Model-Based and Model-Free Control of DC-DC Converters With High-Order Dynamics and Limited Measurements

This article introduces a control design framework for modern dc-dc topologies with high-order dynamics. In particular, model-based and model-free approaches using high-order controllers are introduced. The model-based approach permits the use of a minimum number of sensors, even for converters with a high number of components (e.g., multilevel, quadratic, input/output LC filter-based converters, etc.). This setting does not require estimation of state variables and its gain tuning can be achieved by selecting time-response specifications. The proposed model-free approach control exhibits the same, as well as some additional characteristics. Namely, controllers can be synthesized without the requirement of an explicit dynamic model, and gain tuning with guaranteed stability is directly achieved from measurement data. The latter controller is implemented in discrete time, which facilitates a digital implementation. The proposed approaches are validated through the control design and closed-loop implementation of a sixth-order topology.

Automated summary

This article introduces a control design framework for modern dc-dc topologies with high-order dynamics, including model-based and model-free approaches using high-order controllers. The model-based approach allows for reducing the number of sensors, eliminating the need for state variable estimation, and achieving gain tuning by selecting time-response specifications; the model-free approach enables controllers to be synthesized without an explicit dynamic model and gain tuning with guaranteed stability directly from measurement data.