1. Introduction
In recent years, the worldwide demand for energy has been rapidly growing to support the development of various modern technologies; thus, the vast use of fossil fuels has negatively affected the environment. Therefore, renewable energy (RE)-based distributed power generation (DG) has been intensively researched in the past decade. In the near future, the penetration of various grid-connected DG systems is expected to keep increasing. As a result, utilizing advanced power converter-based compensators to work with RE-based generation systems and existing grids are one of the most important tasks in ensuring the voltage stability and power quality (PQ) of distribution systems regardless of the naturally intermittent characteristic of RE-based power generation. Wind power generation is currently one of the most promising RE sources because it can generate a considerable amount of electrical energy in a short period of time. Commonly used generators for wind turbine generators (WTGs) include permanent magnet synchronous generators (PMSGs) [
1], doubly-fed induction generators (DFIGs) [
2], and brushless wound-rotor doubly-fed generators [
3]. Currently, most research papers on grid-connected PMSG-based WTGs (PMSG-WTGs) found in open literature have focused on advanced control schemes, low voltage ride-through (LVRT) capacity, and operating modes. M. Jamil et al. [
4] suggested that, among commonly used generators in WTGs, PMSGs have more advantages compared with conventional DC and induction generators. The topologies of PMSG-WTG power converters were reviewed, including a thyristor-based inverter, hard-switched pulse with modulation (PWM) converter, multilevel converter, matrix converter, and Z-source inverter. For conventional small- and medium-size WTGs, power interfaces usually adopt the configuration that consists of a diode bridge rectifier and a buck boost converter. Although this configuration has the advantages of a simple circuit, low cost, and easy control, it causes a large amount of current harmonic and lowered power factor (PF) in the three-phase AC side, which leads to voltage pulses that lower PQ and efficiency, and shorten the life of the generator. A good WTG power interface must offer high conversion efficiency, low power consumption, PF correction, and low current harmonic. When a WTG is connected to a grid, appropriate control strategy is necessary to meet the power company’s LVRT specifications. On this basis, this study proposes a distributed control scheme to enhance LVRT capabilities of small- and medium-size PMSG-WTGs using silicon carbide (SiC)-based inverters.
Generally, PQ improvement is necessary for distribution systems embedded with DG systems because of the uncertainty and intermittency of RE sources, which can easily result in various undesirable effects, such as voltage sag. In [
5], a grid-connected direct-drive PMSG-WTG model was simulated for various wind speeds and load conditions using PI-based controller. The results showed a properly regulated DC link voltage and output power. M. H. Qais et al. [
6] reviewed LVRT capability enhancement methods for a grid-connected PMSG driven directly by a variable speed WT (VSWT). Later, the authors used a gray wolf optimizer (GWO) to tune eight PI controllers to improve LVRT capability and MPPT performance. In comparison with the genetic algorithm (GA) and simplex method, GWO algorithm yielded the best convergence, MPPT capability, and LVRT performance during symmetrical and asymmetrical faults [
7]. In [
8], a resistor-type superconducting fault current limiter (R-SFCL) was combined with superconducting magnetic energy storage (SMES) to improve the LVRT capability of a 2.5 MW PMSG-WTG. Improved transient stability and reduced cost for superconducting devices were achieved. P. Xing et al. [
9] proposed a fast compositive control of LVRT for a 2 MW PMSG-WTG by consuming initial excessive energy with a crowbar circuit, and then converting excessive energy into rotor kinetic energy. In [
10], a perturbation-observing nonlinear adaptive control was proposed for a dynamic voltage restorer (DVR) embedded with an energy storage system (ESS). This simple and robust control scheme does not require an accurate system model and full-state feedback. The results showed better performance than conventional fixed-gain vector control (VC) and feedback linearizing control based on an accurate system model. The authors used 10 MW PMSG- and DFIG-based WTGs to verify the control. M. Jahanpour-Dehkordi et al. [
11] combined VC and direct torque control (DTC) using two hysteresis current controllers to achieve fast and smooth LVRT for a 2 MW PMSG-WTG. This method is better than DTC control and VC control in terms of smoothness and speed, respectively. In [
12], the LVRT capability of a 2.5MVA full-scale PMSG-WTG was enhanced with active damping control and DC-link voltage bandwidth retuning. M. Nasiri and R. Mohammadi [
13] enhanced the LVRT capability of a 1.5 MW PMSG-WTG with proposed back-to-back converter controllers and active power limiter without the need for external devices.
