Optimization of Transformation Coefficients Using Direct Search and Swarm Intelligence
This research considers optimization of tap position of transformers in power systems to reduce power losses. Now, methods based on heuristic rules and fuzzy logic, or methods that optimize parts of the whole system separately, are applied to this problem. The first approach requires expert knowledg...
Saved in:
Published in: | Problems of the regional energetics Vol. 33; no. 1; pp. 15 - 23 |
---|---|
Main Authors: | , , |
Format: | Journal Article |
Language: | English Russian |
Published: |
Academy of Sciences of Moldova
01-04-2017
|
Subjects: | |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | This research considers optimization of tap position of transformers in power systems to reduce power losses. Now, methods based on heuristic rules and fuzzy logic, or methods that optimize parts of the whole system separately, are applied to this problem. The first approach requires expert knowledge about processes in the network. The second methods are not able to consider all the interrelations of system’s parts, while changes in segment affect the entire system. Both approaches are tough to implement and require adjustment to the tasks solved. It needs to implement algorithms that can take into account complex interrelations of optimized variables and self-adapt to optimization task. It is advisable to use algorithms given complex interrelations of optimized variables and independently adapting from optimization tasks. Such algorithms include Swarm Intelligence algorithms. Their main features are self-organization, which allows them to automatically adapt to conditions of tasks, and the ability to efficiently exit from local extremes. Thus, they do not require specialized knowledge of the system, in contrast to fuzzy logic. In addition, they can efficiently find quasi-optimal solutions converging to the global optimum. This research applies Particle Swarm Optimization algorithm (PSO). The model of Tajik power system used in experiments. It was found out that PSO is much more efficient than greedy heuristics and more flexible and easier to use than fuzzy logic. PSO allows reducing active power losses from 48.01 to 45.83 MW (4.5%). With al, the effect of using greedy heuristics or fuzzy logic is two times smaller (2.3%). |
---|---|
ISSN: | 1857-0070 |