Resumo:
There is a permanent demand for new computational tools required for
power system operation, supervision, analysis and planning. These software’s
become larger and increasingly complex due to the system’s characteristics, and
when finally finished they are difficult to maintain and to integrate into old
application.
The main objective of this thesis is to consider a new type of multiagent
system’s architecture, capable of to assist the power systems operators in
their work of operation, with special approach for the restoration tasks after a fault.
The multi-agent system’s architecture was divided in two packages of
agents. The first one is composed by "Support Agents", which are responsible for
doing services that assist the decision-making process, as: to establish and to coordinate
the communication between agents, to update the computational model,
interface with the real world, etc. The second package is composed by "Specialized
Agents", that have the specific knowledge used in the analysis and solution of the
problems found in the application domain.
The main benefits obtained with the multi-agent model were:
cooperative behavior, competitive processing, distributed topology and open
architecture.
The cooperative behavior is characterized by dynamic exchange of
information between agents making possible segmentation of tasks, besides allowing
the reduction of the hierarchic characteristics of the decision process. The
competitive processing improves the use of the hardware capacity and decreases the
reply time.
The architecture’s topology considered is distributed, what makes
possible in the abstraction process the system’s division in a lesser number of parts,
limited to functional characteristics. The architecture considered is opened, allowing
the addition of components and the system’s expansion, as part of a bigger system
with the creation of new levels of analysis.
As an example of application of the architecture considered it was
implemented in real cases like power substations and distribution systems of Brazil’s
power utility companies.
