Optimal design of SSSC and crowbar parameters for performance enhancement of Egyptian Zafrana wind farm

Zafarana Wind Park is considered the largest wind farm in Middle East. By the year 2022, Zafarana wind farm will inject 545 MW to Egypt national electric network (ENEN). This wind farm has been connecting to ENEN in stages (eight stages) from ten years ago. Each stage represents a project of wind farm. Hence, the impact of the emergency condition of ENEN on the performance of the installed stage of Zafarana Wind Park should be studied. Moreover, determining the possible method to reduce the side effects of the emergency condition of ENEN on the installed stage of Zafarana Wind Park becomes an important issue. In this paper, a combination between Static Synchronous Series Compensator (SSSC) and crowbar (Cro-SSSC) is used to enhance the performance of the one installed stage of Zafarana Wind Park. The studied stage of Zafarana Wind Park is Zafarana Z1 wind farm (Z1). Zafarana Z1 wind farm represents that first stage of Zafarana Wind Park that was connected to ENEN. Hence, the paper takes Zafarana Z1 wind farm as the studied case to examine the impact of Cro-SSSC. The crowbar is used to protect the power components of the SSSC from the high current during grid fault condition. Also, in this paper multi-objective lightning attachment procedure optimization algorithm (LAPO) and the sine cosine algorithm (SCA) are used to evaluate and tune the control gains of SSSC and the value of crowbar's resistance. This step is important in order to determine the values of the control's gains of SSSC and the value of crowbar resistance which suit the studied system (one installed stage of Zafarana Wind Park). The sliding mode control (SMC) is used as an example of nonlinear control of SSSC. Also, a hybrid of LAPO and SMC is used to improve the performance of Zafarana Z1 during faults. The results of LAPO, SCA and SMC are compared during three-phase fault and single line to ground fault applied to the studied system.


General introduction
At the beginning of this century, Egypt intended to the expansion of the establishment of industrial cities. This has led to an increase in the demand of electrical energy needed for these industrial cities. The main challenge which faces the Egyptian government is how to increase the electric energy production without increase the fuel cost. The expanding in generating electric energy from renewable energy is the best chose to face this challenge. The geographical location of Egypt has granted Egypt advantages such as solar radiation reaches 24 MJ/m 2 [1,2], and wind speed reaches 15 m/s at coastal areas [3]. This engorges Egyptian government to develop plans for the establishment of renewable energy plants. It is known that the generating electricity from wind energy is lower cost than the generating electricity from solar energy. Hence, Egyptian government starts their steps by installing wind farms, especially in Red Sea coast. Hence, this paper focuses on wind farms which have been installed and interconnected with the Egypt national electric network (ENEN).

