An investigation into the utilization of swarm intellingence for the control of the doubly fed induction generator under the influence of symmetrical and assymmetrical voltage dips.

Abstract

The rapid depletion of fossil, fuels, increase in population, and birth of various industries has put a severe strain on conventional electrical power generation systems. It is because of this, that Wind Energy Conversion Systems has recently come under intense investigation. Among all topologies, the Doubly Fed Induction Generator is the preferred choice, owing to its direct grid connection, and variable speed nature. However, this connection has disadvantages. Wind turbines are generally placed in areas where the national grid is weak. In the case of asymmetrical voltage dips, which is a common occurrence near wind farms, the operation of the DFIG is negatively affected. Further, in the case of symmetrical voltage dips, as in the case of a three-phase short circuit, this direct grid connection poses a severe threat to the health and subsequent operation of the machine. Owing to these risks, there has been various approaches which are utilized to mitigate the effect of such occurrences. Considering asymmetrical voltage dips, symmetrical component theory allows for decomposition and subsequent elimination of negative sequence components. The proportional resonant controller, which introduces an infinite gain at synchronous frequency, is another viable option. When approached with the case of symmetrical voltage dips, the crowbar is an established method to expedite the rate of decay of the rotor current and dc link voltage. However, this requires the DFIG to be disconnected from the grid, which is against the rules of recently grid codes. To overcome such, the Linear Quadratic Regulator may be utilized. As evident, there has been various approaches to these issues. However, they all require obtaining of optimized gain values. Whilst these controllers work well, poor optimization of gain quantities may result in sub-optimal performance of the controllers. This work provides an investigation into the utilization of metaheuristic optimization techniques for these purposes. This research focuses on swarm-intelligence, which have proven to provide good results. Various swarm techniques from across the timeline spectrum, beginning from the well-known Particle Swarm Optimization, to the recently proposed African Vultures Optimization Algorithm, have been applied and analysed.