Research on Online Decision-Making Method for Oil Wells Shutting Down or Resetting Based on Dynamic Data-Driven

Abstract

In the unstable and uncertain international situation, how to balance the scale growth and efficiency improvement of oil production is a key issue to ensure national energy security and sustainable development of enterprises. In view of the dual problem of declining oil production and rising production cost in the middle and late stage of oilfield development, this paper focuses on oil well, the smallest unit of oilfield, to improve the economic benefit of single well and optimize the allocation of single well resources. This paper introduces the idea of data-driven decision making and machine learning method, proposes a prediction method of single well economic benefit using long and short-term memory neural network to mine complex structure and potential pattern of massive data, and develops an online decision system based on dynamic data driving to solve the problem that the model method is difficult for managers to use. Numerical example results show that both short-term and long-term memory neural network prediction model of single well economic benefit of short-term forecasting results are good and give full consideration to the well historical data value, break traditional management decision problems, strong paradigm it into the process of petroleum enterprise benefit stable data interactive collaborative decision-making process of learning and resources optimization, assist oil enterprises to avoid low and inefficient investment at source.