Literature survey
There are many articles that studied the issue of enhancing the relation between wind farms and their interconnected grids. These efforts revolve about using flexible AC transmission system (FACTS) to improve the impact of wind farms interconnected grid or using artificial intelligence methods to improve the performance of FACTS integrated with wind farms. In this paper, literature review will focus on the efforts which have been achieved to improve the performance of wind farms based on SCIG (SCIG-WF). A proposed control method based on hierarchal PI-fuzzy-PI controller was discussed in [4] to improve the frequency control of SCIG-WF during an unbalance between load demand and power generated. In Ref. [5], Static Synchronous Compensator (STATCOM) was used to regulate the voltage of SCIG-WF during fault condition. Enhancing the transient stability of SCIG-WF by improving the control system of STAT-COM was introduced in [6]. The artificial neural network (ANN) was proposed to find the optimal value of STATCOM controller in [7]. The optimal place of integrated unified power quality conditioner (i-UPQC) was investigated in [8] to improve the power quality of distribution system interconnected to SCIG-WF. Static Synchronous Series Compensator (SSSC) based on ANN was used to improve different type of wind turbines including SCIG wind farms in [9]. Low Voltage Ride Through (LVRT) capability of SCIG-WF was improved using unified inter-phase power controller (UIPC) in [10]. In Ref. [11], Center-node Unified Power Flow Controller was used to enhance different types of wind farms even SCIG-WF. The crowbar was used to improve the relationship between wind farms and interconnected grid. Most of these articles were based on using crowbar as inner protection to stator of doubly fed induction generators (DFIG) such as in [12][13][14]. In Ref. [15], crower is used as outer protection to improve the impact of wind farm base on SCIG and DFIG. The performance and load flow study of Zafarana wind farm was presented in [16]. The information and communication technology (ICT)-based ANN was investigated in [17] to improve the control system of Zafarana wind farm. Improving the performance of Zafarana wind farm using STATCOM was investigated in [18]. The impact of Harmonic resonance due to overvoltage on the transformers connected the Zafarana wind farm to ENEN was studied in [19].
LAPO algorithm is an efficient physical-based technique which conceptualizes from behavior of Lightening in natural [20,21]. LAPO algorithm was successfully implemented for solving many optimization problems where in [22] LAPO has been used for Optimal Power Flow problem (OPF) solution with STATCOM. The authors in [22] implemented the LAPO algorithm for the wrapper feature selection optimizations. In [22], LAPO algorithm was utilized to determine the most fitted parameters of the UPFC device. In [23], the OPF problem has been solved by the LAPO.
Sine cosine algorithm (SCA) technique is a new population-based optimization technique proposed by Mirjalili [24]. The SCA is also applied for solving many optimization problems where the authors In [25], solved optimization the bend photonics crystal. Attia et al. [26] implemented the SCA for OPF solution. The authors in [27] solved the optimal reactive power dispatch problem with inclusion of series and shunt controller using the SCA.
SMC features with its ability to control the performance of systems without concerning the effect of uncertainty factors. The effect of the phenomena of chattering represents the major disadvantage of this type of control. There are some articles which deal with this issue to overcome the effect of these phenomena such as [28]. In [29], SMC with adjusting output feedback was used to control the power and speed of wind turbines. A hybrid of fuzzy logic control and secondorder SMC was applied to unified power flow controller in [30]. In [31], a developed method of SMC was proposed to reduce the phenomena of chattering and investigate the optimal power extracted from wind.

Contribution of paper
In this paper, the unsatisfactory performance of Zafarana wind farm especially disconnection from ENEN during fault condition is handled using proposed Cro-SSSC. The Cro-SSSC is composed of a combination of crowbar and SSSC. The crowbar is combined with SSSC to protect the SSSC from being damaged due to high current during fault where SSSC is connected in series with the transmission line. In this paper, LAPO and SCA algorithms are used to determine the appropriate values of the crowbar resistance and control parameters of SSSC. These values of the crowbar resistance and control parameters of SSSC are determined to fit with studied system of Zafarana1 wind farm. Also, SMC is proposed to control the SSSC based on the values of the control parameters of SSSC obtained by optimization. The performance of Zafarana1 wind farm with Cro-SSSC when it is tuned by LAPO, SCA and SMC is compared with Zafarana1 wind farm Z1 without Cro-SSSC during three phase and single line to ground faults.  The main feature of Zafarana Wind Park is its construction of several wind farms. This feature facilitates studying the performance of each wind farm separately. In this paper, Zafarana Z1 wind farm has been chosen to represent the studied case. The encouraging reasons for choosing Z1can be summarized as follows: • Zafarana Z1 wind farm is the first stage of Zafarana Wind Park which is already connected to ENEN • Zafarana Z1 wind farm is based on asynchronous generator. This type faces many difficulties to remain connected to the grid during abnormal operation condition of the grid.
Zafarana Z1 wind farm is composed of 50 Nordex-N43 wind turbines. Each wind turbine produces 0.6 MW, 0.690 KV, 50 Hz at 13 m/s rated wind speed. The separated distance between wind turbines is 60 m, and the separated distance between rows is 800 m. 50 Nordex-N43 wind turbines are connected to three feeders (J6, J7 and J10) through 800 MVA, 0.69/22 KV coupling transformers. J6 and J7 contain 17 wind turbines, while J10 contains 16 wind turbines. The three feeders are connected to bus coupler to inject the produced power to a 220 KV transmission system through 75 MVA, 22/220 KV transformer [16][17][18][19]. Figure 2 shows the single line diagram (SLD) of Z1 wind farm.
This SCIG equivalent circuit can be used to write the direct and quadratic (d-q) representation of SCIG with respect to the synchronous as follows [34][35][36]: The electrical torque of SCIG and DFIG in terms of d-q components of current and flux of the stator and rotor can be obtained as follows [34,[37][38][39][40][41]: Equation (4) shows three expressions for electrical torque. The three expressions can be used depending on the objective or the aim. Because third expression depends on the d-q components of stator and rotor currents, the third expression is effective used in designing the control system of wind turbines.

