Improvement of shutdown scheduling to reduce energy not served on the MV network : A case study of the UMEME network

Abstract

This paper presents an algorithm that helps to determine the best times to carry out planned shutdowns on the network within a specific year for maintenance so as to reduce on the Energy Not Served (ENS) while at the same time reducing on commercial losses suffered by UMEME. The algorithm implemented in MATLAB uses historical loading data to establish patterns of power consumption for the customers in order to determine periods of low consumption and taking advantage of these to be the periods to carry out the shutdowns. The case study used is PortBell substation and its 11kV feeders which is part of the UMEME MV (medium voltage) network. It was chosen as the case study because of the nature of the loads connected to it as it was preferred to have a case study that had all the different types of consumers connected to it in order to be able to analyze the different loading patterns. Loading information for the years 2016, 2017 and 2018 is used. The algorithm uses loading values of 2016 and 2017 to determine the most favorable days to shut down in 2018. Energy Not Served (ENS) and ECOST for the new suggested days is compared with the base case scenario, which in this case are values of ENS and ECOST attained after having used the current scheduling system at UMEME. ENS and ECOST is minimized and commercial losses reduced as compared to the base case. The algorithm has an additional function of scheduling activities for the year 2019 using loading values for the 3 previous years 2016, 2017 and 2018.