Three-level midpoint control method for grid-connected inverter

Compared with the traditional two-level three-level power consumption, the sine of the voltage waveform is good, and the harmonic content is low, so the grid-connected performance is superior, but it also has its own shortcomings, that is, the midpoint potential is unbalanced. It is precisely because of the fluctuation of the DC side voltage that the medium and low harmonic content of the inverter output signal is larger than before, which causes the quality of the waveform to decrease. The unbalance of the midpoint voltage will increase the current spectrum wave rate of the inverter output, resulting in increased loss and decreased utilization. If it is severely unbalanced, a capacitor with a large voltage drop will be damaged due to excessive voltage, and the power tube connected in parallel with it will also be damaged.

Through research, it is found that the midpoint clamping structure has the problem of midpoint potential balance. At the same time, both SPWM and SVPWM modulation methods will cause the output to produce AC fluctuations of 3 times the fundamental frequency, which makes the midpoint of the inverter’s DC side The potential fluctuates accordingly. When the capacity of the system is large and the voltage is high, because the load current is large, a relatively large current flows through the neutral point, and the fluctuation of the potential between the capacitors caused by this situation will also increase. This situation will It will seriously damage the capacitor and power tube. The fluctuation of the midpoint potential is closely related to the capacitance of the capacitor. The fluctuation of the potential is inversely proportional to the capacitance value. The smaller the capacitance value, the greater the fluctuation. At the same time, due to manufacturing process and cost constraints, the capacitance value should not be too large. The unbalance is also related to the output voltage amplitude and frequency. The higher the voltage, the higher the frequency, and the more difficult it is to control the balance. The size of the load carried by the inverter output will affect the balance of the midpoint potential. The larger the load, the more serious the midpoint potential imbalance. If there is no effective restraint measures, it is necessary to choose a capacitor with increased capacity and a power switch with increased withstand voltage level to overcome the impact caused by the imbalance of the neutral point potential, but this will increase the investment cost and increase the system’s cost. volume. It can be seen from the above that it is necessary to effectively control the midpoint potential.

(1) Analysis of unbalanced midpoint potential caused by working conditions
The midpoint potential imbalance is mainly caused by the existence of the midpoint current. The following analyzes the impact of the midpoint current on the potential in 27 different switching states.
The zero vector is shown in Figure 1(a), Figure 1(b), Figure 1(c). From the corresponding equivalent circuit, it can be seen that since there is no current flowing through the midpoint of the capacitor, the zero vector state will not cause the midpoint potential unbalanced

Figure 1 Equivalent circuit of load and inverter connection in each state

The large vector is shown in Figure 1(d) and Figure 1(e). The load is directly connected to the positive and negative poles of the power supply, and is not connected to the midpoint 0′.
That is, there is no independent charging and discharging circuit between the midpoint and the positive and negative poles of the DC power supply. This will not cause the capacitor on one side to be charged and discharged separately, so there will be no midpoint imbalance.
The medium vector is shown in Figure 1(f). There is always one phase connected to the midpoint of the three-phase load, so that the load and the two DC voltage divider capacitors C1 and C2 each form a charging and discharging circuit. When the charging and discharging of the two circuits are not synchronized, It will cause the potential of the two voltage divider capacitors connected in series to change, causing potential imbalance.
Small vector, see Figure 1(g), Figure 1(h), Figure 1(i), Figure 1(j), one or two phases of the three-phase load are connected to the midpoint, and the load is only connected to two on the DC side One of the phase series capacitors forms a charge and discharge loop. When the capacitor is charged and discharged, because the total DC voltage Ua is constant, when the voltage on one capacitor increases, the voltage of the other capacitor decreases by the same value. Although the paired redundant small vectors correspond to the same voltage space vector, their effect on the midpoint of the capacitor is opposite.
From the above analysis, the following relationships are summarized.
①The midpoint potential fluctuation has nothing to do with the zero vector and the large vector
②Under the action of a small vector, the load and one of the two voltage divider capacitors form a charging or discharging circuit, which changes the midpoint potential.
③Under the action of the medium vector, the load and the two voltage divider capacitors form a charge and discharge circuit respectively, and the difference of the charge and discharge time will cause the midpoint potential fluctuations.
It can be seen that in the 27 states, only the small vector and the middle vector will affect the midpoint potential of the inverter. The effect of the paired redundant small vectors on the midpoint potential is opposite, so the midpoint balance control can be completed by reasonably allocating the action time of the redundant vectors.
(2) Control strategy of midpoint potential balance
By analyzing the equivalent circuit between the load and the inverter under the action of different voltage vectors, only the general situation of the change in the midpoint voltage of the capacitor can be judged, and the rise and fall of each capacitor voltage depends on the flow direction of the midpoint current. Since ic=C(duc/dt), uc information can be obtained by analyzing ic. The midpoint current i0 and voltage u0 are shown in Figure 2.

Three-level midpoint control method for grid-connected inverter
Figure 2 The relationship between i0 and u0

According to the electrical characteristics of the capacitor, the formula can be listed:

Three-level midpoint control method for grid-connected inverter

Suppose C1=C2=C, then according to the KCL law:

Three-level midpoint control method for grid-connected inverter

From formula (1), we can get

Three-level midpoint control method for grid-connected inverter

From equation 1), i0∝(du0/dt) can be obtained. When u0>0, i0 flows into the midpoint. Conversely, when u0<0, i0 is flowing out
point. It can also be said that i0 causes u0 to change. If i0=0, u0 will not fluctuate. For example, vectors 211 and 100. When the vector 211 works, C1 discharges and C2 charges; when the vector 100 works, C1 charges and C2 discharges. Table 1 lists the influence of redundant vectors on the neutral point potential balance. Therefore, by controlling the action time of the small vector, the fluctuation of the midpoint potential can be minimized.

Redundant small vectorCapacitor voltage change
122.112, 212
100, 110, 010
Table 1 The influence of redundant small vectors on the neutral point potential balance

The action time T1 of the redundant small vector is the sum of the action time of the two small vectors that make the capacitance voltage change opposite, as shown in equation (4)


Tup=T1·[(1+p)/2] ﹣1≤p≤1

Tdown =T1· [(1-p)/2 ] ﹣1≤p≤1

The two capacitor voltages Uc1 and Uc2 are subtracted, and the difference is the change in the midpoint potential. A new variable p (-1≤P≤1) is defined by the difference. If the voltage of C1 is higher than that of C2, P is a positive value, otherwise P is a negative value. The time calculation formulas of Tup and Tdown are formula (2) ~ formula (4). The action time of the positive and negative small vector pairs changes with the change of the midpoint potential difference. For example, when Uref is in the position of Figure 3, Tup=T100, Tdown=T211. When p>0 (the voltage of C1 is higher than that of C2), the formula shows that the action time of T100 decreases and the action time of T211 increases, so that the voltage of C1 decreases and the voltage of C2 increases. During the switching period, the adjustment of the redundant small vector pairs can well suppress the fluctuation of the midpoint potential of the three-level grid-connected inverter and maintain balance

Three-level midpoint control method for grid-connected inverter
Figure 3 Simplification and decomposition of the three-level space vector diagram