(38f) An MINLP Model for the Water Management in Shale Gas Operations

Authors: 
Lira-Barragán, L. F., Universidad Michoacana de San Nicolás de Hidalgo
Ponce-Ortega, J. M., Universidad Michoacana de San Nicolás de Hidalgo
Serna-González, M., Universidad Michoacana de San Nicolás de Hidalgo
El-Halwagi, M. M., Texas A&M University

This work presents a mathematical programing approach for the optimal management of flowback water in shale gas wells in order to meet water demands in hydraulic fracturing operations considering a fixed scheduling for the completion phases of the wells. The objective function is aimed to determine the minimum total costs for the water network constituted by capital and operational costs. Capital costs are associated to the equipment acquisition of storage units, treatment technologies and disposals; whereas the operating costs considers the fresh water cost, operating cost for water treatment and the transportation costs for all transport trajectories required. Additionally, the total amount of fresh water used (environmental concern) is quantified and minimized; in this context, the methodology accounts for the flowback water reuse once it is treated and stored with the purpose to reduce the fresh water consumption, the wastewater streams and the costs. The model involves the seasonal limitations for the fresh water availability, which is an important challenge for some of the existing shale plays.

The proposed design considers that all the flowback water is treated and the outlet streams satisfy the restrictions for the pollutant concentrations in order to be reused and/or the environmental regulations to be disposed. To carry out this task, only treatment units with tested effectiveness in flowback water coming from shale gas wells have been considered. Also, the given scheduling for the fracking operations of each well taken into account can include the simultaneous operation for several fracking crews. The mathematical formulation corresponds to an MINLP multi-period model. Finally, the example problem uses updated information taken of technical reports from some of the most important existing shale regions (Marcellus and Barnette).