Some advanced control strategies for DFIG and PSMG wind turbines can be found in [
14,
15,
16,
17,
18]. In [
14], advanced control strategies were proposed to control the pitch angle, and the rotor and grid side converters of the doubly fed induction generator (DFIG)-based wind turbine (WT) for enhancing the LVRT capability. The converter systems used in the study include a back-to-back converter and a DC chopper. Simulation results showed that with the proposed control means, the LVRT capability of WT generators can be enhanced, and the oscillations in the stator and rotor currents can be effectively reduced. Both low-wind and high-wind speeds cases were investigated to verify the proposed control algorithm. With the same hardware system used in [
14], the same group of authors of [
15] proposed a LVRT control method to achieve the real-time regulation of the rotor’s excessive inertia energy. The reactive power capacity was simultaneously handled to satisfy the grid code. It was demonstrated with simulation results that the control topology or parameters can be changed at different stages of voltage fault period. In [
16], a LVRT control scheme for the permanent-magnet synchronous generator (PMSG) with a back-to-back neutral-point-clamped, three-phase converter was proposed. In the studied grid voltage dip scenario, the proposed controllers designed for generator-side and grid-side converters can work concurrently to meet the LVRT requirement by storing the active power in the mechanical system of WT, and regulating the dc-link voltage at a constant value. Results from both simulation and experimental tests verified the effectiveness of the proposed method. The system stability issues of weak AC grid-connected DFIG-based wind turbines during LVRT were comprehensively investigated in [
17]. To study the instability mechanism of a DFIG with back-to-back converter systems during a weak grid fault, a small signal state-space model was established. The effectiveness of the proposed LVRT control methodologies were validated with simulation and experiments. In [
18], a wind speed combination model (WSCM) was proposed using the field-measured data, which was then used as equivalent wind speeds for each WTG when modeling the entire wind farm. Results from a case study verified the correctness of WSCM and the performance of the proposed algorithm in predicting the LVRT entering status of WT generators in the wind farm.
The use of a static synchronous compensator (STATCOM) for LVRT capability enhancement at the point of common coupling (PCC) was discussed in [
19,
20,
21,
22,
23,
24,
25,
26,
27,
28]. P. Dey et al. [
19,
20] developed a controller based on a pitch angle control and a flux weakening control and a STATCOM to enhance the LVRT capability of a PMSG-WTG. As a result, enhanced LVRT performance, reduced DC-link capacitor overvoltage, and the connection between the WTG and the grid during faults were achieved compared with conventional braking chopper-based LVRT strategy. In [
21], a coordinated current control scheme was proposed for a 100 MW PMSG-WTG with a 20 MW STATCOM for severe LVRT by exchanging control roles between the two converters, such that the DC link voltage was controlled by the machine-side converter (MSC), and the grid active power was controlled by the grid-side converter (GSC). Synchronization was maintained, and reactive power requirement was satisfied. J. Yao et al. [
22] studied a cost-effective capacity configuration strategy for the improvement of LVRT capability of a hybrid wind farm consisting of a 30 MW fixed-speed induction generator (FSIG) in conjunction with a 30 MW PMSG by using the PMSG as a STATCOM during grid fault. For LVRT capability improvement of DFIG-WTGs using STATCOMs, a PI controller with robust control technique was proposed in [
23], where the reliability of the WTG was ultimately increased, and a fuzzy adaptive proportional-integral-derivative (PID) controller was proposed in [
24], with enhanced control accuracy and performance. For self-excited induction generator (SEIG)-WTGs, M. I. Mosaad et al. [
25] proposed a model reference adaptive control (MRAC) of STATCOM, yielding more efficient and more robust performance than that of a GA-tuned PI controller. In [
26], the LVRT capability enhancement of STATCOM for the terminal voltage of a squirrel cage induction generator (SCIG)-WTG was studied in a wind farm in Bizerte, Tunisia. T. Tanaka et al. [
27,
28] used an 80MVar/33kV STATCOM based on multilevel converters for LVRT capability enhancement of non-specific-type offshore WTGs.
