Improved interval method applied to large scale distribution system reliability evaluation

Five deconvolution formulae are deduced in order to obtain sharp bounds of the results of reliability studies. Deconvolution formulae overcome the dependency problem and form the theoretical basis for applying interval analysis to power system reliability assessment. Interval reliability models of Markov systems and an interval-analysis-based reliability evaluation method are proposed. A software package DSIR2002 (interval analysis based distribution system reliability evaluation software) with proprietary copyright has been developed. Extensive investigations on IEEE RBTS test system and a 573-bus distribution system demonstrate that the developed software is valid and effective. All results have shown that the interval-analysis-based approach can serve as a more efficient and faster reliability assessment tool than Monte Carlo simulation and can provide, in an unprecedented flexible manner, much more useful information than traditional sensitivity analysis. It provides a viable and easy-to-use tool for the distribution system planning, engineering and operations.