Two current controlled inverters interfaced to a three-phase grid

The circuit files for this case are in these two links: circuit_files.tar.gz

This case uses the converter topology, output filter design and control strategy used in the previous case (click here). In this case, two Voltage Source Converters (VSCs) are interfaced to the grid. This case is a forerunner for the larger cases that are to follow with several converters and many of these VSCs controlled in a non-identical manner which would be the case in a microgrid. In this case, we begin with both converters being identical in topology and the current control loop. The difference between the VSCs will be that one of them supplies reactive power while the other supplies active power. The VSCs are two level three-wire converters with LCL filters at the output.

The control algorithm remains the same as the previous case. A common PLL generates the phase angle information of the grid for both the VSCs. The current control consists of the basic closed loop control scheme with a Proportional Integral (PI) controller.

To begin with we examine the three-phase voltages with the performance of the PLL in the following plots. The angular frequency of the PLL settles to 377 rad/s corresponding to the grid frequency of 60 Hz. In the plot of the grid voltages, the voltages are seen to have a disturbance at around 0.01s as this is when the inverters begin to inject their currents into the system.

Result 1 Result 2

The next plot shows the current injected by the VSC that supplies the reactive power to the system. The currents in this plot are the final currents injected into the system and not the currents at the output of the inverter. The currents are zero until 0.01s as the disconnect switches in inv2_filter.csv are switched off. It can be seen that the currents have several stages of transients.

Result 3

The first transient occurs when the VSC begins to inject current into the system. This is the time needed by the closed loop current control scheme to settle. This is shown in the next plot and the most obvious characteristic of this transient is the large current spike when the disconnect switches are turned on.

Result 4

The other slower transient is due to the fact that the PLL has not settled and the frequency is still changing. The currents settle to steady values when the PLL has settled at around 0.22s.

The currents of the other inverter that supplies active power to the system are as follows. They have a similar pattern but are smaller in magnitude since the active power demand of the load is lesser than the reactive power. This inverter also begins to inject currents into the system at 0.01s.

Result 5

The currents injected by the two inverters can now be compared. To have a better understanding of the phase angle of the currents, the phase a currents of the two inverters are plotted with the phase a grid voltage that has been scaled down by a factor of 5 so that it can be plotted with the currents. The blue waveform is the phase a grid voltage, the red waveform is the current from the first inverter supplying reactive power and the green waveform is the current from the second inverter supplying active power. As can be seen, the current from the first inverter is lagging behind the voltage by a phase angle of 90 degrees while the current from the second inverter is in phase with the grid voltage.

Result 6

There is another detail that is not so evident. The currents injected by the second inverter contain more oscillations that the first inverter. When the phase a currents of the two inverters are plotted together, this is clearly visible.

Result 7

In this simulation, the control gains of the two inverters are different. The first inverter's current control loop has a smaller integral gain that the second inverter's current control loop. This results in lesser oscillations in the first inverter's current. This can be examined in greater detail while plotting the d and q modulation indices that are the outputs of the two current controllers - m1d, m1q for the first inverter and m2d, m2q for the second inverter.

Result 8 Result 9

The greater oscillations in the second inverter current controller output cause the greater oscillations in the final current injected by the inverter into the grid. The attenuation provided by the LCL filters of the two inverters are quite similar and are shown in the following two plots. The first plot is of the current at the output of the first inverter with respect to the current injected by the inverter into the grid. The second plot shows this comparison of currents for the second inverter.

Result 10 Result 11

This case is a fairly simple case of multiple inverters with similar controllers but with different control objectives since one inverter supplies reactive power while the other supplies active power. We will now gradually progress towards a microgrid where numerous inverters will be controlled in a number of different ways and in some cases, the control will be decentralized without any communications. In this case, since the PLL for both the inverter control schemes was the same, communication is implicit. However, if these two inverters were not close physically, it would not be possible for the same PLL to be the basis for their control schemes.