The artificial neural networks for real-time operation of natural gas production and sale

Abstract

Backpropagation nerual networks (BPNN) is introduced in this paper to explore the non-linear relationship between planned gas supply and actual gas demand from uneven coefficients of natural gas demands of end-users with a case study of North China Branch of Sinopec, aiming to solve the imbalances between gas supply and demand and instabilities of gas operation from uncertainties of natural gas demmand flcutuation with dialy or seasonal change. The research indicates that BPNN could effectively build up complex non-linear map bwteen the planned supply and actual demand in the process of natural gas production and sale in uperstream gas fields, and give the feasible guide to the real-time operation for the gas field proudction and sale, providing a novel intelligent method and new idea for the operation decision-making of natural gas.