Article Preview
TopIntroduction
Recently, with the rising demand of electricity, wind turbine generator system (WTGS) has been attractive and competitive with conventional fossil fuel energy resources as the former is safe, pollution free, inexhaustible and free in term of its natural existence (Sumathi et al., 2015). Wind energy conversion system (WECS) has come up as an important renewable source of alternative energy for the future (Badmasti et al., 2014; Ani, 2017). There are two types of wind turbine which are fixed and variable speed wind turbines (Chen et al., 2009). Variable speed wind turbine (VSWT) is used to maximize energy capture at various wind speeds and provide additional 10%- 15% energy capture than the fixed speed type. It lessens the load on the drive-train and tower structure. The high efficiency, power density, wide speed range, reliability, less maintenance, reduced loss, suitability for grid or standalone system (Koussa, et al.,2015) and full isolation of the permanent magnet synchronous generator (PMSG) generator from the power grid make it preferred in variable speed wind systems (Errami et al., 2015).
As PMSG is used in large capacity for the massive production of electric power; improvement in maximum power point tracking (MPPT) algorithm maximizes the power extracted, thus assists better utilization of the installation. In the WECS, a simple diode bridge rectifier rectifies the 3 phase AC voltage into the DC voltage after which a boost chopper is placed as a part of MPPT controller. Then it feeds to a pulse width modulation (PWM) controlled single level and multilevel inverter so as to convert it into alternating voltage which is sent to the load and grid. MPPT algorithms have been investigated significantly by Thongam & Ouhrouche (2011) for various types of generators. They have used three methods of MPPT: tip speed ratio (TSR) control, Power signal feedback (PSF) control and hill-climb search control (HCS). HCS has been modified for early convergence of MPP (Banu et al., 2018) which reduced oscillations. For small wind turbines MPPT at low wind speed, TSR and PSF control have been shown to be fast in response (Malinowski et al., 2015).
Artificial Neural Network (ANN), Fuzzy Logic (FL), Swarm based and conventional MPPT have been reviewed comprehensively by (Ram et al., 2017). Koutroulis & Kalaitzakis (2012) have also proposed MPPT from calculation of successive output wind power values from measured voltage and current for the controlled DC-DC boost converter. Errami et al. (2012) provided a control strategy for PMSG on MPPT by control of angular speed of rotor. In another literature, the nonlinear backstepping control has been proposed for MPPT in a grid connected system (Errami, 2015). A neural network approach for control of both MPPT and pitch angle for the total range of wind speed has been proposed, but it needs training (Dahbi et al., 2016). Howlander et al. (2014) introduced fuzzy control for power smoothing which was more effective than MPPT control in mitigation of mechanical stress for fault from grid. Combinations of nonlinear and robust control approaches with the conventional methods of MPPT have demonstrated reduction of harmonics and ripples. Tiwari et al. (2018) demonstrate FL control better than PI (Proportional Integral) in 3 phase 4 wire system to reduce harmonics. TSR MPPT control method with integral sliding mode could control chattering in voltage (Yaylaci, 2017). In another work, ANN based estimation of wind speed is done by training (Sabzevari, 2017) which doesn’t need wind speed sensors later on but estimates it. The ripples in power are reduced by fuzzy PI controller, but training takes time. Estimation method of effective wind speed is desirable for proper control to maximize energy extracted. A review has been brought out on control based on wind speed estimation (Jena & Rajendran, 2015). When wind speed measurement facility is available in large scale wind turbine, for short term control, wind estimation is not an issue. A breakthrough on application of statistical analysis for power system such as state estimation considering multi-area power system and for unit commitment problem is presented by (González et al., 2015; Marmolejo & Rodriguez, 2015). The fat tail model used for multi-period unit commitment can be used for improving power extraction for annual wind speed variation.