Extending Stable Parameter Range of LCL- filters for Grid-Connected Converters by Inherent Damping of Model Predictive Control (MPC)

Conventional linear controllers with active damping (AD) cannot stabilise LCL-filters for certain ranges of filter parameters due to resonances. This limits the design space of the filter design and can lead to sub-optimal designs in terms of filter volume and efficiency. When model predictive control (MPC) with a long-prediction horizon is applied, the inherent damping capability of MPC can stabilise the system despite the resonance phenomenon and expands the design space. In this paper, the limited design space due to insufficient damping of linear controllers with AD is analyzed based on capacitor current feedback AD. Exemplary LCL-filter optimization solutions and the resulting Pareto fronts are presented to illustrate the achiev-able volume and/or loss reductions with the extended design space. A novel MPC formulation is proposed to adopt explicit MPC and to enable a real-time implementation even at a high switching frequency. Simulation results validate the presented analysis.