Comprehensive Review on Fault Ride-Through Requirements of Renewable Hybrid Microgrids

Abstract

The world is interested in applying grid codes to increase the reliability of power systems through a micro-grid (MG). In a common practice, the MG comprises a wind farm, and/or photovoltaic (PV) arrays that are integrated with diesel generators and energy storage devices. Fault ride-through (FRT) capability is an important requirement of grid codes. FRT means that the MG is still connected to the grid during numerous disturbances such as faults. This is required to ensure that there is no loss of power generated due to grid faults. Reactive currents must be injected into the grid to increase the power system stability and restore voltage. To enhance FRT for doubly fed induction generator (DFIG) based WT installation, internal control modifications of rotor-side converters and grid-side converters are applied. The solutions that depend on these modifications are traditional and advanced control techniques. Advanced control techniques are needed due to the non-linear nature and less robustness of traditional ones. External hardware devices are also added to improve the FRT of DFIG which are classified into protection devices, reactive power injection devices, and energy storage devices. A comprehensive review of FRT enhancements of DFIG-based WTs, PV systems, and MGs using hardware and software methods is presented in this effort. A classification of FRT of PV systems is characterized plus various inverter control techniques are indicated. Several FRT methods for hybrid PV-WT are presented, with full comparisons. The overall operation and the schematic diagrams of the DFIG-WT with FRT methods are discussed and highlighted. Many Robust control methods for controlling grid connected AC, DC and hybrid AC/DC MGs in power systems are addressed. A total of 210 reported articles were review, including the most up-to-date papers published in the literature. This review may be used as the basis to improve system reliability for those interested in FRT methods. Various traditional and advanced control techniques to improve the FRT abilities are summarized and discussed, including protection devices, reactive power injection devices, and energy storage. In addition, the classifications of FRT hardware methods for DFIG are presented, including grid code requirements.

1. Introduction

Due to the increased demand for energy sources, renewable sources are being used in place of fuels from fossil sources. Fuels from fossil sources are not renewable, are very expensive, and supplies are being exhausted [1]. Resources for renewable energy are photovoltaic (PV), wind, hydroelectricity, geothermal, and biomass energies [2]. FRT enhancement is required because of the problems in managing the frequency, and low voltage may initiate, in the worst case, system collapse.

FRT capability is a significant requirement of grid codes (GCs) which means wind turbines (WTs), PV systems, or MGs remain connected to the grid during and after a fault within a certain limit [3,4]. Disconnection of a WT will cause a greater consumption imbalance, drop in the frequency of the system, and inrush current, so the WT must trip off when the voltage drop exceeds a limit [5]. GCs in a variety of countries are shown in Figure 1a [6,7,8,9]. In Denmark, Ireland, and Germany GCs, voltage profile is defined by the type of fault (symmetrical or asymmetrical). Another national grid code is specified by the depth of voltage drop. Grid code requirements (GCRs) for FRT show allowing voltage fault ($V_f$), fault time $\left(T_f\right)$, recovery time ($T_{rec}$), and recovery voltage ($V_{rec}$) in Figure 1b.

WT generators are permanent magnet synchronous generators [10,11,12], synchronous reluctance generators [13,14], and doubly-fed induction generators (DFIG). DFIG is the most common generator used with wind energy systems due to its wide range of speeds [15,16,17]. Its converter handles 25–30% of its power rating, so power losses are low. DFIG is extremely sensitive to grid disturbances, so FRT enhancement is required.

Internal modifications to controlling rotor-side converter (RSC) and grid-side converter (GSC) are proposed. For example, an enhanced control scheme on GSC by using instantaneous rotor power feedback [18], a coordinated control method depends on adding three controllers (damping controllers, RSC, and GSC) [19], a double degree of internal model control [20], demagnetization current method [21,22], flux vector control method [23] and virtual damping flux [24]. The sliding mode control provides robustness against parameter uncertainties [25]. A fuzzy logic controller provides rapid transient responses with dc-link voltage [26]. A fuzzy logic controller is tuned by the genetic algorithm proposed in [27]. The reference adaptive internal control is tuned by fuzzy logic [28]. A dynamic programming power control plus based on dynamic programming control is proposed in [29]. A vector control scheme based on stator magnetization current is proposed in [30]. A control method based on storing energy during fault and returning it to the grid to keep dc-ink at its rated value is proposed in [31]. (PI-DER) controller based on PI controller and dual-frequency resonant compensator is proposed in [32]. An advanced controller based on feedback linearization is proposed in [33]. Coordinated control of RSC and GSC is proposed in [34]. A power control loop-based hysteresis current regulator is proposed in [35]. A non-linear control scheme for RSC and DC link voltage for GSC is proposed in [36]. A feedforward transient current controller is proposed in [37]. Vector-based hysteresis current controllers are proposed in [38]. A linear quadratic output-feedback control method is applied to GSC and RSC [39]. A new robust controller obtains a very fast dynamic response for GSC [40]. A new controller achieves a constant torque for unbalanced voltage dips by defining a suitable GSC reference current [41]. A control scheme is enabled to control rotor power [42]. A simple controller is established through a switch [43]. A linear quadratic regulator (LQR) is an optimal controller and depends on a group of linear differential equations and matrices [44].

These methods are proposed to enhance FRT for DFIG. Most of them are complicated methods and not efficient, so hardware methods need to be added [45]. Comparisons between control methods are discussed to get high performance and achieve stability of DFIG in the power systems. The main topics covered in this review article revealed in Figure 2.

Hardware methods are protective devices, reactive power devices (RPDs), and energy storage devices (ESDs). Protective devices limit an increase in the rotor current and dc voltage which are a crowbar circuit [46] and a series R-L impedance circuit [47]. DC chopper resistance is proposed to maintain dc voltage at rated value and protect the IGBT switch from over-voltage [48]. A Dynamic braking resistor is an effective protection method for restoring voltage at the load side [49]. A Superconducting current limiter is proposed to limit voltage drop efficiency [50]. Bridge type fault current limiter (FCL) is proposed to eliminate the fault consequences even for zero grid fault [51]. Modulated series dynamic braking resistor (MSDBR) is proposed to limit rotor over current and voltage according to its threshold values. Series GSC is proposed to be capable of off-line operation for high efficiency [52]. RPDs enhance the performance of DFIG by injection of reactive power. Static var compensation (SVC) is a good source of reactive power compensation [53]. Static synchronous compensation (STATCOM) increases the stability of the grid through providing DFIG with more deaccelerated torque [53]. Dynamic voltage resistance (DVR) is an effective method of restoring the voltage [54]. Unified Power Quality Conditioner (UPQC) is a suitable solution for compensating reactive power [55]. ESDs depend on storing active power during faults which are batteries [56], supercapacitor [57], flywheel [58], and compressed air energy storage (CAES) [59].

Several hybrid methods are introduced to improve FRT of DFIG by providing high efficiency and getting the optimum performance to DFIG as crowbar with dc-link chopper [60], crowbar with battery ESD [61], DVR with FCL [62], and DVR with CBFCL [63].

The arrangements of this review paper are as follows: DFIG based on WT system which presents several aerodynamic models of WT and DFIG with different assumptions for modeling RSC and GSC, classifications of FRT capability methods of WT are in Section 2, classifications of FRT capability methods for a PV system is represented in Section 3, Hybrid PV-WT techniques are proposed in Section 4 and summary and discussions are presented in Section 5, respectively. Last, the ultimate concluding remarks along with future prospective are mentioned in Section 6.

