Review of FACTS technologies and applications for accelerating penetration of wind energies in the power system

Conventional energy resource problems made power system planners apply more renewable energy sources (RESs) in the power system. RESs have a significant impact on fulfilling the world's energy demand, protecting the environment at least in the exploration phase, and providing energy reliability as well as security for the new generation of power grids. In this regard, wind power generation is considered a promising solution. However, by increasing the share of wind power generation units in the power sector, some challenges, such as voltage problems, transient stability, increasing system losses, and line congestions can occur. Consequently, an alternative solution is required to accelerate wind power units' connections to the grid by preventing the problems caused by the penetration of wind power. In this regard, Flexible Alternating Current Transmission Systems (FACTS) devices can play an important role while they can improve power grid dynamic performance. This paper presents a review of the optimization techniques applied to accelerate the penetration of wind energy by exploiting FACTS devices. The topic identifies the objectives and optimization formulations, as well as schemes and models available in the published literature. This survey gives noteworthy insights to the researchers to cover available solutions in these regards. The literature in the field is summarized in different aspects, such as FACTS technology, objective functions, constraints, optimization methods, planning or operation purposes, and load flow types. At each aspect, a comparison is derived to recognize the most important issues. The result of the review is the finding of future work and contributions which can be done in the field of wind energy integration acceleration using FACTS devices and the output of the paper can be used as a guide for researchers to find a new path for research in this area.


| INTRODUCTION
Renewable energy sources (RESs), in particular wind energy, are becoming a vital source for power generation in the very early future due to their sustainable features. However, using these renewable energies has its own challenges in the power system. [1][2][3] The voltage regulation problem, transient stability, increasing system losses, and line congestions are some of the most problems with the high integration of RESs, which can be resolved before they become commonplace.
In this respect, the power transmission sector needs new strategies to confront upcoming challenges such as increasing the penetration of renewables and decommissioning thermal power plants, such as nuclear and geothermal. [4][5][6] Besides, the uncertainties issues in wind energies must be managed precisely to avoid the safety problem. Because of the features of the earth, the wind flow regularly shows various features from those in which the characteristic curve of the wind turbines had been designed. 7 Consequently, the grid operators have to deploy innovative solutions that allow them to overcome these challenges.
On this matter, flexible alternating current transmission systems (FACTS) devices can play a significant role to solve the future challenges facing the transmission grid. [8][9][10] These devices can be utilized for energy utilization, power quality improvement, and so on. [11][12][13][14] Additional applications including stability improvement can be attained through such devices. 15,16 The FACTS devices contain a mixture of conventional power system parts (such as transformers and capacitors) and power electronic elements with the ability to increase the AC grid controllability and power transfer ability. 17 Utilizing different combinations of FACTS devices in the network could be one possible cost-effective result for the upcoming problem of the power sector. Thus, selecting the most suitable FACTS devices for the network alongside the proper placement and operation can be a significant issue.
There are many studies on the FACTS technologies and applications in the classic power system. However, there is no specific review paper that collects utilization of FACTS technologies problems and solutions for accelerating the integration of wind energies. This paper reviews several models and methods, which are considered in the published literature. The paper's contribution is to collect papers related to the optimal planning and operation of FACTS devices to improve the dispersion of wind energies in the network. The difference between this review paper with other papers is a comprehensive exploration of the new method band approaches which have been used to resolve these challenges by considering their objective functions, constraints, and uncertainties and also can guide researchers to find which types of optimization model and software/solvers have been used for a specific problem in this area.
The rest of the study is structured as follows. In Section 2, different FACTS types are presented. Section 3 presents different wind power generations. The roles of FACTS devices in wind power integration are given in Section 4. Section 5 describes problem statements. Planning is given in Section 6. The operation of FACT is presented in Section 7. Constraints are described in Section 8. The optimization model, uncertainties, and load flow are discussed in detail in respectively. Finally, software, tested networks, and conclusions are given in Sections 12-14, respectively.

| DIFFERENT FACTS TYPES
To avoid the problems associated with increasing the penetration of wind power into the power grid, the AC grid controllability and power transfer capability have to be enhanced. In this matter, utilizing the different FACTS devices in the power network not only improves the system voltage profile, transient stability, loss reduction, and so forth but also minimizes line overloading problems. In accordance with the connection of FACTS devices to the transmission line, they can be classified into four groups: series, shunt, series-series, and shunt-series controllers.

