Artificial lift methods (ALMs) lift the accumulated fluids from horizontal shale-gas-producing wells and help sustain well performance. An artificial lift infrastructure plan includes the selection of ALMs and their operating schedule. This paper presents two discrete-time large-scale nonconvex mixed-integer nonlinear programming models to solve the artificial lift infrastructure planning problem. Two equivalent mixed-integer linear programming models are formulated using the special structure of the nonlinear terms. A set of valid inequalities is defined to tighten the models and shorten solution times to two orders of magnitude, considering well production limitations. We incorporate endogenous uncertainty in ALM-dependent production rates and exogenous uncertainty in shale gas prices into the models. For a hypothetical case study under only endogenous uncertainties, the value of the stochastic solution is 5%. For the same case study, the exogenous uncertainty in gas prices does not change the optimum solution.

中文翻译：

---

考虑内生和外生不确定性的页岩气水平井人工举升基础设施规划

人工举升方法 (ALM) 从水平页岩气生产井中举升积液并帮助维持井的性能。人工举升基础设施计划包括 ALM 的选择及其运行时间表。本文提出了两种离散时间大规模非凸混合整数非线性规划模型来解决人工电梯基础设施规划问题。使用非线性项的特殊结构，制定了两个等效的混合整数线性规划模型。考虑到油井产量的限制，定义了一组有效的不等式以收紧模型并将求解时间缩短到两个数量级。我们将依赖 ALM 的生产率的内生不确定性和页岩气价格的外生不确定性纳入模型。对于仅存在内生不确定性的假设案例研究，随机解的值为 5%。对于同一个案例研究，天然气价格的外生不确定性不会改变最优解。