2. DFIG Based WT System

The major construction of the DFIG-WT system is WT, drive train system, pitch controller system, speed control system, DFIG, RSC, and GSC which are combined to form a complete control to generate wind power at a constant frequency. The system structure is shown in Figure 3.

2.1. Aerodynamic Model of Wind Turbines

The mathematical model of WTs defines the form of the relationship between the output power and other variables. Design models of WTs depend on the power Coefficient $\left(C_p\right)$ which specifies the type of each turbine, so it has a unique value. Mathematical functions for ($C_p$) may be sinusoidal, exponential, polynomial, alternative, or continuous state observer functions where the exponential-based model is the most effective one. Using WT models, FRT techniques could be enhanced by simulating the effect of devices under various conditions. In

Table 1. Different models for representing the $C_p$ Coefficient of the WT.

Ref	${\mathit{C}}_{\mathit{p}}$ Functions	Expression of ${\mathit{C}}_{\mathit{p}}$	Conclusions
[64,65,66,67,68,69,70,71]	Exponential models depend on variations in the relationship between $\mathsf{\lambda }$ and $\beta$. The output power is controlled through the action of the torque and turbine speed which adjust the rotation of the blades of the turbine	$P\left(w\right)=\frac{\rho }{2}\pi R^2{v}_w^3{c}_p\left(\mathsf{\lambda }\beta \right)$ $c_p\left({\mathsf{\lambda }}_i,\beta \right)=C_0\left(\frac{C_1}{{\mathsf{\lambda }}_i}+C_2\beta +C_3{\beta }^{C_4}-C_5\right)e^{\frac{C_6}{{\mathsf{\lambda }}_i}}+C_7{\mathsf{\lambda }}_i)$ $\frac{1}{{\mathsf{\lambda }}_i}=\frac{1}{\mathsf{\lambda }+x_0\beta +x_1}-\frac{x_2}{{\beta }^3+1}$ This model has eight analyzed models of $c_p\left(\mathsf{\lambda }/\beta \right)$ with different coefficients that define the type of turbine.	This model gives a reliable result with Practical values of $\mathsf{\lambda }$, where his equations describe the behavior of the model effectively. Aguayo et al. model gives a maximum power coefficient which means maximum power can be extracted from the wind. $Cpmax=0.48$ at $\mathsf{\lambda }=8.092)$.
[72,73,74,75,76]	Sinusoidal models depend on variations in the relationship between- $\mathsf{\lambda }$ and $\beta$. The output power is controlled through the action of the torque and $v_w$, and β	$P\left(w\right)=\frac{\rho }{2}\pi R^2{v}_w^3{c}_p\left(\mathsf{\lambda }\beta \right)$ $\mathsf{\lambda }=\frac{R{\omega }_r}{v_w}$ $c_p\left(\mathsf{\lambda }/\beta \right)=\left[a_0+a_1\left(b_0\beta +a_2\right)\right]\sin \left[\frac{\pi \left(\mathsf{\lambda }+a_3\right)}{a_4+a_5\left(b_1\beta +a_6\right)}\right]+a_7\left(\mathsf{\lambda }+a_8\right)\left(b_2\beta +a_9\right)$ This model is satisfied by five analysed models which differ in coefficients.	This model handles with impractical $\mathsf{\lambda }$. The Bouallegue et al. model gives $Cpmax=0.5405$ at $\mathsf{\lambda }=4.972$. In these models, period and amplitude are varied from each other.
[77,78]	Polynomial models depend on specific speed ($\mathsf{\lambda }$), where the angle of attack of the blade WT is constant. This model is used in turbines with lower power.	$P\left(w\right)=\frac{\rho }{2}\pi R^2{v}_w^3{c}_p\left(\mathsf{\lambda }\beta \right)$ $\mathsf{\lambda }=\frac{R{\omega }_r}{v_w}$, $c_p\left(\mathsf{\lambda }\right)={\displaystyle \sum }_{i=0}^{i=n}{a}_i\mathsf{\lambda }{ }^i$ Where the maximum order is n = 7. Four models are found which are third-order, fourth-order, fifth-order, and seventh-order.	The third-order model gives $Cpmax=0.52$ at $\mathsf{\lambda }=9.799$ which is very simple and easy to carry out. Fourth-order gives $c_p$ out of limits of Betz law so it is un possible to use it.
[79]	Alternative function	$P\left(w\right)=\frac{\rho }{2}·\pi ·R^2·v_w^3·c_p\left(\mathsf{\lambda }\beta \right), \mathsf{\lambda }=\frac{R·{\omega }_r}{v_w}$ $c_p\left(\mathsf{\lambda }\right)=\frac{1}{0.95\surd 2\mathsf{\pi }}·e^{\left(\frac{0.1296}{1.805} {\mathsf{\lambda }}^2-\frac{2.0736}{1.805}\mathsf{\lambda }+\frac{8.2944}{1.805}\right)}$ $c_p\left(\mathsf{\lambda }\right)=\frac{1.8e^{-0.18\left(\mathsf{\lambda }-4\right)}}{\left(1+e^{-0.18{\left(\mathsf{\lambda }-4\right)}^2}\right)·\left(1+e^{-1.26\left(\mathsf{\lambda }-4\right)}\right)}$	This model is subject to $c_p$ limited magnitude for values of $\mathsf{\lambda }$ which is out of experimental limits.
[80]	Continuous state observer is used to estimating $c_p$ in WT which is connected to separately excited dc generator	$P\left(w\right)=\frac{\rho }{2}·\pi ·R^2·v_w^3·c_p\left(\mathsf{\lambda }\beta \right), \mathsf{\lambda }=\frac{R·{\omega }_r}{v_w}$ $\frac{d\omega }{dt}=\left(\frac{\rho \pi R^2{v}_w^3{c}_p}{2j\omega }-\frac{\gamma i_a{i}_f{k}_{1 }}{j}-\frac{B\omega }{j}\right)$	This observer helps in controlling MPPT design. It is used with WECS which deals with different generators.
[81]	The new electricity system cascade analysis	$P\left(w\right)=\frac{\rho }{2}·\pi ·R^2·v_w^3·c_p\left(\mathsf{\lambda }\beta \right), \mathsf{\lambda }=\frac{R·{\omega }_r}{v_w}$ $c_p\left(\mathsf{\lambda }\right)=\frac{G·\mathsf{\lambda }·\left({\mathsf{\lambda }}_o-\mathsf{\lambda }\right)}{a^2+{\left({\mathsf{\lambda }}_o-\mathsf{\lambda }\right)}^2}$	a, G, ${\mathsf{\lambda }}_o$: constant of regression process.
[82]	This model depends on adopting piecewise function	$P\left(w\right)=\left\{\begin{array}{c}0 v_t\ge v_o, v_t \le v_i\\ P_r\left(\frac{v_w^3-v_i^3}{v_r^3-v_i^3}\right) v_i\le v_w\le v_r\\ P_r v_r\le v_t\le v_o \end{array}\right\}$	$P_r$ represents WT rated power, $v_r$ represents rated wind speed of WT; $v_i$ cut-in wind speed of WT, $v_o$ cut-out wind speed of WT.
[83]	Multi-objective self-adaptive differential evolution algorithm	$P\left(w\right)=\frac{\rho }{2}\pi R^2{v}_w^3{c}_p\left(\mathsf{\lambda }\beta \right)$ $P\left(w\right)=\left\{\begin{array}{c}0 v_w\ge v_o, v_t \le v_i\\ P_r\left(\frac{v_w^2-v_i^2}{v_r^2-v_i^2}\right) v_i\le v_w\le v_r\\ P_r v_r\le v_t\le v_o \end{array}\right\}$	$v_i$= 2.5 m/s, $v_r$ = 13 m/s, and $v_o$ = 25 m/s.
[84]	The output power of WT is a polynomial function which is a function of wind velocity.	$P\left(w\right)=-2\times e^{-5}{v}_w^6+0.001v_w^5-0.0155v_w^4+0.1058v_w^3+0.763v_w-1.9152$	In this model, outpower is extracted when wind speed is above 3.5 m/s (cut-in speed) until a rated wind speed of 9.5 m/s is reached. For wind speed exceeding 25 m/s, the turbine has to be stalled to prevent structural damage.
[85]	The electrical power extracted from WT is limited by the range of operation between the cut-in and cut-out wind speed	$P_w=P_r\left\{a_1\exp \left[-{\left(\frac{v_t-b_1}{c_1}\right)}^2\right]+a_2\exp \left[-{\left(\frac{v_t-b_2}{c_2}\right)}^2\right]+a_3\exp \left[-{\left(\frac{v_t-b_3}{c_3}\right)}^2\right]\right\}$	$a_1$, $b_1$, and $c_1$ are regression coefficients.
[86]	Multiple polynomial regression with a maximum relative error of about 14%	$P_w=\frac{\rho }{2}\pi ·R^2·v_w^3·c_p\left(\mathsf{\lambda }\right)·{\eta }_{°}$ $c_p\left(\mathsf{\lambda }\right)=\left[-2.025\times {10}^{-7}{v}_w^6+1.926\times {10}^{-5}{v}_w^5-7.421\times {10}^{-4}{v}_w^4)\right]+1.483\times {10}^{-2}{v}_w^3-0.162v_w^2+0.887v_w-1.508$	Variable speed, constant speed, pitch-controlled WTs, and blade number determine this model’s form. This model needs fewer parameters only $c_p$ and ${\eta }_{°}$.
[87]	$\alpha$ is the Hellman coefficient	$P_w=\left\{\begin{array}{c}\frac{\rho }{2}\pi R^2{v}_w^3{c}_p\left(\mathsf{\lambda }\right){\eta }_{pv } v_i\le v_w\le v_o \\ o v_w\ge v_o, v_t \le v_i \end{array}\right.$ $\frac{v_2}{v_1}={\left(\frac{h_2}{h_1}\right)}^{\alpha }$	$c_p$ = 0.42, $\rho =\mathrm{kg}/m^3$, $\alpha$ = 0.25, $h_1$ = 10 m and $h_2$ = 40 m.

