Surrogate-based optimization of hydraulic fracturing in pre-existing fracture networks

Mingjie Chen*, Yunwei Sun, Pengcheng Fu, Charles R. Carrigan, Zhiming Lu, Charles H. Tong, Thomas A. Buscheck

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

20 Citations (Scopus)

Abstract

Hydraulic fracturing has been used widely to stimulate production of oil, natural gas, and geothermal energy in formations with low natural permeability. Numerical optimization of fracture stimulation often requires a large number of evaluations of objective functions and constraints from forward hydraulic fracturing models, which are computationally expensive and even prohibitive in some situations. Moreover, there are a variety of uncertainties associated with the pre-existing fracture distributions and rock mechanical properties, which affect the optimized decisions for hydraulic fracturing. In this study, a surrogate-based approach is developed for efficient optimization of hydraulic fracturing well design in the presence of natural-system uncertainties. The fractal dimension is derived from the simulated fracturing network as the objective for maximizing energy recovery sweep efficiency. The surrogate model, which is constructed using training data from high-fidelity fracturing models for mapping the relationship between uncertain input parameters and the fractal dimension, provides fast approximation of the objective functions and constraints. A suite of surrogate models constructed using different fitting methods is evaluated and validated for fast predictions. Global sensitivity analysis is conducted to gain insights into the impact of the input variables on the output of interest, and further used for parameter screening. The high efficiency of the surrogate-based approach is demonstrated for three optimization scenarios with different and uncertain ambient conditions. Our results suggest the critical importance of considering uncertain pre-existing fracture networks in optimization studies of hydraulic fracturing.

Original languageEnglish
Pages (from-to)69-79
Number of pages11
JournalComputers and Geosciences
Volume58
DOIs
Publication statusPublished - Aug 2013

Keywords

  • Fractal dimension
  • Global sensitivity
  • Hydraulic fracturing
  • Optimization
  • Surrogate model

ASJC Scopus subject areas

  • Information Systems
  • Computers in Earth Sciences

Fingerprint Dive into the research topics of 'Surrogate-based optimization of hydraulic fracturing in pre-existing fracture networks'. Together they form a unique fingerprint.

Cite this