Control Approach Of A Grid Connected Dfig Based Wind Turbine Using Mppt And Pi Controller

. A double-fed induction generator (DFIG) has been frequently utilized in wind turbines due to its ability to handle variable-speed operations. This study investigates the real parameters of the Mitsubishi MWT 92/2.4 MW wind turbine model. It performs and implements grid-connected variable-speed turbines to control the active and reactive powers. Moreover, it presents a vector control strategy for DFIG for controlling the generated stator power. The unique feature of the approach proposed in the study is the comparison between two control techniques - the Maximum Power Point Tracking (MPPT) algorithm and the Proportional-Integral (PI) controller - for regulating DFIG based wind turbine systems. Thus, the result demonstrates that the performance of the MPPT technique provides strong robustness and reaches steady-state much faster than the PI controller with variable parameters. To the contrary, a typical PI controller gives a fast response when tracking the references of DFIG magnitudes. The effectiveness of the overall sys-tem is tested by MATLAB simulation.


Introduction. Problem Definition
Wind energy is a popular renewable source of energy, and its use is increasing steadily across the world.Wind technology is promoted in several nations through different government programs and market mechanisms.An alarming advancement have been observed in the wind power system across the world in the last decades with increasing rotor diameters and utilization of advanced power electronics that operate at a variable speed [1].In comparison to other forms of wind power system, DFIG is found to be the most extensively /studied machine.The study in [2] demonstrates an overview of grid-connected wind energy conversion system (WECS) in details.Precise detection of the grid voltage angle has been noted to be an essential task in most of the control techniques in the GSC [2] and [3].Furthermore, it has been recommended that all advanced wind turbines must include a faster and efficient tracking equipment to determine the frequency and stator voltage of DFIG during the grid synchronization.
The study in [3] provides a review on the issue of effective estimations of wind speed (EEWS).The main EEWS strategies and their implementations are thoroughly addressed.The study demonstarted that an accurate estimations of the wind speed is important for energy capture.Improving the effectiveness of vector control (VC) strategy has attracted many researchers.
In several studies [4], [5], [6], and [7], the authors presented that wind turbine-based DFIG provides decouple control of active or reactive power of the system, resulting in good dynamic performance, higher efficient energy output, and good power quality.Recall that DFIG is more difficult to control compared to the typical induction generator.Hence, the rotor currents are regulated by the power electronics converter in the rotor circuit, which results in the control of DFIG [8].VC is a popular strategy for controlling DFIG-based wind turbines [9] and [10].The technique presented in [11] introduces a new integrated frequency control technique that dynamically integrates inertial and pitch angle control to enhance the performance of DFIG in power system frequency control.Different studies have been conducted investigating DFIG control; for instance, a novel controller has been introduced that allows the DFIG system to positively contribute to grid operation [12].Further, a comparative analysis between robust sliding mode control (SMC) and typical PI controller in DFIG based WECS has been presented [13].Their simulated results demonstrate that the proposed SMC provides a faster response with very little steady state error in comparison with the PI.Moreover, comparing PI and fuzzy control based DFIG wind turbine, it is obvious that fuzzy control is more robust against machine parametric perturbations and delivers faster convergence [14].According to the study in [15], by using a fuzzy-PI controller, the settling time and the value of peak overshoot have reduced, and variations are damped down faster compared with the typical PI controller.Besides, the transient response given by fuzzy-PI controller has also proven to be better than typical PI controllers.The study in [16] introduces a dual-loop control technique to enhance the dynamic performance of DFIG that may be subjected to grid disturbance.This technique can quickly deteriorate the fluctuations of stator flux and successfully mitigate the impact of stator transient flux on the performance of the DFIG.Further, in [17] the comparative investigations of control strategies have been considered in DFIG system.It states that direct torque control (DTC) is quicker than VC in terms of transient response, that can be very beneficial and provide a lot of capacity when specific control manipulations are required.The research in [18] investigates a coordinated control approach for DFIG with a DC connection.The experimental findings suggest that a control approach is available for the proposed power grid configuration.It concludes that the DC/DC converter is the most important component in WECS for the integration of the DFIG system.From [19], it can be realized that the study places a PI controller before the conventional DTC or direct power control (DPC) blocks to improve the dynamic performance.Consequently, the study shows that DTC has higher operat- ing performance than DPC since DTC directly controls the torque.
Numerous control techniques for DFIG, such as DPC, VC approach, and DTC, have been suggested throughout the years [20].The VC strategy has advantages among other control techniques; for example, it gives reduced harmonic distortion and fewer power ripples [1], [10], and [21].However, VC has been applied only for PI control strategy in the literature [22] and [23].But, in our study PI controller and MPPT techniques has been performed.Moreover, none of the aforementioned literature analyzes the performance and control of the magnitudes of grid-connected DFIG system using PI and MPPT controller in terms of settling time.Here, the DC-link capacitor is used to separate RSC and GSC.The DC link stores the energy in the capacitor and the purpose of this capacitor is to keep the voltage terminals constant.
This study assesses the real parameters of the Mitsubishi MWT 92/2.4MW wind turbine model integrated with a grid connected DFIG system using the VC strategy.This study is organized into several sections.Section I briefly presents the overview of the wind turbine-based DFIG system.Wind turbine modeling has been developed in Section 2. .The VC strategy of both sides -RSC and GSC are given in Section 3. .Two types of control method schemes -PI controller and MPPT for DFIG have been demonstrated in Section 4. .The simulation result has been discussed in Section 5. , and finally, conclusions are drawn which are stated in Section 6. .

