1 Introduction
With the continuous depletion of traditional energy resources and the ongoing development of new energy technologies, photovoltaic (PV) power generation systems have become an important component of the power energy system. As the penetration power of PV power generation increases, it will bring new challenges to the stability of the power grid. When a transient fault occurs in the power grid, if large-capacity PV power generation units cannot maintain the voltage and frequency of the grid, it will have a significant impact on the stability of the power system.
When power plants with fluctuating characteristics are connected to the grid on a large scale, the lack of Low Voltage Ride-Through (LVRT) capability poses a significant threat to the safe and stable operation of the power grid. Therefore, LVRT capability is particularly important. Low Voltage Ride-Through (LVRT) refers to the ability of a photovoltaic (PV) power generation system to maintain continuous operation without disconnecting from the grid when a power system fault or disturbance causes a voltage drop in the grid, supporting grid recovery until the voltage returns to normal levels, thus traversing the low-voltage region. In wind power systems both domestically and internationally, new grid operation guidelines have been established, stipulating that wind power systems must possess LVRT capability as a condition for grid connection. In domestic PV power generation systems, relevant guidelines have been established that require LVRT capability. The State Grid Electric Power Research Institute has taken the lead in researching and standardizing LVRT and related detection devices. Based on LVRT, and according to actual grid requirements, it has proposed grid connection requirements for PV inverters with zero voltage ride-through capability.
This paper introduces the low voltage ride-through control and analyzes the relevant requirements for zero voltage ride-through. It then conducts research on the control of zero voltage ride-through experiments and presents the relevant experimental results waveforms.
2 Technical Requirements
According to the "Technical Regulations for Photovoltaic Power Stations Connected to the Power System" GB/T19964-2012 standard, when a power system accident or disturbance causes a voltage drop at the grid connection point of a photovoltaic power station, the photovoltaic power station can ensure continuous operation without disconnecting from the grid within a certain voltage drop range and time interval.
Figure 1 shows the low voltage ride-through requirements that photovoltaic power plants should meet.
(1) When the voltage at the grid connection point of the photovoltaic power station drops to 0, the photovoltaic power station should be able to operate continuously without disconnecting from the grid for 0.15s;
(2) When the voltage at the grid connection point of the photovoltaic power station drops below curve 1, the photovoltaic power station can be disconnected from the grid.
Figure 1 Zero-voltage ride-through capability requirements for photovoltaic power plants
3 Dynamic reactive power support capability
When different types of faults occur in the power system, if the grid connection test voltage of the photovoltaic power station is entirely within the area of the voltage outline in Figure 1, the photovoltaic power station should be guaranteed to operate continuously without disconnecting from the grid; otherwise, the photovoltaic power station is allowed to disconnect. The test voltages for different fault types are shown in Table 1.
Table 1 Low Voltage Ride-Through Test Voltage for Photovoltaic Power Stations
Fault type | Test voltage |
Three-phase short circuit fault | Grid connection point line voltage |
Two-phase short circuit fault | Grid connection point line voltage |
Single-phase short circuit fault | Grid connection point phase voltage |
For photovoltaic power stations in a photovoltaic power station cluster that are connected to the grid at a voltage level of 500kV (or 700kV) through a 220kV (or 330kV) photovoltaic power generation collection system, when a short-circuit fault in the power system causes a voltage drop, the dynamic reactive current injected into the grid by the photovoltaic power station should meet the following requirements:
(1) The response time of dynamic reactive current shall not exceed 30ms from the moment the grid connection point voltage drops.
(2) From the start of the dynamic reactive current response until the voltage recovers to 0.9 pu, the dynamic reactive current IT injected into the power system by the photovoltaic power station should track the voltage change at the grid connection point in real time and should meet the following requirements:
4. Phase-locked loop processing
A traditional three-phase phase-locked loop typically consists of a phase detector, a loop filter, and an oscillator. The structure of a traditional three-phase phase-locked loop is shown in Figure 2.
