Abstract:
Losses in distribution transformers are estimated as 30% of overall transmission and
distribution losses. It is further estimated that the losses in all of the world’s electrical
distribution systems are about 1715TWh. One-third of these losses occur in the
distribution transformers. In this thesis, a mathematical model is done and a new
objective function is proposed for minimization of losses in a distribution transformer.
This thesis presents a loss-reduced optimal design of three phases, 200kVA, 15/0.4kV,
50Hz, oil-immersed, core type distribution transformer. Two optimization techniques
namely, Brute force search and Genetic algorithm are used in MATLAB® to obtain an
optimum design of a distribution transformer that has minimum losses which met the
requirements and constraints. The loss of a distribution transformer designed using both
algorithms is compared, and then a comparison is made with transformer manufacturers’
used design based on analytical method. The results from the optimization algorithm
show that the design reduces the total losses on the existing distribution transformer
selected for the study from 2,756.50W to 1,994.14W by 762.36W, thus representing a
percentage reduction of 27.66%. If this saving is applied to the existing 49, 200kVA
distribution transformers of the Jimma town route of the case study area, the saving will
be 37,355.64W. If the optimally designed transformer is to be implemented on a larger
scale across the electrical distribution networks of Ethiopia, the magnitude of savings
would be huge. Moreover, the designed distribution transformer using the optimization
algorithm has a slightly increased material cost, but it has the lowest total cost. Thus, the
design is cost-effective.