| Series controllers
The two most common series-connected FACTS devices are the thyristor-controlled series compensator (TCSC) and the static synchronous series compensator (SSSC) as shown in Figure 1. The TCSC which is a series FACTS device in which a capacitor is placed in series with the transmission line and a thyristor-controlled reactor (TCR) is connected in parallel with the capacitor. By regulating the thyristors firing angle, the capacitive reactance is varied over a wide range, therefore line reactance gets modified. The TCSC can operate in capacitive or inductive states based on the firing angle. 18 The SSSC consists of a VSC linked to the transmission line by a series coupling transformer. 19

| Shunt controllers
The two most used FACTS devices are the static VAr compensator (SVC) and the static synchronous compensator (STATCOM) as shown in Figure 2. The SVC includes a bank of TCR in parallel with a bank of thyristor-switched capacitors (TSC). The TCR uses VAr from the system to lessen the voltage if the reactive load is capacitive. Alternatively, if the reactive load of the system is inductive, the TSC banks are switched in to inject reactive power, which in turn raises the system voltage. The thyristor pairs operate as switches which permit the SVC to respond rapidly to network fluctuations. 20

| Series-series controllers
These exist as a mixture of series devices; the series device affords series compensation for the line and transfers active power between the transmission lines as well. The most used device is the interline power flow controller (IPFC) as given in Figure 3. 21

| Series-shunt controllers
These exist as a mixture of series and shunt devices together, whereas the shunt part injects current into the network. On the other hand, the series part injects voltage in series in the line. In addition to the advantage of FACTS devices, their cost is also an important parameter for the utility's leadership. The most used device is the unified power flow controller (UPFC) as depicted in Figure 4. 22 In this matter, having an idea regarding the cost of FACTS devices can help the designer to find an optimal solution The cost of FACTS devices can be divided into two categories: device investment cost, operation, and maintenance costs. Device investment costs depend not only on the installation rating but also depend on special requirements (e.g., communication with the regional control center, control and protection system, etc.). However, the maintenance cost of the FACTS device is depending on the installed device size and local ambient pollution. Moreover, the operation cost of FACTS devices is negligible as the changing of set points or operation modes can be done locally or remotely. Table 1 provides an idea regarding the average cost of some of the FACTS  devices. 23 Figure 5 demonstrates the different categories of FACTS devices as well as their generations.

| DIFFERENT WIND POWER GENERATION
A wind turbine contains the main tower, blades, and nacelle as shown in Figure 6. 24 In the following different types of wind turbines are presented.

| Types of wind turbines
There exist two basic types of wind turbines:

| Horizontal axis wind turbine
All the components including the blades, shaft, and generator are mounted on the upper part of a tower as demonstrated in Figure 6, and the blades face the wind. In the study by Zaki et al., 25 a horizontal axis wind turbine is utilized to present the study considering the shape of small wind turbines.

| Vertical axis wind turbines
The vertical axis is displayed in Figure 7. The wind turbine is near the ground, different from the horizontal where everything is on a tower. 24 In the study by Li

| Asynchronous generators
Induction generators include fixed-speed induction generators (FSIGs) with squirrel cage rotors (SCIGs) and doubly-fed induction generators (DFIGs) with wound rotors. In the study by Benbouhenni and Bizon, 27 a commanding method based on the direct field-oriented control technique for asynchronous generators is offered. SCIGs generators were used in early fixed-speed wind turbine designs by Yasmeena and Das 28 and Sule et al. 29 The squirrel cage induction generator and geared DFIG generator are shown in Figures 8 and 9. DFIG is basically a standard, wound rotor induction machine with its stator windings directly connected to the grid and its rotor windings connected to the grid by a converter. 28 Modeling and control of stator active and reactive power for DFIG is presented in the study by Kelkoul and Boumediene. 30 In addition, the mentioned control system has been compared with super-twisting algorithm.

| Synchronous generators
Coupling a synchronous generator to the grid is done through full-size power converters. The generator is either connected directly to the turbine rotor operating at low rotating speeds or uses a single-stage gearbox and operates at medium speeds. 31 A deadbeat predictive control strategy for this wind application is offered by Abdelrahem et al. 32 Moreover, an extended Kalman filter is used to enhance the robustness of the deadbeat predictive control strategy, observe the rotor speed and position of the permanentmagnet synchronous generator, and reduce the harmonic distortion in the stator current. A typical direct drive synchronous generator is shown in Figure 10.

