A mixed integer programming model for optimal fault section estimation in power systems

• A combinatorial two-stage methodology for fault section estimation is proposed.
• The MIP model provides a ranking of events as a guide to power restoration.
• MIP model provides robust solutions in the absence of alarms.
• Protection statistics are considered by local classifier to produce better solutions.

This paper proposes a methodology to deal with the fault section estimation problem in electrical power systems. The main motivation for this study lies in the fact that operators of control centers usually are subject to information overload during great contingencies. The fault diagnosis is formulated as an optimization problem and solved through two stages: classification of events at equipment level and the fault section estimation. The first stage consists of a heuristic based on Bayes’ theorem that provide the input to the second stage where a mixed integer programming model is solved by commercial package in order to determine the faulted section. Several fault scenarios from a real Southern Brazilian subsystem were considered to validate the methodology. The results show that the proposed method can correctly identify the faulted section even in case of multiple faults or in case of improper operation of protective devices.