Wind Turbine Model
Normally, wind turbines generate electricity by converting mechanical energy provided by the kinetic energy of the wind [21].Here, the wind turbine model comprises several parts: a wind speed, an aerodynamic system, a mechanical or drive train system, and a generator.A typical block diagram of a wind turbine is given in Fig. 1.

Wind Speed Model
Wind speed tends to vary depending on the nature of the environment and it fluctuates randomly over time.This explains the fact that it has a significant impact on the electromagnetic torque and hence it seems to have a major impact on the power generated by the three blades [24].Thus, to simulate the dynamic performance of wind turbine model, the wind speed should be taken into consideration.In this work, wind speed can be modeled by adding the following four constituents: where v wa (t), v wr (t), v wg (t) and v wt (t) are the constant, turbulence, ramp, and gust constituents, respectively.The gust constituent can be modeled to represent unusual transient increases in wind speed and can be formulated as follows: where T sg and T eg are start and end time of gust constituents and A g is gust amplitude.Eventually, the turbulence constituent can be indicated by a signal with the following power density [24].
where l is turbulence scale, h is the height and l = 20h, that has maximum height of 300 meter; z 0 is the raggedness length value for different landscape type as it can be seen in Tab. 1 [24].
Tab. 1: Values of zo for various landscape type [24].Since we know all the other values, the next stage is to generate a signal with power spectrum.Considering that the P Dt is close to the responses of first-order filter, then the suggested transfer function is as follows:  Where p and K are formulated as follows: And K 1 and K 2 are expressed as: The calculated signal of the power spectrum is defined as:

Aerodynamic Model
Normally, the aerodynamic system of a wind turbine calculates the electromagnetic torque as shown in Fig. 2 and is given as [20].
Where, ρ, R, V w , and C t are air density, radius (meter), wind speed and torque coefficient of wind turbine, respectively.The power coefficient is given as: with The tip step ratio can be obtained as:

Drive Train System
In some literature [25] it is possible to model the drive train system in a 2-mass where the stiffness of the shaft links the rotor of DFIG to the turbine mass as it can be seen in Fig. 3. Drive train contains of a turbine, lowspeed shaft, generator, gearbox, and high-speed shaft [26] and [27].
For instance, the torque (T m ) and stator reactive power (Q s ), can be formulated as follows: and Where: θ t , θ m , ω t and ω m are turbine angle, generator angle, turbine and generator angular speed, respectively; τ t is torque supplied to rotor shaft, and τ m is generator torque; P is number of poles; M is magnetizing inductance.

Generator Model
For a proper understanding of the behaviors of a generator or DFIG system, it is important to use a basic and special model such as rotating 2-phase dq which is provided by Park transform technique [28].Fig. 4 represents the power circuit of DFIG.The dynamic equation of the system is given as a fourth order state space based synchronous dq representation, stator and rotor voltage equations can be given [22]: The stator and rotor fluxes are determined: Stator active and reactive powers are given:

Vector Control Strategy
Controlling alternating current (AC) machines may be generally categorized into two, namely scalar and vector controls.The implementation of scalar control is simple and may produce an approximate steady state response, especially when the dynamics are slow.For this purpose, to achieve a higher accuracy and better dynamics and also good response during steady state, VC strategies have to be applied [23].The VC strategy focuses on RSC and GSC control.

