阅读说明：本技术 带阻感负载三电平逆变器的改进模型预测控制方法 (Improved model prediction control method for three-level inverter with inductance-resistant load ) 是由 卞婉春 卜京 张飞云 孙莹 郑铭洲 夏凡吴双 夏星星 蒋建斌 于 2019-08-29 设计创作，主要内容包括：本发明公开了一种带阻感负载三电平逆变器的改进模型预测控制方法。该方法主要针对阻感负载之下,基于模型预测的二极管箝位型三电平(NPC)在一个周期之内需要完成对27个开关状态的电流、电压预测和目标函数的计算量比较大的问题,提出将目标函数中的电流预测值用电压预测值代替,使得预测控制的目标变为在27个电压矢量中选择与参考矢量最为接近的一个,在一定程度上减少了运算量,提高了效率,同时也不影响其性能。(The invention discloses an improved model prediction control method for a three-level inverter with a resistive load. The method mainly aims at the problem that under a resistive load, the current and voltage prediction of 27 switch states and the calculation amount of an objective function are large when a model prediction-based diode clamping type three-level (NPC) needs to be completed in one period, and the current prediction value in the objective function is replaced by a voltage prediction value, so that the target of prediction control is changed into one closest to a reference vector from 27 voltage vectors, the calculation amount is reduced to a certain extent, the efficiency is improved, and the performance of the method is not influenced.)

1. An improved model prediction control method of a three-level inverter with a resistive-inductive load is characterized by comprising the following steps:

1) detecting three-phase load current and two voltage dividing capacitor voltages at a direct current side aiming at an inverter load model;

2) setting a reference current and performing phase angle compensation; setting an initial time optimal switch state and an initial objective function optimal value g _ opt;

3) converting the three-phase load current in the step 1) from abc coordinates to alpha beta coordinates, and performing clark transformation;

4) carrying out t on the current and the DC side capacitor voltage according to a discrete formula by using the optimal switch state at the initial moment_k+1Predicting the time, and simultaneously calculating a voltage vector;

5) using t in step 4)_k+1Predicted value of time, predicting t according to discrete formula_k+2The dc side capacitor voltage at a moment;

6) calculating improved objective functions g under 27 switch states by using the voltage vector value in the step 4) and the predicted value of the direct-current side capacitor voltage in the step 5);

7) comparing the improved objective function value in the step 6) with the initial objective function optimal value g _ opt, selecting a small value between the improved objective function value and the initial objective function optimal value as a new objective function optimal value, outputting the switching state at the moment, and finally applying the new switching state to drive the inverter.

2. The improved model predictive control method of the three-level inverter with the inductive load according to claim 1, is characterized in that: the inverter load model in the step 1) comprises an NPC inverter, an RL load, a direct-current voltage source, two voltage division capacitors and a driving module.

3. The improved model predictive control method of the three-level inverter with the inductive load according to claim 1, is characterized in that: in the step 2), the reference current adopts sine wave, a numerical value is set, and the reference current is calculated according to the formula (1) by using a phase angle compensation method

Wherein the content of the first and second substances,is t_k+2The reference current at the time of day is,is t_kThe reference current at time, ω ═ 2 π f, f is the frequency, and Ts is the sampling time.

4. The improved model predictive control method for the three-level inverter with the inductive load according to claim 1, wherein in the step (3), a clark formula is used to simplify the calculation, and the conversion formula is as follows:

wherein i_α,i_βFor the current i in a two-phase coordinate system_a,i_b,i_cThe current is in a three-phase coordinate system.

5. The improved model predictive control method of the three-level inverter with the inductive load according to claim 1, is characterized in that: in the step (4), according to the characteristics of the three-level inverter with the inductance-rejection load, t is obtained by adopting the discrete formulas (3) and (4)_k+1Predicted current of time of dayAnd the predicted voltage of the capacitorCalculating t by backstepping according to the formula (3)_k+1Voltage vector of time of day

Where R is the resistance on the load side, L is the inductance on the load side, Ts is the sampling time, i_c1(k)、i_c2(k) Is t_kCapacitor current u on the time DC side_α,β(k) Is the output voltage of the inverter for the kth cycle; wherein i_c1(k)、i_c2(k) Calculated according to the following formula:

wherein i_sa,sb,sc(k) Is t_kLoad current at time H_sa,sb,scIs the on-off state at that moment.

