VNS ALGORITHM FOR OPTIMIZED ALLOCATION OF WIRELESS SMART CONCENTRATORS IN LOW VOLTAGE ELECTRIC NETWORKS
With the advancements in sustainable technologies and the need for modernization of distribution networks, electric power utilities are installing Wireless Smart Concentrators (WSCs) along low-voltage distribution networks. These concentrators communicate with Wireless Smart Meters (WSMs) of end consumers to maintain a database that assists in energy billing and defining new market structures.
In extensive distribution networks, the installation of concentrators becomes a challenge for planners as it can lead to significantly high installation costs. To address this issue, specialized literature suggests the application of computational intelligence techniques to conduct optimized allocation studies of WSCs in low-voltage distribution networks. However, in extensive low-voltage networks, various signal propagation environments can hinder the communication between WSCs and WSMs. This work presents a computational intelligence algorithm based on the Variable Neighborhood Search (VNS) technique for the optimized allocation of WSCs in extensive low-voltage distribution networks in urban areas. This algorithm aims to minimize the investment required for WSC installation in extensive low-voltage networks while ensuring communication coverage for all WSMs. Thus, the signal propagation environment between WSMs and the available types of WSCs within a specific area is considered through a propagation coefficient that characterizes the coverage radius of the concentrators. In this context, the proposed algorithm is applied to a low-voltage feeder in Denmark comprising 15 nodes and 45 consumers, using WSCs based on ZigBee (ZB) technology. This algorithm is compared with other computational intelligence techniques used in specialized literature. 100 executions of the proposed algorithm are conducted to allocate WSCs in a low-voltage distribution network in Denmark. Following these executions, a histogram is developed to display the frequency of the best result found by each metaheuristic. The proposed algorithm found the best solution 80 times compared to 60 times found by the compared metaheuristics. The result of the proposed algorithm can assist low-voltage electrical network planners in identifying optimal locations for WSC installations, ensuring adequate coverage for WSMs in urban areas with a high concentration of vertical infrastructures or tree cover.