Novel control strategy for three-phase AC/DC PWM rectifier

A novel direct adaptive control strategy using a B-spline neural network for the three-phase AC/DC (alternating current/direct current) voltage source pulse width modulation (PWM) rectifier is proposed, which realizes a sinusoidal AC input current and unity power factor. The control law is constructed based on the B-spline neural network directly, and the training of the network is on-line. An on-line learning algorithm is proposed for the neural network, and the space vector modulation method is then used to determine the switching instants of the rectifier, which makes the whole control strategy more suitable for real-time application and adaptive situations. The whole control strategy is also realized based on the digital signal processing (DSP) technique. Computer simulation and experimental results are both used to verify that the system with the proposed control strategy not only guarantees the good stability, but also provides good transient response and unity power factor.