Tesis
Heurística matemática para resposta emergencial no sistema de distribuição de energia elétrica com ilhamento imprevisto
Fecha
2018-02-23Autor
Schmitz, Magdiel
Institución
Resumen
The main objective of the electric power utility is the uninterrupted supply of energy,
however, this is not always possible due to the actions of nature, or even from equipment
or human failures. These interruptions are called emergency orders, which must be
serviced by repair crew in order to restore normal operation of the electrical system. Given
a set of emergency orders, the operations center must find a crew scheduling in order to
attend all orders and minimize the total cost of service. Before the system repair step, the
restoration step is performed, in order to supply the de-energized loads belonging to the
healthy parts of the network. One of the ways to supply them is through distributed
generation (DG), such as solar and wind, with high penetration in the electric system in
recent years, operating in islanded mode. The island operation plays an important role in
the restoration step of the electric power system, as it allows to reduce the total
interruption time. It allows analyzing the impact of customer cost, with interrupted power
supply, depending on the time of day, varying with the capacity of the DG generation curve
and the load curve. In this way, this work proposes a new matheuristic for emergency
response for the emergency orders scheduling problem (EOSP) of the electric distribution
systems, considering unforeseen islanding. The model has the objective of minimizing the
total attendance cost, being for this model the costs of energy not supplied and those
related to the penalties, associated to the importance of consumers. The matheuristic will
serve as a management tool for decision-makers to determine an appropriate scheduling
strategy. In this work the computational implementation of the proposed model is
presented, with the realization of two case studies in the IEEE 33-bus system, modified
with DGs insertion. Computational results show the effectiveness of the model developed
by reducing the total attendance cost in relation to the case without islanding and the
feasibility for applications in other distribution systems.