Load independent adaptive current control strategy for PWM inverters

A high efficiency neural implementation of a current control scheme for a three-phase VSI-PWM inverter is analysed. The new structure offers the advantage of a control which is independent of the load type and is only based upon an equivalent circuit of the real load consisting for each phase of a resistor, an inductor and a voltage source (R, L, e). The scheme is based on a predictive current neural controller using a state-space observer. The observer is able to reconstruct the non-inductive voltage of the load (Ri+e) on the basis of the currents and voltages across the load, while a PWM control signal generator produces the switching pattern for the transistors in the inverter. The inductance L is identified online by a neuronal inductance estimator which ensures optimal current control across the load