Title: Prediction and optimisation of electricity market clearing price in Turkey by using machine learning methods

Authors: Murat Ince; Ahmet Kabul; Mesut Aksoy

Addresses: Department of Computer Technologies, Isparta University of Applied Sciences, Isparta, Turkey ' Mechanical Engineering in Technology Faculty, Isparta University of Applied Sciences, Isparta, Turkey ' Energy Systems Engineering in Technology Faculty, Isparta University of Applied Sciences, Isparta, Turkey

Abstract: This research aims to reduce price instability in the market clearing price (MCP) in Turkey by estimating MCP using machine learning techniques based on production resource-based data. The model will balance market prices by shifting from a price-based to a resource-based approach, minimising the price of electricity units by decreasing imported energy production and increasing domestic and renewable energy production. Thus, in this study, the effect of MCP on electricity unit prices and forecast values until July 29, 2023, was compared. By using past year data between 2014 and 2022, the MCP price in 2023 is determined. As a result of artificial neural network prediction, the average MCP value for 2023 was revealed 85.9 USD. The best results were obtained with artificial neural network (ANN) (R² = 0.8827, RMSE = 0.0309 and MAE = 0.0223). Also, the model predicts estimated 2023 energy production by incorporating real-time production values from energy resource production data. The performance indicators of the implemented forecasting methods increase efficiency in future production forecasts and contribute to accurate pricing in energy purchases. [Received: February 17, 2024; Accepted: July 7, 2024]

Keywords: market clearing price; MCP; energy efficiency; machine learning; regression; Turkey; artificial neural network; ANN.

DOI: 10.1504/IJOGCT.2025.146520

International Journal of Oil, Gas and Coal Technology, 2025 Vol.37 No.4, pp.438 - 466

Received: 08 Feb 2024
Accepted: 07 Jul 2024
Published online: 02 Jun 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article