_p、K_iRespectively, a proportional coefficient and an integral coefficient of the proportional PI controller.

And step 3: for the current reference value i obtained in the step 2^* _dref、i^* _qrefComparing the current value obtained by sampling at the power grid side, if the deviation is less than 0.5A, adopting a PI control mode for the current inner loop, executing the step 4 and ending; if the deviation is larger than 0.5A, the current inner ring adopts a fixed frequency model prediction mode, and step 5 is executed;

the inner ring adopts PI control as follows:

calculating the predicted current value of the next period:

wherein i_α(k) And i_β(k) Is three-phase sampling current at the beginning of the k (k is 1,2,3 …) th control period under the αβ coordinate system, f_αiAnd f_βi(i ═ 1,2,3) is the increment of the instantaneous current under Clark transformation, t_i(i is 1,2,3) is the corresponding inverter output voltage vector v_iThe action time of (c);

J＝(E'_α)²+(E'_β)²

and

in addition

Finally, the action time t in step 5 is calculated_iComprises the following steps:

wherein the content of the first and second substances,

if a traversing optimization method is directly adopted according to the steps, each group of switch sequences needs to be calculated, the calculated amount of the acting time for solving the optimal switch sequence is greatly increased, and therefore an improved optimization strategy is adopted to obtain the voltage vector value of the sector;

assuming a target output voltage vector of a grid-connected inverter

The reference current can be effectively tracked, and fig. 5(b) shows the positions of the center vector of 6 sectors and the target output voltage vector.

Can be adjusted by aiming at a target voltage vector

The sector is positioned in an optimizing mode, and then the action time of 3 voltage vectors is adjusted by using the latest voltage vector sequence. According to the cosine theorem of the triangle, the central vectors of the sectors I, II, III, IV, V and VI can be obtained from the geometrical relation and are transformed into the following vectors by Clark:

wherein the content of the first and second substances,

In the sector, the objective function is:

by finding the minimum value of J', the target vector can be obtained

And further calculating the action time of the three switching sequences corresponding to the sector through SVPWM.

Under the condition that the three-phase grid-connected inverter is provided with a common resistive load, the numerical value of resistance and inductance can be given due to manufacturing process, abnormal working state, special temperature and the like; model mismatch further increases model error with unmeasured bias, and Δ R and Δ L represent positional deviations of resistive and inductive parameters, as shown in fig. 7, in view of these uncertain disturbances. Error compensation control is employed to eliminate this.

step 7.1: adopting a two-dimensional fuzzy controller to calculate the inductive current error w at the time k_kAnd rate of change of error w_k-w_k-1The output quantity delta w of the fuzzy control is obtained as the input of a two-dimensional fuzzy controller;

w_c(k)＝w_k(k)-w_k(k-1)

wherein i_α(k) And

α sampled value of current and reference value, w, at time k below axis_kIs the inductor current error at time k, w_c(k) The rate of change of the inductor current error at time k;

in order to improve the control precision of the controller, the input quantity and the fuzzy subset of the input quantity on the domain interval are defined as { large Negative (NB), medium Negative (NM), small Negative (NS), Zero (ZO), small Positive (PS), medium Positive (PM) and large Positive (PB) }7 language values.

Fuzzifying input and output, and establishing a fuzzy subset:

w_k＝{NB NS ZO PS PB}

w_c＝{P O N}

Δw＝{NB NM NS ZO PS PM PB}

in order to ensure that the basic domains of argument of each variable can be better covered by fuzzy, the domains of argument of input quantity and output quantity are set as follows:

w_k＝{-10 -5 0 5 10}

w_c＝{-5 -3 03 5}

Δw＝{-2 -1.5 -1 -0.5 0 0.5 1 1.5 2}

as a core of the fuzzy control, it is important to establish a reasonable and effective rule, and the fuzzy control rule table is shown in table 2, and the rule is as follows:

1) when the error is a positive large value and the error change rate is a positive value, the output is a large positive value so as to reduce the error as soon as possible;

2) when the error is positive and large but the rate of change of the error is negative, the output should be a small positive value or 0.

TABLE 2 fuzzy controller rule Table

When fuzzy control is carried out, a plurality of control rules are subject to the deduction calculation, and then control output is obtained by combining all deduction results obtained by calculation; in order to obtain the output of the controlled system, the fuzzy set must be defuzzified, and the method for defuzzifying adopts an area barycenter method to obtain the fuzzy control output quantity delta w.

Squaring the error to prevent error inundation, and then storing the past 2n fuzzy control output quantities delta w in a DSP (digital signal processor) to be used as a database for obtaining a virtual reference current compensation quantity and a weight factor M;

Δw²＝[Δw²(k-2n) Δw²(k-2n+1) ... Δw²(k-1)]

Δi_ref＝[Δi_ref(k-n) Δi_ref(k-n+1) ... Δi_ref(k-1)]

then, one can obtain:

step 7.3: using the latest set of error data [ Δ w ]²(k-n) w²(k-n+1) ... Δw²(k-1)]^TAnd a virtual reference current variation Δ i_ref(k) Error replaces [ Δ w ] in the weighting factor M equation described in step 7.2²(k-2n) w²(k-2n+1)... Δw²(k-n-1)]^TAnd Δ i_ref(k-n) to obtain a weight factor M for the latest iteration;

Δi_ref(k)＝M[Δw²(k-n) Δw²(k-n+1) ... Δw²(k-1)]^T

The method comprises the following steps:

inductive current reference value representing k +1 th sampling of power grid side；

The method aims at controlling the current, enables the predicted optimal switching sequence to control 3 switching states in one switching period to form a switching sequence and controls the action time of each switching period when the model predicts that the current is directly controlled to be the research direction. Fig. 5(a) shows a space vector diagram corresponding to 8 switching states of the two-level grid-connected inverter.

