Forecasting natural gas demand in Istanbul by artificial neural networks method and planning of city gate stations

Abstract

Purpose: This study aims to ensure that the city gate stations are planned at the right time and place to ensure natural gas supply security in possible bad scenarios with natural gas demand forecasting. Theory and Methods: In this study, daily and hourly natural gas demand for Istanbul's Anatolian and European sides are estimated by using Artificial Neural Networks. Parameters affecting natural gas usages such as the number of consumers, average daily temperature, minimum daily temperature, official holidays, and heating degree days have been determined. Using the data obtained from the year 2008 to the end of 2018, the forecasting model created by the MATLAB software estimates the natural gas demands up to 2027 according to the coldest day of Istanbul in the last century, which occurred on 9 February 1929, with the minimum daily temperature of-16 degrees C and the average daily temperature of-7 degrees C. As a result of this study, it is decided which natural gas city gate station will be constructed with natural gas demand forecast. Results: According to the results of this study, the natural gas demand forecast values were found to be 0.98 Rsquare and the optimum city gate stations were determined according to these results. Conclusion: According to the natural gas demand estimation results made with artificial neural networks, it is seen that there is a need for additional natural gas city gate stations until 2023 for the Anatolian side of Istanbul and until 2024 for the European side. According to the result, the ideal locations of natural gas city gate stations were determined with Synergi Gas software by making velocity and pressure analyses and considering their proximity to the natural gas transmission line.