A tailored ADMM approach for power coordination in variable speed drives

Abstract

Variable speed drives (VSDs) are AC/DC/AC power converters that drive electric machines at variable frequency. In a VSD the electric energy is first rectified to DC, then stored in a DC-link, and finally inverted back to the desired AC signal. The energy stored in the DC-link can be kept within tight bounds when the rectifier and the inverter coordinate their actions. This is achieved by utilizing model predictive control (MPC), which translates the control task into a sequence of optimization programs. In this paper we show how to utilize the structural properties of the underlying system to solve these optimization programs efficiently. We decompose the VSD into the rectifier, the DC-link and the inverter, and we use the alternating direction method of multipliers (ADMM) to exploit this decomposition. We show that our proposed method behaves favorably when compared to an interior point method, or a conventional ADMM approach.