A novel digital‐signal‐processor‐based maximum‐power‐point tracking control design for a vertical‐axis wind‐turbine generation system using neural network compensator

Abstract

This paper presents a novel maximum‐power‐point tracking algorithm of a vertical‐axis wind‐turbine (VAWT) generation system using neural network compensator on the basis of a digital signal processing chip. First, the mechanical power versus rotation‐speed curve of a 3‐kW VAWT generator was constructed by means of the measurement data in various wind speeds. Because the slope of the wind‐turbine output power is a non‐linear function of the generator output current, air density, wind speed, and the characteristics of the wind‐turbine generator, the analytic solution for the maximum power is difficult to obtain due to its complexity, nonlinearity, and uncertainties of parameters. The innovation of this study is to obtain the maximum power of the wind‐turbine system using a recurrent neural network (RNN) by transferring the maximum‐power‐point tracking problem into a unit‐step‐command regulation problem despite the variation of the wind speed, ambient air density, and the load electrical characteristics. The uncertainties in the system are compensated by the RNN, which is composed of three‐layer networks with three neurons and feedback loops in the hidden layer for capturing the characteristics of the wind‐turbine generator system. Without the necessity of wind speed sensor, it only needs the generator output current and rotation speed as the inputs of the input layer. This structure containing only six neurons totally is simpler than the existing work. The duty cycle of the boost converter is determined by a proportional‐integral controller, the parameters of which are determined via a genetic algorithm by the minimisation of a performance index function with the help of MATLAB simulation tool. From the simulation and experimental results, the validity of the proposed control scheme was verified under various wind speed conditions. The performance efficiency and the tip speed ratio in each case of the experiment is near the optimal values of 0.28 and 4.3, respectively. The output power achieves 2048.4 W at the wind speed of 11 m/s in the steady state.

Uncertainty analysis of wind speed of wind turbine

“…Ghaedi et al 4 considered the dependence of different component failure rates on wind speed changes, and used Monte Carlo simulation to evaluate the reliability of power systems including large wind farms. Tsai et al 5 used the measured data under different wind speeds to build the mechanical power and speed curve of the 3 kW VAWT generator, converted the maximum power point tracking problem into the unit step command regulation problem, and used recurrent neural network (RNN) to obtain the maximum power of the wind turbine system. Mohammad et al 6 summarized that a kind of uncertainty refers to the lack of certainty in probability assessment.…”