Stochastic dynamic programming approach for proactive replacement of power cables

The purpose of this paper is to present a stochastic dynamic programming based model to solve the optimization problem of cable replacement. The proposed methodology can be implemented on cables with known failure distribution and insulation degradation level; the methodology to estimate both of the elements is based on previously developed Non-homogenous Poisson Process model (NHPP) and stochastic degradation model, respectively. The model gives the sequence of decisions for each year of the planning horizon such that it optimizes the overall cost and improves the reliability by lowering the frequency of unplanned outage. The model was tested on an unjacketed XLPE cable.