Bayesian spatial reliability model for power transmission network lines

The paper is devoted for the problem of assessment of the dependence of power transmission lines outages on the geographical position in the network. Authors present the model based on Poisson-gamma random field for the purpose of capturing the information about the strength of correlation between the power transmission network outages and geographical positions. The model presentation is followed by the analysis of the model and application to the North American power grid. The main finding is that outages of the power grid are geographically correlated and this effect should be incorporated into the overall reliability assessment of the grid to obtain more accurate estimates. It is proved in this paper, that taking into account dependency of line outages on the geographical location, enables to obtain more accurate results as compared to classical Poisson and to hierarchical Poisson-Bayesian models.