基于高斯过程回归和信息分析法的常压塔操作优化

Operational optimization of an atmospheric column using gaussian process regression and information analysis

针对常压塔操作优化问题,通过高斯过程回归建立常压塔的元模型,并用信息分析法进行迭代计算,最终获取常压塔的最优操作条件。用2个实验验证了该算法的有效性:(1)固定常压塔三侧线产品的全塔效益最大化;(2)全塔综合效益最大化。从结果中可以看到,采用高斯过程回归建立常压塔模型进行优化,能够提高常压塔的经济效益。

The paper mainly discussed operational optimization of the atmospheric distillation columns（ADCs）, the models were builded by using gaussian process regression, then the information analysis was used for iterative computations. It could eventually obtain the best operating conditions of ADCs. To verify the effectiveness of the algorithm, two case studies were carried out： （1） the net profit of ADC was maximized with fixing production rates for three side-draws; （2） the net profit of ADC was fully maximized. The results showed that it could increase the net profit of ADC by using the algorithm.