Static Synchronous Series Compensator (SSSC)
The Static Synchronous Series Compensator (SSSC) is the most famous member of series FACTS devices family. The function of SSSC is based on injecting a voltage in series with the transmission line and quadratic to the line current.
The injected voltage has the ability to achieve two functions: enhancing the power flow by changing the overall impedance of the transmission line and regulating the voltage by controlling the reactive power at the connection point [42]. Figure 4 shows the equivalent circuit and The SLD of SSSC. From the SLD of SSSC, the SSSC can be connected to any one of the two buses of transmission line. One of the two buses is the sending bus (m), and the other bus is the receiving end bus (n). From the equivalent circuit of SSSC, the current equation can be written as follows [42]: where YL 1/ZL and represents the admittance of the transmission line. Vm and Im the voltage and current of bus m. Vn and In the voltage and current of bus n. The active power and reactive power at the receiving end bus n are given by [43]:    Fig. 5 The SLD of SSSC control system

SSSC control system
5 Lightning attachment procedure optimization algorithm LAPO algorithm is one of the recent physical based algorithms. LAPO algorithm mimics the lightning phenomena in nature where a huge amount of electric charges are accumulated inside the cloud which leads to increasing the electrical potential, and then the lightning will strike ground throughout several downward and upward leaders. Lightning attachment is based on four stages which are: (1) Air breakdown (2) Downward movement of Lightning channels (3) Upward leader motion, and (4) Strike point. The steps of the LAPO algorithm can be summarized as follows: Step 1: Initialization The initial points of lightning downward leaders (trial spots) are generated randomly as follows: where X i in is the trial point. X min is the minimum limit of X , and X max is its maximum value. rand denotes a random value within [0,1]. The fitness functions of the initial points can be found as follows: Step 2: The next jumping All locations of the obtained points are averaged, while their objective functions are given as follows: F avr obj(X avr ) X avr is the averaged point. F avr represents the objective function of the averaged point. The locations of current points are updated based on the average point where to update the point i, a random solution j is selected where i j. Then the obtained solution is compared with the selected solution as follows: (17) Step 3: Section Diminishing The updated solution will remain continuous if its value is better than the previous value.
Generated branch will remain continuous if the critical value is less than the electric field of the new test point; otherwise it will fade, and the section diminishing is expressed as follows: Step 4: Upward Leader Movement This step defines the movement of upward leader. The updated point based on movement of upward leader is given by where X i best represents the best solution, while X i wor st represents the worse solutions. S is an exponent operator is given by where t and t max are the current iteration number and maximum number of iterations.
Step 5: Strike point When the down leader and the up leader combined together. the lightning is paused at a certain point which known as strike point.

Sine cosine algorithm (SCA)
SCA is a well algorithm that mimics the sine and cosine functions variation [24]. The movements of populations in the SCA around best solution based on sine and cosine functions are given in (21) as follows: where t represents the current iteration number. X t best is the best value among the population. A 2 ,A 3 and A 4 are random where k is a constant value.t max is the maximum iterations number. It is worth mentioning that A 1 regulates the trends direction or position of the populations to go outward or inward the best solution as illustrated in Fig. 7.

