Oil production forecasting using higher order neural networks and reservoir history matching with the application of adaptive genetic algorithm

Abstract

Ever increasing input material cost coupled with environmental concerns has put pressure on profits of oil industry. There is a serious concern to plan, develop and exploit the oil and gas reserves in an efficient and optimal way. Production forecasting and reservoir modeling can provide vital inputs to efficient management of this hugely important energy source. Building reliable numerical reservoir models that incorporate all the geological, geophysical, geochemical, and petrophysical data of the reservoir available through petroleum exploration process, can help mitigate this problem. Since the reservoirs are highly heterogeneous and nonlinear in nature, it is often difficult to obtain accurate estimates of the spatial distribution of rock properties representing the reservoir and corresponding production profiles. Petroleum engineers always seek to construct reservoir models which are capable of consistent production forecasts such that further reservoir development in terms of locating new wells, recovery strategies (primary, secondary and tertiary), and surface facilities can be optimally designed.