Transformerless solar photovoltaic grid tie inverter and its control technology

Influence of low frequency harmonics of grid voltage and dead zone effect

1. The influence of low frequency harmonics of grid voltage

The equivalent model of the grid-connected inverter is shown in Figure 1, in which the output voltage uinv of the bridge arm of the inverter is equivalent to a controlled voltage source (the specific form is related to the inverter control). Using the superposition theorem, it can be known that if the grid voltage contains low-frequency harmonic voltage, the corresponding low-frequency spectrum wave current must be generated in the inductor.

The transfer function from grid voltage to grid current is

Figure 2 shows the schematic diagram of the amplitude-frequency characteristic curve of formula (1.1), and the current regulator is a PI regulator. The magnitude of the current harmonics generated by the low-frequency harmonics of the grid voltage is equal to the magnitude of the voltage harmonics multiplied by the magnitude of equation (1.1) at the corresponding frequency.

When the content of low-frequency spectrum waves in the grid voltage is high, the incoming current must be seriously distorted, and it is difficult to meet the grid access standard.

2. Dead zone effect

Since the power device has a certain turn-on and turn-off time, in order to prevent the pass-through phenomenon of the power tube of the same bridge arm of the inverter, it is necessary to artificially insert a period of dead time between the complementary driving signals. This paper takes the oH5 topology grid-connected inverter as an example to briefly explain the influence of dead zone on the inverter output waveform.

The driving logic of the oH5 topology is shown in Figure 3, in the positive half cycle of the power grid, S4 is always in the off state, and only the dead time needs to be inserted between S1 and S5; while in the negative half cycle of the power grid, S6 is always off, and only the dead time needs to be inserted between S1 and S3.

In addition, in order to prevent S1 and S2 from short-circuiting C_dc1, a dead time needs to be inserted between S1 and S2. Since s2 only affects the leakage current, it will not be discussed in detail here. Enlarge the segment from t_a to t_b in Figure 3, and add the drive logic after dead zone is shown in Figure 4, where t_d is dead zone time.

As can be seen from the above figure, S5 is turned on after t_d time, and turned off in advance of t_d time. In addition, S1, S3-S6 all use the same power devices. Under the condition that the driving circuits are completely consistent, different power tubes can basically be switched on and off at the same time, and the error time is very short. Therefore, the following analysis considers that S1 and S6 are turned on and off at the same time in the positive half cycle of the power grid; similarly, S1 and S4 are turned on and off at the same time in the negative half cycle of the power grid. Regardless of the switching time of the power tube for the time being, that is, it is considered that the turn-on and turn-off are instantaneously completed, then in the positive half-cycle working stage of the power grid:

(1) When the inverter side current i_L1 is positive, that is, when i_L1 is in the same direction as the grid voltage u_g, in the power transmission stage, i_L1 flows through S1, S3 and S6, and at the moment when S1 and S6 are turned off, iL1 freewheels through the anti-parallel diodes of S3 and S5. At this time, the dead time between S1 and S5 has no effect on the output voltage of the bridge arm.

(2) When i_L1 is negative, that is, when i_L1 is opposite to u_g, in the power transmission stage, i_L1 flows to the DC side through the anti-parallel diodes of S1, S3 and S6, until the moment when s5 is turned on, i_L1 is freewheeling through the anti-parallel diodes of s3 and S5. At this time, the dead time between S1 and S5 prolongs the time when the output voltage of the bridge arm is +U_dc, as shown in Figure 5. Among them, u_inv is the output voltage of the bridge arm of the inverter under ideal conditions, and u_{inv_1} and u_{inv_2} are the output voltages of the bridge arm when i_L1 is positive and i_L1 is negative, respectively.

In the negative half cycle of the power grid, S3 is turned on after t_d time, and turned off in advance of t_d time, and the output voltage of the bridge arm changes between (-U_dc, 0). The analysis of the dead time effect is similar to the positive half cycle of the power grid: when i_L1 is negative, that is, when i_L1 and u_g are in the same direction, the dead time has no effect on the output voltage of the bridge arm; when i_L1 is positive, that is, when i_L1 and u_g are reversed, the dead time prolongs the time when the output voltage of the bridge arm is -U_dc.

It can be seen from the above analysis that if the inverter works with a unit power factor, the dead time will not affect the output voltage of the bridge arm, that is, it will not affect the incoming current. However, in fact, iL1 contains a lot of high-frequency ripples, so that , always fluctuates positively and negatively near the zero-crossing point, and the output voltage of the bridge arm of the inverter will always fluctuate between the two states of u_{inv_1} and u_{imv_2}. Since the duty cycle is very small near the zero-crossing point of the modulation wave (ie, the grid voltage zero-crossing point), the driving signal is very narrow, and the dead time cannot be ignored relative to the on-time of the power tube, and may even exceed the theoretical on-time of the power management. At this time, the current control is in an uncontrolled state, and the incoming current is prone to distortion. In addition, unlike the inverter unit power factor operation, when the inverter non-unit power factor operation, because the amplitude of the incoming network current is often large when the modulation wave crosses zero, the current distortion caused by the zero-crossing dead zone will be more serious.