Loading...

Forecasting Residential Natural Gas Consumption in Tehran Using Machine Learning Methods

Khazaei, Armin | 2022

147 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54856 (46)
  4. University: Sharif University of Technology
  5. Department: Energy Engineering
  6. Advisor(s): Maleki, Abbas
  7. Abstract:
  8. According to increasing energy demand in Iran and the world, the role of natural gas as a relatively clean and cost-effective source has received more attention. Given the high share of the residential sector in the country's natural gas consumption, providing a model for forecasting the demand of this sector is of great importance for policy makers and decision makers in this field. In the present study, we employ three popular methods of machine learning, support vector regression, artificial neural network and decision tree to predict the consumption of natural gas in the residential sector in Tehran according to meteorological parameters (including temperature, precipitation and wind speed), natural gas prices in the residential sector and Gas consumption with lag 12 months, was used as model inputs. The results showed that the artificial neural network model has the highest accuracy in predicting gas consumption and finally from this model to predict natural gas consumption in the residential sector of Tehran in the next 5 years with three scenarios of fixed gas price, 10% price increase per year and A 20% price increase per year has been used. It is predicted that in 1403, the amount of gas consumption in Tehran will increase by more than 12% and will reach more than 9’600 million cubic meters per year. The results also show that rising prices will not play a significant role in reducing gas consumption by the residential sector
  9. Keywords:
  10. Support Vector Regression ; Machine Learning ; Natural Gas ; Energy Demand ; Artificial Neural Network ; Decision Making Tree ; Residential Customer

 Digital Object List

 Bookmark

No TOC