| ROLES OF FACTS DEVICES IN WIND POWER INTEGRATION
To overcome the impacts caused by wind speed fluctuations, FACTS devices should be used. 33 Series devices can be utilized on the transmission system while shunt devices can be implemented either in the transmission  network, generation side, or load buses. Different roles of FACTS devices in wind power integration include: Reactive power compensation: FACTS devices can be connected at the point of common connection (PCC). They can also be used to alleviate voltage variation of the wind generator. 34 Power system stability improvement: Wind power plants are placed in remote areas and power is evacuated using long transmission lines. Shunt FACTS devices applied at the PCC of such a plant greatly enhance system voltage stability. 35 Voltage ride-through capability: Wind turbine generations (WTGs) were usually disconnected from the grid in fault conditions, to avoid any possible damage to wind turbines. With increased penetration of wind power, sudden disconnection of WTGs could result in poor grid stability. FACTS devices can be utilized to guarantee reliable power delivery to the grid during abnormal operating situations. 36 Power flow control capability: With the power flow control ability presented by FACTS devices, transmission bottlenecks can be evaded by shifting the power from the congested lines to the underutilized lines nearby. 19 Power quality improvement: FACTS devices can be used to diminish power quality problems, such as interruptions, poor power factor as well as harmonics. 37

| PROBLEM STATEMENTS
By increasing and connecting wind penetration infrastructure to the grid, some troubles such as congestion of transmission lines, voltage problems, and increasing system losses will come out. 38 New FACTS technologies have the potential to control each power line dynamically based on the real-time needs of the grid. Therefore, utilizing various FACTS devices can help grid operators mitigate the aforementioned problems. In addition, by optimal placement and operation of the new FACTS devices into the grid, grid operators can lower the investment, which is required to accommodate higher penetration of wind energy and resolve the problems, which are faced due to these reasons. In addition, the time of bringing a wind project online in regions with weak transmission systems can be shortened. Therefore, to increase the level of wind power plants in the grid by utilizing various FACTS devices, constructing an accurate planning and operation optimization framework is becoming essential.

| PLANNING
A planning study aims to give an investment plan for a power system network covering certain planning over a planning horizon. 39 In this regard, the definition of objective functions, constraints, uncertainties, and algorithms is becoming crucial to make an appropriate planning decision.

| Decision variable
In the optimization model, decision variables are the variables that the decision-makers needed to decide about it. The decision variables considered in the published literature for the planning are illustrated in Table 2. Many authors, [40][41][42][43][44] have considered only the FACTS devices parameters such as locations and rates as the main decision variable for their studies.
However, in the studies by Ziaee et al., 49,50 authors rather than FACTS device's locations and rates, also have considered locations of the line switches as other parameters. Because the optimal placement of line switches is also vital to minimize the total cost of customer service outages as well as the investment cost of line switches.
Moreover, in the study by Shen et al., 51 researchers have considered both the transmission expansion planning (TEP) and FACTS devices locations and rates as their main decision variables. In the study by Mokhtari et al., 52 the performance of the TCSC as a FACTS device is investigated.
In this regard, determining whether and where to include FACTS devices among the possible reinforcement options in TEP processes become important. Also, the generated active/reactive power and the voltage magnitude and angles at each bus are the most common state variables considered by the researchers in the published literature.

| Objective functions of the planning problem
It is essential and vital that the designers search for an optimal solution that fits the objectives under a set of constraints. Various criteria are considered for optimal settlement and sizing of FACTS to increase the penetration of wind power in the network. Below we summarize the related research works and the objective functions in the published literature.

| Congestion alleviation
The congestion problem has to be minimized as much as probable, which has been investigated in several papers. 53,54 In this regard, as an important example, Reddy 47 GHADIMI ET AL. considered overload minimization as one of its objective functions by subtracting the power flow of the line and the maximum power flow that can flow in that line. A thyristor-controlled series electrical device using a particle swarm algorithm is used to alleviate the congestion transmission network in the study by Nireekshana et al. 55

| Transmission loss
Minimizing transmission losses and providing the bus voltages within satisfactory bounds are vital in the power system. Moreover, system loss minimization can lead to more profit. Reddy and colleagues 43,47 considered the transmission loss minimization as an objective function. In the study by Nadeem et al., 56 an optimal sizing/sitting of the FACTS device is presented.

