Comprehensive Investment Framework for Optimal Exploitation of Liquid-Rich Unconventional Reservoirs and Natural Gas Distribution

Abstract

Abstract A techno-economic framework is proposed for the optimal exploitation of unconventional shale reservoirs and the delivery of shale gas products within the Marcellus region. A 3-D compositional reservoir model is constructed to simulate realistic drilling, completion, stimulation (DCS) strategies and pad production processes. A feed-forward neural network (FNN) is implemented to understand re-frac well candidature so as to aid development of alternative development strategies throughout a project planning horizon. In addition, optimal well configurations and DCS strategies for well-pad development are developed in line with best petroleum engineering practices. Simulation results are incorporated in a long-term strategic planning model used for determining the optimal shale gas supply chain network and its operations. For this, a mixed-integer non-linear programing (MINLP) formulation is developed to maximize net present value (NPV). Natural gas demand is predicted by running long short-term memory (LSTM) neural networks based on the historical data of critical predictors. Results of this framework depict the influence of reservoir engr., completion engr., and machine learning in predicting project profitability and establishing optimal strategies for shale gas production and distribution.