Neural network-based controller for voltage PWM rectifier

In order to meet more stringent distortion and power factor requirements in AC/DC rectifiers, PWM voltage source topologies have been proposed as an alternative to phase controlled rectifiers. However, the characteristics of these power converters are nonlinear and conventional fixed structure PI-type regulators cannot be optimized for all operating conditions. Neural network-based controllers, among others, can handle such nonlinearities. This paper investigates the feasibility and performance of an adaptive control scheme using neural networks for PWM rectifiers. The controller provides unity power factor. The investigation is carried out on a neural network controller implemented on a TMS320C30 DSP and driving at 1.5 kVA voltage source rectifier. Results show that close to unity power factor is obtained and that fast transient response is achieved