Multi-objective optimal reconfiguration of distribution network

Abstract

Configuration of distribution system can be changed manually or automatically, by changing the status of the respective switching elements, with the aim of reducing power losses, increase system reliability, or improving the power quality. When changing the status of switching equipment it is necessary to satisfy the requirement for the radial and connected structure of the distribution network. Using the single criteria optimization it is possible to improve one of the characteristics of the distribution network, on the other hand by using multicriteria optimization it is possible to find a network configuration that enhances multiple distribution system characteristics at the same time. In this paper, a modification of the multi-criteria Gray Wolf optimization algorithm is proposed in order to create an efficient algorithm that can be implemented in the management functions of smart grid concept of modern distribution systems. The proposed reconfiguration algorithm was tested on standard symmetrical IEEE 33 test distribution network.

Reconfiguration of distribution system with distributed generation using an adaptive loop approach

“…The loop is iterative function that examines if the considered network is connected. In each iteration this function calculates absolute sum of each element of a corresponding column in matrix A, and nodes whose sum is equal to 1 determine which branch (row) will be removed from the matrix A [8]. The iterative procedure ends when there are no more nodes whose sum is equal to 1.…”

“…The elements of the set L s3 are [9,8,7,1], while the elements of the set L r3 are [12,14,6,5,13,7,1]. Applying (6), a set L c3 , the elements of which are [5,6,8,9,12,13,14,15], is created. Closing the switch in branch number 15 and selecting any allowed value for the third control variable, a connected radial network is created at which further optimization calculations can be made.…”

“…While on the other hand, metaheuristic methods enable better, spread search of the so-lution space and, depending on the applied method, they can find solution that is close enough to the global optimum. Therefore, meta-heuristic methods have been used more frequently in recent years for solving this problem: ant colony algorithm [7], gray wolf optimizer [8], particle swarm optimization [9], cuckoo search algorithm [10], genetic algorithm [11], etc. In addition to meta-heuristic methods, in recent years, the number of papers utilizing other artificial intelligence tools such as fuzzy logic and clustering algorithms [12,13] has also increased.…”

“…Validation of generated solution control variables: (a) -the first, (b) -the second, (c) -the third, and (d) -correction of unacceptable control variablethe sets L s2 =[7,1] and L r2 =[5,6,14,12,15,9,8,7,1] are created. Applying (6), a set of allowed candidates for second control variable, L c2 =[5,6,8,9,12,13,14,15]…”

Fault detection in power grids based on improved supervised machine learning binary classification

“…Consequently, feature extraction methods are crucial to separate these signals and to accurately diagnosis fault signals [2]. Many approaches were proposed to handle feature extraction problem such as transformation [3,4], dimensionality reduction [5], and modal transformation (Clarke transformation) [6,7].…”