Data driven techno-economic framework for the development of shale gas resources

Abstract

This work presents a data driven techno-economic framework that combines upstream, midstream and downstream operations to optimize profitability of shale projects. This framework is illustrated using a shale gas supply chain structure planned for development and integration in an over-supplied gas market. Field development strategies are developed based on applying machine learning techniques to an existing field. Alternative development strategies are implemented on the integrated production-modeling platform RESOLVE–REVEAL to simulate hydrocarbon/water production. Long-short term memory (LSTM) neural networks are developed to predict gas demand and freshwater availability. Long-term strategic planning is achieved using a mixed-integer non-linear programming (MINLP) formulation. Results indicate a net present value (NPV) of 205.56 MMUS$ for optimal infrastructure design, gas and liquids transportation and distribution, and water management after 54-well integration. Additionlly, results provide an optimal gas storage schedule that supports shale asset profitability. Application of this techno-economic approach improves profitability projections for shale enterprises.