Rotor Side Converter Control
The main idea behind the control of RSC is to keep rotational speed constant regardless of wind speed.The rotor voltage can be defined as: The voltages in dq references can be written as: The vector control of the RSC design is represented in Fig. 5.

Grid Side Converter Control
The function of GSC is to keep the DC-bus voltage (V bus ) referred to the stator constant.In our work V bus is given 1150 V.The dynamics of the voltage in the grid side can be represented as follows: The voltages in dq references may be given as: Vector control for the GSC design is indicated in Fig. 7.

Proposed Control Approach
This section presents two types of control method of DFIG that has been implemented in the system simulations.

MPPT Technique
A MPPT control strategy is significant since it assures a variable speed system, which maximizes power production throughout a particular wind speed.Consider the curve shown in Fig. 6, the variable speed system is operating at a level.Once the value of wind speed changes from V v1 to V v2, then operating point and torque change to b and T t−b , respectively.The regulator produces torque that corresponds with the maximum power curve (level c), which is less than T t−b .Thus, this increases the rotor speed of the turbine until it reaches level c (equilibrium point).
Rotor speed reference at optimum lambda is given by: The electromagnetic torque (T em ) can be given: Simplifying Equation ( 28) gives: Equation ( 29) leads to Fig. 8

PI Controller
It is very important to create PI regulators.Fig. 10 illustrates the turbine coefficient as a function of lambda and power curve.Here, the suggested power curve (Fig. 10) is verified by comparing the values of power of the DFIG system with different wind speed.
The specifications of wind turbine system are given in Tab. 2.

Result and Discussion
In this study, the real parameters of MWT 92/2.4MW wind turbine of the Mitsubishi have been implemented for simulation study.Hence, the overall system has been verified by MATLAB/Simulink environment.Furthermore, the grid voltage and switching frequency are constant throughout the simulation, i.e., voltage is The DFIG system operation is first conducted using typical PI and MPPT controller, then the comparison of the overall system with the two controllers is performed in terms of settling time in Tab. 3.For instance, the reference speed is set to be 1500 rpm (157.08 rad/sec) the simulation of PI controller started with 90% of reference speed (see Fig. 12), the rotor speed is operating as sub-synchronous (motor mode) up to 4 sec, then at 4 sec the operation is changed to synchronous mode (100% of the reference speed), later at the simulation time (t s = 6s) the operation is changed to hyper-synchronous (generator mode) with the same rated torque.Similarly, the electromagnetic torque is controlled according to the synchronous speed.Fig. 14 presents the quadrature rotor current i qr that is changing according to the electromagnetic torque, because torque is proportional to the quadrature rotor current.
One DFIG system is found to be sensitive to the grid disturbances.Normally, the rotor side of DFIG is connected to the grid via converter.Thus, when a    grid fault occurs the converter will be damaged.However, this can be addressed through the use of some hardware protections such as crowbar protection and DC choppers.One of the mainy, advantage of DFIG system integrated with the wind turbines is that the stator voltage and frequency remain constant, regardless of how the speed of the wind blows on the turbine rotor.At this point, the wind speed profile is set to be 8.5 m/s.Furthermore, the wind speed model is designed by adding four constituents such as constant, turbulence, ramp, and gust constituents.The gust constituents are used to describe the unusual sudden changes in wind speed.Equation (4) leads to the wind speed profile represented in Fig. 15.Taking the RSC into consideration, the rotor speed is set to 136.5 rad/sec at steady state and the wind speed is modified to 11.5 m/s at 4 seconds, As a result, the rotor speed increased to 196.4 rad/sec, as shown in Fig. 16.Similarly, Fig. 17 presents the electromagnetic torque at wind speeds of 8.5 m/s and 11.5 m/s, resulting -5500 Nm and -11500 Nm of generated torque, respectively.Accordingly, the generated power is modified to -0.75 MW and -2.25 MW, respectively.The generated torque and power are negative because the machine is operating as a generator convention.Fig. 18 represents the quadrature rotor current (i q ), and Fig. 19 shows the stator active power which modifies according to i q .Fig. 20 indicates the direct rotor current i d r, which is set to 850 A at 6 seconds and reactive stator power is controlling and modifying according to i d r as shown in Fig. 21.
In GSC, the DC-bus voltage is set to be 1150 V, as shown in Fig. 22, DC bus voltage is maintaining constant at 1150 V at the steady-state with the change in torque.Fig. 23 represents grid reactive reference               (Q g_ref ), which, for instance, is set to be zero up to t s = 6s.and changed to -400 kVAR and is maintaining and tracking the desired value.Furthermore, Q g_ref can be changed according to the desired grid code.In Fig. 24, the quadrature grid current is changing according to the change in Q g_ref .Fig. 25 and Fig. 26 represent direct grid current (i dg ) and three phase stator voltages, respectively.As shown from the rotor speed (by MPPT) in Fig. 16, as the wind speed increases, torque and generated power also increases.However, in this work the wind speed doesn't go beyond 11.8 m/s.For instance, in case it exceeds that value, the pitch controller will take place to limit the maximum extracted power.Even though, pitch controller is not implemented, instead parameters of Mitsubishi MWT 92 have been carried out in MATLAB (see Fig. 10).Thus, the result shown in Fig. 27, Fig. 28, Fig. 29 and Fig. 30 has been compared and verified with Fig. 10.In this study, the rotor speed response, generated torque, quadrature rotor current and direct rotor current were mainly investigated.In the literature, optimum control of a DFIG system using differential evolusionary strategy are presented to improve the performances of DFIG systems during perturbation [29] and [31].Accordingly, by using fuzzy-PI controller in [15] and [30], the settling time and the value of peak overshoot have reduced, and variations are damped down faster when compared with typical PI controller.Besides, the transient response given by fuzzy-PI controller has also proved to be better than typical PI controller.Compared to the aforementioned literature with this work, as illustrated in above results (Fig. 27, Fig. 28, Fig. 29 and Fig. 30), it can be seen that there is higher overshoot/undershoot in the PIcontroller representation.Further, this higher overshoot and undershoot can be ruduced by tunning the PI-gains and current loops until the desired values are  obtained.Meanwhile, the performance of DFIG based wind turbine with the MPPT control improves in terms of settling time as given in Tab. 3.Moreover, it can be observed that the MPPT control has less perturbations and provides strong robustness and reaches the steady state faster with variable parameters.Additionally, active power of the DFIG system matches with the power curve shown in Fig. 10.

