Application of neural networks to the optimal control of three-phase voltage-controlled inverters

A novel concept of application of neural networks for generation of optimal switching patterns in voltage-controlled inverters is described. The pulsewidth modulator (PWM), based on hardware analog neural networks, responds with high accuracy to any desired value of the modulation index. No microprocessor or large memory typical for the existing optimal PWM schemes is required. Theoretical analysis of the proposed so-called sparse neural network is provided, and experimental results that confirm the high performance and technical advantages of the developed modulator are presented.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>