| Wind curtailment cost
Due to weak grids, some of the wind sources with zeroemission and low cost have to be turned off for a while to avoid congestion in the grid. In this case, the output of renewable resources cannot be used and the result would be waste. Zhange and colleagues 40,42,44 considered wind energy curtailment cost minimization in their objective functions such that every unit of wind curtailment is attached to penalty cost. Consequently, wind curtailment would be minimized by curtailment cost reduction. 57

| Generation cost
Many researchers 41,42 have considered the minimization of generation costs as their objective function. In this matter, high priority is given to the generators with the lowest generation cost and consequently, more benefits can be obtained for consumers of electricity. On the other hand, Alhasawi and Milanovic 41 also applied the net present value as the objective function where the savings have to be maximized over the device's lifetime. The cost of generation and FACTS devices is considered as an objective function in the study by Pati and Karajgi, 58 also the optimal location and size of multiple FACTS are identified.

| Spinning reserve cost
The spinning reserve can be defined as the ability of the system to quickly increase in generation in case of loss of a major generator unit or transmission equipment. 59 Therefore, the amount and location of the spinning reserve find important to decrease the system cost.
Since FACTS devices can control power flows across a control area interface. So, the mentioned feature can greatly optimize the amount and location of the spinning reserve. Consequently, increasing the utilization of lowcost generation units causes a decrease in spinning reserve cost. In this regard, Sang and Sahraei-Ardakani 42 has considered a spinning reserve cost minimization as one of its objective functions.

| Load shedding cost
To meet load requirements, the system has to be adequately designed to avoid load shedding in normal and contingency conditions. 60 Zhang and colleagues 40,50 considered the minimization of load shedding in their objective functions by assigning of cost coefficient for lost load in the system. In the study by Taher et al., 61 load shedding reduction and load ability enhancement of the power system using shunt FACTS devices is established. However, it is worth mentioning that the load-shedding cost is highly dependent on the importance of the load.

| Power loss cost
Increasing the penetration of wind energies in the power system can change the total power system loss. 62,63 Therefore, to reduce power loss, the FACTS devices can be potentially utilized. Xu et al. 46 exploited the power loss cost minimization in its objective function. This has been performed by assigning the coefficient associated with electricity price. 64

| New line investment cost
Adding new transmission lines to the existing lines has been at all times a strategic decision. By increasing the penetration of wind energies in the system, the complexity of this strategic decision has increased significantly. 65,66 In this regard, Shen et al. 51 considered two planning problems (e.g., TEP and FACTS device optimal placement problems) in a single problem. To solve the problem of correlation between wind and demand scenarios, this study exploited both planning problems. Moreover, a probabilistic TEP model with FACTS aims to improve network flexibility is established by Yuan et al. 67 6.2.9 | FACTS costs Many papers seek to minimize the FACTS investment costs in their objective function by assigning the cost coefficients for the candidate FACTS devices in the model. Enhancing of system parameters considering FACTS devices cost in a 24-bus power system including a wind farm research data is presented by Sharbaf and Shojaei. 68 On the other hand, Shen et al. 51 considered the operation cost of these devices as well. However, Ziaee and Choobineh 49 also considered the cost of maintaining FACT devices.

| OPERATION
Flexible power flow control provided by the FACTS devices can offer an opportunity for better exploitation of the existing transmission system. By increasing the amount of wind penetration in the grid, these devices can play a significant role to deliver power to the load without compromising the security and reliability of the grid. Therefore, to fully utilize the existing transmission system, an accurate control scheme has to be established for the optimal setting of these devices.

| Objective functions of the operation problem
By growing the number of renewable sources in the grid, the operating challenges of the power system have been increased. Therefore, finding a proper way to better utilization of the existing power system gained importance. In this regard, modern technologies like FACTS devices can provide technical solutions to address the new operating challenges. Below summarizes the objective functions of related research works in the published academic literature. It has been also demonstrated in Table 3.

