COMPEL 2020: Performance Limits of Differential Power Processing

Conference: COMPEL 2020
Presenter: Ping Wang
Abstract: This talk explores the performance limits of differential power processing (DPP) for large-scale modular dc energy systems with stochastic loads. An analytical stochastic model is developed to estimate the average power loss of a DPP topology under probabilistic load distributions. A scaling factor S[.] is introduced to describe how power loss scales as the system size or load power variance increases. The average power losses of several example DPP topologies are analyzed and compared against conventional dc-dc converters given the same total switch die area and magnetic volume. The performance limits for various DPP topologies are derived and verified by Monte-Carlo simulations in SPICE, and the results indicate that the ac-coupled DPP converter stands out from all the representative DPP topologies discussed here in terms of the lowest power loss. The talk provides an analytical framework to evaluate the performance of different DPP topologies methodically, offering insights for the design of DPP systems with large-scale stochastic loads.