DSP implementation of neural network-based controller for voltage PWM rectifier neural

PWM voltage source rectifiers feature high power factor and low line current harmonic distortion, and can therefore meet the more stringent requirements proposed for rectifiers, unlike thyristor line commutated rectifiers. However, they exhibit no-linear characteristics and conventional linear controllers cannot be optimized for all operating conditions, particularly if parameters change with time or are ill-defined. This paper shows that neural network based controllers offer a number of advantages, among which are: (a) adaptive features; (b) fast and consistent transient responses; and (c) close to unity power factor. Experimental results on a 5 kVA rectifier controlled by a TMS320C30 DSP are used to illustrate the feasibility of the proposed control techniques.