Research Projects

Specialized equipment in medical and space applications require compact power electronic converters for their power supply requirements. Fast switching is essential to achieve high power density. Such critical applications also require output voltages/currents having specific time-varying profiles with wide operating ranges, fast dynamic responses, unipolar/bipolar waveforms, and even pulsed waveforms in certain cases. Stringent regulatory norms exist for EMI and equipment performance in such applications. These requirements make the converter control highly challenging. Presently simple PI/PID controllers and ON-OFF controllers are used for such applications due to implementation simplicity. However, they have issues at wide operating ranges and have variable switching frequency respectively. Predictive control is suitable for such multi-objective control requirements and fast responses as compared to most other control techniques. However, the inherent issues with predictive control namely non-zero steady state error, model accuracy dependency, poor robustness to parametric variations, and variable switching frequency have limited its wide use in critical power converter applications. The optimization step of the cost function is computationally demanding, limits the achievable switching frequency and constraints the prediction horizon to two or three typically. Presently, predictive control is challenging for 50kHz switching and beyond. Therefore, the existing predictive control framework is not feasible for most high-frequency power converters (500kHz and above) in critical applications, despite its potential for multi-objective control. The recent works by the PI and his team have attempted to address a few of these challenges in predictive control of a three-phase Vienna rectifier system, and the results have shown promise. Motivated by these results, this proposal aims to develop a novel predictive control framework for high-frequency power converters suited for critical applications with challenging control objectives. Two novel approaches will be explored to develop the proposed control framework. Fixed-frequency switching operation will be ensured using an appropriate pattern of the control actuation. Considering the digital implementation aspects, the loop delays, and the limited available time for computations at high frequencies, the stability of this proposed framework will be explored using time-delay system concepts such as Lyapunov Krasovskii functionals. The proposed work is divided into three-stages. In the first stage, the proposed control framework will be developed analytically and validated in simulations for various power converters. In the second stage, two converter prototypes will be developed using SiC/GaN devices and experimentally validated for up to 500kHz using the proposed control - a three-phase AC to programmable bipolar DC output isolated converter with PFC capability, and an H-bridge DC-DC Converter. The first converter is a programmable power supply, while the H-bridge converter will be used to validate standard configurations of buck, boost, and buck-boost operations. Stability studies on these converters using time-delay system approaches will be done. In the final stage, the proposed control framework will be modified to achieve up to 1 MHz switching frequency in the converters and will be experimentally validated. The last stage is expected to be highly exploratory in nature. The outcomes of this research are expected to address the current challenges of predictive control in power converters, thus increasing its applicability for high-frequency converters and critical power supplies with stringent control requirements. Further, it will pave way to future indigenous technology development of specialized high-performance power supplies for medical, space and other critical applications in India.