Hybrid Semi-mechanistic Well Performance Modeling and Prediction

Open Access
Author:
Aniemena, Chigozie C
Graduate Program:
Petroleum and Natural Gas Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
December 16, 2013
Committee Members:
  • Yilin Wang, Dissertation Advisor
  • Yilin Wang, Committee Chair
  • Turgay Ertekin, Committee Member
  • Derek Elsworth, Committee Member
  • James Terry Engelder, Committee Member
Keywords:
  • Well Performance Prediction
  • Production Forecasting
  • Reserves Assessment
  • Hydraulic Fracture Design Optimization.
Abstract:
A well performance prediction method is developed with the advantage of increased time efficiency relative to reservoir simulation in single well reserves assessment and hydraulic fracture design optimization. Mechanistic limitations of decline curve analysis are also overcome. A theory of internal decline resistance is proposed and coupled with the concept of loss ratio linearity to characterize the decline exponent based on the physics of production decline. Components of internal decline resistance are identified as transient contacted volume and dynamic decompression effects which together justify a hyperbolic time dependent decline exponent model. As the arising differential equation cannot be solved in terms of elementary functions, parametric equivalence and direct substitution are used to arrive at a reduced order production decline model rooted in the Gaussian hypergeometric functional class – hypergeometric production decline (HG) model. Representative validation of the HG model is performed using simulated production data from simple to complex single well reservoir models. The condition necessary for hypergeometric production decline is also identified. Rate transient identities are developed from analytical solutions and coupled with a customized multi-variable non-linear regressional optimization algorithm. The result is a mechanistic model calibration framework within which the GHG model parameters are given as functions of reservoir system properties. In the generalized case of radial flow, the quadratic tendency of the loss ratio in transient flow justifies a linear time dependent decline exponent model. This linear decline exponent model is combined with the hyperbolic decline exponent model resulting in a quasi-hyperbolic decline exponent and the generalized hypergeometric production decline (GHG) model. The GHG production decline model coupled with the mechanistic model calibration framework constitute a reduced order well performance prediction method – the GHG method. The practical benefit of the GHG method is demonstrated within the context of hydraulic fracture design optimization where thousands of well performance predictions are performed in seconds. Such time efficiency is practically impossible using numerical reservoir simulation run on typical desktop computing power usually available to practicing reservoir engineer. The GHG method is neither a decline curve analysis tool nor a replacement for numerical reservoir simulators. It is a hybrid of both methods serving the niche purpose of expedited single well reserves assessment and hydraulic fracture design optimization while preserving mechanistic integrity.