08-10-2012, 11:25 AM
NEW EVOLUTIONARY TECHNIQUE FOR OPTIMIZATION SHUNT CAPACITORS IN DISTRIBUTION NETWORKS
NEW EVOLUTIONARY TECHNIQUE.pdf (Size: 115.56 KB / Downloads: 30)
INTRODUCTION
It is known that the flow of reactive current in an
electric network produces supplementary power loss and
increases the voltage drop. Comparing to transport network,
the distribution network have low voltage and a
large current which produce the losses by Joule effects
relatively higher than in the transport networks (more
than 13% of the losses). To improve the load flow, the
quality of energy and to avoid as well a new investment
on building a new grid, we have to reduce the losses by installing
shunt capacitors in the appropriate places. In the
literature we can find many different optimization techniques
in away to optimize locations, sizes and numbers
of capacitors.
K. Prakash and M. Sydulu [8] have proposed an algorithm
based on Particle Swam Optimization, a metaheuristic
technique, for estimating the shunt compensation
level necessary to improve the voltage level and reduce
the active power losses. For finding the optimum
placements they have used the loss sensibility factors.
M. D. Reddy and V. C. V. Reddy [4] have proposed
a method with two levels to find the placement and the
size of shunt capacitors. They used a fuzzy algorithm to
search the optimal placement and real genetic algorithm
to find the optimal size of capacitors.
POPULATION AND CODING
The population research methods are performed among
the population constituted by several potential solutions.
Generally, the size of the population is between 30 and
200 individuals. The individuals coding is an essential parameters
for the method. They are represented in chain
structure, as chromosomes, containing genes or characters
of predetermined alphabet. There is different way for
coding a solution. In our study, individual is represented
by two parts of distinct chromosome. The first part receives
only binary values for coding candidate position
state. If the element in position i is equal to 1 then a capacitor
is connected to the bus I. Otherwise, the bus does
not receive a capacitor. The second part includes integer
numbers which represent the battery size indices [2]. Each
part of chromosomes has N positions. N represents the
buses number of the network. The following figure gives
an example of solution for a network with 9 buses.
EVOLUTION PROCESS
The stochastic optimization techniques are base on a
balance between intensification of the research and its diversification.
The intensification permits the research of
better quality solutions basing on previous found solutions.
And the diversification sets the strategies which
allow to investigate a large space of solutions. If the balance
is not observed, the convergence will move fast toward
local minimums (without diversification) or a long
investigation (without intensification).
CONCLUSION
Through the comparison between the obtained results
and previous results presented by different authors, we
notice clearly that the proposed new technique have given
a best and further optimization of numbers, sizes and locations
of the shunt capacitors. Consequently, we get further
minimization of the annual cost of the active power
losses and more improvement of different buses voltages.
The buses were classified on decreasing sensitive order in
function with their contributions in total losses in the network.
The attribution of the size is depending on sensitive
order of each emplacement in the initial population. The
evolution process is based on the intermediary recombination
of individual chosen in function of the objectives.
The presented new technique is full of promises.