Optimizing distributed generation allocation to minimize power loss in distribution systems using DBO-APCNN approach

This paper proposes a hybrid approach for optimizing DG allocation to minimize power loss in distribution systems. The proposed hybrid approach combines the Dung beetle optimizer (DBO) and attention pyramid convolutional neural network (APCNN). Hence, it is named as DBO-APCNN. While the APCNN is used to predict power loss in the distribution system, the DBO approach determines the best location for DG systems (DGs).To lessen voltage drops, power losses, and costs, it is important to position distributed generators (DGs) in distribution networks as ideally as possible. The major aim of the proposed method is to increase voltage stability and reduce power loss. The proposed techniques is tested in the IEEE 33 and IEEE 69 bus radial distribution systems (RDS) using the MATLAB platform, and its effectiveness is compared to a number of other existing techniques, including particle Swarm optimization and the Cuckoo Search Algorithm Salp Swam Algorithm. The DBO-APCNN technique findings in the lowest power loss of 35.3 MW with DG sizes of 526.9 MW, 934.9 MW, and 648.16 MW placed at nodes 15, 28, and 24, The DBO-APCNN method results in the lowest cost of 2,114,898$. This indicates that DBO-APCNN is more efficient than existing methods.