6. The improved model control method of the three-level inverter with the inductive load according to claim 1, is characterized in that: in the step (5), t is obtained according to the formula (6)_k+2Time-of-day DC-side capacitance predicted voltage

Where R is the resistance on the load side, L is the inductance on the load side, Ts is the sampling time,is t_k+1The current predicted value at the time of day,is t_kCapacitor current u on the time DC side_α,β(k +1) is the output voltage of the inverter for the k +1 th cycle;is t_k+1The predicted current at that time and the predicted voltage of the dc-side capacitance.

7. The improved model control method of the three-level inverter with the inductive load according to claim 1, wherein in the step (6), the components of the objective function comprise: current tracking, midpoint potential balance, common-mode voltage reduction and switching frequency reduction, and an objective function is expressed in the form of absolute error:

in the formula (I), the compound is shown in the specification,andis the voltage vector corresponding to the switch state at t_k+1Real and imaginary parts of time, u_jα(k +1) and u_jβ(k +1) is the voltage vector at t_k+1Real and imaginary parts of, lambda_dc、λ_nAnd λ_cmAre a midpoint potential balance weight factor, a weight factor to lower the switching frequency and a weight factor to reduce the common mode voltage,andis t_k+2Predicted value of capacitor voltage at time, n_cIs the switching times of the switching tube when switching from the current state to the switching state at a time in the future, u_cmIs a function of the common mode voltage, both calculated according to the formula (8);

n_c＝|S_a(k)-S_a(i_opt)|+|S_b(k)-S_b(i_opt)|+|S_c(k)-S_c(i_opt)|

Technical Field

The invention belongs to the technical field of power electronic control strategies and new energy, and particularly relates to an improved model control method of a three-level inverter with a resistive-inductive load.

Background

The diode-clamped three-level-based three-.

The diode clamping type three-level inverter has two major problems, namely neutral point potential balance and common mode voltage, and the service life of the inverter is shortened and potential safety hazards are caused due to the two problems. At present, most scholars mainly start from the two aspects, and consider the hardware and software aspects respectively, the advantage of the software is greater than that of the hardware, so the control strategy is the focus of the broad scholars. The scholars propose a plurality of control methods for solving the two problems to solve one problem independently, and the model predictive control can put the two problems into an objective function as control targets so as to achieve a good control effect, but the calculation amount is increased correspondingly.

Disclosure of Invention

The invention aims to provide an improved model control method of a three-level inverter with a resistive load, which solves the problem that the calculated amount is not considered in a control mode utilizing model prediction in the traditional three-level inverter with the resistive load.

The technical solution for realizing the purpose of the invention is as follows: an improved model prediction control method of a three-level inverter with a resistive-inductive load comprises the following steps:

1) detecting three-phase load current and two voltage dividing capacitor voltages at a direct current side aiming at an inverter load model;

2) setting a reference current and performing phase angle compensation; setting an initial time optimal switch state and an initial objective function optimal value g _ opt;

3) converting the three-phase load current in the step 1) from abc coordinates to alpha beta coordinates, and performing clark transformation;

4) carrying out t on the current and the DC side capacitor voltage according to a discrete formula by using the optimal switch state at the initial moment_k+1Predicting the time, and simultaneously calculating a voltage vector;

5) using t in step 4)_k+1Predicted value of time, predicting t according to discrete formula_k+2The dc side capacitor voltage at a moment;

6) calculating improved objective functions g under 27 switch states by using the voltage vector value in the step 4) and the predicted value of the direct-current side capacitor voltage in the step 5);

7) comparing the improved objective function value in the step 6) with the initial objective function optimal value g _ opt, selecting a small value between the improved objective function value and the initial objective function optimal value as a new objective function optimal value, outputting the switching state at the moment, and finally applying the new switching state to drive the inverter.

