Research on pre-synchronization control strategy of the micro-grid with multi-VSG

1 Introduction

Virtual synchronous generator (VSG) controls the rotation and damping characteristics of the synchronous generator by simulating the mechanical and electromagnetic characteristics of the synchronous generator and controlling the grid connected inverter. With the development of VSG technology, it is expected to be widely applied in the field of microgrid in the future. Like traditional generators, VSG also needs to synchronize the frequency and phase of VSG with the grid.

Reference [1-6] point out that in order to improve the efficiency and flexibility of distributed generation,more and more grid connected inverters require the dual mode operation capability of islanding and grid connection. By simulating the external characteristics of the synchronous machine, VSG should not only operate in normal mode of power grid, but also run off the network when the power grid fails, in order to ensure the highquality power supply of the important load. Reference [2-8,10-13] on the basis of the characteristics of the VSG network, the pre-synchronization method is adopted to achieve seamless switching between VSG and off grid. A VSG control strategy of a lithium battery grid connected power generation system with a PQ controlled PV grid connected inverter makes up a small micro network to achieve seamless switch off of the network. In reference[4-12], analyses of VSG are compared with droop control.Reference [5-13] propose a pre-synchronization control strategy based on virtual power and frequency quadratic control for virtual synchronous generators, but all aimed at VSG's stand-alone grid connection. However, the operation of the micro-grid islanding is a multi-source parallel operation mode. The inverter power supply with one or several voltage source external characteristics needs to be used as the amplitude-frequency support of the micro-grid.When multiple micro-grids are provided by the VSG for the micro-grid, and when the amplitude-frequency is supported,the pre-synchronization process is rarely reported.

According to controlling the frequency of distributed power supply, the micro network structure under the peer control can control their different phases among the different distributed power sources to achieve stable power output. But the pre-synchronization process is often accompanied by the adjustment of frequency and phases,which can cause the readjustment of power output in inverters under the peer control. The over-current ability of power electronic devices is weak, so improper presynchronization process may cause the power oscillation between parallel VSG and then the inverters will run into the over-current state. In order to solve the problem, this paper proposes a pre-synchronous control-strategy based on the phase identification. And it effectively improves the negative effect of grid pre-synchronization on VSG parallel operation. We set up a micro-grid under the peer control of two VSGs, we analyze the principle of the process of parallel grids’ pre-synchronization, and we also analyze the influence of VSG’s power output under peer control of presynchronization.

2 Pre-synchronization principles

This paper selects the second-order mathematic model of synchronous generator and draws lessons from the working principle of parallel synchronous generator of traditional synchronous generator. The VSG presynchronization principle is shown in Fig.1 below.

In Fig.1, kp is the primary FM coefficient, kfi/s is the introduced second FM feedback.

The schematic diagram of its phase pre-synchronization method is shown in Fig.2.

When ωVSG > ωg and the voltage phase is ahead of the grid, the output of PI regulator is less than zero after the reverse. The VSG angular frequency is reduced by superimposing on the angular frequency instruction, and the phase and frequency are synchronized until the phase angle difference is zero. When ωVSG < ωg, and voltage phase advance in grid, PI regulator output instruction to angular frequency reversely overlaing VSG angular frequency is further reduced, until the VSG phase lag in the power grid, PI regulator output after not being greater than zero. Adjustment of the VSG angular frequency increases the last phase and frequency to achieve the synchronous state with power grid. On the contrary, it is similar to the above adjustment process.

Fig.1 VSG pre-synchronization control schematic

Fig.2 Schematic diagram of phase difference pre-synchronization method

The stator electrical equation and rotor motion equation of VSG can be expressed as:

Definition H=JωN, Dp=DωN. ra is stator armature resistance; Xd is synchronous reactance; Ė is excitation electromotive force; İ is the armature current; J is the moment of inertia of the synchronous generator; D is the damping coefficient; ωN is the rated angular velocity of the grid; Pm and Pe are the SVG’s active output setpoint and the real active output, respectively.

