Design and Performance Comparison of PI and Adaptive Current Controllers for a WECS

Abstract

This work proposes a novel controller in a stationary reference frame for permanent magnet synchronous generator (PMSGs) of grid-connected wind turbines. The wind energy conversion systems (WECS) adopt a Back-To-Back converter system with voltage source inverter (VSI) and a phaselocked loop to track continuously the fundamental frequency and its phase from voltage node that is connected to the system generation. One of the central themes of this paper is the development of a control scheme that uses a B-spline artificial neural network for controllers tuning when the system is subjected to disturbances. The B-spline neural network must be able to enhance the system performance and the online parameters updated can be possible. This work proposes the use of adaptive PI controllers to regulate the current, frequency and DC-link voltage. Results of the dynamic behavior of the whole WECS are presented along with different cases subject to various disturbances. The simulations show the feasibility and robustness of the proposed control schemes for PMSG based wind turbines. 2002. His current research interests are in modeling and control of wind energy systems and motor drives with computational intelligence techniques.

An Adaptive Speed Control Approach for DC Shunt Motors

“…In the initialization step, a layer formed with two basis functions is defined as shown in Figure 2. This configuration is proposed due to the presence of a single error signal, and good control performance criteria obtained in previous works [15][16][17]. The shape of these basis functions permits a soft neural network output response, and different shapes are previously analyzed with superior orders, but the performance is similar with the restriction that a greater number of calculations is required.…”

“…Each configuration has a different performance and the number of operations depends on the type and order of the basis function. This ANN structure has been explored in other dynamic systems showing a good performance [15][16][17]. The data set used in this designation is presented in Table 1.…”

“…In this application, it has been settled at 1 × 10 −3 . There are some applications of adaptive controllers based on B-spline neural networks where how to define the base functions, neural network structure and a training rule are explained [15][16][17][18]. A clearly simple structure facilitates the form of implementation and adaptation to different systems; in addition, the number of neurons, structure and shape of base functions have similarity in all these cases; therefore, the same structure can be extended for systems of different characteristics.…”

“…At the same time, they consider the complexity of the physical system and provide a realistic control with less computational time for an efficient and robust control in a wide operating range. In this regard, B-spline neural networks (BSNN) are a particular class of neural networks that have exhibited important results in various practical physical systems [15][16][17][18][19][20]. This paper presents the design and performance assessment of an adaptive tracking control approach based on B-spline neural networks with the capabilities required for real-time applications of DC shunt motors, without requiring a detailed mathematical model and values of the parameters of the nonlinear dynamic system.…”

Influence of failure rate and downtime in annual energy produced on a horizontal axis wind turbine

“…Manwell et al [12], apontam que o método de classes de vento fornece outra maneira de visualizar e estudar a produtividade de um aerogerador. Os dados devem ser separados em intervalos iguais de velocidade, conhecido como classes de velocidades, ou velocidades discretizadas, a exemplo da metodologia proposta por Douglas et al [13] para estimativa preliminar de produção de energia eólica offshore do litoral brasileiro .…”