Forecasting Monthly Residential Natural Gas Demand Using Just-In-Time-Learning Modeling

TL;DR

This paper introduces a novel Just-in-Time-Learning Gaussian Process Regression method for more accurate monthly natural gas demand forecasting, demonstrating improvements over traditional models in Turkish cities.

Contribution

The paper proposes a new JITL-GPR approach with a unique feature representation for demand data, enhancing forecast accuracy over existing time series models.

Findings

JITL-GPR outperforms SARIMA and ETS models in forecast accuracy.

The method effectively captures demand patterns using a 2-D grid feature representation.

Forecast errors are reduced compared to traditional and state-of-the-art models.

Abstract

Natural gas (NG) is relatively a clean source of energy, particularly compared to fossil fuels, and worldwide consumption of NG has been increasing almost linearly in the last two decades. A similar trend can also be seen in Turkey, while another similarity is the high dependence on imports for the continuous NG supply. It is crucial to accurately forecast future NG demand (NGD) in Turkey, especially, for import contracts; in this respect, forecasts of monthly NGD for the following year are of utmost importance. In the current study, the historical monthly NG consumption data between 2014 and 2024 provided by SOCAR, the local residential NG distribution company for two cities in Turkey, Bursa and Kayseri, was used to determine out-of-sample monthly NGD forecasts for a period of one year and nine months using various time series models, including SARIMA and ETS models, and a novel proposed machine learning method. The proposed method, named Just-in-Time-Learning-Gaussian Process Regression (JITL-GPR), uses a novel feature representation for the past NG demand values; instead of using past demand values as column-wise separate features, they are placed on a two-dimensional (2-D) grid of year-month values. For each test point, a kernel function, tailored for the NGD predictions, is used in GPR to predict the query point. Since a model is constructed separately for each test point, the proposed method is, indeed, an example of JITL. The JITL-GPR method is easy to use and optimize, and offers a reduction in forecast errors compared to traditional time series methods and a state-of-the-art combination model; therefore, it is a promising tool for NGD forecasting in similar settings.