The methodology of the research
The main aim of this paper is to enhance the performance of Z1 wind farm as a part of Zafarana wind bark using SSSC. The SSSC is connected in series with transmission line which means that seriousness is great if SSSC is subjected any damage due grid faults. The main challenges are how protect the SSSC from damage and find the value of the control parameters of the SSSC which suit the studied system (Z1 wind farm). The crowbar is used as outer protection for SSSC where SSSC is connected from the side of Z1 wind farm of the bus coupler, while the crowbar is connected from the grid side of the bus coupler as shown in Fig. 8. The next table shows the protection system of the studied. The main target of the protection system is to protect the wind farm from over voltage, current and speed and from under voltage, current and speed. The delay time of each type of protection is shown in Table 3.
In this paper, the value of AC and DC voltage regulator of SSSC's control system and the value of crowbar resistance are determined using lightning attachment procedure optimization algorithm (LAPO) and sine cosine algorithm (SCA). LAPO and SCA can call the MATLAB-Simulink model of Z1 wind farm of Zafarana in every iteration. The optimization of LAPO and SCA is based on three objective functions. The three objective functions are: 1. Minimizing the error signal (Δp) which represents the difference between the reference signal of the active power and the accrual active power during fault at the bus coupler.
p p re f − p actial dt  Fig. 8 The SLD of the proposed method and studied system 2. Minimizing the error signal (Δi) which represents the difference between the reference signal of the current and accrual current during fault at the bus coupler.
i i re f − i actial dt (24) 3. Minimizing the error signal (Δv) which represents the difference between the reference signal of the voltage and accrual voltage during fault at the bus coupler. v The next function represents the objective function of LAPO and SCA The values of PI regulators of SSSC (K pvac , K ivac , K pvdc , K ivdc ) are kept in the range [1e −5 -1000]. The value of crowbar resistance (R crow ) is kept in the range [0.1-100].

Sliding Mode Control (SMC) applied to SSSC
SMC) is a type of modern nonlinear control. The importance of (SMC) is its ability to ensure that the control system works efficiently without being affected by uncertain factors [44].
In [45], SMC to control the grid and rotor side converters of DFIG based on fuzzy control to tune PI controller. In this paper, SMC is proposed to control the AC and DC voltage regulators of SSSC based on LAPO to tune PI controller. As shown in Fig. 9, the proposed control signals (u smc_vac and u smc_vdc ) are add to AC voltage and DC voltage regulators of SSSC, respectively. To investigate the SMC of AC and DC voltage regulator, the following steps will be followed. The dynamic equation of V qsssc is defined as follows: (28) where F is the disturbance occurred during operation condition. The error signal is given by (29) By adding the additional control signal of SMC (u smc_vac ) and substituting Eqs. (29) into (10), Eq. (10) can be written as follows: V qsssc K pvac e vac + K ivac e vac dt − F +V qsssc + u smc_vac (30) The switching equation can be defined as follows [45]: The state space of Eq. (31) and derivative of S can be defined as follows: In order to ensure that the system will follow the sliding surface, u smc has two components u smc1 and u smc2 ; where u smc_vac u smc_ac1 + u smc_ac2 . u smc_ac1 represents the ideal condition of the sliding mode, and u smc_ac2 is used to reduce the effects of chattering. The ideal condition of the sliding mode will occur when S 0. u smc_ac1 will be equal to the next equation.
In order to reduce the effects of chattering, u smc2 will be defined as follows [28]: By substituting Eqs. (35) and (36) into (30), the V qsssc can be written as follows: (38) By the same way, the SMC of DC voltage regulator can be investigated and Eq. (9) can be written as follows:

