Neural network basedpipelined-parallel generation of PWM signals suitable to prive three single phase UPS

Abstract EN

A reliable technique has been proposed to generate a real time pulse width modulation (PWM) signals in order to drive three – single phase uninterruptible power supply (UPS).

The PWM patterns have been generated using field programmable gate array (FPGA) and based on selective harmonic elimination method.

These patterns are used to drive the six of switching power transistors of the voltage source inverter to produce three– single phase UPS.

In order to solve the problem of the complexity of the nonlinear transcendental equations, an intermediate steps have been taken, using artificial neural network (ANN).

This will overcome the problem of the off line solution and therefore obtaining the required data to solve the obstacle solution from off line to on line.

Therefore, the trained ANN is implemented in a parallel hardware by using FPGA.

The benefits of using FPGA to perform ANN are promising and the technique becomes very attractive.

It allows a real time, simple, fast, reliable and efficient design with low hardware costs.

Finally generating selective harmonic elimination pulse width modulation (SHEPWM) patterns as a real time signals are become visible.