Adaptive directive neural network control for three-phase AC/DC PWM converter

A novel control strategy is proposed: adaptive B-spline neural network for three-phase AC/DC voltage source converters, which realises a sinusoidal AC input current and unity power factor. Compared with other PWM techniques, neural network control provides an excellent component of a nonlinear system and is adaptive enough to fit the environment change. An on-line B-spline neural network is used because of its local weight updating characteristic, which has the advantages of fast convergence speed and low computation complexity. This is very important for real-time control applications. Both simulation and experimental results are presented to verify the effectiveness of the proposed control strategy.