Design of a LQR-Based Boost Converter Controller for Energy Savings

Digital power management systems allow power converters to operate with multiple changes in voltage reference. This asks for a new paradigm that shifts from achieving faster response to achieving best efficiency during multiple transients, with transient being eventually time constrained. State-space-based design of the control system fits perfectly this goal, since it is a time-domain design tool. It can further be enhanced with a linear quadratic regulator (LQR) optimization, configured with converter loss equation as cost function. The LQR mathematics guarantees a solution to the algebraic Riccati equation that produces the best system efficiency for the converter system during multiple transients. This article documents the design of the LQR control system having converter loss as a cost function and demonstrates the advantages of this method. The closed-loop control software yields very simple, running with a 250 kHz sampling frequency capability on a general use Microchip microcontroller platform. While a loss reduction of 2.5% is demonstrated by experiment in our setup, the advantage depends on the actual system work profile.