, the various models are compared and concluded.

Mechanical power produced by WT is as shown in (1). The tip speed ratio can be calculated based on the ratio between linear turbine speeds to wind speed which is shown in (2).

$$p_m=\frac{\rho }{2}\pi ·R^2·v_w^3·c_p(\mathsf{\lambda }\beta )$$

(1)

$$\mathsf{\lambda }=\frac{R·{\omega }_r}{v_w}$$

(2)

2.2. Doubly Fed Induction Generator

DFIG is the extremely generator used in wind systems which operate at sub and super synchronous speed. Its stator associates directly with the grid, while the rotor associates to a back-to-back converter [88,89,90]. DFIG and its conversion system are in Figure 3. The drawback of DFIG is increasing maintenance cost due to its slip rings that have a negative effect on power systems efficiency [91]. It is extremely sensitive to voltage fluctuation as its stator is connected directly to the grid [92]. The mathematical model of DFIG according to d-q synchronize rotating reference frame is presented in [93,94,95].

2.3. Conversion System of DFIG

In DFIG back-to-back converters, active power and reactive power are controlled to achieve variable speed operation, as shown in Figure 3 [96]. DFIG converters are controlled through vector control. It gives effective control for RSC and GSC in the synchronous d-q reference frame. RSC connects to rotor winding to control electromagnetic torque through the q axis and stator reactive power through the d axis. Three-phase rotor current ($i_{rabc }$) is become rotor current in the-d-q axis ($i_{qr}, i_{dr}$). Reference power ($p_s^{*}$) which is got from the MPPT curve is compared with the measured value ($P_S$) to get $i_{qr}^{*}$. Reactive power is controlled by comparing $Q_s^{*}$ with $Q_S$ to get $i_{dr}^{*}· i_{qr}^{*}$ and $i_{qr}$ are compared to get the PI controller signal to get $V_{\mathrm{dr}}$ and also the same for $i_{dr}^{*}$ and $i_{dr}$ to get $\mathrm{Ver}$. $V_{\mathrm{dr}}$, $V_{\mathrm{qr}}$ signals are passed to PWM to get the switch signal of the RSC.

GSC regulates the dc voltage at rated value and controls reactive power which is passed to GSC and the grid. The reference value of dc-link voltage ($v_{dc}^{*}$) is compared with the measured value of dc-link voltage to get $i_{dg}^{*}$. $Q_g^{*}$ is compared with $Q_g$ to get $i_g^{*}$. $I_{qg}$ is opposed to $i_{qg}^{*}$ to get the PI controller signal to get $v_{qg}$. The same is for $i_{dg}$ and $i_{dg}^{*}$ to get $v_{dg}$. The current regulator adjusts the error signals $v_{qg},v_{qg}$ and sends them to PWM to get the signal of the GSC switch. RSC and GSC can be considered as two controllable voltage sources so, internal control modifications can be implemented. Different assumptions to control RSC and GSC are discussed in

Table 2. Different models of RSC and GSC.

Ref	Controlling (RSC)	Controlling (GSC)	Conclusions
[7,53,98,99]	Voltage-oriented vector control (VOC) controls the RSC of DFIG, where the stator voltage is aligned to the d-axis of the rotating frame for the decoupling control of the active and the reactive power.	GSC controls the dc-link voltage where the d-axis of the rotating reference frame is aligned with the positive sequence of grid voltage.	VOC proposes a good steady-state operation for DFIG-WT. VOC is the most commonly used control technique in modern DFIG-WT
[24,27,54,55,56]	Stator flux-oriented vector control (FOC) controls RSC which is achieved through rotor speed $({\omega }_r)$ and $i_{qr}^{*}$. It controls the output of DFIG by controlling the q component of RSC	GSC control-dc-link voltage where $Q_g^{*}=0$ as DFIG converter handles with 25–30% of its rating.	FOC reduces the oscilation damping when the excitation current is increased.
[16]	Direct torque control (DTC) technique controls the RSC of DFIG to control reactive power and generator speed by controlling electromagnetic torque	GSC controls dc-link voltage where $i_q^{*}=0$.	DTC regulates rotor currents through PI controllers. Its performance has a fast response, but it has a constant switching frequency which produces ripples in current and reduces the robustness of the mechanical gearbox.

[97,98,99,100,101].

2.4. Classification of FRT Methods for DFIG-WT

RSC governs the speed of DFIG and the reactive power. GSC governs the dc voltage and keeps it at its rated value to ensure active power is exchanged between RSC and GSC to the grid. DFIG is extremely sensitive to the grid fault as there is a direct connection between its stator and the grid [91,92,93,94,97]. Therefore, different methods are proposed to provide enhancement the FRT capability of DFIG which are protection devices, reactive power injection devices, ESDs, traditional control methods, and advanced control methods. The classifications of FRT capability methods are depicted in Figure 4.

