Deep Learning Based Transient Stability Assessment for Grid-Connected Inverter

Droop control based grid-connected inverters are facing with transient stability problem. At present, the major analysis method for that problem is time-domain simulation, which can get a result in minutes. However, that time consumption can't meet the on-line requirement. In this paper, deep learning theory is applied in transient stability assessment for grid-connected inverter as a data driven framework. Testing results show that the system's stability conclusion can be obtained in microseconds and the prediction accuracy can reach more than 99%. And the effectiveness of proposed algorithms is verified by specific simulation cases.