Abstract The current scheme for developing shale reservoirs necessitates special considerations while estimating the reserve. While reservoir characteristics lead to an extended infinite acting flow regime, completion schemes could result in a series of linear flows. Therefore, the initial linear flow does not have to be followed by a boundary-dominated flow. Overlooking this observation leads to unphysical Arps' exponents and overestimations of the Estimated Ultimate Recovery (EUR). We are proposing a workflow to overcome these challenges and honor the inherited uncertainty while using the classic Arps (1956) hyperbolic forecast. Our workflow starts with identifying the current flow regime of the well where two intermediate flow regimes are considered, namely Linear-post-Linear (LPL) and Linear Post Linear Post Linear (LPLPL) flow regimes. We deterministically forecasted the boundary-dominated wells and probabilistically forecasted the rest. We used the distribution of the current flow regime in the field to forecast the transient wells. The well features of the infinite/semi-infinite wells are stochastically sampled from the field database combining well features of the boundary-dominated wells. After that, Monte Carlo simulation is employed to probabilistically estimate the EUR. We constrained the well life by an economical limit or a maximum of 40 years. We selected Bone Springs formation as a field study of this workflow. We found that the LPL is the common current flow regime. A reduction in Arps’ exponent is observed when well experience an intermediate linear flow before boundary-dominated flow. Significant number of interference wells are identified through their Water/Oil Ratio (WOR) signature. We also studied the evolution of EUR which suggests that more than 75% of the production history is needed for the deterministic methods to provide reliable estimates of the EUR. We generated heat maps of the area of interest to summarize the EUR and remaining reserve results.

中文翻译：

---

页岩井最终采收率估算的蒙特卡罗模拟与生产分析

摘要 目前的页岩储层开发方案在估算储量时需要特殊考虑。虽然储层特性导致扩展的无限作用流态，但完井方案可能导致一系列线性流动。因此，初始线性流不必跟随边界控制流。忽视这一观察会导致非物理的 Arps 指数和估计最终回收率 (EUR) 的高估。我们提出了一个工作流程来克服这些挑战并尊重继承的不确定性，同时使用经典的 Arps (1956) 双曲线预测。我们的工作流程首先确定井的当前流态，其中考虑了两种中间流态，即线性后线性 (LPL) 和线性后线性后线性 (LPLPL) 流态。我们确定性地预测了边界控制的井，并概率性地预测了其余的井。我们使用现场当前流态的分布来预测瞬态井。无限/半无限井的井特征是从结合边界控制井的井特征的现场数据库中随机采样的。之后，采用蒙特卡罗模拟对 EUR 进行概率估计。我们通过经济限制或最长 40 年来限制油井寿命。我们选择 Bone Springs 形成作为此工作流程的现场研究。我们发现 LPL 是常见的电流流态。当井在边界主导流之前经历中间线性流时，观察到 Arps 指数的减少。通过其水油比 (WOR) 特征识别出大量干扰井。我们还研究了 EUR 的演变，这表明确定性方法需要超过 75% 的生产历史才能提供可靠的 EUR 估计值。我们生成了感兴趣区域的热图来总结欧元和剩余储备结果。