A Data-Driven Approach to Power Converter Control via Convex Optimization

A new model reference data-driven approach is presented for synthesizing controllers for the CERN power converter control system. This method uses the frequency re- sponse function (FRF) of a system in order to avoid the problem of unmodeled dynamics associated with low-order parametric models. For this particular application, it is shown that a convex optimization problem can be formulated (in either the H∞ or H2 sense) to shape the closed-loop FRF while guaranteeing the closed-loop stability. This optimization problem is realized by linearizing a non-convex constraint around a stabilizing operating point. The effectiveness of the method is illustrated by designing a controller for the SATURN power converter which is used in the Large Hadron Collider, in injector machines, and for pulsed applications at CERN. Experimental validation in the frequency-domain is also presented.