A Unified Controller for a High-Frequency-Link (HFL) Inverter

High-frequency-link (HFL) inverter topologies featuring dc-link-capacitor-less solution have gained tremendous significance due to their higher power density and modular design yielding a superior solution to several conventional inverter approaches for renewable energy systems. Among the HFL inverters, those with pulsating-dc links eliminate the need for bulky dc-link capacitors in contrast to the HFL fixed-dc-link inverters. However in isolated topologies, operating without a buffer stage, performing different control actions such as voltage regulation, MPPT, and loss mitigation in different stages of the inverter independently leads to degraded overall performance. In addition, stiffness of the HFL inverter is reduced against source / load transients because there is no intermediate energy storage. To address the first issue, a most feasible solution is to perform all of the control actions in a single stage using an optimization-oriented algorithm. Regarding the second issue, a high-bandwidth power conversion is required to provide fast enough dynamic response for the power stage. However, this requires a high switching frequency, which enhances the switching loss of the inverter. Yet another approach is to control the switching sequences of the inverter in such a way so that an optimal dynamic response is achieved without degrading the efficiency. In this Dissertation a unified controller is proposed, which simultaneously meets different control-performance requirements at disparate time scales (i.e., slow- and fast-time scales) under wide operating conditions. The key advantage of the proposed control over conventional model-predictive and sliding-mode controls is the ability of the proposed control to restrict the control search space to only reachable switching sequences, which are obtained using a composite-Lyapunov function. This feature enables the proposed controller to dynamically change the switching sequences of the inverter under stability bound. Further, a reduction in the search space for control sequences reduces the time to execute the optimal switching-sequence-based controller. Finally, the feasibility of the inverter when subjected to device fault is also demonstrated using a modulation based switching-sequence control.