Compared with the prior art, the invention has the following remarkable advantages: (1) on the premise of not influencing dynamic response, the operation amount is reduced by optimizing the objective function. (2) Meanwhile, current tracking, midpoint potential balance, common-mode voltage reduction and switching frequency reduction (3) are considered in the index function, and a two-step prediction mode is adopted to solve the delay problem caused by sampling and calculation and perform phase angle compensation on the reference current.

Drawings

Fig. 1 is an inverter load diagram.

Fig. 2 is a flowchart of the overall implementation of the optimization method.

Fig. 3 is a diagram of midpoint potential balance.

Fig. 4 is a graph of common mode voltage.

Fig. 5 is a current tracking diagram.

Detailed Description

The invention relates to an improved model control method of a three-level inverter with an inductive load, which is characterized in that the current of a three-phase load and the capacitance voltage of a direct current side are sampled according to an inverter load model; determining reference current, realizing phase angle compensation, and calculating a voltage vector; coordinate conversion is realized on the current of the three-phase load; predicting load current and direct-current side capacitor voltage by adopting a discretization model and a model two-step prediction method; and constructing a target to be controlled into an objective function, and performing corresponding calculation, so as to select a switching state driving inverter which enables the objective function to be optimal.

The method comprises the following specific steps:

first, sampling. And sampling the current on the load side of the three-level inverter by using a three-way current detection device, and detecting the capacitor voltage at the voltage dividing capacitor on the direct-current side by using a two-way voltage detection device.

In a second step, a reference current is determined. And (4) automatically setting a numerical value, performing phase angle compensation by adopting the following formula, and calculating a voltage vector.

Where ω is 2 pi f, f is the frequency, Ts is the sampling time,is t_k+1The reference vector of the time of day,is t_kReference current vector at time instant.

And thirdly, transforming coordinates. The coordinate transformation is performed on the three-phase load current using the following formula.

Wherein i_α,i_βFor the current i in a two-phase coordinate system_a,i_b,i_cThe current is in a three-phase coordinate system.

And fourthly, predicting the first step by the model. The load-side current and the dc-side capacitor voltage are respectively predicted by the following formulas, and a voltage vector is calculated.

Where R is the resistance on the load side, L is the inductance on the load side, Ts is the sampling time, i_α,β(k) Is t_kCurrent measurement at a time, i_c1(k)、i_c2(k) Is t_kCapacitor current u on the time DC side_α,β(k) Is the output voltage of the inverter for the kth cycle;andis t_k+1The predicted current at that time and the predicted voltage of the dc-side capacitance.

And fifthly, predicting the second step by the model. The dc-side capacitor voltage is predicted using the following equation.

Where R is the resistance on the load side, L is the inductance on the load side, Ts is the sampling time,is t_k+1The current predicted value at the time of day,is t_kCapacitor current u on the time DC side_α,β(k +1) is the output voltage of the inverter for the k +1 th cycle;is t_k+1Predicting current and direct-current side capacitance predicted voltage at the moment;andis t_k+2The predicted current at that time and the predicted voltage of the dc-side capacitance.

And sixthly, performing an objective function. The calculation is performed using the following objective function with multiple control objectives.

n_c＝|S_a(k)-S_a(i_opt)|+|S_b(k)-S_b(i_opt)|+|S_c(k)-S_c(i_opt)|

In the formula (I), the compound is shown in the specification,andis the voltage vector corresponding to the switch state at t_k+1Real and imaginary parts of time, u_jα(k +1) and u_jβ(k +1) is the voltage vector at t_k+1Real and imaginary parts of, lambda_dc、λ_nAnd λ_cmAre a midpoint potential balance weight factor, a weight factor to lower the switching frequency and a weight factor to reduce the common mode voltage,andis t_k+2Predicted value of capacitor voltage at time, n_cIs the switching times of the switching tube when switching from the current state to the switching state at a time in the future, u_cmIs a function of the common-mode voltage, S_a,b,cIs a three-phase switch of the inverter.

