Towards Smarter Energy Planning: District-Based Natural Gas Consumption Forecasting in Istanbul with BI-GRU

Abstract

With the increasing demand for energy, accurate forecasting of natural gas consumption in large cities has gained strategic importance. This paper compares the performance of various forecasting models using natural gas consumption data of districts in Istanbul, the most populous city in Türkiye. Using a time series analysis approach, the deep learning models Long Short Term Memory (LSTM), Gated Recurrent Unit (GRU), Bidirectional LSTM (BI-LSTM), Bidirectional GRU (BI-GRU) and Recurrent Neural Network (RNN) are compared with machine learning models Linear Regression (LR), Ridge, Support Vector Regression (SVR), Random Forest (RF) and Gradient Boosting algorithms. Model performance was evaluated with R² and Mean Absolute Percentage Error (MAPE) metrics. According to the results, the BI-GRU model provided the highest accuracy with a MAPE of 12.98% and an R² value of 0.8902. LSTM, BI-LSTM and RNN models also showed high success. In contrast, linear regression-based models performed poorly with low R² and high MAPE values. Among the machine learning methods, only the Random Forest model gave a strong result with an R² of 0.9063, while the MAPE value remained high. As a result, deep learning models, especially BI-GRU, better capture short-term consumption fluctuations and provide better forecasts compared to classical models.

Keywords

• Energy consumption
• Natural gas
• Deep learning
• Machine learning

References

1. Armstrong, J. S., & Collopy, F. (1992). Error measures for generalizing about forecasting methods: Empirical comparisons. International Journal of Forecasting, 8(1), 69–80.
2. Arslantürk, V., & Şirin, B. T. (2025). Machine learning approaches for energy management in public buildings. Black Sea Journal of Engineering and Science, 8(5), 1415–1428.
3. Azadeh, A., Zarrin, M., Rahdar Beik, H., & Aliheidari Bioki, T. (2015). A neuro-fuzzy algorithm for improved gas consumption forecasting with economic, environmental and IT/IS indicators. Journal of Petroleum Science and Engineering, 133, 716–739. https://doi.org/10.1016/j.petrol.2015.07.002
4. Başarslan, M. S. (2025). M-C&M-BL: A novel classification model for brain tumor classification: Multi-CNN and multi-BiLSTM. The Journal of Supercomputing, 81, Article 502. https://doi.org/10.1007/s11227-025-06964-x
5. Beyca, O. F., Cayir Ervural, B., Tatoglu, E., Ozuyar, P. G., & Zaim, S. (2019). Using machine learning tools for forecasting natural gas consumption in the province of Istanbul. Energy Economics, 80, 937–949. https://doi.org/10.1016/j.eneco.2019.03.006
6. Chicco, D., & Jurman, G. (2020). The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation. BMC Genomics, 21(1), Article 6.
7. Demirel, Ö. F., Zaim, S., Çalışkan, A., & Özuyar, P. (2012). Forecasting natural gas consumption in İstanbul using neural networks and multivariate time series methods. Turkish Journal of Electrical Engineering and Computer Sciences, 20(5), Article 4. https://doi.org/10.3906/elk-1101-1029
8. Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT Press.

9. Graves, A., & Schmidhuber, J. (2005). Framewise phoneme classification with bidirectional LSTM networks. Proceedings of the IEEE International Joint Conference on Neural Networks (IJCNN), 2047–2052.
10. Haksevenler, G. (2025). Towards carbon neutral cities: An insight for Istanbul, Türkiye. Sustainable Development, 33(1), 717–732. https://doi.org/10.1002/sd.3154
11. Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780.
12. Karadağ, İ., & Sağtaş, K. (2025). Energy consumption forecasting with artificial intelligence models. Black Sea Journal of Engineering and Science, 8(6), 1780–1793.
13. Kizilaslan, R., & Karlik, B. (2008). Comparison neural networks models for short term forecasting of natural gas consumption in Istanbul. Proceedings of the 1st International Conference on Applied Digital Information and Web Technologies (ICADIWT), 448–453. https://doi.org/10.1109/ICADIWT.2008.4664390
14. Kong, W., Dong, Z. Y., Jia, Y., Hill, D. J., Xu, Y., & Zhang, Y. (2019). Short-term residential load forecasting based on LSTM recurrent neural network. IEEE Transactions on Smart Grid, 10(1), 841–851. https://doi.org/10.1109/TSG.2017.2753802
15. LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444.
16. Marino, D. L., Amarasinghe, K., & Manic, M. (2016). Building energy load forecasting using deep neural networks. Proceedings of the IEEE IECON 42nd Annual Conference of the Industrial Electronics Society, 7046–7051. https://doi.org/10.1109/IECON.2016.7793413
17. Potočnik, P., Govekar, E., & Grabec, I. (2007). Short-term natural gas consumption forecasting. Proceedings of Applied Simulation and Modelling, Palma De Mallorca, Spain.
18. Schuster, M., & Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE Transactions on Signal Processing, 45(11), 2673–2681.
19. Stambouli, A. B. (2011). Fuel cells: The expectations for an environmental-friendly and sustainable source of energy. Renewable and Sustainable Energy Reviews, 15(9), 4507–4520. https://doi.org/10.1016/j.rser.2011.07.100
20. Yang, Z., Yang, D., Dyer, C., He, X., Smola, A., & Hovy, E. (2016). Hierarchical attention networks for document classification. Proceedings of the North American Chapter of the Association for Computational Linguistics (NAACL).
21. Zanella, A., Bui, N., Castellani, A., Vangelista, L., & Zorzi, M. (2014). Internet of Things for Smart Cities. IEEE Internet of Things Journal, 1(1), 22–32. https://doi.org/10.1109/JIOT.2014.2306328