摘要
基于Aspen HYSYS软件建立了与某炼油厂有限实际生产数据相吻合的连续催化重整装置的机理模型。考虑多种生产可能性,扩展数据范围得到完整的装置产品预测数据集;与常用的BP神经网络作对比,采用训练速度快、预测精度高、适合非线性过程建模的LightGBM决策树模型,以该催化重整装置的4个反应器的温度和循环氢流量为特征变量,分别以戊烷、二甲苯、C 6、重整汽油、氢气的流量和氢气纯度为目标建立了6个单目标数据驱动产品预测模型。通过对特征变量和目标之间的相关性分析,进行10折交叉验证,得到了特性变量的重要度排序,从而针对不同生产目标找出影响最大的操作变量。结果表明,使用LightGBM建立模型的预测准确度比BP神经网络的预测准确度有大幅度提升。
Based on the Aspen HYSYS software platform,the mechanism model of a continuous catalytic reforming unit was established based on limited production data from a refinery.With considering various production status,limited production data could be extended to a more complete data for product prediction.Furthermore,the LightGBM decision tree model,which is suitable for nonlinear process modeling,was used to model the catalytic reforming unit.Compared with the BP neural networks commonly used in current similar studies and with the same training speed,the LightGBM models have shown higher prediction accuracy through the 10-fold cross validation tests.The temperatures of the four reactors and the recycle hydrogen consumption were selected as characteristics and operational variables,and product flowrates of pentane,xylene,C 6,gasoline,hydrogen and the purity of hydrogen were used as objectives.Six single objective data-driven product prediction models were established.Finally,the ranks of feature importance were obtained by correlation analysis between the characteristics and objectives.The simulation results could provide decision guidance for refinery’s online optimal operation.
作者
刘禹含
曹萃文
LIU Yuhan;CAO Cuiwen(Key Laboratory of Ministry of Education of Chemical Process Advanced Control and Optimization Technology,East China University of Science and Technology,Shanghai 200237,China)
出处
《石油学报（石油加工）》
EI
CAS
CSCD
北大核心
2020年第4期756-766,共11页
Acta Petrolei Sinica(Petroleum Processing Section)
基金
国家自然科学基金项目(61673175,61573144)资助。
