Random Space Vector PWM for Acoustic Noise Cancellations in Voltage Source Inverter PMSM Electrical Vehicle Drive System : Random Space Vector PWM for Acoustic Noise Cancellations in Voltage Source Inverter PMSM Electrical Vehicle Drive System

Abstract

Pulse Width Modulation (PWM) for Voltage Source Inverters (VSI) has grown in recent years. For any motor drives with low common mode voltage and high DC-link voltage usage, a new age PWM methods have been developed in current era. In PWM inverters the external noise tricky. To suppress those multifrequency noise the chokes/filters have been proposed and connected in the outside of the VSIs. However, these methods suggest higher installation drive cost. Despite the fact that the suppression filters are easily get damage and it has its own age limit. In EV drive applications, this technique uncovers extensive promises. Therefore, the interest of EV manufactures is to find out the solution through inverter PWM itself. The random pulse width modulation (RPWM) is the most promising method amongst all the PWM methods as it has owning characteristics to avoid EMI noises [1]. The RPWM is the ideal PWM for handling induction motor drives because of its simple approach and inexpensive cost, making it one of the most efficient PWM schemes. In other side, space vector PWM(SVPWM) is superior in terms of inverter DC-link voltage utilizations. Likely all recently developed PWMs are having their own advantages and boundaries. Considering the EV drive requirements; they need high DC-link (battery voltage utilization), lesser harmonics and lesser noises. Hence the new interest to hybrid the PWM methods is getting famous these days. The most widely used technique was the random pulse position pulse width modulation, which involved changing the switching events in an operational frequency cycle at random with the modulated carrier frequency [2,3]. If SVPWM is utilised to incorporate RPWM, the alteration is accomplished by the researcher by modifying the slope of the carrier triangle or the angle of the space vector reference [4]. The random signal is employed in RSPWM instead of the carrier triangular wave form modification to provide the switching action. There are PWMs is suggested through spreading multifrequency harmonics. Nevertheless, these methods ignore the impact of acoustic noise and the use of inverter DC-link voltage. The generating the asymmetric carrier wave is important for RPWM and it is digitally implemented without the use of an external circuit [5].  More unique states are required to generate a random bit number, which necessitates a larger number of consecutive digital states. However, as the number of digital circuits grows, therefore does the expense and difficulty of implementation. As a result, the researcher suggested using a linear feedback shift register (LFSR) to increase randomization. The LFSR code generator, also known as a pseudo-PWM code generator, operates by leveraging a binary number's digital logical process.

    This paper focuses to use RPWM and SVPWM combinedly (called RSVPWM) to generate the PWM to reduce the acoustic noise and more unitization of DC-link voltage. The SVM works using multicarrier (carrier waves with distinct fixed frequencies) that are chosen using a random binary bit generator. The suggested RSVPWM generates pulses with a randomised triangular carrier (10 ± 2.5 kHz), whereas the traditional RPWM approach uses a fixed frequency triangular carrier and a random pulse location. This is work, the random binary is generated using two FPGA-based PRBS bit (8 bit and 16 bit) generators. The SVM was created using the same FPGA controller and uses the PRBS binary selector block to generate random carriers. A three-phase VSI linked 3 HP, 250V Battery, EV grade PMSM motor is used to validate the proposed PWM performance.  A six switch (Power MOSFET – SCH2080KE) inverter power module (IPM) is used as an inverter drive to evaluate the experimental feasibility of the developed RSVPWM. The simulation and hardware findings reveal that the VSI and motor function similarly to the traditional MCRSVPWM, however the noise power spectra of current, voltage, dominant harmonic components, and acoustic noise spectra are lowered when compared to previously published RPWM approaches.