2.4.1. Protection Devices

Conventional protection techniques are vital to limit the RSC overcurrent and maintain dc-link voltage at its desired value. Crowbar circuit (CB), braking chopper resistor (BCR), DBR, and FCLs are the protection devices as shown in Figure 5. CB is preferred in DFIG-WT systems as it is the simplest method to protect RSC from overvoltage, easy operation, and control [102,103]. Three schemes of CB are proposed which are antiparallel thyristor with resistance (CB1), IGBT bridge with resistance (CB2), and diode bridge rectifier with IGBT with resistance (CB3). CB1 is the cheapest type as it used a thyristor switch. It offers instant turn-on/off. The disadvantage of this scheme is the turn-off characteristic which needs waiting for rectifying zero current because the rotor current has a large dc component. CB2 limits over current and protects RSC. This scheme is more complex and expensive than clamping circuitry [45,104]. CB3 saves cost and its design is easier to control as one switch is used. It is a successful scheme for FRT enhancement of DFIG as it eliminates short circuit current [105].

In the presence of a CB circuit linked to the rotor circuit, DFIG behaves like a squirrel cage induction generator, and more reactive power is consumed from the grid during fault [46]. The suitable value of crowbar resistance is depended on DFIG performance. The suitable ratio of $R_{crowbar}/R_{rotor side converter }$ is 10 for 9 MW DFIG-WT [47,106,107]. The suitable value of crowbar resistance can be calculated by the following equation [90].

$$R_{cr}=\frac{\surd 2\left(v_{max}{\omega }_s{L}_S\right)}{\surd \left(3.2v_s^2-2v_{max}^2\right)}$$

(3)

BCR connects to the dc-link to dampen the oscillation. It protects the converters without disconnecting them during disturbances. In this technique, resistance connects to the IGBT device and PI controller to control the switching signal [8,108]. DBR is a resistor that connects in parallel with a switch to control its resistance. It links to the stator side of DFIG to limit the stator over the current [109]. It increases the time of operation of RSC to avoid disconnection during a fault. CB circuit reduces torque fluctuations [108]. It is very sensitive to the switching delay so it is essential to use Feed Forward or Feed Back control technique to enhance the FRT of DFIG [110].

MSDBR is a resistor that connects to anti series power electronic switches at each phase to avoid synchronizing in stator voltage. PWM signals control it with a fixed switching frequency, which is the most cost-effective strategy [111].

Super conduction FCL (SFCL), capacitive bridge FCL (CBFCL), switch type FCL (STFCL), and resonant type FCL (RTFCL) are the most effective types of FCL [112]. SFCL locates in two positions stator side and rotor side where SFCL on the stator side has a good performance. Types of SFCL are inductive type(L-type), resistive type (R-type) and hybrid between(L-type and R-type) solid-state FCL (SSFCL) [112]. L-type is effective in reducing voltage drop, R-type is effective in consuming energy and SSFCL has both characteristics of R-type and L-type which protects DFIG during fault [51,113]. SFCL does not need any control technique, so it provides self-regulating behavior [114].

CBFCL is modified to inject reactive power by capacitive impendence to satisfy the requirements of GCs of DFIG-WT by supporting grid connection points. CBFCL application is the effective method to decrease short circuit current if the fault occurs and provides the needed reactive power [100,115]. The main merit of resonant type FCL (RTFCL) lies in its impedance, which is negligible in normal operation. Managing the first peak of a fault current is possible [116]. Developments in the use of wind energy have led to the use of SSFCL, which is the most commonly used system, that also has a high cost. CBFCL has a high cost, but it significantly outperforms other types. Comparisons between protection devices are given in

Table 3. Comparisons between protection devices.

Ref	Device	Advantages	Disadvantages
[46,117]	Crowbar	Prevents RSC from overcurrent by making short circuits on rotor winding. It is very simple and effective in symmetrical faults	The DFIG consumes a greater amount of reactive power, voltage recovery is inhibited and it behaves like a squirrel cage induction generator. The low value of resistance causes increasing electromagnetic torque oscillation which causes more stress on the drive train
[46,48]	DC chopper resistance	Maintains dc voltage at a constant value and tends to protect IGBT from overvoltage	It does not affect rotor current damping. This method dissipates the power in the dc-link
[46]	Modified dc chopper resistance	Three semiconductor switches are used to insert chopper resistance in series or parallel connection with the dc-link. It is used to limit overcurrent on rotor side and increase voltage in the dc-link according to its threshold values	High cost
[49,118,119]	DBR	Effective protection method that restores voltage at the load side. The RSC continues to operate, unlike the crowbar circuit	Avoids dc-ink overvoltage but it needs a control technique to overcome frequent faults. Difficult to synchronize the stator voltage. It is very sensitive to switching delays. It is not limit fault current effectively
[50,120]	SFCL	In DFIG, SFCL connects in series with the rotor winding or stator winding. It is used in transmission lines. It has a fast response. It limits voltage drop efficiency. It does not need adding impedance in normal operation	Needs a special complex cooling system for the superconductors
[99]	Switch type FCL	It connects to the dc side of RSC to limit the overcurrent of RSC. It solves the problems of crowbar protection as it eliminates the fault consequences even for zero grid voltage. It does not use superconducting inductance, so it is less cost	Affected by temperature and current density, and it needs compensation devices
[51,115,121]	BCFCL	Combines protection tasks for DFIG and provides reactive power compensation. It has low cost, low losses, and low voltage drop in limiting fault current	High cost

.

2.4.2. Reactive Power Injection Devices

The risk of voltage collapse for weak power networks and injecting reactive power is an important demand of GCs which are led to the use of reactive power injection devices to help DFIG-WT in voltage recovery. Series compensation devices SVC, the static synchronous compensator (STATCOM), shunt compensation device (i.e., DVR, and hybrid compensation device of UPQC and shunt compensation and series compensation together) are reactive power injection devices.

SVC consists of a thyristor-controlled reactor (TCR) and a thyristor-switched capacitor (TSC) to provide and absorb reactive power to the power systems by lag and leading current as shown in Figure 6A [122]. STATCOM is composed of IGBTs linked in series with a voltage source converter (VSC) [123]. SVC and STATCOM provide a fast steady-state and transient voltage control but they don’t achieve complete protection for DFIG, so ESDs are needed to add.

DVR is a nonlinear device that comprises a resistance linked in series with the grid terminal through a coupling transformer and in parallel with the dc-link of the converters with ESD which is the effective method to enhance FRT of DFIG from descending the line voltages during disturbances in voltage supply [54]. The drawback of DVR has required a converter with a high power rating. This problem can be solved by reducing the grid converter’s reference power. The active power can be obtained from the following equation [90]:

$$p_{DVR}=\left(\frac{v_{pcc}-v_f}{v_{pcc}}\right)·p_{load}$$

(4)

DVR with CBFCL is introduced in [63]. DVR is modified by adding an ESD to enhance the FRT capability of DFIG as shown in Figure 6B.

UPQC is an active power filter that consists of a shunt and a series compensation device as shown in Figure 6C. In the past, it can be achieved by six switches but now it is achieved by ten semiconductors switches VSC. The extra three switches are used with the same VSC in shunt term when a voltage of the common coupling point is V_PCC > 0.85 pu and they are used in series when V_PCC < 0.85 pu. This method provides DFIG with a more reactive current during a fault where GCRs are achieved [55]. Comparisons between RPDs are arranged in

Table 4. Comparisons between RPDs.

