Design optimization of distribution transformers by minimizing the total owning cost

Abstract

Summary This paper presents a new optimization problem for the design of distribution transformers under the strict requirements on energy efficiency imposed by modern standards and directives. The proposed methodology is devoted to the minimization of the total owning cost of the transformer while meeting a set of constraints that includes all the phenomena implied in the transformer operation. In particular, both electromagnetic and thermal aspects are considered and their relationships with electrical and geometrical variables are accurately detailed. Furthermore, the power losses in both core and windings are accurately derived and the limitations imposed by the standards are taken into account by imposing proper constraints. The generality of the proposal allows considering different technological solutions in material used for the core. The results proposed in the numerical applications showed the effectiveness of the proposed procedure and permitted highlighting the feasibility of different solutions.

Implementation of tree pruning method for power transformer design optimization

“…In Yadollahi and Lesani, a new heuristic algorithm (HA) combined with Mixed Integer Nonlinear Programming method has been proposed, and also, some modifications to conventional Power Transformer Design Optimization (PTDO) procedure. Most of the optimization methods presented in literature focused on the optimal design problem for distribution transformers . Mixed Integer Nonlinear Programming is compared with Harmony Search, DE, and GA for the transformer design optimization problem in Amoiralis et al In many of aforementioned methods, in order to limit the Search Space size, number of design variables are reduced, for example, design variables as the height and width of wires are ignored in previous studies …”

“…Having reviewed most works tackling optimal design of transformers, one can came to following conclusions: Most of the hitherto published works addressed the optimal design of distribution transformers These works did not consider voltage regulating winding and its variables as genes of the main chromosome Nearly all previous works used optimization methods with random nature, meaning that they produce different solutions for each separate run.…”

“…Nearly all previous works used optimization methods with random nature, meaning that they produce different solutions for each separate run. This obviously means that they have problem finding the optimal minimum These works only consider a limited number of design variables.…”

“…These works only consider a limited number of design variables. However, there are many other contributing variables which need to be considered …”

“…Most of the optimization methods presented in literature focused on the optimal design problem for distribution transformers. 5,[14][15][16][17][18] Mixed Integer Nonlinear Programming is compared with Harmony Search, DE, and GA for the transformer design optimization problem in Amoiralis et al 17 In many of aforementioned methods, in order to limit the Search Space size, number of design variables are reduced, for example, design variables as the height and width of wires are ignored in previous studies. 5,[16][17][18] Having reviewed most works tackling optimal design of transformers, one can came to following conclusions:…”