Robust Predictive Control of Variable-Speed Wind Turbine Generator by Self-Tuning Regulator

The parameter variation problem in controller design for wind turbine generator (WTG) systems due to wind speed fluctuations is addressed in this paper. The wind speed is generated by an auto-regressive and moving average (ARMA) model while drive-train components are modelled as a two- mass system described by non-linear equations for accurate dynamic performance assessment. The control of the unit is exercised through a multi-input, multi-output (MIMO) self- tuning regulator (STR) that uses the local generator speed to produce appropriate control signals for compensating parametric uncertainty. The parameters of the STR are adjusted online for quality improvement of the power supply under turbulent wind. Simulation results show the proposed controller causes the closed- loop wind turbine to behave as a simple first order power filter, where power output is a low-pass filtered version of the incident wind power. Further, the control scheme achieves the regulation of torsional dynamics without involving converter controls.