1
Department of electrical and computer engineering, Hamedan University of Technology, Hamedan, Iran
2
Division of Energy Systems Engineering, Hamedan University of Technology, Hamedan, Iran
10.61882/jgeri.2026.2096518.1128
Abstract
Accurate forecasting of residential natural gas consumption is critical for ensuring supply reliability, informing infrastructure development, designing optimal tariff structures, and implementing effective demandside management policies. In this study, Hamedan—one of the coldest and highest gas-consuming provinces
in Iran—was selected as a case study to evaluate the predictive capabilities of advanced deep learning models. Four state-of-the-art architectures, namely BLSTM, LSTM, CNN, and MLP, were trained on 11 key
predictive features. Data augmentation techniques were incorporated to improve generalization and robustness. The proposed models achieved exceptional predictive accuracy (R² = 0.97, CC = 0.98), demonstrating
their ability to capture complex, non-linear consumption patterns. These results highlight the significant
potential of AI-driven forecasting to optimize energy distribution, advance sustainable resource planning,
and facilitate evidence-based decision-making in the natural gas sector.
Keywords: Residential gas consumption, Neural networks, Energy sustainability, Deep learning models,
Shahdoosti, H. R, Fattahi, A. & Mohammadi, P. (2026). Prediction of Residential Natural Gas Demands based on Deep Learning. (e743559). Journal of Green Energy Research and Innovation, (), e743559 https://doi.org/10.61882/jgeri.2026.2096518.1128
MLA
Shahdoosti, H. R, Fattahi, A., & Mohammadi, P. "Prediction of Residential Natural Gas Demands based on Deep Learning" .e743559 , Journal of Green Energy Research and Innovation, , 2026, e743559. doi: 10.61882/jgeri.2026.2096518.1128
HARVARD
Shahdoosti, H. R, Fattahi, A., & Mohammadi, P. (2026). 'Prediction of Residential Natural Gas Demands based on Deep Learning', Journal of Green Energy Research and Innovation, (), e743559. doi: 10.61882/jgeri.2026.2096518.1128
CHICAGO
Shahdoosti, H. R, Fattahi, A. & Mohammadi, P., "Prediction of Residential Natural Gas Demands based on Deep Learning." Journal of Green Energy Research and Innovation, (2026): e743559, doi: 10.61882/jgeri.2026.2096518.1128
VANCOUVER
Shahdoosti H. R, Fattahi A., Mohammadi P. Prediction of Residential Natural Gas Demands based on Deep Learning. JGERI. 2026;():e743559. doi: 10.61882/jgeri.2026.2096518.1128
