This paper proposes a methodology based on the unconstrained binary programming (UBP) model and an optimization immune algorithm (IA) to estimate fault sections in electric power systems. The UBP model is formulated using the parsimonious set covering theory for associating the alarms of the protective relay functions informed by the SCADA (supervisory control and data acquisition) system and the expected states of the protective relay functions. The IA is developed to minimize the UBP model and to estimate the fault sections as quickly and reliably as possible. The proposed methodology is tested using part of the South-Brazilian electric power system. The control parameters of the IA are set to reach the maximum computational efficiency and reduction of the processing time. The results show the methodology's potential to estimate fault sections in electric power system control centers in real-time.