Ref	Device	Advantages	Disadvantages
[53]	SVC	Simple structure and good source for reactive power compensation. Its bus voltage is controlled and TSC reduces the power losses.	Controlling reactive power depends on voltage level. Installation and maintenance are very high in cost. Due to the fast response of SVC, the system has unstable voltage oscillations
	STASTOM	After the voltage is recovered, STACTOM provides DFIG with more deaccelerated torque which increases the stability of the grid. It provides higher currents at low voltage compared with SVC.	Lower response time than SVC. It does not protect RSC from overcurrent. High mechanical stress.
[54]	DVR	It restores the voltage when a fault occurs by using ESD. Other protective devices are not needed during operation. It is an effective solution when used with ESD.	Needs a high power rating from the power converter.
[55]	UPQC	Controls the active and reactive power. Suitable solution for compensating reactive power. Fast response.	Requires a high rating of the DC-link capacitor.

.

2.4.3. Energy Storage Devices

ESDs contribute to keeping the balance between generation and consumption in DFIG by absorbing or consuming power in dc-link during fault [124,125]. The bidirectional dc-dc converter helps in charging and discharging power in ESDs according to the duty ratio to achieve efficient power transfer in two directions and soft switching to reduce power losses [126,127,128]. CAES is used with a large-scale grid system if a suitable site is available [129]. Comparisons among ESDS are announced in

Table 5. Comparisons between ESDs.

Ref	Device	Advantages	Disadvantages
[58]	Flywheel	Efficiency is 85–95%, less maintained, high response time, and not affected by repeating charge and discharge.	High cost and a short life span.
[56]	Battery	Efficiency 75–95%, free maintains cost, and depth of charge 80%. It is the most effective way to store electricity by controlling its state of charge.	Short life cycle if the battery is discharged deeply. It has a slow time response so it cannot provide frequency support.
[97,114,115]	SMES	Cyclic efficiency 90–95%, large power density, response time is very short, and unlimited charging and discharging cycle.	It has a high capital cost. To ensure efficient application of SMES, suitable power system locations must be selected carefully in the power system.
[57]	Supercapacitor	Efficiency 85–95%, an important source of reactive power, long lifetime not affected by charging and discharging rate, and high power density.	Less energy density.
[59]	CAES	Efficiency 80–90%, and higher power density.	High capital cost.

.

Flywheel Energy Storage (FES) is a mechanical device that stores kinetic energy in a rotating mass to generate electrical power by an electrical motor. Low-speed FESS and high-speed FESS are types of FESS that are rotated at a speed less than 6 × 10³ rpm and the other rotates in the range between 10⁴–10⁵ rpm [130,131,132,133].

SMEs is a DC control device that stores electrical energy into an electromagnetic field when the direct current has flowed in a superconductor coil [134,135].

There are two types of SMEs: low-temperature and high-temperature types that are used with grid applications [136,137,138,139,140,141].

SC is an important source of reactive power when the voltage level is decreased than its rated value. Choosing the optimal value of SC is not an easy task as the large size of SC does not ensure the best solution but has a high cost. Equations are used to determine the suitable value of SC are in [57,142,143].

BES is the most widespread type of ESD which can charge or discharge electrical power by using a bidirectional DC-DC converter [144,145]. A lithium-ion battery is the most commonly used to enhance the dynamic performance of DFIG according to decision matrix (DM) criteria in [8,146,147,148]. Comparisons between ESDs are available in the Table 5.

2.4.4. Traditional Control Techniques

As protection devices, RPDs, and ESDs are unsatisfactory to enhance the performance of DFIG only, Traditional control techniques are proposed to control parameters of DFIG. Traditional control techniques are explained in Figure 3.

The pitch controller keeps the output power at the rated level when the wind speed increases the rated speed by using a PI controller. As the PI controller has limited bandwidth, an advanced fuzzy control technique is used with the pitch control technique to decrease its response time.

Hysteresis current controller technique [HCR] is utilized to obtain the switching signal by nonlinear feedback to enhance FRT of DFIG by providing a fast transient response [96]. This switching signal is determined by logic states [149]. This strategy reduces the switching frequency of RSC as it relies on instantaneous current measurements. It is more used than the model predictive control technique as it has a low order current harmonics [150]. It suffers from a high average switching frequency.

The feed-forward current controller is updated to a traditional vector control scheme which is depended on state estimation. The rotating speed of stator flux is calculated to use in feed-forward current control to reduce transient rotor current [151]. This method reduces torque ripple by injecting transient compensation to the current controller. Decoupling DQ current term and stator voltage coupling term are added to the output of the PI controller as a feed-forward current regulator to improve the FRT of DFIG [37]. Comparisons between traditional control techniques are presented in

Table 6. Comparisons between traditional control techniques.

Ref	Control Method	Advantages	Disadvantages
[158,159,160,161,162,163,164]	Blade pitch angle control	It protects WTs from damage during grid fault through the PI controller. This controller is a traditional controller which is improved by fuzzy control technique	It depends on generator speed not depends on generator power. It takes a longer time (13 s) to return the system to steady-state condition.
[165,166]	Hysteresis current vector controller	Simple technique to control active and reactive power. It helps in limiting peak current. It is not sensitive to the system parameters.	It may not provide DFIG with suitable compensation. Its voltage quality is not favorable. It causes large oscillations in the output current.
[37]	Feedforward control technique	Its load voltage is feedback to the voltage controller. Therefore, it is a simple technique and the stability of the power system is increased.	Its response is slow. It has a steady-state error. It is not robust to external disturbance
[102]	Vector control	High performance for DFIG in steady-state operation	It is not an effective control technique when severe voltage sages have occurred as its control technique is very sensitive to the parameter values of DFIG.

[152,153,154,155,156,157,158,159,160,161,162,163,164,165,166,167,168].

2.4.5. Advanced Control Techniques

For the nonlinear performance of the DFIG during a fault, advanced control techniques are illustrated to improve the FRT capability of the DFIG-WT.

The sliding mode controller is the nonlinear controller. It depends on the disturbance observer which is based on discontinuous control signals to change the dynamic performance of DFIG [153]. It eliminates fluctuations in electromagnetic torque [154]. This method is more robust, with no steady-state error and fast response [155,156].

Artificial intelligence and fuzzy logic are suitable solutions to improve the performance of nonlinear systems as they do not need to know the exact parameters of the system as traditional PI controllers. Its drawbacks are that it is difficult to calculate the maximum point or sharp point because of are more points nearly and knowing to map the input and output space [157]. New control technique for DFIG using fuzzy logic which is dependent on the P&Q method and FB method to get PID controller parameters [98]. It is very effective than the sliding mode technique in smoothing active and reactive power.

LQR is an optimal controller that is dependent on linear differential equations and matrices. It consists of two cascaded controllers to obtain the suitable value of PI by PWM. It monitors the reduced voltage on the DFIG side converter and deals with a multivariable controller based on the system linearization model [44]. Comparisons among advanced control techniques are depicted in

Table 7. Comparisons between advanced control techniques.

