xpertsystems/oil011-sample

OIL-011是一个合成溢流和井喷场景数据集（样本），用于井控机器学习、防喷器分析、溢流检测建模和安全培训数据生成。该样本覆盖了12个全球盆地和10个地层类别的1,500口井，包含115,251行数据，其中108,000个时间点的钻井遥测数据链接在13个表格中。数据集包括井主数据、钻井时间序列、溢流事件、井喷场景、防喷器操作、节流管汇日志、气体流入剖面、压井操作、警报和警告、设备故障、安全响应日志、事故根本原因和场景标签等文件。数据生成基于行业标准（如API、IADC、IOGP、NORSOK、SINTEF等）进行校准和验证，总体验证得分为100/100（A+级）。数据集适用于溢流检测、防喷器可靠性机器学习、井喷升级预测、溢流类型分类、警报延迟回归、根本原因分析分类、压井方法选择和风险等级评分等多种用例。OIL-011 is a synthetic kick and blowout scenario dataset (sample) for well-control machine learning, BOP analytics, kick-detection modeling, and safety-training data generation. The sample covers 1,500 wells across 12 global basins and 10 formation classes, with 115,251 rows including 108,000 timepoints of drilling telemetry linked across 13 tables. The dataset includes files such as wells master, drilling timeseries, kick events, blowout scenarios, BOP operations, choke manifold logs, gas influx profiles, kill operations, alarms and warnings, equipment failures, safety response logs, incident root cause, and scenario labels. Data generation is calibrated and validated based on industry standards (e.g., API, IADC, IOGP, NORSOK, SINTEF), with an overall validation score of 100/100 (Grade A+). The dataset is suitable for various use cases including kick detection from timeseries, BOP reliability ML, blowout escalation prediction, kick type classification, alarm acknowledgment-delay regression, root cause analysis classification, kill method selection, and risk level scoring.