Data-driven multi-inverter cooperative control for voltage tracking and current sharing in islanded AC microgrids

A data-driven cooperative control scheme utilizing additional local and network sensor information is proposed for voltage tracking and current sharing under a cyber-physical system framework. An information-based inverter controller is developed as a replacement for the physical mechanism or model-based method because of uncertainties related to the inverter application scenario, system structure and specifications. To overcome the disadvantages of the reference voltage set-point strategy, an integrated cascade controller is adopted in which the outer loop is a learning controller with a noncausal filter, used to achieve high-precision voltage tracking, and the inner loop is a current consensus regulator implemented by means of nonlinear proportional-integral-derivative (PID) control, used to realize instantaneous value smoothing and output current consensus among a number of inverters. The designed controller does not rely on the plant model, structure or parameters and does not use the droop mechanism; it relies on fully data-driven and distributed (without root nodes or leaders) techniques. This controller design scheme can be uniformly applied for different kinds of distributed generation unit to achieve flexible and expandable microgrid deployment. Plug-and-play and load change capabilities, as well as tolerance to communication constraints, are also verified on the MATLAB/Simulink platform.