基于矿物组分的岩石抗钻特性评价及应用

Evaluation and Application of Rock Anti-drillability CharacteristicsBased on Mineral Composition

PL区块沧浪铺组地层岩石非均质性强,钻井提速困难,严重制约了油气勘探开发进度。因此,精确评价地层岩石的抗钻特性对优选钻头、提高钻井时效具有重要意义。本文采用BP-ANN模型建立了基于矿物组分含量的地层抗钻参数预测模型,首先基于实验数据进行皮尔逊相关性分析,筛选模型输入变量;其次利用Leave-One-Out-Cross-Validation(LOOCV)数据集划分方法,以均方误差(MSE)、决定系数(R 2)和批次(Epochs)三种指标对BP-ANN模型进行训练并评估其性能;最后使用训练后的BP-ANN模型对该区域沧浪铺组地层岩石进行抗钻特性评价,在已有钻头的基础上,依据岩石抗钻参数优化钻头设计并应用于现场。研究表明:①训练后的BP-ANN模型可以准确预测PL区块PT101井沧浪铺组地层岩石的抗钻参数,预测精度大于90%;②基于地层岩石抗钻特性的钻头优化方案在PT101井取得了良好的应用效果,相较于邻井同层位钻头,平均机械钻速提高约110%,单个钻头进尺超过200 m,现场提速效果明显。文章研究结果可以准确评价地层岩石的抗钻特性,为钻头选型、优化提供理论依据。

The heterogeneity of rock formations in the PL block of the Canglangpu Formation poses challenges to drilling acceleration,significantly impeding the progress of oil and gas exploration and development.Therefore,the precise evaluation of the rock formation′s anti-drillability characteristics holds crucial significance for optimizing drill selection and improving drilling efficiency.This paper employs BP-ANN to establish a predictive model for anti-drillability parameters based on the mineral composition of the rock.The study begins with Pearson correlation analysis based on experimental data to screen model input variables.Subsequently,the Leave-One-Out-Cross-Validation(LOOCV)dataset splitting method is utilized,and the BP-ANN model is trained and evaluated for performance using metrics such as mean square error(MSE),determination coefficient(R 2),and epochs.Finally,the trained BP-ANN model is applied to assess the anti-drillability characteristics of the Canglangpu Formation in the PL block.Building upon existing drill bits,the rock′s anti-drillability parameters are used to optimize drill bit design,which is then implemented in the field.The research findings indicate that:The trained BP-ANN model accurately predicts the anti-drillability parameters of the Canglangpu Formation in the PT101 well within the PL block,with a prediction accuracy exceeding 90%.The drilling optimization strategy based on the anti-drillability characteristics of the rock formation demonstrates favorable results in the PT101 well.Compared to neighboring wells with the same stratigraphic position,there is an average mechanical drilling speed increase of approximately 110%,and individual drill bits achieve a penetration depth exceeding 200 meters,showcasing a significant on-site acceleration effect.The research results can accurately evaluate the rock anti-drillability characteristics of geological formations,providing a theoretical basis for drill bit selection and optimization.