Most of the above reviewed applications of STATOM for LVRT capability enhancement adopted the configuration of centralized compensation, where a single STATCOM module was employed at the PCC to take charge of the control goal. However, this configuration is expensive, and the required system capacity is normally high. Moreover, if a single malfunction occurs in the STATCOM, then the LVRT capability of the system may deteriorate drastically. In this regard, this study aims to propose a distributed LVRT compensator (LVRTC) for every separate WTG. The use of distributed LVRTCs increases reliability and flexibility, while reducing the required capacity and overall costs. In the next section, LVRT specifications will be explained followed by the proposed distributed LVRTC and WTG system configuration. The mathematical models required for the controller design using a dq-axis current decoupling algorithm and quantitative design of the three-phase inverter controller are explained in
Section 3. In
Section 4, typical results of simulation in a computer software environment and hardware tests using TI’s DSP as control core are presented. Finally, the conclusion of this work is summarized in the last section.
3. Controller Design for LVRTC
To achieve a better dynamic performance, the proposed LVRTC on a three-phase inverter adopts a dual-loop control scheme, where the inner loop controls the output currents, and the outer loop controls the DC voltage and reactive power. When the WTG outputs power, the DC voltage controller regulates the DC link voltage to follow the voltage command, such that the link stays unaffected by the control of reactive power. Under normal circumstances, higher power factor and low-harmonic currents are required to reduce losses, and improve system stability. With a proper mathematical derivation, system voltage and current equations can be decoupled into d and q axes, as shown as follows:
where
ud,q represents the inverter switching node dq-axis voltage command,
vd,q represents grid dq-axis voltage,
Lf is the filter inductance,
denotes the equivalent resistance of the inductor.
Ifd,q represents dq-axis filter current, and
ed,q represents inverter switching node dq-axis voltages. After derivation, we obtain the plant transfer function and dq-axis current controller transfer function as follows:
where
kiP represents the proportional gain,
kiI represents the integral gain, and
Ti represents the integral constant and is defined as
kiP/
kiI. The inner loop current control structure is shown in
Figure 5.
As shown in
Figure 5, the close-loop transfer function of the current controller can be derived as follows:
where
Tj represents the time constant, and is defined as
Lf/
rl. Letting
Td =
Tq =
Tj, we obtain the following:
The DC voltage controller is constructed as shown in
Figure 6, where the DC voltage error is used to obtain the current command through controller
GvPI. Controller and plant transfer functions are expressed as follows:
where
is the dc capacitor of the inverter.
kvP represents proportional gain,
kvI represents integral gain, and
Tv represents integral constant, and is defined as
kvP/
kvI. The control structure is shown in
Figure 6.
Based on
Figure 6, the close-loop transfer function of the DC voltage controller can be expressed as shown in Equation (12). As a result, the complete inverter controller can be obtained as shown in
Figure 7.
As shown in
Figure 2 and
Figure 7, when the grid voltage falls below 90% of the nominal value, the reactive current control command will be switched to LVRT mode, and perform reactive power command tracking control. In this study case, the system specifications of the inverter are as follows: system rating: 2 kVA; grid voltage
VL-L: 220 V/60 Hz; DC voltage
Vdc: 400 V; filter inductance
Lf: 3 mH; filter inductor resistance
rl: 3.5 Ω; DC link regulator capacitance
Cd: 560 μF; and switching frequency
fSW: 18 kHz. According to these specifications, the plant transfer functions of the DC voltage loop and current loop are as follows:
In this application case, a dual-loop control scheme is used to achieve a better voltage regulation on the DC bus. Theoretically, the outer loop (the DC voltage controller) should have a smaller bandwidth than the inner loop (the current controller). In practice, these parameters are properly chosen to achieve the desired dynamic performance [
30]. Here, the close-loop bandwidth of the DC voltage controller is set at 500 Hz, and the bandwidth of the current response is set at 1/10 times the switching frequency, which is 1.8 kHz. This approach yields proportional and integral gains of the current control loop:
kdP =
kqP = 33.93, and
kdI =
kqI = 39584. Next, Equation (12) can be expressed in the form of a second order standard equation, with damping ratio
ξ, and undamped natural frequency
ω0, as shown as follows:
where
kvP = 2
ξω0Cd,
Tv = 2
ξ/
ω0, and,
kvP =
kvP/
Tv.
