Gradient Descent Optimization Based Parameter Identification for FCS-MPC Control of <i>LCL</i>-Type Grid Connected Converter

Aging and temperature changes in the passive components of an LCL -filter grid connected converter system (GCCs) may lead to parameter uncertainties, which can turn influence its modeling accuracy for finite-control-set model predictive control (FCS-MPC). The presence errors will change resonance point deteriorate power quality current, degrading active damping performance. In this situation, there is a serious possibility that GCCs malfunction automatically disconnect from grid, causing great challenges stability. To solve problem, first, prediction error analysis FCS-MPC due presented. Second, achieve high fast filter estimation utility, adaptive online identification method based on gradient descent optimization (GDO) has been proposed. Finally, further reduce searching time needed by optimal iteration step, variable step root-mean-square-prop (RMSprop) GDO Experimental studies -GCCs prototype laboratory have conducted validate effectiveness proposed method.