Prediction of power transmission line losses with artificial neural networks

The demand for electrical energy, which is a part of our lives, is increasing day by day. The inadequacy of renewable energy resources and the rapid depletion of renewable energy resources direct scientists to new alternative studies, while also requiring the reduction of losses in energy production, transmission and distribution lines. In order to reduce losses, the factors affecting losses must be correctly identified and calculated. Since energy transmission line repair takes a long time and is costly, it is important to plan and renew energy transmission lines well. In this context, in order to plan energy transmission lines efficiently, the factors affecting energy transmission line losses must be determined correctly and the losses must be calculated well. In this study, in order to estimate the losses occurring in the energy transmission line, the voltage value of the line, the type of conductor used, the line length, the pressure level, the temperature value, the smoothness factor and the load current are used as network inputs, and the energy transmission line loss is determined as the network output. A 1-output ANN model was developed using MATLAB software program. In this study, using feedforward back-propagation ANN, the number of hidden layers, the number of neurons in these layers and other training parameters were optimized with the Levenberg Marquardt (LM) algorithm throughout the training process. By using MATLAB/nntool to calculate the losses, the effectiveness of this input parameter on the power transmission line loss was successfully estimated by changing 1 input parameter at certain intervals, both for all parameter inputs at the same time and providing that 6 input parameters remain constant.