Figure 2. Principle of a traditional three-phase phase-locked loop
First, the three-phase voltage is transformed to the dq coordinate system. To ensure that the d-axis component is completely in phase with the voltage vector, the q-axis voltage reference value Uq_ref is set to 0, and the q-axis voltage deviation value Uq_err is used as the angle deviation output by the phase detector. The frequency is obtained by the PI controller output, and then the angle value is obtained by an integral stage. When only the positive-sequence fundamental component exists in the grid voltage, its steady-state value in the dq coordinate system is a DC value. Phase-frequency locking can be achieved by controlling the q-axis component to zero.
When an imbalance fault occurs in the power grid, the grid voltage contains not only positive-sequence components but also negative-sequence and zero-sequence components. For a three-phase system without a neutral wire, the zero-sequence electromotive force is generally not considered. In this case, the grid voltage can be expressed as:
In the formula, u+ and u- represent the magnitudes of the positive-sequence and negative-sequence voltages, respectively, and θ0- represents the initial phase angle of the negative-sequence voltage relative to the positive-sequence voltage. After a 3/2 transformation, the voltage expression in the αβ stationary coordinate system is obtained as follows:
The positive and negative sequence components of the grid voltage can be represented in the αβ stationary coordinate system as follows:
By transforming to the synchronous coordinate system using the dq transformation, we can obtain...
The formula above can be used to obtain...
When the phase-locked loop is working normally, wt≈θ, therefore the above formula can be expressed as:
As can be seen from the above equation, in the positive-sequence synchronous coordinate system, the positive-sequence voltage component is converted into a DC component, and the negative-sequence component is converted into a component with twice the grid frequency. Because traditional PI regulators can only control DC components without steady-state error, the negative-sequence voltage component will cause a large disturbance to the output of the phase-locked loop.
To address the shortcomings of traditional three-phase power grid phase angle detection methods under power grid fluctuation and fault conditions, this paper proposes a novel digital phase-locked loop, the structure of which is shown in Figure 3.
Figure 3 Improved phase-locked loop structure
To eliminate disturbances caused by grid voltage imbalance, a phase-shifting controller is used to eliminate disturbances at twice the grid frequency. The basic principle of the phase-shifting controller can be expressed as follows:
As can be seen from the above equation, disturbances at twice the grid frequency can be eliminated by adding them after a 1/4 grid cycle delay. In the control system, the phase-shifting controller must be represented in discrete form, and its discrete structure is shown in Figure 4.
Figure 4. Discrete structure of phase-shift control
In Figure 4, N represents the number of samplings within one power grid cycle. After passing through the phase-shift controller, the positive-sequence component of the proposed phase-locked loop output is...
Under grid imbalance conditions, it is also necessary to accurately obtain the amplitude and phase information of the negative sequence component of the grid voltage. The negative sequence component is obtained in the negative sequence synchronous coordinate system. The relationship between the positive and negative sequence components of the grid voltage and the positive and negative sequence synchronous coordinate system is shown in Figure 5.
Figure 5. Relationship between positive and negative sequence voltages of the power grid in the dual synchronous coordinate system.
The principle of negative order coordinate transformation can be expressed as follows:
When the phase-locked loop is working normally, wt≈θ, therefore the above formula can be expressed as:
As can be seen from the above equation, in the negative-sequence synchronous coordinate system, the negative-sequence voltage component is converted into a DC component, and the positive-sequence component is converted into a component with twice the grid frequency. Therefore, the disturbance of twice the grid frequency can also be eliminated by phase-shift control. After phase-shift control, the expression for the negative-sequence voltage in the negative-sequence synchronous coordinate system can be obtained as follows:
5 Control methods
When the grid voltage drops suddenly, the potential difference between the grid voltage and the grid-connected converter's output voltage increases abruptly. The full-power converter cannot adjust its output voltage instantly, causing a sudden increase in the output current to the grid. When the grid voltage drop is significant, the output current can increase to 2-3 times the normal output current. If proper low-voltage ride-through control is not implemented, this can lead to overcurrent in the full-power converter and overvoltage on the DC bus. The sudden increase in the output current to the grid also causes an increase in the output current of the photovoltaic power source, deviating from its maximum power point (MPP). If MPP tracking is still used at this time, positive feedback may occur, causing the current to continue increasing. Therefore, upon detecting a grid voltage drop, MPP tracking should be stopped immediately.