Conclusion
The VC approach has been designed to control the rotor and grid sides of the DFIG system independently.This control strategy reduces harmonic distortion and produces fewer power ripples.The overall system consists of two cases.In Case I, PI controller is applied to the rotor side of DFIG system.Hence, the rotational speed is operating from sub-synchronous to hyper-synchronous mode with a constant input torque.Therefore, a constant rotor speed and constant electromagnetic torque are obtained for a constant reference speed.In case II, MPPT closed loop strategy has been considered.In this strategy, constant rotor speed and generated torque are obtained for a variable wind speed, as a result, the generated power is obtained, and confirmed by comparing it with the extracted power (Fig. 10).Further, the rotor side is able to control the stator's active and reactive power.On the other hand, the grid side controls grid reactive references, which can be controlled according to the grid codes.
The grid side is also used to keep the DC bus voltage constant.Thus, PI controllers are effective for fast tracking of current references, and the current loops have been tuned until the desired value is obtained.
On the other hand, MPPT control has fewer perturbations, provides strong robustness, and reaches a steady state faster with variable parameters.From the results, it is clearly obvious that the designed MPPT is efficient, reliable, and gives better performance than PI controller.
In the future, the researchers can focus on designing the pitch controller to achieve good performance at much higher wind speeds.

Tab. 2 :
Values of zo for various landscape type [24].Nominal power (Ps) 2.4 MW The Rated torque (Tem) 12732 Nm.Rotor speed I (n) 157.08 rad/sec Rotor speed II (nand frequency is 4000 Hz.Fig. 11 represents the Simulink model of PI controller.

Fig. 27 :
Fig. 27: Steady state response of rotor speed with MPPT and PI controller at 8.5 m/s.

Fig. 28 :
Fig. 28: Steady state response of torque with MPPT and PI controller at 8.5 m/s.

Fig. 29 :
Fig. 29: Steady state response of quadrature rotor current with MPPT and PI controller at 8.5 m/s.

Fig. 30 :
Fig. 30: Steady state response of direct rotor current with MPPT and PI controller at 8.5 m/s.