| Capacity margin
Some FACTS devices like TCSC can control the power flow by altering the reactance of the existing lines. 74,75 In this matter, Yang and Hug-Glanzmann 69 considers the maximization of the minimum value of the capacity margin of transmission lines as an objective function. The stability margin in the problem of FACTS devices location using artificial bee colony is presented by Kamarposhti et al. 76 Also, the results of an artificial bee colony are compared with genetic and particle swarm algorithms. In this matter, the FACTS devices are trying to push away flows from their capacity limit and pull in more flow to the lines which has an adequate capacity limit.

| Generation cost
Sang and colleagues [70][71][72] have considered the minimization of system operating costs as an objective function. In this respect, by giving the priority to the generators with the lowest generation costs, more benefits can be obtained for the consumer of electricity.

| Wind curtailment
Sang and colleagues 70,71 considered minimizing wind curtailment cost as one of their objective functions. However, Nasri et al. 73 only considered the minimization of wind curtailment uninterested in its cost in the objective function.

| Reserve energy deployment cost
To avoid load shedding, the system operators keep in service large amounts of generations in reserve to conserve the network for an event, such as loss of generators, sudden load changes, and so forth. 59 Nonetheless, holding the generators in reserve conditions is costly. Therefore, the amount and location of the spinning reserve find it important to decrease the system cost. In this regard, Sang et al. 70 has considered a reserve energy deployment as one of its objective functions.

| Power loss
By increasing the penetration level of wind farms into the grid, the changes happen in the active power loss of the grid. 62,63,77 In this matter, Ref 78 considered the minimization of power loss as one of its objective functions uninterested in its cost in the objective function.

| Unserved load cost
In the study by Li et al., 71 authors have used the minimization of unserved load cost as an objective function, on the other hand, in the study by Nasri et al., 73 researchers have considered only the unserved load minimization regardless of its cost in the objective function.

| CONSTRAINTS
The constraints have been considered in objective functions optimization to guarantee that the design conditions are within the limits during the finding of location, size as well as operation of FACTS devices to improve penetration of wind energy. Several power system constraints and utility capacity limitations have been considered by the researchers. The power balance, bus voltage, transmission line flow limits, FACTS limits, generators limits and transmission line loss constraints are the most common constraints considered by the researchers in the published literature.

| Load flow balance
The total loads plus the losses in the system must be equal to productions from conventional and renewables units. This constraint has been considered in almost all the studies.

| Voltage profile and angle
The voltage limitations have been considered in the majority of studies. This constraint restricts the bus voltages of all buses in the system to the upper and lower acceptable levels. An improvement of voltage profile stability of the Iraqi power system using the optimal value of FACTS devices is established by Salman et al. 79

| Line thermal limits
The maximum capacity of each individual transmission line cannot be higher than the defined maximum power that can flow through the lines. Adewolu and Saha 80 investigates the performance of FACTS according to various approaches of sitting for existing transfer ability enhancement considering line thermal limitation.

| FACTS number and size
The number of FACTS devices cannot exceed the maximum number of FACTS allowed to be installed. This limitation is referred to as the number of FACTS devices that can be installed in the planning studies. Moreover, the FACTS size limitation is also considered in the planning studies. FACTS devices are optimally sized and placed in the transmission system to enhance voltage stability using different optimization algorithms by Rajasekaran and Muralidharan. 81

| OPTIMIZATION MODEL
The planning and operation of FACTS devices to allow the penetration of more wind energies in a grid is presented before. This section will discuss the constructed optimization models in the published literature. The optimization models and evolutionary approaches used in the literature can be classified into the following groups: genetic algorithm (GA)-based optimization programming, differential evolution (DE) algorithm, mixedinteger linear program, particle swarm optimization (PSO), opposition-based bacteria dynamic algorithm, and mixed-integer nonlinear optimization.

| Mixed-integer linear programming
If some of the decision variables in the optimization problem are discrete, then an optimization model becomes a mixed-integer program. Zhang and colleagues 40,42,46,49,50,70 has formulated its optimization problem as a mixed-integer linear programming (MILP). Different optimization approaches can exploit the MILP. Two examples come as follows; the framework in Zhang et al. 40 is constructed using bilevel optimization to include the market-clearing condition in its optimization model. In this study, the upper levels seek to find the optimal location of the FACTS devices, however, the lower level aims to find an amount of load shedding as well as wind curtailments for different load-wind scenarios. Besides, Ziaee and Choobineh 49 has used two stages of stochastic programming to model the planning and operation problem. In this regard, the decision related to investing in FACTS as well as switching devices is considered as a first stage. On the other hand, the status of the switch and the compensation level of the FACTS is considered the second stage in this study.

