Sintonia dos controladores para aerogeradores de velocidade variável com geradores de indução de dupla alimentação utilizando um algoritmo meta-heurístico de inteligência coletiva

Abstract

The increase in wind power integration into power systems is currently a general trend in many countries and the Variable-Speed Wind Turbine (VSWT) Equipped with a Doubly Fed Induction Generator System with back-to-back power converters are most often used for this purpose. There is the need for its control system to operate properly to ensure stability and to achieve the desired performance when they are subjected to different transient disturbances, which is a difficult taste due to nonlinearities and strong coupling between electrical and mechanical variables in this type of wind turbine. In this context, this work aims to tune the gains of the PI controllers that make up the control system using a collective intelligence meta-heuristic algorithm for two normal perturbations in the operation of a wind farm: a wind variation and electrical fault. For that, it was modeled the wind turbine taking advantage of detailed modeling, considering the stator voltage oriented control structure, with active and reactive power control loops and their respective current controls, all equipped with PI controllers. Therefore, a set of objective functions was generated penalizing the most important variables directly from the dynamic simulation. Next, a multi-objective PSO algorithm has been enhanced, to be used as a tuning tool. Then the algorithm was applied to the VSWT for Single-machine infinite bus system for wind speed variation and electrical fault. Finally, the robustness of the tunings was tested on the in Single-machine infinite bus system and a 32 nodes feeder. The PSO showed good ability to exploitation and intensification of the search space reaching the preset tuning requirements in all cases. Stability and satisfactory performance results were obtained for transient disturbances thanks to key variables penalty in the objective functions. The tuning resulted robust enough for variations of feeder parameters and operating conditions besides the proper stiffness coefficient of the turbine generator mechanical coupling, whenever the wind turbine was subjected to disturbances considered in the tuning. It is also noted that exploring the objective functions and the algorithm itself can attain a large improvement in the performance of the tuned wind turbine control system for wind variations and electrical faults