Abstract Drilling optimization can provide significant value to an oil and gas project, especially in a low-price environment. This is generally approached by optimizing the rate of penetration (ROP) of the well, which may not always be the best strategy. Two additional strategies (or models) can be used to optimize a well – torque on bit (TOB) response to reduce vibrations at the bit or mechanical specific energy (MSE) to reduce the energy used by the bit. This paper evaluates these three models for drilling optimization based on several criteria. Models for ROP, TOB and MSE are built using a data-driven approach with the random forests algorithm using drilling operational parameters such as weight-on-bit, flow-rate, rotary speed, and rock strength as inputs. The drilling models are optimized using a meta-heuristic optimization algorithm to compute the ideal drilling operational parameters for drilling ahead of the bit. Machine learning is used to develop these models since these models are coupled which enable calculation of interaction effects. Results show that optimizing the ROP model leads to a 28% improvement in ROP on average, however, this also increases the MSE and the TOB which is undesirable. Optimizing the MSE model results in a (smaller) increase of ROP (20%). This is accompanied by a decrease in MSE (by 15%) and decrease in TOB (by 7%) which may result in longer bit life and additional savings over time. Hypothesis testing has been used to ensure that all simulations conducted in this paper show statistically significant results.

中文翻译：

---

用于钻井优化的耦合机器学习模型评估

摘要 钻井优化可为油气项目提供重要价值，尤其是在低价环境中。这通常是通过优化井的钻速 (ROP) 来实现的，这可能并不总是最好的策略。两种额外的策略（或模型）可用于优化油井 - 钻头扭矩 (TOB) 响应以减少钻头的振动或机械比能 (MSE) 以减少钻头使用的能量。本文根据几个标准对这三种模型进行钻井优化评估。ROP、TOB 和 MSE 模型是使用数据驱动方法和随机森林算法构建的，使用钻井操作参数（例如钻压、流速、转速和岩石强度）作为输入。使用元启发式优化算法优化钻井模型，以计算用于钻头前方钻井的理想钻井操作参数。机器学习用于开发这些模型，因为这些模型是耦合的，可以计算交互效果。结果表明，优化 ROP 模型可使 ROP 平均提高 28%，然而，这也增加了 MSE 和 TOB，这是不希望的。优化 MSE 模型会导致 ROP (20%) 的（较小）增加。这伴随着 MSE（降低 15%）和 TOB（降低 7%）的降低，随着时间的推移，这可能会导致更长的钻头寿命和额外的节省。假设检验已被用于确保本文中进行的所有模拟都显示出具有统计意义的结果。