Switching Loss Reduction of Half-Bridge ZVS-PWM Inverter with Artificial
Neural Network Controller
Abstract
In the development of power electronics, stationary power converting
devices like AC-DC converters and DC-AC inverters are used in various
industrial areas. The performance of DC-AC inverter is enhanced using
some controller techniques like PI, PR and PID, however it contains some
limitations like high complexity and high switching loss. The
application ranges of Sinusoidal Pulse Width Modulation (SPWM) are
wide-ranging, and it is essential to reduce its switching losses (SWLs),
also it includes higher switching frequencies, high frequency harmonics
generation and high attenuation. Moreover, the hard switching mechanism
used by the traditional inverter design suffers from a more significant
power loss, and it seems difficult to operate under high frequency.
Thus, this paper presented a new design named as an Artificial Neural
Network (ANN) Controller with soft-switching (SS) technique to minimize
the switching loss and enhance the converter efficiency. The zero
voltage switching- pulse width modulation (ZVS-PWM) inverter is designed
and the inverter gate pulse is tuned by ANN controller for reducing the
switching losses and harmonics. The advantage of ANN is the ability to
work without knowing the switching condition. For that, ANN controller
is used here for reducing the switching loss in ZVS-PWM inverter. The
MATLAB platform is used to implement the proposed work and the
performances are analysed in terms of input current, voltage, THD,
efficiency and so on. The proposed design results are compared with
other existing works to exhibit the superiority of the proposed inverter
design.