摘要
目前,天然气脱酸工艺采用人为优化调整少量关键参数的模式很难达到全局最优。通过Aspen Hysys软件搭建了典型脱酸工艺流程,采用接口技术将Hysys软件和Matlab软件交互连接,以Hysys软件作为模拟平台,Matlab软件作为优化和数据库存储平台,在目标函数和约束条件的限制下,通过粒子群算法和径向基神经网络实现对多个决策变量的优化和自动调整。结果表明,所建的Hysys模型可以很好地反映脱酸工艺生产规律;在净化气达到二类指标的基础上,优化后MDEA循环量、MDEA溶液质量分数、原料气温度、再生塔塔板数等均有所下降,单位综合能耗值较优化前下降22.9%;再生塔回流比从3.5降至3.2,优化后回流比可保证再生塔的各层塔板工作在最优区间。HYSYS与MATLAB联用的模式可为天然气处理工艺参数的实时调整和优化提供参考。
In view of the current natural gas deacidification process,only a few key parameters are manually optimized and adjusted,which can not achieve the global optimal problem.The typical gas deacidification process was built by Aspen Hysys software.The interface technology was used to interactively connect Hysys software and Matlab software.Hysys software was used as the simulation platform,and Matlab software was used as the optimization and database storage platform.The decision variables are optimized and adjusted automatically by particle swarm optimization and radial basis neural network,which reduces the influence of human factors.The results show that the Hysys model can well reflect the production law of deacidification process.On the basis of the purification gas reaching the second class index,the MDEA circulation quantity,MDEA solution mass fraction,raw material temperature and the number of regenerated column plates all decreased after optimization,and the unit comprehensive energy consumption decreased by 22.9%compared with that before optimization.The reflux ratio of the regeneration tower is reduced from 3.5 to 3.2,and the optimized reflux ratio can ensure that each layer plate of the regeneration tower works in the optimal interval.The results can provide practical reference for real-time adjustment and optimization of natural gas treatment process.
作者
周嘉琦
ZHOU Jiaqi(The fourth oil production plant of Petro China Huabei Oilfield Company,Langfang,Hebei 065000,China)
出处
《世界石油工业》
2023年第2期50-56,共7页
World Petroleum Industry
关键词
天然气
脱酸工艺
HYSYS
MATLAB
粒子群算法
优化调整
natural gas
deacidification process
Hysys
Matlab
particle swarm optimization
optimization and adjustment