VSG pre-synchronization process is shown in Fig.1.The micro-grid voltage q-axis component Uq is reduced from 0 to form a PI regulation and feedback to the VSG angular frequency, so that the micro-grid voltage vector is in phase with the grid voltage vector and has the same angular frequency.

3 Influence of pre-synchronization on parallel VSG

3.1 The steady-state analysis

According to the VSG power sharing principle, it’s obvious that when multi-VSG is running in parallel, the active power distribution is:

According to the rotor motion equation of VSG, when the multi-VSG works in the steady-state:

When the pre-synchronization process goes into the steady state: ω=ωg, where ωg is the voltage angle frequency of power grid. Therefore, the active power output by the presynchronization VSG is changed as its angular frequency changes. The amount of change in active power of other VSGs are taken by the pre-synchronization of VSG. The amount of active power change is:

The i doesn’t include the number of pre-synchronization VSGs. Therefore, the pre-synchronization VSGs must have enough power capacity to take the active power change∆ P caused by pre-synchronization.

3.2 Transient process analysis

Angular frequency satisfies the equation ω1=ω2=ω3…when multi-VSG is in a steady-state. When the inductance of the wire is much larger than the resistance, which is R<<X and α≈90°, the active power circulation output from DG1 to DG2 is:

Fig.3 Multi-VSG parallel simplified map

where δ is the phase difference between DG1 and DG2. It is known from the power system stability that power system loses its stability when δ≥90°. Therefore, the following equations must be satisfied when the pre-synchronization starts:

where t1 is the time when pre-synchronization starts, and t2 is the time when it completes.

There are three different conditions where voltage vector is in the different place of synchronous rotating coordinate system in Fig.4. The synchronous rotating coordinate system is established with the grid angular frequency as the synchronous angular frequency, and the micro-grid voltage vector rotates at angular velocity ωg.Therefore, the rotational angular velocity of the microgrid voltage relative to the synchronous rotating coordinate system is △ω = ω-ωg.

Fig.4 The vector diagram of pre-synchronization process

In Fig.4 (a), we assume that ω < ωg, and the voltage vector of VSG is located in the first and second quadrants of the synchronous rotating coordinate system and ignore the stage ① ② after the angle of the concussion process.It is approximately considered that the adjustment process is divided into two stages. At this point, Uq>0, the PI regulator is adjusted to form a negative feedback, so that the frequency of the VSG corner is further reduced until it reaches a negative value φ<0, Uq> 0. Then it forms a positive feedback and the positive feedback increases the VSG angular frequency to achieve synchronization. In Fig.4 (b), we assume that ω < ωg, the voltage vector of VSG is located in the third and fourth quadrants of the synchronous rotating coordinate system, and |φ| is large.when Uq < 0, after the formation of PI feedback, positive feedback makes VSG angular frequency increase, so VSG’s phase and frequency achieve synchronization with the grid. However, as the large |φ| goes through the PI regulator, there appears a big excessive regulation. If |φ| is a small phase, the excessive regulation will be small. Just as the above analysis, it can be reasoned out that the best presynchronization time is when Uq > 0 and |φ| is small.

The best closing time of pre- synchronization switch is when ω ≥ ωg and Uq is a positive number, or ω ≤ ωg and Uq is a negative number.

As Table.1 shows, the best closing time of presynchronization switch can be recognized by the sign of Uq,the angle frequency of micro-grid, and the angle phase |φ|.If these variables meet the conditions, pre-synchronization dynamic process will be Fig.4(c), so that the active power shock between multi-VSGs can be reduced. Because the range of ɵ and ɵg changes periodically in [0,2π], it is difficult to determine if the phase difference in the control program is positive or negative. And Uq is the positive voltage synthesis vector in one and two quadrants, while it is negative in three and four quadrants. Therefore, using Uq and |φ| to control the φ value before pre-synchronization within a certain range will effectively solve the problem of over-adjustment of δ value according to the relationship between the frequency value of micro-grid and power grid.The value of |φ| in this paper is shown in Table 1.

