Resumo:
This thesis deals with voltage collapse phenomenon in power systems. To identify the
critical buses, bifurcation theory is employed. The load flow model adopted here allows a
better identification of saddle-node bifurcations.
A computational program was developed using the “Continuation Method”. This
program is shown and discussed, with special attention to the performance of the
computational method. A brief description of load flow “Decoupled Method” is also
shown. This load flow resolution method is incorporated into the “Continuation Method”,
and all the advantages and disadvantages are widely discussed. In particular, all the
aspects related to accuracy and the computational times performance are analyzed.
Theoretical aspects are also discussed. This discussion is due to the real need in reducing
the computational time associated with the “Decoupled Continuation Method”. For this
sake, stopping criteria are tested, and they are all based in the identification of the null
eigvalue at the saddle-node point.
The influences of LTC adjustments are also studied, and the effects in the load margin
and critical buses identification are addressed.
A methodology to assess the most severe contingency outages is proposed and the
“Continuation Method” is used as a benchmark.
Finally, the “Continuation Method” is used to plot the QV curve considering some
stopping criteria, and a discussion about the results is carried out. The gain observed
renders this application as effective.
