A fitting procedure for probability density functions of service restoration times. Application to underground cables in medium-voltage networks

Abstract

Distribution companies have the responsibility to provide a quality service to their customers, according to the existing regulation. Reliability issues, such as power outages, are registered in databases for a quantitative evaluation of this quality. This paper uses one of these historical records to make a statistical analysis of service restoration times, applied to the particular case of underground cables in medium voltage networks. An algorithm is proposed to fit the raw data to the probability density functions typically used in reliability analysis. The best-fitted distribution is determined in each case according to the information provided by a set of goodness-of-fit tests. Different groups are considered for the elements of the systems, concerning their functionality and voltage level. The presented procedure is applied to an electrical network with more than 350 feeders. Results have been obtained globally, showing that the observed service restoration time is lower than the estimated maximum limit in 98.00% of cases. The probability functions provided by the proposed algorithm can be used to improve the accuracy of the reliability models for the electric power system.