Long-Term Recurrent Convolutional Network-based Inertia Estimation using Ambient Measurements

Conventional synchronous machines are gradually replaced by converter-based renewable resources. As a result, inertia, an important time-varying quantity, has substantially more impact on modern power systems stability. The increasing integration of energy resources imports different dynamics into traditional systems; therefore, the estimation system inertia using mathematical model becomes difficult. In this paper, we propose novel learning-assisted based long-term recurrent convolutional network (LRCN) that uses wide frequency and phase voltage measurements. proposed approach non-intrusive probing signal to perturb collects ambient measurements with phasor measurement units (PMU) train LRCN model. Case studies conducted IEEE 24-bus system. Under signal-to-noise ratio (SNR) 60dB condition, achieves accuracy 97.56% mean squared error (MSE) 0.0552. Furthermore, low SNR 45dB, is still able achieve high 93.07%.