| Nonlinear programming
When constraint equations or objective function equations are nonlinear, the optimization problem becomes an NLP problem. Pati and colleagues 43,48,72,73 formulated their optimization problems as an NLP problem. In nonlinear programming, the mixed-integer approach is a method that briefly is discussed here. mixed-integer nonlinear programming (MINLP) is the area of optimization that addresses nonlinear problems with continuous and integer variables.
Li et al. 71 shows the formulation for this type of optimization problem. However, formulating a nonlinear problem so that a solver will be able to compute a good solution and do quickly is important. A good formulation for an NLP problem typically involves several things, including identifying functional initial values, adjusting variable bounds, and scaling variables and equations. Various factors to consider are techniques for blocking degenerate cycling and the possible benefits of preventing expressions in nonlinear functions. Although this approach may have some problems such as convergence time and inaccuracy but solving this problem through a highcapacity processor and probably using artificial intelligence and GA can be considered as an open issue for future research plans. Reddy 47 has used an opposition-based bacterial dynamics algorithm (OBDA) for solving the congestion management approach based on the optimal power flow (OPF) problem in a dynamic situation. The OBDA uses a primary (associated) and two secondary bacteria (opposite associated). By randomly initializing the primary bacterium and generating an associated bacterium by moving the primary to one dimension among the upper/lower bands. Furthermore, creating the opposite one via moving the primary into an identical dimension but in the contrary direction. The primary bacterium accelerates to a better fitness position between two secondary bacteria.

| Particle swarm
The main advantage of the PSO algorithm with respect to other optimization algorithms is its inherent simplicity. It is a derivative-free algorithm and does not require gradient information to converge to a solution. These features of the PSO algorithm give the flexibility to deal with different objective functions. In the study by Prasanthi and Sindhu, 45 an OPF has been performed by utilizing the PSO algorithm to find the locations and rates of FACTS devices simultaneously.

| Genetic algorithms
To solve optimization problems in power system studies, one of the powerful numerical algorithms that have been used in the literature is GA which are suitable for complex nonlinear models where finding the location of the global optimum is a difficult job. GA method is used to consider problems that cannot be modeled accurately. Alhasawi and Milanovic 41 employed OPF and GA to solve the problem of optimal placement of the FACTS devices. Also, Teh et al. 44 considers various objective functions (e.g., wind curtailment cost, total social cost, and SVC operation cost) by using a multi-objective optimization algorithm. In this research paper, a determination of the final optimal solution is performed by the fuzzy decision-making method.

| DE algorithm
One type of evolutionary algorithm for optimization problems is DE. Shen et al. 51 utilized a DE algorithm to handle its optimization model.

| UNCERTAINTIES
The production of wind energies depends on many factors such as time and weather, and as a result, follows a stochastic principle. Therefore, it needs a statistical analysis in which production estimates should be linked with the probability of occurrence. Several methods are utilized in the studies to model the wind speed, such as Weibull distribution, the autoregressive moving average (ARMA) model, and so forth. In various studies, 46,47,53,72 authors have utilized the Weibull distribution as an approximation of wind speed distribution. The Weibull distribution is often used to characterize wind regimes because it has been found to provide a good fit with measured wind data. However, in the study by Teh et al. 44  A fuzzy logic technology and numerous approaches are combined for various applications because of its capability to fine-tuning of control variables. 82,83 However, in the paper, 84 fuzzy controllers are designed for FACTS devices for the improvement of power quality.
Another research work is related to the neuro-fuzzy controller (NFC) which has a faster response than conventional controllers. Neuro-fuzzy control is emerging as a complement to conventional controls because it can deal with the uncertainties in the power system effectively.
The advantage of NFC over a conventional controller is that they do not need a perfect mathematical model. NFC functions with imprecise inputs, handles nonlinearity, imitates the human decision-making process, and can be implemented with more success in complex systems than conventional controllers. 85,86 In addition, considering the production uncertainties of wind power is necessary for optimal planning and operation of the FACTS devices in the network.
There are many papers that researched uncertainties in the estimation of wind energy production. But the uncertainties considered in the published literature for the planning and operation of FACTS devices to increase the penetration of wind energies to the power systems are limited.
There are several methods, such as Monte Carlo simulation (MCS), robust, point estimation method (PEM), and scenario-based optimization method that can be used to recognize the uncertainty in the forecasting models. But, many of the studies in the published literature have utilized the scenario-based method to consider the uncertainties.