And seventhly, driving the inverter. And (4) applying the optimal switching state calculated in the steps as a gate driving signal of the switching tube to drive the inverter.

The invention is further described below with reference to the accompanying drawings.

The invention relates to an improved model prediction control method of a three-level inverter with a resistance-inductance load, which not only solves the problems of the middle point balance and the common mode voltage which cannot be avoided by the three-level inverter, but also fully considers the effects of reducing the switching frequency and tracking the current, and provides an optimization scheme for the problem of large calculated amount of the three-level inverter, and the target is achieved mainly by optimizing an objective function, and the specific optimization method comprises the following implementation steps:

first, as shown in fig. 1, a three-way current detection device is used to sample the current on the load side of the three-level inverter, and a two-way voltage detection device is used to detect the capacitor voltage at the voltage dividing capacitor on the dc side.

And secondly, determining a reference current, and performing phase angle compensation by adopting the following formula.

Where ω is 2 pi f, f is the frequency, Ts is the sampling time,is t_k+2The current vector at the moment in time,is t_kThe current vector at the moment in time,is t_k+1The voltage vector at the moment.

And thirdly, carrying out coordinate transformation on the three-phase load current by referring to the following formula.

Wherein i_α,i_βFor the current i in a two-phase coordinate system_a,i_b,i_cThe current is in a three-phase coordinate system.

And fourthly, respectively predicting the load side current and the direct current side capacitor voltage in a model prediction mode according to the following formula, and simultaneously calculating a voltage vector.

Where R is the resistance on the load side, L is the inductance on the load side, Ts is the sampling time, i_α,β(k) Is t_kCurrent measurement at a time, i_c1(k)、i_c2(k) Is t_kCapacitor current u on the time DC side_α,β(k) Is the output voltage of the inverter for the kth cycle;andis t_k+1The predicted current at that time and the predicted voltage of the dc-side capacitance.

And fifthly, predicting the DC side capacitor voltage in the second step by using the following formula.

Where R is the resistance on the load side, L is the inductance on the load side, Ts is the sampling time,is t_k+1The current predicted value at the time of day,is t_kCapacitor current u on the time DC side_α,β(k +1) is the output voltage of the inverter for the k +1 th cycle;is t_k+1Predicting current and direct-current side capacitance predicted voltage at the moment;andis t_k+2The predicted current at that time and the predicted voltage of the dc-side capacitance.

And sixthly, calculating by using the following objective function with multiple control targets, compared with the traditional method: the current, voltage prediction and objective function calculations for 27 switch states that need to be completed within a cycle become the one that is closest to the reference vector is selected among the 27 voltage vectors, reducing the amount of calculations on the basis of not affecting performance.

n_c＝|S_a(k)-S_a(i_opt)|+|S_b(k)-S_b(i_opt)|+|S_c(k)-S_c(i_opt)|

In the formula (I), the compound is shown in the specification,andis the voltage vector corresponding to the switch state at t_kReal and imaginary parts of time, u_jα(k) And u_jβ(k) Is the voltage vector at t_kReal and imaginary parts of, lambda_dc、λ_nAnd λ_cmAre a midpoint potential balance weight factor, a weight factor to lower the switching frequency and a weight factor to reduce the common mode voltage,andis t_k+1Predicted value of capacitor voltage at time, n_cIs the switching times of the switching tube when switching from the current state to the switching state at a time in the future, u_cmIs a function of the common-mode voltage, S_a,b,cIs a three-phase switch of the inverter.

And seventhly, the inverter is driven by using the optimal switching state calculated in the steps as a gate electrode driving signal of the switching tube, the control purpose is achieved, the midpoint potential balance is shown in figure 3, the common-mode voltage is inhibited as shown in figure 4, and the current tracking effect is shown in figure 5.

Those skilled in the art to which the invention pertains will appreciate that various modifications and alterations may be made without departing from the spirit and scope of the invention. Therefore, the protection scope of the present invention should be determined by the appended claims.