Dynamic stability analysis of power grid in high proportion new energy access scenario based on deep learning

Abstract

The “carbon peaking and carbon neutralization” scheme formulated to solve the power energy shortage makes the new energy units incorporated into the power grid on a large scale. The randomness and volatility caused by the high proportion of new energy access have a certain impact on the dynamic stability of the power grid. Based on the dynamic stability theory, this paper applies the dual-stream CNN algorithm in the deep learning technology to model, takes the power of each node and line as the input and the key eigenvalue as the output, quickly identifies the main oscillation modes of the power system, makes a qualitative evaluation. Combined with the actual power grid analysis, the interval oscillation frequencies of the system are 0.57 Hz and 0.53 Hz respectively, which belong to strongly damped oscillation and can be quickly subsided after being excited. It can be concluded that the dual-stream CNN algorithm can provide reliable technical support for subsequent dispatching operation, and is conducive to improving the safe and stable operation level of power system.