Distribution network reconfiguration using hybrid heuristic - Genetic algorithm
Paper i proceeding, 2017

This paper presents algorithm for optimal reconfiguration of distribution networks using hybrid heuristic genetic algorithm. Improvements introduced in this approach make it suitable for real-life networks with realistic degree of complexity and network size. The algorithm introduces several improvements related to the generation of initial set of possible solutions as well as crossover and mutation steps in genetic algorithm. Since the genetic algorithms are often used in distribution network reconfiguration problem, its application is well known, but most of the approaches have very poor effectiveness due to high level of individuals' rejections not-fulfilling radial network constraints requirements and poor convergence rate. One part of these problems is related to ineffective creation of initial population individuals. The other part of the problem in similar approaches is related to inefficient operators implemented in crossover and mutation process over created set of population individuals. The hybrid heuristic-genetic approach presented in this paper provides significant improvements in these areas. The presented algorithm can be used to find optimal radial distribution network topology with minimum network losses or with optimally balanced network loading. The algorithm is rested on a real size network of city of Dubrovnik to identify the optimal network topology after the interpolation (connection) of a new supply point.

Författare

Damir Jakus

University of Split

Rade Cadenovic

University of Split

Mia Bogdanovic

University of Split

Petar Sarajcev

University of Split

Josip Vasilj

Chalmers, Signaler och system, System- och reglerteknik

2nd International Multidisciplinary Conference on Computer and Energy Science, SpliTech 2017, Split, Croatia, 12-14 July 2017

8019262
978-953290071-2 (ISBN)

Styrkeområden

Energi

Ämneskategorier

Annan elektroteknik och elektronik

ISBN

978-953290071-2

Mer information

Senast uppdaterat

2018-04-05