To control the inrush current that occurs when the grid voltage drops, this paper adopts a control method that directly adjusts the feedforward compensation voltage Ud. The specific analysis is as follows: Ud is compensated to the output of the d-axis current regulator after filtering. When the grid voltage drops, Ud also drops, thereby reducing the output voltage of the grid-connected converter and controlling the current flowing to the grid. However, Ud obtained after the grid voltage undergoes an abc/dq transformation needs to pass through a low-pass filter, affecting the system's real-time response speed and failing to effectively suppress current increases. Considering this, this paper adjusts Ud in real-time by detecting the degree of the grid voltage drop, accelerating the adjustment speed of the grid-connected converter's output voltage, and thus controlling the current flowing to the grid and stabilizing the DC bus voltage. The specific control method is shown in Figure 6.
Figure 6 Low Voltage Ride Control Block Diagram
6. Analysis of Experimental Results
Based on the requirements of the LVRT standard above, verification and implementation were carried out according to the existing drop source platform conditions and software programs:
After improving the phase-locked loop (PLL), the d and q components of the grid voltage after park transformation were observed. During normal operation, the d-axis component was stable, while the q-axis component was near 0. Under balanced voltage dips, the d-axis component remained stable, but the q-axis component was larger, affecting low-voltage ride-through control. The q-axis component became even larger under unbalanced voltage dips. Normal operation was possible at 90% and 75% voltage dip depths, but failures and shutdowns occurred at other voltage dip depths. The positive and negative sequence components were unstable. Analysis revealed that the program only filtered and extracted the positive sequence component, neglecting the negative sequence component, which exhibited significant variations. To address this, separate extraction of the positive and negative sequence components is necessary.
When the grid voltage drops, the control voltage output of the inverter needs to be changed in time. In order to avoid grid disconnection, the control voltage needs to track the grid voltage quickly and accurately. Therefore, vector control is generally used, that is, the three-phase AC quantities are transformed into the positive sequence rotating coordinate system and the negative sequence rotating coordinate system respectively, and then the positive and negative sequence currents are controlled separately.
Currently, common methods for extracting positive and negative order components can be divided into analytical methods and filtering methods. Analytical methods are faster but are too sensitive to harmonics or abrupt changes. Filtering methods obtain positive or negative order components by transforming coordinates and then filtering, resulting in more stable results. However, the addition of filtering introduces a certain delay, making the process slower.
The three-phase instantaneous values are transformed into a positive-sequence rotating coordinate system using the Park transform, yielding the dq-axis components of the positive-sequence and negative-sequence voltage vectors in this system. The dq-axis components in the positive-sequence rotating coordinate system consist of the DC quantity corresponding to the positive sequence and the AC quantity corresponding to the negative sequence, rotating clockwise at a second harmonic angular velocity. Similarly, the dq-axis components in the negative-sequence rotating coordinate system consist of the DC quantity corresponding to the negative sequence and the AC quantity corresponding to the positive sequence, rotating clockwise at a second harmonic angular velocity. After filtering, the positive-sequence components in the positive-sequence rotating coordinate system and the negative-sequence components in the negative-sequence rotating coordinate system are obtained.
The following are the experimental waveforms of a 500kW photovoltaic grid-connected inverter (the top two lines represent the positive-sequence d-axis and q-axis components; the bottom two lines represent the negative-sequence q-axis and d-axis components).