Optimal values of Cro-SSSC obtained by LAPO and SCA
In order to obtain optimal values of Cro-SSSC using LAPO and SCA, the properties and features of MATLAB program are used to link M-file scripts of LAPO and SCA with proposed MATLAB-Simulink model of Zafarana Z1 wind farm integrated with Cro-SSSC. Table 4 shows the optimal values of proposed Cro-SSSC. Also, the optimal values of proposed Cro-SSSC when SMC is used as an example of nun-liner control method. Figure 10 shows the active power of Zafarana Z1 wind farm during three-phase fault when it connected with and without P of Z1 with Cro-SSSC LAPO P of Z1 with Cro-SSSC SCA P of Z1 without Cro-SSSC P of Z1 with Cro-SSSC SMC Cro-SSSC. As shown in Fig. 10 when the Zafarana Z1 wind farm is integrated with Cro-SSSC, the active power generated by is 21 MW during the fault period. This means that it has not separated from the ENEN due to the performance of Cro-SSSC. Unlike the non-use of Cro-SSSC, the active power of the Zafarana Z1 drops to zero MW. This means, the Zafarana Z1 suffers from complete separation from the ENEN. Also, Fig. 10 shows that the Cro-SSSC has almost the same performance when it is tuned by LAPO and SCA. Also, Fig. 10 shows when SMC is used as nonlinear control method to control Cro-SSSC has improved the active power produced by Zafarana Z1 and injects more power to ENEN during the fault. There must be a reason for the dropping in active power of Zafarana Z1 wind farm when there is no connection to Cro-SSSC. To discover this reason, the behaviors of voltage and current have to be observed and studied. Figures 11 and 12 show the voltage and current of Zafarana Z1 wind farm during fault condition.

Obtained results during three-phase fault
As shown in Fig. 11, when the Cro-SSSC has not been used, the voltage of Zafarana Z1 wind farm is dropped to zero. On the other hand, when of Zafarana Z1 wind farm is integrated with the Cro-SSSC, the voltage is only dropped to 0.8 pu and 0.98 pu when SMC is used. Also, the performance of Cro-SSSC when it is tuned by LAPO or SCA is almost the same. As shown in Figs. 10 and 11, both of active power and voltage were dropped to zero from the first moment of fault occurrence. Hence, the time of disconnection of Zafarana Z1 wind farm from ENEN can be observed. The time of disconnection of Zafarana Z1 wind farm from ENEN can be observed by studying the behavior of current as shown in Fig. 12. Figure 12 shows that when Zafarana Z1 wind farm is integrated with Cro-SSSC, the lowest value of current is 0.8 pu for both LAPO and SCA. When SMC is used, the lowest value of current is 1.09 pu which sets in the allowable limits of protection system. On other hand, when the Cro-SSSC has not been used the current of Zafarana Z1 wind farm is increased to 2.5 pu at the begging of fault then decreased to 1.5 pu. After the occurrence of fault by 0.1 s (at time equal to 25.105 s) the current is dropped to zero. This means that the protection system has disconnected the Zafarana Z1 wind farm from after the delay time of the fastest protection. According to Table 3 the fast protection is the under voltage protection which will disconnect the Zafarana Z1 wind after 0.1 s, this means that the disconnection is due to the under voltage. The behavior of current at the connection bus will affect the reactive power absorbed by the Zafarana Z1 wind farm ENEN. This is shown in Fig. 13 which shows the reactive power of the Zafarana Z1 wind farm at the connection bus.
When the Zafarana Z1 wind farm is integrated with the Cro-SSSC which is tuned by LAPO, SCA and SMC, the absorbed reactive power of the Zafarana Z1 wind farm is decreased during fault. The reactive power of the Zafarana Z1 has returned to steady state value much faster when Cro-SSSC is controlled by SMC. This decreasing in the absorbed reactive power is consistent with the decreasing in current in case of Zafarana Z1 wind farm is integrated with the Cro-SSSC. Consequently, the increase in current of Zafarana Z1 wind farm without suing Cro-SSSC will cause increase in the absorbed reactive power by Zafarana Z1 wind farm during fault. This is cleared in Figs. 12 and 13. The performance of current and reactive power of Zafarana Z1 wind farm during fault will reflect on the rotor speed of SCIG wind turbines which are used in Zafarana Z1 wind farm. Figure 14 shows the rotor speed of SCIG wind turbines. As it is cleared from Fig. 14 when the Cor-SSSC has not been used, the rotor speed is increased as the absorbed reactive power by Zafarana Z1 wind farm increases. In an inverse way, when the Cro-SSSC is used, it can be said that the change in rotor speed is not noticeable, especially when Cro-SSSC is controlled by SMC.
The impact of Cro-SSSC on parameters of Zafarana Z1 wind farm is almost the same in both case of tuning (LAPO and SCA). One the other hand, when Cro-SSSC is controlled using SMC based on optimization LAPO, this impact is almost non-existent. This means the change in the voltage of DC bus of SSSC will be almost the same in both case of tuning (LAPO and SCA). Figure 15 shows the voltage of DC bus of Cro-SSSC will be almost the same in both case of tuning (LAPO and SCA). As shown in Fig. 15, the impact of using LAPO to tune Cro-SSSC on voltage of DC bus of Cro-SSSC is very close to the impact of using SCA to tune Cro-SSSC.
Also, Fig. 15 shows the ability of the SMC to make the Cro-SSSC follows the path of the sliding surface. Figures 16 and 17 show the active power and current of Zafarana Z1 wind farm at the connection bus. As it can be observed, in all cases the Zafarana Z1 wind farm is still connected to the grid. The worst case happened in the absence When LAPO and SCA are used to tune Cro-SSSC, the performance of Zafarana Z1 is almost identical and better than the case of absence of Cro-SSSC. Also, Figs. 16 and 17 show that in the case of Cro-SSSC controlled by SMC based on LAPO, the change in active power and current due to single line fault is very small. According to [46], the absence of one phase will increase the voltage in the other two phase, and hence, the root mean square (rms) value of voltage will be higher than the nominal value; this will happen if the reactive power is increased. This is shown in Fig. 18 which shows voltage of Zafarana Z1 wind farm and Fig. 19 which shows the reactive power of Zafarana Z1 wind farm (single line to ground fault). In the absence of single phase and increasing in absorbed reactive power the rms of voltage will increase in the case of LAPO, SCA and SMC due to the presence of Cro-SSSC. On other hand, when Cro-SSSC is not used, the voltage is decreased and the reactive power is injected from the capacitor bank connected to SCIG wind turbines of Zafarana Z1 wind farm. From Fig. 20, it can be observed that change in DC voltage of Cro-SSSC is from 0.995 to 0.005 pu. Also, the change in case of SMC is lower the LAPO and SCA. This is due to the impact of SMC which drives the