Table 1 Pre-synchronization switch closed condition

Angle frequency state Uq |ϕ| Whether to close the pre-sync switch ωω +g 17[0,] [ ,2]9 9 π ππ∩Y ωω —g 17[0,] [ ,2]9 9 π ππ∩N

4 Simulation verification

In order to verify the correctness and validity of the phase identification method, a micro-grid and offgrid model shown in Fig.3 is constructed in MATLAB/Simulink. DG1 and DG2 are two lithium battery energy storage and grid-connected power generation systems with same capacity. The DC voltage is 700 V and the VSG control strategy with same control parameters runs in parallel mode when leaving the grid. Simulation parameters are shown in Table 2.

Table 2 Simulation parameters

Parameter P=P =m m 1 2 0.9 L= mH f 1 D=D =1 2 2 p p C µ F f =500 H=H=1 2 0.2 S= kW B 100 R=R= Ω1 2 0.1 L L mH 1 = 2 =1

As shown in Fig.5 (a), the beginning of the simulation is ω1 < ωg. At 0.3 s, the micro-grid voltage is located at 45° in the first quadrant. In Fig.5 (b), the micro-grid voltage is located in the first quadrant 50° and open presynchronization. Pre-synchronization process is shown in Fig.2 of the analysis.

Fig.5 Using two conventional DG presynchronization method active distribution process

The simulation results show that when the phase angle between the micro-grid voltage vector and the grid voltage vector is larger, the PI regulator of the pre-synchronization module outputs a larger value and the angular velocity of the micro-grid voltage is adjusted more rapidly, causing the partial overshoot to finally reach the steady state. It is noteworthy that, when the micro-grid voltage phase is located in the second and the third quadrants, the overshoot increases, causing the power oscillation value to be larger.

Fig.6 shows the dynamic process of active assignment of two DGs when using phase identification method.Before the start of synchronization, the two DGs are equally divided into 200 kW and 0kvar loads. when ω1 = ω2< ω at 0.3 s, micro-grid lagging behind the grid voltage phase 18° in line with pre-synchronization conditions,starts the pre-sync switch. Two-inverter active adjustment process is gentle, and no over-current phenomenon occurs.Due to the existence of δω, the power redistribution△P caused by pre-synchronization is assumed by DG1.At 2.5 s, when entering a new steady state, micro-grid achieves large power grid frequency, and phase presynchronization. At 3 s, grid switch is closed , and the presynchronization controller is disconnected, as DG1, DG2 output active power Pm1 = Pm2 = 90 kW in accordance with the active given value and the power grid shares the power required by the load. At this point, due to the role of large power grid clamp, micro-grid operation reaches at a rated frequency of 50 Hz.

Fig.7 shows the voltage pre-synchronization process when the phase identification method is adopted. The VSG voltage lags behind the grid voltage when the pre-synchronous voltage is turned on, then enters the synchronous process and runs stably.

Fig.6 Using phase identification method 2 DG active allocation of dynamic process.

Fig.7 Voltage pre-synchronization

5 Conclusion

Micropower source pre-synchronization process with multi-VSG control strategies is required to consider the power distribution between multiple VSGs and power redistribution occurs in parallel VSG due to pre-synchronization feedback. The larger the presynchronization closing micro-grid voltage vector and grid voltage phase difference, the worse active redistribution dynamic performance. Serious loss of stability, even converter over-current could happen. Based on the above analysis, a pre-synchronization control strategy based on phase identification is proposed, which is able to effectively improve the active process of active redistribution, to ensure the normal operation of the converter to prevent the parallel VSG instability.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (51677174); the State Grid Science & Technology Project (Title: The key technology research of client-side distributed energy storage flexible coordination control and joint operation).