xpertsystems/oil028-sample

OIL-028 — 合成油罐存储数据集（样本）是一个免费、模式相同的预览版本，来自XpertSystems.ai的企业油罐存储数据集，用于库存预测、油罐利用率优化、溢油预防、油罐完整性机器学习、SCADA遥测分析和存储终端操作。该样本覆盖了500个油罐，涉及6种产品类型（原油、柴油、汽油、航空燃料、液化天然气、液化石油气）和3种油罐类型类别（浮顶、固定顶、低温），包含294,500行数据，分布在6个表格中，时间跨度为168小时（1周）的每小时时间序列。OIL-028的独特特征是质量平衡耦合的库存累积——每小时库存根据每个油罐的公式“inventory_t+1 = clip(0, capacity, inventory_t + inflow_t - outflow_t)”演变，产生真实的时间序列动态，并在0%和100%容量处具有现实的边界裁剪。OIL-028 — Synthetic Tank Storage Dataset (Sample) is a free, schema-identical preview of XpertSystems.ais enterprise tank storage dataset for inventory forecasting, tank utilization optimization, overfill prevention, tank integrity ML, SCADA telemetry analytics, and storage terminal operations. The sample covers 500 tanks across 6 product types (Crude Oil, Diesel, Gasoline, Jet Fuel, LNG, LPG) and 3 tank type categories (Floating Roof, Fixed Roof, Cryogenic), with 294,500 rows linked across 6 tables spanning 168 hours (1 week) of hourly time-series. OIL-028s distinctive feature is the mass-balance-coupled inventory accumulation — hourly inventory evolves as inventory_t+1 = clip(0, capacity, inventory_t + inflow_t - outflow_t) per tank, producing genuine time-series dynamics with realistic boundary clipping at 0% and 100% capacity.