Digital control techniques for switching power converters

Digital control methods for switching power converters offer greater robustness, more flexibility to changing operating characteristics, and better system performance than conventional techniques, which are often model-limited and only work well in a small range of conditions. Digital controllers are broadly classified into five generations, from 0 through 4. Generation 4 methods, such as the three techniques proposed in the present work, use new system formulations to achieve advanced control objectives. The first proposed technique is a singular perturbation analysis that provides a theoretical foundation for time-scale separation. If a buck, boost, buck-boost, or flyback converter meets a simple requirement, then inductor current operates on a fast time scale while the capacitor voltage changes on a slow time scale. This separation enables other control techniques. The second new technique employs a Kalman filter to create a sensorless power factor correction (PFC) controller. The proposed method uses voltage measurements in a switching power converter to eliminate the need for current sensing. An experimental converter that meets regulatory requirements validates the system. Finally, an online optimization method, discrete-time ripple correlation control (DRCC), is shown to automatically operate a switching power converter at an optimal point, such as maximum power from a source. DRCC is derived, stability is proven, and an application to a photovoltaic system is demonstrated experimentally. These three techniques together form a toolbox for future control applications.