Data-Driven Estimation of Failure Probabilities in Correlated Structure-Preserving Stochastic Power System Models

We propose a data-driven approach for propagating uncertainty in stochasticpower grid simulations and apply it to the estimation of transmission linefailure probabilities. A reduced-order equation governing the evolution of theobserved line energy probability density function is derived from theFokker--Planck equation of the full-order continuous Markov process. Our methodconsists of estimates produced by numerically integrating this reducedequation. Numerical experiments for scalar- and vector-valued energy functionsare conducted using the classical multimachine model under spatiotemporallycorrelated noise perturbation. The method demonstrates a more sample-efficientapproach for computing probabilities of tail events when compared with kerneldensity estimation. Moreover, it produces vastly more accurate estimates ofjoint event occurrence when compared with independent models.