Date of Award
Doctor of Philosophy (PhD)
This dissertation proposes new mixed-integer optimization models and computational strategies for optimal offshore oil and gas field infrastructure planning under fiscal rules of the agreements with the host government, accounting for endogenous uncertainties in the field parameters using a stochastic programming framework. First, a multiperiod mixed-integer nonlinear programming (MINLP) model is proposed in Chapter 2 that incorporates field level investment and operating decisions, and maximizes the net present value (NPV). Two theoretical properties are proposed to remove the bilinear terms from the model, and further converting it to an MILP approximation to solve the problem to global optimality. Chapter 3 extends the basic deterministic model in Chapter 2 to include complex fiscal rules maximizing total contractor’s (oil company) share after paying royalties, profit share, etc. to the host government. The resulting model yields improved decisions and higher profit than the previous one. Due to the computational issues associated with the progressive (sliding scale) fiscal terms, a tighter formulation, a relaxation scheme, and an approximation technique are proposed. Chapter 4 presents a general multistage stochastic MILP model for endogenous uncertainty problems where decisions determine the timings of uncertainty realizations. To address the issue of exponential growth of non-anticipativity (NA) constraints in the model, a new theoretical property is identified. Moreover, three solution strategies, i.e. a k-stage constraint strategy; a NAC relaxation strategy; and a Lagrangean decomposition algorithm, are also proposed to solve the realistic instances and applied to process network examples. In Chapter 5, the deterministic formulations in Chapter 2 and 3 for oilfield development are extended to a multistage stochastic programming formulation to account for the endogenous uncertainties in field sizes, oil deliverabilities, water-oil-ratios and gas-oil-ratios. The Lagrangean decomposition approach from Chapter 4 is used to solve the problem, with parallel solutions of the scenarios. To improve the quality of the dual bound during this decomposition approach, a novel partial decomposition is proposed in Chapter 6. Chapter 7 presents a method to update the multipliers during the solution of a general twostage stochastic MILP model, combining the idea of dual decomposition and integer programming sensitivity analysis, and comparing it with the subgradient method. Finally, Chapter 8 summarizes the major findings of the dissertation and suggests future work on the subject.
Gupta, Vijay, "Modeling and Computational Strategies for Optimal Oilfield Development Planning under Fiscal Rules and Endogenous Uncertainties" (2013). Dissertations. 231.