Esquema de alívio de carga por subfrequência visando a menor perda de receita

Abstract

The Regional Load Relief Scheme (ERAC) is made from the underfrequency protection relay, this scheme acts as protection for the Electric Power System (SEP) since sustained under frequencies in the electrical grid can cause blackouts or shutdowns of large power blocks. The ERAC is composed of five cutting stages in which a certain amount of load is disconnected from the SEP as a function of the value and time of the underfrequency. The problem analyzed in this research consists of the occurrence of percentage load cuts higher than the minimum required by the standard. This can occur due to the static method of selection and parameterization of the current frequency relay existing in electrical substations and disregard of the hourly variations of the load of each Automatic Recloser (RA). The conventional method for choosing loads for each ERAC stage consists of the manual selection of the RA that will be switched off for each stage. This work is justified by the dynamic search for the RAs to be turned off, assuming that the load shedding percentages can be closer to the limits stipulated by current regulations. The general objective was to choose the loads automatically, using an algorithm for monitoring the instantaneous loading of the electrical system, in order to reduce the revenue losses of power distributors, as well as to promote load flexibility. The methodology consisted of the optimized choice of the RAS seeking to decrease the revenue loss of the distributors and the turnover of the disconnected loads. The result proposed a better choice of loads in the disconnecting stages of each RA of several substations of a hypothetical micro-region for ERAC. The conclusion showed that the algorithm maintained adequate load shedding, not exceeding two percentage points (2%) of the shedding value requested by the standard and, for the most part, choosing the RAs within the time established by the National Electric System Operator (ONS). The main recommendation is to analyze the performance of the algorithm in cases of data collection and decision in real-time, using appropriate hardware for this application.