Ref	Control Method	Advantages	Disadvantages
[175]	Sliding mode operation	Fast response and provide robustness against parameter uncertainties to improve the transient performance of DFIG	It isn’t use for back-to-back converters and is used for RSC only as it is more complex. High-frequency switching
[158]	Fuzzy logic, genetic algorithm	Good performance and extend rapid-transient responses in dc-link voltage	They cannot track the maximum power point. They cannot face the impact at the different wind speeds
[44]	LQR	High accuracy for the system parameters, rapid convergence, and fast response.	It is more complex and has a steady-state error
[169,176]	Deadbeat controller	Computes the rotor voltage and applies it during voltage sage. This controller is designed by using a stator field controller	Not robust and it is used for a short sampling time
[177,178,179,180]	Model predictive control	It predicts the future behaviour of the controlled variables, which deals with multi-input multi-output problems. It gives high performance and fast response as it does not use a cascade control. It is a simple method as GSC is replaced by a dc power source for simplicity	Sensitive to parameter problems and its performance depends on accurate information about DFIG, which may not be available in a practical system. Needs some optimization, which reduces the computational efficiency
[181]	Nonlinear adaptive backstepping control	Calculate the parameters of any nonlinear system TrackS the rotor speed to optimise the extracted output power to control stator reactive power	Needs an additional observer to measure flux, which can’t be measured.

.

2.5. Hybrid Advanced Control Techniques

DFIG is not effective in avoiding oscillation due to the disturbances when using PI control because of its bandwidth limitations but this strategy is impacted by the leakage inductance of the stator and rotor of DFIG. The limitations of hardware devices are illustrated. Therefore, different strategies based on hardware devices with control techniques are introduced.

Heightened state-feedback predictive control (HSFC) is proposed in [167]. It enhances the FRT of DFIG by calculating the suitable rotor voltage to avoid the back electromotive force. Model predictive control based on BTFCL is proposed in [137]. It is an efficient method to decrease fault current and improve the performance of DFIG by compensating reactive power. In [58], a hybrid current control technique is proposed. PI current regulator is used in normal conditions and vector control-based hysteresis current control is used to limit over current. This technique has a low switching frequency, and its transient response is very high; however, its steady-state operation isn’t optimal. In [98] P&Q optimization and feedback controller are combined which is depend on PI controller. P&Q optimization and FB technique are tuned by fuzzy and genetic algorism to improve DFIG performance. In [159], fractional order the PI controller is applied to compensate pitch control loop at a rated value to improve the performance of DFIG. In [168,169], fuzzy logic depending on the deadbeat controller is proposed. The deadbeat controller uses a discrete-continuous equation to calculate input until the output reaches the desired value and sends an error to the fuzzy logic controller. This strategy is more robust but it has high switching losses. In [170], fuzzy logic based on particle swarm optimization is proposed. This technique can deal with the nonlinear system. It has fewer switching losses and protects the system from overvoltage. In [63], voltage control of DVR by the Feed-forward technique is proposed. It achieves a good transient and steady-state response. It supports the grid with good reactive power during fault and after a fault.

In [54], DVR with vanadium redox flow BES is proposed. DVR associated with DFIG in series and parallel with dc-link with BES. BESD can smooth fluctuations in active power and maintenance the dc-link in the desired value and DVR can compensate for the grid voltage drop. The VRB has a large capacity, long life, low cost, low maintenance requirements, and fast response in wind conversion applications. In [61] combined hysteresis control technique with Crowbar protection is proposed and BES device. The hysteresis technique reduces the operation time of the crowbar to protect RSC. Various Hybrid ESDs are proposed due to their high efficiency and getting optimum performance. Lithium-ion battery-based SMEs is proposed in [146,171], flywheel-battery ESD [172], compressed air–flywheel [173], and battery-super capacitor [174].

Table 8. Effect of enhancing FRT of DFIG when the fault occurred.

FRT Capability Methods	Limiting Stator Current	Decrease Rotor Current	Reduce Voltage Drop	Maintain Dc-Link Voltage	Reactive Power Support	Reduce Oscillations of Active Power	Decrease Oscillation Voltage	Reducing Torque Fluctuations	Effective for Symmetrical Fault	Effective for Asymmetrical Fault
Crowbar [46,117]	ꭗ	✓	ꭗ	✓	ꭗ	ꭗ	ꭗ	✓	✓	ꭗ
DC link chopper method [48]	ꭗ	ꭗ	ꭗ	✓	ꭗ	ꭗ	✓	ꭗ	ꭗ	✓
Modified dc chopper resistance [46]	ꭗ	✓	ꭗ	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	✓
DBR [118,119]	✓	ꭗ	ꭗ	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	✓
SFCL [50,120]		✓	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ
Switch type FCL [99]	ꭗ	ꭗ	ꭗ	ꭗ	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ
BCFCL [51,115]	✓	ꭗ	✓	ꭗ	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ
SVC [53]	ꭗ	ꭗ	ꭗ	ꭗ	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ
STASTOM [53]	ꭗ	✓	ꭗ	ꭗ	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ
DVR [54]	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ
UPQC [55]	ꭗ	ꭗ	ꭗ	ꭗ	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ
ESDs [58]	ꭗ	ꭗ	ꭗ	✓	✓	✓	ꭗ	ꭗ	ꭗ	ꭗ
Crowbar with SDBR [108]	ꭗ	✓	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	✓	✓	✓
Crowbar assembled with series R-L [47]	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	✓	✓
Crowbar assembled with dc-link chopper [47]	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	ꭗ	✓	✓

shows the effect of enhancing FRT on DFIGs connected to the grid during a fault condition.

3. Modelling of PV System

The output power of PV systems is affected by several factors, so various equations used to model PV system in

Table 9. Represented models of PV cells.

