Research Projects

In this work, a 5 kW experimental prototype of a multilevel multiphase IM (ML-MPIM) drive setup is to be developed for an Electric Vehicle (EV) application. Due to multiphase IM in EVs, a high power or torque density, improved motor efficiency, and better fault-tolerant capability can be achieved at a reduced per-phase stator current. The literature review shows that a six-phase motor is better in terms of lower line voltage and stator current requirements, improved power density, efficiency, and reliability than a five-phase motor. Further, using a multilevel inverter (MLI) compared to the conventional two-level inverter enhances the stator voltage quality (i.e., lower harmonics and THD) even at lower switching frequencies and reduced switches' dv/dt stresses can improve inverter efficiency. Thus, a multilevel multiphase inverter can give the best results, such as better fault-tolerant capability, higher power density, better system efficiency, smaller torque ripples, lower DC-link capacitors, etc. This work selects a three-level six-phase inverter for a six-phase IM, and its performance is compared with a similar five-phase system for cost and complexity analysis. Under normal operating conditions, a Space Vector PWM (SVPWM) with a direct torque control (DTC) technique is developed for the multiphase induction motor (MPIM). The DC-link voltage can be adjusted to control the speed of the motor so that the actual AC voltage at the stator phase and line voltage of the MPIM is a staircase nature with a maximum number of voltage levels to maintain the good quality voltage and reduce harmonics even at low speed. Moreover, optimized stator flux control algorithms have also been developed for MPIM drives to minimize the rotor core losses, which enhance motor efficiency considerably, especially under light load operations of the EV. The fault identification in the inverter switches and phases using AI and ML enhances complex systems' reliability, efficiency, and safety by leveraging advanced computational methods to detect, classify, and predict faults in real time. The AI/ML technique involves processing large volumes of data from sensors and operational logs to identify patterns, anomalies, or deviations that indicate system malfunctions. AI and ML techniques enable automated, accurate, and scalable solutions for fault diagnosis, reducing downtime, improving maintenance strategies, and minimizing the risk of critical failures across applications like electric powertrains and multilevel inverters. The project is also expected to develop fault-tolerant control algorithms for MPIM drives under open-phase faults. The proposed work aims to improve the post-fault performance of MPIM driven by the DTC strategy under faulty operations that ensure smooth and reliable operation, reduce torque ripple, and maintain high system efficiency. Finally, a low-cost prototype design is needed for cost-effective product development for Indian EV applications.