1 results listed
Corresponding author E-mail: hasanozcan@karabuk.edu.tr
Abstract— This study aims to investigate the suitability and
fidelity of three optimization metaheuristics applied to a simple
thermal system costing problem and to discuss their generic
comparison by taking into account the cost assessment of an air-
cooling unit. Undertaken thermodynamic plant consists of an air
chiller plant, which requires additional heat reservoir to keep the
air temperature at a desired level. The temperature level of this
reservoir is kept constant by using a cooling tower. A pre-cooler
is also introduced as a black-box mathematical model for
enhanced chiller performance. Three stochastic population-based
metaheuristics, namely particle swarm optimization (PSO),
differential evolution (DE) and backtracking search algorithm
(BSA), are applied for many case studies throughout the studied
system and the results are validated with Lagrange multipliers
method (LM) as the only direct search algorithm. Optimization
results suggest that a successful optimum cost is easily achievable
by using each algorithm with a trade-off. While BSA provides a
useful amount of minimum costs for all considered cases by
reporting the best system parameters selection, PSO and DE
algorithms perform faster to reach the optimal value at the
specified
solution
space
with
a
requirement
of
suitable
.
initialization
1.st International Conference Energy Systems Engineering
ıcese'17
Hasan OZCAN
Leandro dos Santos COELHO
Mehmet Özdemir