Non-Invasive Closed-Loop System Identification of an Active Rectifier

Abstract

Industrial drive systems are characterized by dynamic peak loads and regenerative braking. These short-term loads lead to oversized supply infrastructure and low conversion efficiencies for prevalent partial loads. Energy storage systems (ESS) can be integrated into the DC link between rectifier and multiple inverters to increase energy efficiency and cover peak loads. To design a controller for the ESS, the underlying controller and process have to be modelled using physical laws and measured data. However, for each application this would mean having to extensively tune and adapt the controller to a specific manufacturing machine to ensure a stable, safe and energy-efficient operation. A self-tuning controller could solve this problem by automatically identifying the process and controller model of the closed-loop system. This paper aims to implement non-invasive closed-loop system identification for retrofitting an energy storage system to a machine tool. The validation of the approach shows that the actual DC link capacity and the controller parameters of the active rectifier can be identified reliably, e.g. with deviations of only 1.5 %.