Conclusion
This research addresses the problem of disconnection of Zafarana wind farm from Egypt national electric network (ENEN). To solve this problem, a combination of SSSC and crowbar (Cro-SSSC) was used in this paper. Due to the fact that the SSSC is connected in series, the crowbar was used as a safeguard for SSSC from the high current during grid fault. LAPO and SCA were used to evaluate the appropriate values of crowbar's resistance and control gains of SSSC, which is suit the Zafarana wind farm associated with Cro-SSSC. Also, SMC based on LAPO was used as nonlinear control. The performance of Zafarana Rotor speed of Z1 with Cro-SSSC SMC Rotor speed of Z1 with Cro-SSSC LAPO Rotor speed of Z1 with Cro-SSSC SCA Rotor speed of Z1 without Cro-SSSC wind farm associated with Cro-SSSC during three-phase fault and single line to ground fault were examined when Cro-SSSC was tuned by LAPO, SCA and when it controlled by SMC. Also, this performance is compared with Zafarana wind farm without Cro-SSSC. The results show that the using Cro-SSSC is not only capable of keeping the continuous connection of Zafarana wind farm but also can keep its current, active power and rotor speed at acceptable limits especially when it is controlled by SMC based on LAPO.

Recommendations
In this paper, the MALAB Simulink model was built based on the actual data of Zafarana Z1 wind farm. The parameters of Nordex wind turbine -N43 which are used in Zafarana Z1 wind farm are given in Table 2 [16-19, 32, 33]. Hence, it is expected that the results of this research will be very close to the actual application. The authors recommend using Cro-SSSC controlled by SMC base on LAPO in solving the problem of disconnection of Zafarana Z1 wind farm from ENEN.
Funding Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecomm ons.org/licenses/by/4.0/.