Artificial Neural Network based Cost Estimation of Power Losses in Electricity Distribution System

Goren G., Dindar B., Gül Ö.

4th IEEE Global Power, Energy and Communication Conference, GPECOM 2022, Cappadocia, Turkey, 14 - 17 June 2022, pp.455-460 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/gpecom55404.2022.9815721
  • City: Cappadocia
  • Country: Turkey
  • Page Numbers: pp.455-460
  • Keywords: Artificial neural networks, energy efficiency, electricity distribution losses, load loss factor, forecasting
  • Istanbul Technical University Affiliated: Yes


© 2022 IEEE.Electrical energy demand is increasing day by day with developing technology and increasing population. The limitation of resources reveals the importance of efficient use of energy. While the electrical energy is delivered to the final consumer, losses occur in the transmission and distribution grids. These losses may be due to technical and non-technical reasons. Higher losses cause more energy to be produced to compensate for this loss, making the system heavy loaded. It increases the cost of electricity and reduces its quality. For these reasons, Turkish Energy Market Regulatory Authority (EMRA) determines the loss-theft ratio of distribution companies with the ceiling price application, which is a penalty-reward method, in order to increase the performance of distribution networks. Due to these regulations, distribution companies make investments in order to improve their loss-theft ratio. The location and cost of losses must be known to predict investment. Due to the variable loads in the networks, it is difficult to measure and store the losses continuously. It requires a large amount of data and it is a time consuming process. The main purpose of this study is to estimate the losses in the distribution grid using artificial neural networks (ANN) instead of calculating them. Thus, an economical and fast methodology has emerged for distribution networks. The forecast results and the actual energy losses in this study are very close. In this study, firstly, the losses in the electricity distribution networks were calculated by using the load loss factor (LLF). With the obtained data set, future predictions were made using artificial neural networks. The costs of the estimated and calculated lost energies were calculated and compared.