Résumé:
The demand for electric energy has become increasingly enormous, while the rapid
growth of electricity production has continued, particularly with regard to renewable energy
sources (RES). Recorded RES distribution systems are becoming more important and should
play a bigger role in the near future.
Connecting RES facilities to distribution networks involves addressing several technical
challenges for grid operators, who perceive renewable energies as a source of complications
for the electricity system. Congestion is one of the most complicated issues that can arise
when integrating renewable energy into the electrical system that cannot be left without
treatment.
For that, this work examines a corrective method of an overloaded electrical system in
the presence of a wind farm integrated in a distribution network, taking into account the cost
of congestion. Attempts were done in order to eliminate the overloads and monitor the power
flow on the lines using as the first suggestion the UPFC universal power flow controller. The
second proposed methodology is the ELM Extreme Learning Machine. This algorithm is
extremely fast, having excellent performance in terms of generalization. The fundamental
point of this successful study is the relief of congested structures. In addition, other objectives
are achieved: improvement of the voltage profile, avoidance of load shedding, reduction of
losses and congestion cost. The proposed congestion management methods were tested for an
Algerian power system with 22 busbars at the city of Adrar. The results presented in this work
were discussed and compared. These results show an improvement in the behavior of the
electrical network.