Ref	PV Power System Equations	Conclusion
[182]	$P_{pv}={\eta }_{pv}·A_{pv}·G_{\left(t\right)}$ $G_{\left(t\right)}=I_d·\left[\cos \left(\theta \right)+C·{\cos }^2\left(\frac{\varnothing }{2}\right)+\rho \left(\cos \left(x\right)+C\right)·{\sin }^2\left(\frac{\varnothing }{2}\right)\right]$ Where ${\eta }_{pv}$ is the PV system efficiency, Id is the direct normal irradiance, θ is the angle between the tilted surface and the solar rays, C is the diffuse portion coefficient, ρ is the reflection index and χ is the zenith angle. $P_{pv}$: instanteneous solar power generation	Do not take into consideration the influence of temperature on PV power generation.
[183]	$P_{pv}={\eta }_{pv}·A_{pv}·G_{\left(t\right)}$ ${\eta }_{pv}={\eta }_{pv-ref}·{\eta }_{MPPT}·\left[1-\beta \left(T_c-T_{c-ref}\right)\right]$ $T_c=T_a+\left[\frac{NOCT-20}{800}\right]·G_{\left(t\right)}$ $G_{\left(t\right)}=G_{\left(b\right)}·\left(\cos \left({\theta }_{pv}\right)+{\cos }^2\left(\frac{{\varnothing }_{pv}}{2}\right)·\sin \left(x\right)+{\rho }_{pv}·(\cos \left(x\right)+C)·{\sin }^2\left(\frac{{\varnothing }_{pv}}{2}\right)\right)$	where ${\eta }_{MPPT}$ is 1 in this model, $T_{c-ref}$ is taken 25 °C, $\beta$ ranges from 0.004 to 0.006 per °C for silicon cells and is set to 0.0048 in this model.$NOCT$ is defined as 45 °C. $T_a$ is 25 °C
[184]	$P_{pv}={\eta }_{pv}·A·G_{\left(t\right)}$ ${\eta }_{pv}={\eta }_{pvref}·A·\left[1-k_t·\left(T_C-T_{Cref} \right)\right]$ $T_C=T_a+\left[\frac{NOCT-20}{800}\right]·G_{\left(t\right)}$	${\eta }_{pvref}$ = 12%, $k_t$ = 0.33%, $T_C$ = 25 °C, and $NOCT$ = 45 °C.
[83]	$P_{pv}=P_{pvref}·\frac{G}{G_{ref}}· \left[1+k_t\left(T_{amb}+\left(0.0256\times G\right)\right)-T_{cref}\right]$	$G_{ref}$ is 1 kW/m2; and $k_t$ is a constant = −3.7 × 10−3 (1/°C)
[185]	$P_{pv}=\left\{\begin{array}{c}\begin{array}{c}{P}_{ref}\left(\frac{R^2}{R_{srs}{R}_{cr}}\right) 0\le R\le R_{cr}\\ P_{ref}\left(\frac{R}{R_{srs}}\right) R_{cr}\le R\le R_{srs}\end{array}\\ P_{ref} R_{srs}\le R\end{array}\right\}$	R: solar radiation factor, $P_{ref}$: rated power of PV array, $R_{cr}$: Specified radiation point, and $R_{srs}$ Standard test conditions for radiation. $R_{cr}$ = 150 W/m2,$R_{srs}$ = 1000 W/m2 and $P_{ref}$ = 260 W
[186]	$P_{pv}=f_{pv}·p_{pvref}·\left(\frac{G_T}{G_{ref}}\right)$	$f_{pv}$: derating factor.
[84]	$P_{pv}=-1.69e^{-10}{G}^4-1.47146e^{-7}{G}^3+2.2301e^{-5}{G}^2+1.358G-0.89025$	Output power of PV system is computed by polynomial function at any given solar irradiation $\left(G\right(w/m^2$)).
[87]	$P_{pv}=p_{pvref}·\frac{G}{G_{ref}}·{\eta }_{MPPT}$	The out power of PV system is still around maximum power
[187]	$P_{pv}=N_{pv}·{\eta }_{pv}·p_{stc}·\frac{G}{G_{ref}}·\left(1+\left(T_c-T_{stc}\right)·T_c\right)$
[188]	$P_{pv}\left(t\right)=N_{\mathit{pv}}·p_{\mathit{pvref}}·\left(\frac{G}{G_{\mathit{ref}}}\right)·\left[1+k·\left(T_c-T_{cref}\right)\right]$
[82]	$P_{pv}\left(t\right)=N_{pv}·p_{pv-ref}·{\eta }_{pvconv}·\left(\frac{G}{G_{ref}}\right)·k·(1+\left(T_c-T_{rcef}\right)$	In this model, the extracted power of PV is depended on sunlight intensity and temperature.

. Using PV models, FRT techniques could be enhanced by simulating the effect of devices under various conditions.

3.1. Conversion System of PV System

The conversion system of grid-connected PV systems has increased interest which can be either a one-stage or a two-stage system [189]. In single-stage, PV arrays are linked directly to the DC side of the inverter which has several disadvantages as the inability to handle a small range of input voltage, poor power quality, and reduced power capacities. In two-stage systems, the PV array is associated with the inverter, and then, dc-ac inverter inverts the available DC power to AC power which is called the second stage [190]. In the past, this strategy can be implemented through transformer-less inverter. The inverter control methods can be current control or voltage control which is shown in Figure 6. Current control methods are the most utilized methods which are divided into linear and non-linear control methods [191]. A comparison between different control techniques is in

Table 10. Comparisons between different control techniques.

Ref	Current Control Techniques	Advantages	Disadvantages
[189,191,192]	Proportional integral (PI) implemented in dq frame	Good filtering. Easy hardware implementation. Easy current control technique. Good dynamic response.	Not provide harmonics compensation Not provide Steady-state error removal
	Proportional-resonant (PR) implemented in $\alpha \beta$ frame	High gain in resonance frequency. High dynamic response. Provide good removal in steady-state error Provide good harmonic compensation.	Hardware complexity
	Repetitive current (RC) implemented in $\alpha \beta$ frame	Good removal in steady-state error. High gain in resonance frequency Easy reference current tracking High order harmonics compensator	Slow dynamic response Low stability
	Dead-beat (DB)	Provide current regulation. High sensitivity. Robustness. High dynamic and fast response.	Implementation in high frequency DSP. High delay Only applicable to active filters.
	Hysteresis control	Good stability, High dynamic, provides good transient response Individual load parameters Robustness	Varying modulation frequency, High complexity controller, resonance is occurred in the grid voltage
[189,191]	Predictive control	Minimize forecast error Good precision control Low harmonic & noise Good dynamic response	Poor performance under variable parameters Hardware complexity High sampling rate
[189]	Sliding mode control	Simple controller, Robustness Good stability, High speed controller	Chattering Discontinuous control function
	Feedback linearization	Good stability, High accuracy Good dynamic response	Complex controller High computational process
	Back stepping control	Robustness, High stability Applied Easy control, High efficiency	Gains can be adjusted
	Fuzzy logic control	Global controller, High efficiency, Easy integration with conventional controllers, Noisy-friendly controller Low overshoot & oscillation, Fast convergence, and its Parameter not sensitive	Complex controller It depends on fuzzy rules
	Genetic algorithm	Provide good optimization Suitable for complex problems	High computational time

. See Figure 7 for various inverter control techniques.

The conversion system of PV units is divided into 4 schemes: (i) centralized inverters, (ii) string inverters, (iii) multi-string inverters, and (v) AC cell or AC module inverters [192].

3.2. Classification of FRT Methods for PV Power System

FRT of PV system can be divided into two groups modified control type and adding external devices as shown in Figure 8. MIC is the highly efficient FRT capability method followed by the FACTS and ESS methods. To guarantee grid support during FRT, the MIC method injects active and reactive power control into the inverter. A comparison between various FRT techniques is in

Table 11. Comparisons between various FRT techniques of PV system.

Ref	Technique	Advantages	Disadvantages
[193]	FCL	High ability to limit increasing AC current Increase the grid stability The complexity is medium	Excessive AC current can be highly restricted Combined with other methods to enhance the FRT capability of PV system
[194]	SDBR	Low maintenance and high reliability Comparatively fewer switches than FACTS Complexity is medium	Weak in reactive power control Can’t with stand voltage fluctuations The cost is medium
[195]	BES	Simplest protection device Protect the inverter from increasing voltage Long cycle life and -Low cost	It mixed with other techniques to enhance the FRT capability of PV system
[53]	SVC	Simple structure Bus voltage is controlled and TSC reduces the power losses.	Controlling reactive power Installation and maintenance costs are very high It has unstable voltage oscillations
[196]	STATCOM	Needs many switches has not ability to provide active power Control efficiently the reactive power Fast response during disturbances Decrease the negative sequence of the voltage Complexity is high	Reduce the negative voltage sequence. Comparatively less disturbances than SVC High cost
[197]	SSSC	Less complex and low cost Comparatively better device from cost and performance than TCSC
[198]	DVR	Inject reactive power provide voltage stability in weak system Complexity is high	Reactive control relies on the voltage As a result of the fast response, voltage oscillations are unstable—High cost
[199]	BESS	Capable of storing excess energy Reduces the amplitude of AC current Medium-complexity Approximately 0.2 s are required, and the system is ensured to be safe	Short life span fluctuation in DC parameters Short life cycle Need uniform maintenance High cost
[200]	SCESS	Limit overvoltage Provide injection reactive current Long cycle life and medium-complexity Faster and smooth response	The specific energy is relatively low Short term voltage stability and high cost
[201]	BFCL	Response is robust and No spikes voltage drop is low in PV side and grid side During and after a fault, it reduces stress on semiconductor devices	High cost
[201]	STFCL	Response is robust during the fault initiation and clearing time	Needs a special complex cooling system for the superconductors
[202]	MIC	Efficient to meets the FRT requirements No additional hardware—Less cost Complexity is low and most efficient method	Some power is lost during grid fault
[203]	PSO	Flexibility and simplicity High efficiency in MPPT Fast response during fault The complexity is medium	Some overshooting and oscillation is occurred Complex dynamic systems Operate in more instinctive way Low cost
[204]	FLC	Active and reactive current components are controlled Limits over-voltage and over-current during faults. Complexity is medium and low cost.	Complex dynamic systems Operate in more instinctive way
[201,204]	DCL	Does not need extra protective inverter and other devices	DC-link voltage and reactive current injection problems are not addressed
[205]	FLS	Complexity is low and low cost Maintains fault current limits of the inverter Enhances FRT capability without requiring external devices	DC-link voltage and reactive current injection problems are not addressed
[206]	ADL	Decreases increasing voltage during disturbance Controls the positive and negative sequence currents during fault conditions Complexity is low and the cost is medium	DC-link voltage fluctuation is occurred Redundant AC-current is addressed