| LOAD FLOW TYPES
Load flow studies are essential for future planning as well as for determining the system's operation. In the academic literature, both the AC and DC load flows have been utilized. But, the calculations by Sang and colleagues, 42,49,50 were based on DC OPF mode in which reactive power is ignored.
Another main solution for load flow analysis which is described in the literature review is fuzzy logic-based power flow. In this method, the real and reactive power  mismatch are the unknown variables. However, the voltage magnitudes and angles are the output of the system. It demonstrates that the fuzzy logic-based power flow solution is much faster, and simple to be analyzed than the conventional methods of power flow solutions. 87,88 12 | SOFTWARE The software as well as the solvers used in the literature can be found also in Tables 4 and 5. It should be noted here that the utilized optimization solvers in the published literature were YALMIP and CPLEX in the MATLAB environment. 89,90 The TOMLAB/SOL environment in MATLAB is also utilized in the academic literature as well. Moreover, the general algebraic modeling system (GAMS) is used as an optimization modeling programming tool that is linked with many different solvers, such as CPLEX, CONOPT, and DICOPT. 91 13 | TEST NETWORK In the end, a few authors have implemented their models to the IEEE-one area, 6-bus system, and regional 86-bus system, respectively.

| CONCLUSIONS AND FUTURE DIRECTIONS
This paper presents the general background, objectives, constraints, algorithms, and uncertainties for optimal placement, sizing, and operation of FACTS devices to increase the penetration of the wind power sources in the power network. In addition, this paper also investigates load flow types, software, and test network in the literature. As a matter of fact, the design and deployment of a specific optimization method to improve the penetration of wind energies is a challenging problem and cannot be easily considered only one aspect of the system. Therefore, different types of objectives, constraints, and uncertainties that have been used to solve these issues are collected in this paper. In this respect, these optimization solutions are expected to be a viable alternative, which can properly take into account all the valuable information and parameters of the problem. For planning and operation, various methods including analytical and meta-heuristic approaches are reviewed in detail. For integrating more wind power systems into the power system, the use of FACTS devices is necessary. In this regard, the UPFC and IPFC can provide appropriate control reactive and real power with bus voltage. The SVC and STATCOM are effective in reactive power compensation. TCSC has a suitable ability to vary the impedance of the transmission line while the SSSC controller provides active power control. For uncertainty modeling also several approaches are reviewed in detail. Among various methods, scenariobased stochastic programming is utilized to address decision-making problems under uncertainty in which unknown parameters are signified through a finite set of scenarios. However, the drawback of stochastic programming is that its performance depends on the information on the probability distribution function of the uncertain data. Furthermore, the probability of the problem is only assured for the input scenarios considered. Consequently, an adequate number of scenarios are needed. Nevertheless, a large number of scenarios can lead to a computational challenge. Another approach to cope with decision-making challenges under uncertainty is robust optimization. It is modeled using decision variables within a pre-defined uncertainty set. However, robust optimization is a very complex method and multi-stage process that could provoke intractability issues. The MCS is very accurate but faces a computational burden. The use of the PEM increases the speed of the running, but it is not very precise.
This survey remarks that the obtained results which are achieved in the literature, are highly dependent on the objective functions considered (i.e., by changing the objective functions, the outcomes would be completely different), so defining the appropriate objective function is inevitable. Moreover, the survey of the literature also mentions that the results of the previous studies are highly dependent on the types of considered scenarios in the tested network.
Specific issues for future works include: (a) utilizing the latest techniques to model the renewable energy uncertainties; (b) developing an evaluation index according to different features of the FACTS devices; (c) developing the generalized model for planning and also the operation of FACTs devices that can be applied to any system involving high penetration of wind energies; and (d) investigation of the influence of transmission