Natural gas is transported through pipeline networks such as transmission and distribution to the final consumers. Local natural gas companies should make the continuous delivery of natural gas under any ambient temperature and demand. Natural gas demand varies due to ambient temperature variation (ATV) in different seasons or even during a day. The continuous supply of natural gas could be achieved only if the response of natural gas pipeline network to ATV is fully studied and understood. This study investigated the response of a typical natural gas distribution pipeline network to ATV. Firstly, a multilayer perceptron neural network model has been developed to forecast natural gas demand at any ambient temperature. Secondly, a new approach has been presented to simulate the natural gas distribution pipeline network in unsteady conditions. This approach is developed to predict the response of the distribution pipeline to ATV. The city of Semnan, Iran, was selected as the case study. Natural gas consumption in 4 coldest days of a year was extracted from the metering points. According to the results, the forecast data for natural gas demand has about 1% division compared with the actual values. Also, the node pressures significantly dropped on the coldest day of the year due to the increase in natural gas demand. In addition, the effect of natural gas composition on node pressure investigated. Results show that the natural gas with higher molecular weight has a lower pressure in all network nodes.

Keywords