Voltage monitoring based on ANN-aided nonlinear stability analysis for DC microgrids

Due to the low inertia of the DC microgrid, the DC bus voltage is prone to drop or oscillate under disturbance. It is also challenging to supervise the stability of a DC microgrid since it is a highly nonlinear dynamic system with high dimensionality and randomness. To tackle this problem, this paper proposes a new method using ANN-aided nonlinear dynamic stability analysis for monitoring the DC bus voltage, which is combined with two steps. The first step is to establish six corresponding nonlinear accurate discrete iterative models of six switching modes of the PV-battery-load-based DC microgrid system, based on the Poincare map theory, in order to judge the stability quantitatively with a promoted stability margin index. The second step is to use artificial neural networks (ANNs) to forecast the operating mode of the system when random changes occur in environmental circumstances and load power; this will aid the first step in being efficient and adaptable while determining stability cases. And the employed ANNs are trained with the datasets, including the circuit data, ambient temperature, irradiance, and load power, which are generated by MATLAB/Simulink simulation. Theoretical and simulation analyses are carried out under different operating conditions to validate the proposed method's efficacy in judging the DC microgrid's destabilizing oscillation and stable running.