xpertsystems/oil023-sample

OIL-023是一个合成催化剂降解数据集，专注于石油天然气下游炼化领域的催化剂失活建模、活性衰减预测、积碳沉积预测、再生周期优化、催化剂剩余使用寿命预测和更换经济性机器学习。该样本数据集包含50个反应器，覆盖11个工艺单元和10个全球区域，包含164,608行数据，分布在13个表格中，模拟了365天的每日运行情况。数据基于Bartholomew (2001)的失活机制和Arrhenius动力学原理生成，具有物理耦合特征，如催化剂活性与积碳的强负相关性。数据集适用于表格分类、回归和时间序列预测任务。OIL-023 is a synthetic catalyst degradation dataset for catalyst deactivation modeling, activity decay prediction, coke deposition forecasting, regeneration cycle optimization, catalyst RUL prediction, and replacement economics ML in the oil and gas downstream refining sector. The sample covers 50 reactors across 11 process units and 10 global regions, with 164,608 rows linked across 13 tables spanning 365 days of daily simulation. Data is generated based on Bartholomew (2001) deactivation mechanisms and Arrhenius-style kinetics, featuring strong physics coupling such as inverse activity-coke correlation. It is suitable for tabular classification, regression, and time-series forecasting tasks.