Neuro-fuzzy controlled hybrid PWM method for two-level inverter fed three phase induction motor drive

This paper proposes a neuro-fuzzy controlled three-zone hybrid space vector pulse width modulation (HSVPWM) method for a two-level voltage source inverter (VSI) fed induction motor. The three-zone HSVPWM improves the performance of induction motor compared to popular literature reported methods. Hence a pre-learned neuro-fuzzy controller (NFC) based HSVPWM is designed to directly obtain three phase duty ratios by taking rotation angle and change in rotation angle of reference voltage space vector (RVSV) as inputs. The purpose of NFC in designing three phase duty ratios is to improve the firing strength of the pulses fed to the two-level inverter compared to HSVPWM method by adjusting control gains at inputs and output for the NFC. Moreover, the practical implementation of HSVPWM method at higher switching frequencies is difficult because it requires sector identification along with zone of operation of RVSV. In NFC based HSVPWM, without identifying sector and zone of operation of RVSV, duty ratios are generated and is independent of switching frequency. Thus, it can be implemented practically for higher switching frequencies. Three-zone HSVPWM and NFC based HSVPWM methods are simulated and implemented in real-time using STM32F49I ARM cortex M4 processor at 5 kHz switching frequency for a IGBT based two-level VSI fed 1 HP induction motor.