ξ and
ω0 can be chosen according to the desired dynamic performance. Letting
ξ = 1 and
ω0 = 500 yields the following results:
kvP = 0.56, and
kvI = 140.
Figure 8 shows the Bode plots of the DC voltage loops.
4. Simulation and Implementation
In this study, the POWERSIM simulation software is used in both simulation and hardware implementation, and a personal computer equipped with an i7-9700 CPU, @ 3.00 GHz is used as a control desk. The sampling frequency used in the hardware design of the proposed LVRTC is 50 kHz. It is the same as that of the switching frequency used.
Figure 9a shows the schematic architecture of the simulated LVRTC system connected to the power grid. For the simulation scenario, the grid voltage variation sequence is planned as follows: 0.5 pu, 0.6 pu, 0.7 pu, 0.8 pu, 0.9 pu, 1.0 pu, 1.2 pu, and then 1.0 pu. The corresponding reactive power commands for the LVRTC are generated according to the grid code. The simulation results are shown in
Figure 10.
Figure 11 shows the detailed steady-state waveforms of
Figure 10. It can be observed that the LVRTC successfully feeds appropriate reactive current to the grid during faults.
Figure 12 shows the results of the experimental tests using the constructed 2 kVA SiC-based three-phase inverter as shown in
Figure 9b, and a programmable AC power supply emulating the grid. In this study, the TI’s DSP (TMS320F28335) is used as the control core for the proposed LVRTC hardware, in which six SiC switching devices using ON Semiconductor’s NTHL060N090SC1 with the driver integrated circuits of Si8271 and the sensor devices for sensing currents and voltages are used to facilitate a flexible experimental system. Regarding the task of DSP’s programming, the SimCoder tool of POWERSIM software (2021) is used to convert the AD modules, various functional blocks, and controllers into the required DSP control program. Then, the DSP control program can be burned into the DSP chip. In the hardware implementation stage, the control desk (PC) is used to communicate with the DSP, and the controller parameters can be adjusted online to achieve the best control performance and accuracy.
Figure 9c shows the DSP programming steps in hardware implementation of LVRTC. In this study, the identical system condition used in the simulation case is implemented here in the hardware test.
Figure 13 shows the detailed steady state waveforms of
Figure 12. It can be clearly observed that the implementation results are in good agreement with the simulation results.
5. Discussion
With the growth of wind power generation over the past decade, various LVRT technologies have been proposed to ensure the safe and reliable operation of WTGs and power systems. Some advanced LVRT control algorithms reviewed in this paper and reported in many other similar papers in the literature can be categorized into two groups, i.e., the protection strategies- and the control strategies-based methodologies. However, most of the reported control strategies were implemented by the WTG built-in converters with complex controllers. It should be noted that with a limited hardware capacity, during the period of fault, the real power output capability of the built-in converters of WTGs must be dynamically limited to meet real-time reactive power requirements, and comply with strict grid codes. Based on the LVRT grid code, when the voltage dipped over 0.5 pu, the total current capacity of inverter system of the WTGs should be used to output the reactive power. This can strongly limit the control flexibility of a real power restoration scheme in the post-fault period. In this paper, we have demonstrated a distributed LVRT compensator (LVRTC) control scheme that can, to some degree, solve the above problem, and increase the system control flexibility. In addition, with the proposed distributed LVRTC control scheme, the WTG in a wind farm has greater freedom in performing its protection schemes, or regulating its power flow to ensure system stability.