Evaluation of Impact of Low Discrepancy Sequences on Predictive Reliability Assessment of Distribution System

Abstract

Reliability evaluation of power distribution systems is a key aspect to assess system performance in terms of interruptions. Probabilistic evaluation methods are widely used for reliability analysis to handle uncertainties. These methods become computationally burden with increase in size of the power system. Quasi-Monte Carlo (QMC) is an advanced Monte Carlo (MC) method to improve the accuracy and computation time. This paper studies the impacts of Low Discrepancy Sequences (LDS) on sampling of failure and repair rates in Monte Carlo simulation (MCS) based approach. Low Discrepancy or Quasi-Random sequences samples the failure states more uniformly than a pseudo-random sample. This study investigates the reliability performance of the distribution system for Van Der Corput (VDC) and Halton sequences. The proposed Quasi Random Monte Carlo Simulation (QRMCS) algorithm is validated with analytical and MCS methods using IEEE RBTS test system. Further the predictive reliability assessment is carried out for a practical 11kV Indian radial distribution system. Results demonstrate that the QRMCS method converges faster than MCS for a specific level of accuracy.