Design and Analysis of a FPGA-Based Controller for GaN-Based Three-Phase Inverters: Efficiency, Performance, and Reliability Optimization
Date Issued
2024-12
Author(s)
Huseynli, Mujgan
Abstract
The main objectives of the research thesis are the design, simulation, and analysis of a FPGA
controlled GaN based 3-phase grid-tied inverters in terms of efficiency and performance. High
frequency switching applications provide significant challenges for conventional inverters
based on MCUs due to their slow processing speeds, sequential execution, and increased
sensitivity to EMI. Despite their small size, efficiency, and need for precise control, inverters
based on GaN are fast. FPGA is proposed in this work as they can execute computations
quickly, adapt to unique hardware needs, and handle operations in parallel.
In order to analyze functionality of both FPGA and MCU, we employ FIL and PIL simulations.
Traditional MCUs have difficulties with high-frequency switching in PIL simulations because
of their high resource usage and long execution durations. On the other hand, based on FIL
simulations, FPGA is one of the best options for real-time control. System performance can be
improved by smart resource allocation. The study recommends sampling at rates higher than
the switching frequency to avoid distorted waveforms, erroneous control signals, and inferior
performance. FPGA-based waveform quality and PWM signal precision improve as switching
frequency increases. Finally, to enhance waveform quality and system efficiency, research
could explore advanced modulation techniques such as Space Vector Modulation (SVM) or
Model Predictive Control (MPC).
The subsequent phase of this research may involve transitioning from simulation to empirical
testing; this would need the assembly of the circuit using genuine GaN devices, drivers, and
passive components, followed by an assessment of the system's performance in a practical
environment. Integrating AI & ML methods for predictive control and real-time optimization
of inverter performance might also be a good option, and it might enhance not only system
efficiency but also stability. This research will help to demonstrate the potential of FPGA for
high-frequency GaN inverters, by proving reliable and efficient aspects of power electronics
and industrial applications.
controlled GaN based 3-phase grid-tied inverters in terms of efficiency and performance. High
frequency switching applications provide significant challenges for conventional inverters
based on MCUs due to their slow processing speeds, sequential execution, and increased
sensitivity to EMI. Despite their small size, efficiency, and need for precise control, inverters
based on GaN are fast. FPGA is proposed in this work as they can execute computations
quickly, adapt to unique hardware needs, and handle operations in parallel.
In order to analyze functionality of both FPGA and MCU, we employ FIL and PIL simulations.
Traditional MCUs have difficulties with high-frequency switching in PIL simulations because
of their high resource usage and long execution durations. On the other hand, based on FIL
simulations, FPGA is one of the best options for real-time control. System performance can be
improved by smart resource allocation. The study recommends sampling at rates higher than
the switching frequency to avoid distorted waveforms, erroneous control signals, and inferior
performance. FPGA-based waveform quality and PWM signal precision improve as switching
frequency increases. Finally, to enhance waveform quality and system efficiency, research
could explore advanced modulation techniques such as Space Vector Modulation (SVM) or
Model Predictive Control (MPC).
The subsequent phase of this research may involve transitioning from simulation to empirical
testing; this would need the assembly of the circuit using genuine GaN devices, drivers, and
passive components, followed by an assessment of the system's performance in a practical
environment. Integrating AI & ML methods for predictive control and real-time optimization
of inverter performance might also be a good option, and it might enhance not only system
efficiency but also stability. This research will help to demonstrate the potential of FPGA for
high-frequency GaN inverters, by proving reliable and efficient aspects of power electronics
and industrial applications.