Page 295 - Decision Making Applications in Modern Power Systems

Particle swarm optimization applied Chapter | 10 255

10.7 Methodology for particle swarm optimization
application to reactive power dispatch considering
tangent-vector-based generation selection
In this section, the methodology for PSO algorithm is described to perform the
reactive power dispatch to reduce the system electrical losses. For this purpose,
the IEEE 118-bus test system is employed considering renewable generation,
wind and solar photovoltaic, and variable demand profile during a day period.
The renewable generation is considered in PQ mode, which allowed its
representation as negative load in the load flow problem. The generated
powers are assessed for an entire day period. The wind sources contribution
is obtained from a wind speed profile [20], while for solar photovoltaic gen-
eration, temperature and solar irradiation profiles [21] are considered.
For the allocation of these renewable sources, the system is considered in
the base case and two candidate groups, respectively, considering voltage
stability and loss sensitivity criteria, as in [22]. The first group consists of
the critical buses of the system for allocation of wind sources, whereas the
second is composed of the sensitive to losses for allocation of solar sources.
The groups’ identification is performed as depicted in Section 10.4.
After RES penetration, the loss reduction is performed by a two-steps
process. First, the dispatchable generators identification is performed as
described in Section 10.5. Second, these generators are provided to the PSO,
and the optimization process is performed (Section 10.6).
This process is described in Fig. 10.3.
The PV and QV curves are plotted, and active and reactive load margins
are determined for each interval before and after the optimization process. In
addition, active power losses are also presented.
A comparison can be made by the means of dispatch benefit, allowing a
superior performance once the increment for each generator is limited. This
benefit corresponds to the ratio between the increases of the active load mar-
gin by the total dispatch increment.

10.8 Results and analysis

To verify the methodology effectiveness, five scenarios corresponding to
distinct periods of the day are selected. These scenarios are presented later,
where Pg wt denotes the wind generation power, Pg pv presents the solar gen-
eration power, and fc describes the demand profile.

Scenario 1: corresponds to interval A1, dawn period, without renewables
penetration, and fc 5 0.8922 p.u;
Scenario 2: corresponds to interval A11, morning period, with penetration of
both renewables, Pg wt 5 2.4 p.u. and Pg pv 5 0.0768 p.u. and fc5 1.0406 p.u;
Scenario 3: corresponds to interval A17, afternoon period, with penetra-
tion of both renewables, Pg wt 5 2.4000 p.u. and Pg pv 5 1.1744 p.u. and
fc 5 0.8614 p.u;