ConvLSTM based Real-time Power Flow Estimation of Smart Grid with High Penetration of Uncertain PV

Abstract:
A modern smart grid tends to have increasingly various uncertain renewable generations. The power flow estimation using existing methods with a large number of measurements in real-time could be time-consuming and computationally expensive. This paper proposes an efficient deep learning approach, a convolutional long short-term memory (ConvLSTM) model, to estimate the real-time power flow of the smart grid considering the high penetration of uncertain PV generations and synchrophasor data. The performance and effectiveness of the proposed ConvLSTM with cross-validation techniques are examined with the modified IEEE 14 bus test system. The state measurement is employed to obtain the accurate actual power flow results for offline training. The proposed ConvLSTM with time-series cross-validation provides more accurate results than DNN and LSTM models in three uncertain PV scenarios.