.

4. Classification of FRT Methods for Hybrid PV-WT (PV-WT HRS)

PV-WT hybrid renewable system (HRS) can achieve complementary daily and annual power patterns in comparison to PV systems only or WT systems only. Several approaches enhance MG systems which are distributed generation (DG) limitations, modifying protection systems, and adding external devices.

Table 12. Several approaches for enhancing FRT of MG.

Ref	Description	Control Method	Conclusions
[207]	The PMSG and PV arrays are connected to the same DC-link capacitor through the turbine-side converter and PV side converter respectively. The DC-Link is associated with grid through GSC and a step-up transformer.	Adaptive direct output control	Regulate the GSC output current through saving the power imbalance between the DC and AC terminals. Smoothing the net generated power of the HRES.
[208]	Switched modulated power filter compensator (MPFC) is one D-FACTS devices	Effective solution for compensating reactive power and harmonic reduction. The MPFC relies on a combination of shunt capacitor bank along with tuned arm filter. MPFC’s shunt admittance can be adjusted using two complementary switching pulses, controlled by a dynamic multi-loop error-driven PID	Appropriate solution to overcome the shortcomings of the inverter and provide better performance especially with induction motor (IM) loads Maintains the voltage total harmonic distortion (THD) within the acceptable limits Additional costs will be added Withstand longer fault durations. Resolves some poor power quality issues and gives flexibility in operation
[49]	Hybrid PV/WT using DVR	Adding device	Effective techniques to restore the voltage during disturbance
[209,210]	PV/WT based on SC energy-storage device	Adding device	Decreases power fluctuations under variations in wind speed and PV irradiance SC provides a fast response for power regulation
[123]	Hybrid PV/wind using STATCOM	Adding device	After the voltage is recovered, it provides more deaccelerated torque, which increases the grid stability

shows several approaches for enhancing the FRT of MG. The classification of voltage control methods of hybrid PV-WT to improve FRT is depicted in Figure 9. Control strategies are needed for controlling the operation of grid-connected MGs in renewable power systems [which are shown in Figure 10.

5. Summary and Discussions

Here, the overall operation and the schematic diagram of the DFIG-WT with FRT methods are discussed. A comparison of different power coefficient models is presented. Different functions are proposed for power coefficient, where the sin function obtains the maximum power coefficient but the exponential function effectively describes the WT model with a practical speed.

Several external devices are incorporated into DFIG-WT to enhance the FRT of DFIG. Various solutions such as protection devices, reactive power injection devices, and energy storage are discussed. This review helps WT constructors to choose the most suitable method, but there is a standing handicap of cost.

Due to the demerits of hardware devices, control techniques have obtained more attention. They decrease the cost of hardware devices and improve the FRT of DFIG. They are favored in wind farms under construction by helping to overcome disadvantages. They have a low cost. Control techniques are very effective in light voltage drops but in severe voltage dips, hardware devices should be added.

DFIG is frequently used with WT due to its variable speeds. Its converter can operate at 25–30% of its rated power. VOC, FOC, and DTC are different approaches to modeling RSC and GSC. VOC is the most popular technique because it provides good performance at steady-state operation. DTC performance has a fast response and produces ripples in the current. FOC reduces the damping of the DFIG-WT system.

Crowbar(R-impedance) protection enhances FRT by shorting the rotor through resistance. It is installed between the slip rings between the RSC and the winding rotor. Generally, this strategy is easy to implement because its components are controlled and depend on an ON-OFF switch that can be found at a low cost but, if it is removed later, more reactive power will be absorbed. It cannot generate reactive power for the grid when it is active. R-crowbar is active, DFIG excitation control is lost as RSC is disconnected and magnetization current is obtained from the grid which converted DFIG to a squirrel cage induction generator.

Adding a series (R-L) branch to R-crowbar is maintained DFIG excitation control, which obtains magnetization current from stator winding but osculation in rotor current remains. This technique maintains excitation voltage before and after the fault because it operates as a voltage divider between the RSC and rotor winding. It does not affect transient over-voltage. The R-L branch contributes to smoothing the transition period between initial fault and correction, thereby increasing the lifespan of the RSC and the dc-link.

The crowbar technique integrated with the (R-L) branch, integrated with the DBR, and integrated with the dc chopper all is contributed to decreasing the operational time of the crowbar and reducing the risk of turning DFIG into SCIG. Therefore, the hybrid techniques can improve the transient response of DFIG-WT, but they increase the system complexity, cost, and decrease the reliability of the power system. Crowbar with BES is the suitable solution to enhance FRT as ESDs can store access power in the dc link and provide the system with reactive power.

Using SDR can weak the operation RSC as it avoids DC link overvoltage. FCL enhances FRT of DFIG through limiting fault current and back EMF. Enhancement material of SFCL increases the efficiency of FCL. CBFCL protects the DFIG-WT system from overvoltage and is modified by capacitive impedance to provide reactive power.

Reactive power injection devices are connected to PCC between DFIG-WT and the grid to inject the required reactive power during voltage dip to keep the nominal voltage at its nominal value. SVC has a superior improvement of the FRT capability of DFIG than STATCOM. They have a high cost and they don’t efficient only so ESDs are needed. DVR is a nonlinear device that protects sensitive loads during disturbances. An ESD is added to the inverter to produce the desired voltage.

6. Conclusions and Future Prospective

In this literature survey, the overall operation and the schematic diagrams of the DFIG-WT with FRT methods were addressed. Classifications of FRT of PV systems were presented. Robust control methods for controlling grid connected AC, DC and hybrid AC/DC MGs in power systems were also discussed. Various inverter control techniques of PV systems were reviewed. In addition, several FRT methods for hybrid PV-WT were reviewed. Future trends focus on using hybrid methods and testing them to determine their reliability on DFIG based wind turbines. Classifications of FRT of PV systems were characterized. A total of 210 reported articles were reviewed, and the most up-to-date articles published in the last five years found in the literature were discussed. This review article may be used as basis for those are interested in improving the capability of FRT methods. Various traditional and advanced control techniques to improve FRT were summarized and discussed, including protection devices, RPDs, and ESDs. Furthermore, classifications of FRT hardware methods for DFIG were presented, along with grid code requirements.