Figure 7 shows the three-phase voltage balance drop when the voltage drop depth is 1V (i.e., zero voltage crossing drop depth). The test waveforms are at 70% rated power. The waveforms show that both positive and negative sequence dq components are very stable, and the q-axis component is small, almost zero. The negative sequence component is essentially zero.
Figure 8 shows the test waveforms at a single-phase voltage drop depth of 1V and 70% of rated power. The waveforms show that both positive and negative sequence dq components are very stable, with the q-axis component being small, essentially near zero. The negative sequence component is essentially zero.
Figure 7 Three-phase voltage balance drop
Figure 8 shows the single-phase voltage sag.
7. Actual Test Analysis
The DC side of the photovoltaic inverter is connected to a simulated PV power supply. The test inverter, a 500kW grid-connected photovoltaic inverter independently developed by Jiuzhou Electric, was verified on a low-voltage drop test platform. The effective value of the grid line voltage is 270V, and the grid frequency is 50Hz. According to the requirements of the standard GB/T19964-2012 "Technical Regulations for Photovoltaic Power Stations Connected to the Power System", when a power system accident or disturbance causes a voltage drop at the grid connection point of the photovoltaic power station, the photovoltaic power station can ensure continuous operation without disconnecting from the grid within a certain voltage drop range and time interval. The table below shows the voltage drop values and time specifications.
Voltage rating (UN: rated voltage) Voltage value (V) Time (ms)
Voltage rating ( UN : rated voltage) | Voltage value (V) | Time (ms) |
0U N | 1 | 150 |
When the photovoltaic inverter is operating at full power, the grid voltage momentarily drops to zero, which has a significant impact on the operation of the photovoltaic inverter. Therefore, this test selected the inverter to operate at 70% of its rated power, and the three-phase grid voltage on the grid-connected side of the photovoltaic inverter dropped to the most severe operating condition shown in the table above. The waveforms of the three-phase and single-phase voltage drop tests using the LVRT control method are shown in Figure 9 and Figure 18, respectively. The yellow line represents the DC bus voltage of 500V/div; the green line represents the grid line voltage of 200V/div; and the purple line represents the A-phase grid current of 500V/div.
Figure 90%UN shows the three-phase voltage drop.
Figure 1 shows the single-phase voltage drop of phase A at 100% UN.
As shown in Figures 9 and 18, when using LVRT control, the inverter output current achieves a smooth transition, effectively suppressing current rise and ensuring continued grid-connected operation. The inverter's output active power decreases, and the accumulation of intermediate energy leads to an increase in the intermediate DC voltage. The PV analog power supply output power decreases, and the DC-side current decreases, which is beneficial for LVRT implementation. The PV analog power supply output power decreases until it equals the grid-connected power, returning to a stable operating state. When the grid voltage returns to normal, the intermediate DC voltage decreases with the increase in grid-connected power, the PV analog power supply output power increases, and the DC-side current increases until normal operation is restored. During grid faults, the inverter sends a certain amount of reactive power to the grid to support the grid connection point voltage, contributing to the recovery of the grid voltage.
The test results verified that the positive and negative sequence separation LVRT control strategy can effectively suppress the rise in inverter output current caused by the grid zero voltage drop, so as not to shut down and disconnect from the grid due to overcurrent protection, thus ensuring the reliability of the photovoltaic inverter grid-connected operation.
8. Conclusion
This paper introduces the zero-voltage drop control method of LVRT (Low Voltage Reduction) for photovoltaic (PV) grid-connected inverters, and analyzes in detail the grid-connected modes of PV inverters before and after grid faults. The amount of reactive current to be sent to the grid is determined based on the depth of the grid voltage drop. Experimental results verify the effectiveness of the LVRT control method, providing theoretical and experimental guidance for large-scale PV power generation grid connection.