In order to reduce the switching times and harmonic content, according to the minimum switching loss principle, a zero vector concentration method is adopted, and finally 5-segment SVPWM is adopted, so that only 3 switching states are switched in each switching period, and the harmonic content is not too large. The reference vector voltage UREF is located at a position and a switching sequence shown in table 1.

TABLE 1 location of UREF and switch switching sequence alignment

In accordance with the above principles, fig. 8 shows the ideal case of output current i when the vector sequence of output voltages is active during one sampling period_αAnd i_βIs the output current i_αAs shown in FIG. 8(a), the output current i_βFIG. 8(b) shows the variation of (A).

In the optimal control of the selection switch sequence, it can be seen from fig. 8 that the output voltage vector sequence corresponding to the switch sequence includes i_αThe increasing voltage vector also includes the decreasing voltage vector, likewise i_βThis situation is also satisfied. It is ensured that the actual current change resulting from the conversion and sampling between the instants k +1 and k after one period is close. According to the 5-segment SVPWM, only one arm needs to be operated when switching states are switched 3 times. At this time, only the commutation time t of the three voltage vectors vi calculated according to step 6 is needed₁，t₂，t₃And i belongs to {1,2,3}, so that the dynamic and steady-state control of the grid-connected inverter can be realized.

Claims

1. An inverter control system based on ACSF-MPC and PI dual-mode switching is characterized by comprising the following components:

the data control module 1 adopts a PI control mode for the current inner loop;

the data control module 2 adopts a fixed frequency model prediction control mode for the current inner loop;

2. The method for controlling the inverter by using the ACSF-MPC and PI dual mode switching based inverter control system as claimed in claim 1, comprising the steps of:

step 1: collecting a current value and a direct current side voltage value of an inverter on an alternating current side, and obtaining an alternating current side current amount and a direct current side voltage amount through an average filtering module of a DSP (digital signal processor);

step 2: voltage outer loop control is adopted;

and 5: the current and voltage obtained in the step (1) and the system parameters obtained in the previous control period are brought into a constant frequency current prediction model, and the predicted current value of the next control period is calculated;

3. The method for controlling the inverter of the ACSF-MPC and PI dual mode switching based inverter control system as claimed in claim 2, wherein the procedure of step 2 is as follows:

given a current reference value, V, for the q-axis_dcFor sampling the DC side voltage, K_p、K_iRespectively, a proportional coefficient and an integral coefficient of the proportional PI controller.

4. The method for controlling the inverter of the ac sf-MPC and PI dual-mode switching based inverter control system according to claim 2, wherein the process of obtaining the inverter output voltage vector by performing PI regulation by subtracting the current reference value from the sampled grid-side current value in step 4 is as follows:

the inner ring adopts PI control as follows:

5. The method for controlling the inverter of the ACSF-MPC and PI dual mode switching based inverter control system as claimed in claim 2, wherein the procedure of step 5 is as follows:

calculating the predicted current value of the next period:

6. The method for controlling the inverter of the ACSF-MPC and PI dual mode switching based inverter control system as claimed in claim 2, wherein the procedure of step 6 is as follows:

J＝(E'_α)²+(E'_β)²

and

in addition

wherein the content of the first and second substances,

In the sector, the objective function is:

wherein the content of the first and second substances,

the current increment value corresponding to the central vector αβ coordinate system of each sector and the central vector value of the sector in the αβ coordinate system are respectively calculated by the corresponding three vectorsMean value and Clark transformation:

wherein the content of the first and second substances,

And then the acting time of three voltage vectors can be obtained by utilizing SVPWM (space vector pulse width modulation).

7. The method for controlling the inverter of the ACSF-MPC and PI dual mode switching based inverter control system as claimed in claim 2, wherein the procedure of step 7 is as follows:

Δw²＝[Δw²(k-2n) Δw²(k-2n+1) ... Δw²(k-1)]

Δi_ref＝[Δi_ref(k-n) Δi_ref(k-n+1) ... Δi_ref(k-1)]

then, one can obtain:

step 7.3: using the latest set of error data [ Δ w ]²(k-n) w²(k-n+1) ... Δw²(k-1)]^TAnd a virtual reference current variation Δ i_ref(k) Error replaces [ Δ w ] in the weighting factor M equation described in step 7.2²(k-2n) w²(k-2n+1) ...Δw²(k-n-1)]^TAnd Δ i_ref(k-n) to obtain a weight factor M for the latest iteration;

Δi_ref(k)＝M[Δw²(k-n) Δw²(k-n+1) ... Δw²(k-1)]^T

The method comprises the following steps: