Performance of a vector control for DFIG driven by wind turbine: real time simulation using DS1104 controller board

Received Oct 7, 2018 Revised Feb 14, 2019 Accepted Mar 1, 2019 In this research paper we investigate the modelling and control of a doubly fed induction generator (DFIG) driven in rotation by wind turbine, the control objectives is to optimize capture wind, extract the maximum of the power generated to the grid using MPPT algorithm (Maximum Power Point Tracking) and have a specified reactive power generated whatever wind speed variable, the indirect field oriented control IFOC with the PI correctors was used to achieve such as decoupled control. To validate the dynamique performance of our controller the whole system was simulated using dSPACE DS1104 Controller board Real Time Interface (RTI) which runs in Simulink/MATLAB software and ControlDesk 4.2 graphical interfaces.


INTRODUCTION
In recent years, the world has known a rapid evolution in various sectors that means the increase in demand on electrical energy; The uses of traditional fossil and nuclear energy resources pollute the environment, for this reasons countries are focused on development and exploitation of renewable energy sources such as biomasses, wind energy and solar energy which can offer the opportunity to produce the electrical energy in a clean way and help to save the environment [1]. Wind energy is one of the most effective and clean resource, it presents a promising and important source of energy with low cost, so many researchers have been focused on wind energy systems to develop new technologies in order to optimize the capture wind and the electrical power produced [2,4].
In wind farms the most installed generator is the Doubly-Fed Induction Generator DFIG, This machine presents several advantages; it makes possible to better use the wind energy resources by offering a large range variation of speed about ± 30% around the synchronous speed (hyper-synchronous and hyposynchronous mode) [4]. The power converters connected to the rotor are dimensioned to pass a fraction of the nominal power produced by the generator (20%-30%), consequently, the reduction of losses in power electronics converters and cost compared to others generators [3][4][5][6]. To optimize the performance and effectiveness of the aero-generators and extract the maximum of the power generated to the network, several solutions and control technologies have been studied in the literature [3][4][5][6][7]. The objectives of this research paper is to extract the maximum of the electrical power generated from the wind turbine for each wind speed variation and validate the dynamic performances of the system by simulation using DS1104 Controller Card which runs in Simulink/MATLAB environment and ControlDesk graphical interface. To control the generator we have used the PI controller for his easy maintenance, good reliability and simplicity of implementation compared to other nonlinear controllers [3][4].
In this paper we start by the modeling of the wind energy conversion system; the modeling of the wind turbine and the DFIG in the Park reference d-q, in the second section we present the indirect field oriented control IFOC with the PI corrector to control the rotor side converter (RSC), then we present how to extract the maximum of the power using the MPPT algorithm (Maximum Power Point Tracking), in the fourth section the real time simulation results of the system using DS1104 card was presented, finally a conclusions are given.

MODELING OF WIND ENERGY SYSTEM
The structure studied in this paper is based on the wind energy conversion system based on a doubly fed induction generator with gearbox; the stator of the machine is connected directly to the network, while the rotor is connected to the grid via two powers converters Figure 1.

Wind turbine model
The modeling of the wind energy system is given by the following [6-7]: The power coefficient Cp depending on the blade pitch angle β and the Tip Speed Ratio (TSR) λ are defined as follows [7]:

1005
Where Pv, Paer , ρ, S, R, Ωt and v are the power of the wind turbine, the aerodynamic power available on the rotor, the aerodynamic torque, the air density (ρ =1.22 kg/m3), the circular surface swept by the turbine, the radius of the turbine, the turbine tpeed (rad/s) and the wind speed (m/s), respectively. The fundamental equation of dynamics is given by [8][9]: Where Ωmec, J, Tg, Tem and Cf are the mechanical rotor speed (rad/s), the total inertia, the generator torque, the electromagnetic torque and the coefficient of viscous friction, respectively.

Mathematical model of DFIG
The mathematical model of the stator and rotor voltages of the doubly fed induction generator in the Park reference (d-q) is given by the following [ Where Ls, Lr are the cyclic stator/rotor inductance, Lm is the maximum of the mutual inductance and Rs, Rr are the stator/rotor resistance. The stator active/reactive powers & the electromagnetic torque of the DFIG are given by [11][12]:

Converter model
The Rotor Side Converter (RSC) controlled by the Pulse with Modulation (PWM) signals is modelled by a matrix. The matrix that connects the rotor voltages of the DFIG (Vra, Vrb, Vrc) and the control signals provided from the PWM module (Sa, Sb, Sc) is constructed as follow [3]: Where: Udc is the DC link Voltage.

CONTROLLERS MODEL 3.1. Indirect field oriented control (IFOC)
To control the reactive power and the Torque of the generator (DFIG) independently we have to apply the vector control in order to ensure a decoupling between the variables of the machine [13]. In this paper, we chose a two phases d-q linked to the rotating reference frame, and the stator flux is oriented along the axis d, the expression of the flux becomes [14]: Assuming that the flux stator is oriented along the axis d and constant at the permanent regime and the stator resistance Rs of the generator is neglected, the equations of the stator and rotor voltages can be expressed as [14][15]: The expressions for the stator powers & the electromagnetic torque of the DFIG are given by [16]: Fundamentally, the indirect field oriented method using the proportional integral correctors consists of two control loops [2]. The first loop serves to calculate reference rotor currents (Irq*, Ird*), this control loop gets its reference from the maximum power point tracking algorithm (MPPT) in order to optimize the power produced from the generator for each wind speed changes. The second loop is used to determine the rotor voltages references (Vrq*, Vrd*). According to (11), (12) and (13) we concluded that Vrd* and Vrq* are the variables of the rotor voltages to impose on the DFIG to get the rotor currents, the electromagnetic torque and the reactive stator power desired. The proposed model is given in Figure 2.

Proportional integral (PI) controller design
The synthesis of the proportional integral (PI) regulators is divided into two steps; the first step serves to regulate the currents and the second step to regulate the powers/torque. The rotor current loop is shown in Figure 3.
The compensation method is used to eliminate the zero of the transfer function H(p), to ensure the performance of the compensation method the parameters of the transfer function most be known, we choose: The Open Loop Transfer Function H(p) becomes: The Closed Loop Transfer Function G(p) is giving by the following equation:

Maximum power point tracking (MPPT Control)
To extract the maximum power generated it is necessary to provide the adequate electromagnetic torque which is used to vary the mechanical rotor speed of the DFIG for each wind speed variation [6]. In order to estimate the value of the wind speed which gives the maximum of the power we can fixed the TSR to its optimum value λopt that corresponds to the maximum power coefficient Cp-max, According to (1) and (2) it's can be possible to deduce the expression of the reference electromagnetic torque as a function of the optimum TSR λopt and mechanical rotor speed Ωmec [3][4][5][6]: Where G is the gain of the multiplier wich adjusts the turbine speed to the rotor speed of DFIG. The block diagram in Figure 4 presents the model of the turbine and the MPPT Algorithm.

DSPACE DS1104 CONTROLLER BOARD
The dSPACE DS1104 Controller card is a cost effective system with a real time processor, it can be mounted in a personal computer with a 5V PCI slot. The function models are easily turned on the DS1104 system with Real Time Interface (RTI), to configure the Input & Output of the system studied graphically the ControlDesk interface can be used, insert the blocks diagrams and generate the model code via SIMULINK software; the real time simulated models are compiled, downloaded, and started automatically [17]. This increases systematically the productivity and reduces development costs and times. Figure 5 shows the block diagram of the DS1104 Controller Card used in our laboratory which consists of a main processor PowerPC 64-bit floating point processor with a CPU Clock 250 MHz, a slave DSP I/O and a master PPC I/O subsystem [17][18].

RESULTS AND DISCUSSIONS
The simulation of the whole system is performed using the Simulink/MATLAB software. The Table 1 resumes the parameters of the wound induction generator A4222 (1.5 kW), after doing the experimental determination of the parameters in our Laboratory.

Simulation results
To improve the performances of the system we apply a variable wind speed profile, it is modeled by a sum of several harmonics around an average speed of 10.5 (m/s) presented in Figure 7, despite of the variations of the wind speed the power coefficient Cp is keeps at its maximum value Cp-max= 0.49  Figure 9 with Figure 10 we notice that the decoupling between the currents rotor Ird and Irq is ensured and the tracking response is well done using PI correctors. Figures 9 and 10 illustrate the reactive stator power Qs and the electromagnetic torque Tem, they track perfectly their reference values against wind speed variations, the electromagnetic torque Tem follows exactly the reference Tem* calculated from the MPPT method, thus a small variation of the wind introduces a great variation on the torque, the reactive power stator Qs shown in Figure 12 is fixed at 0 VAR to have a unity power factor on the stator side and optimize the quality of the electrical energy produced to the grid.    Figure 13 illustrates that the currents stator Is for the three phases are sinusoidal with fs=50 (Hz). Figure 14 gives the spectrum analysis of the Total Harmonic Distortion (THD) of the stator current for the phase (a) Isa, as we see THD=6.65%, the harmonics appearing on the injected stator current Isa to the grid are minimized that implies the wind energy provide the optimal electrical energy to the network.

CONCLUSION
The work was devoted to control a Doubly Fed Induction Generator applied in a wind energy conversion system. After modelling the whole system we have adopted the Indirect Field Oriented Control (IFOC) to control the reactive stator power and the electromagnetic torque independently, then the MPPT Algorithm to extract the maximum of the power generated to the network.
The Wind Power Generation System was modelled and simulated with a variable speed operation for a power of 1.5 (kW). To validate the proposed control model we have performed a simulation used dSPACE RTI1104 Real Time Interface. The results obtained under ControlDesk 4.2 graphical interface were satisfactory that ensure the performances of the control system studied.