A Combined Graph Theory–Machine Learning Strategy for Planning Optimal Radial Topology of Distribution Networks. Issue 13 (11th May 2022)

• Record Type: Journal Article
• Authors: Gunturi, Sravan Kumar; Sarkar, Dipu; Sumi, Lilika; De, Abhinandan
• Is Part Of: Electric power components and systems. Volume 49:Issue 13/14(2021)
• Page Start: 1158
• Page End: 1168
• Publication Date: 2022-05-11
• Subjects: voltage stability -- network reconfiguration -- graph theory -- machine learning -- smart grid; Electric machinery -- Periodicals; 621.3104205
• DOI: 10.1080/15325008.2022.2050444
• Languages: English
• ISSNs: 1532-5008

Abstract

Abstract: A mesh type distribution network is often configured into a radial structure to offer a variety of practical benefits, such as lower short-circuit current, low operating costs, and low construction costs. This paper presented a novel technique for optimal planning of radial distribution networks based upon a combination of Graph Theory and Machine Learning (ML) approaches. A network comprises several sectionalizing and tie-type switches. The conversion of a mesh type system into a radial one involves the selection of the best set of switches on the feeder lines that are to be opened or closed. Out of all switching combinations, we have provided the optimal radial combination, consequently the maximum voltage stability of the network with less active and reactive power losses. The proposed system also preserves the system's radial nature, with out islanding any load by a simple coding scheme. To test the capabilities of the proposed technique, we carried out simulations on the 30-node test distribution system and the results are encouraging.