基于大数据的高含硫天然气脱硫工艺优化

为了解决高含硫天然气脱硫工艺中脱硫选择性差、能耗高等问题,提出了基于大数据的高含硫天然气脱硫工艺优化方法。首先,通过工艺流程分析,发现对性能指标有显著影响的决策参数,建立无迹卡尔曼滤波神经网络动态模型,获知了脱硫工艺的潜在规律;然后,针对原脱硫工艺中H_2S、CO_2过分脱除问题,采用偏好多目标优化的方法,分别以H2S浓度逼近2.5 mg/m~3、CO_2浓度逼近2%为目标函数,采用非支配性排序遗传算法对模型进行多目标优化,获得了最佳工艺参数。采集某高含硫天然气净化厂脱硫单元2014年1—12月的生产数据,取前80%数据作为训练集,后20%数据作为测试集,进行了仿真实验。结果表明:1所建立的动态模型能够较好地反映脱硫工艺生产规律;2优化结果建议适当降低一级吸收塔温度,提高二级吸收塔温度,提高闪蒸罐压力,并减少胺液循环量;3优化后净化气中H_2S浓度将由0.62 mg/m~3提高至3.22 mg/m~3,CO_2浓度由1.19%提高至1.99%,脱硫选择性显著提高;4相对胺液循环量下降16.67%,蒸汽消耗量减少,净化气产率提高0.8%,总体实现了增产节能降耗的目的。

Optimization of acid gas sweetening technology based on big data

In this paper, an optimization method based on big data was proposed to improve the desulfurization selectivity and reduce the energy consumption of traditional acid gas sweetening technologies. At first, decision-making parameters which have significant effects on the performance indexes were identified by analyzing the sweetening process. Then, a dynamic model of unscented Kalman filter neuralnetwork was built to describe the potential rules of the sweetening process. And finally, a preference-based multi-objective optimizationwas adopted to address the issue of excessive removal of H₂S and CO₂ in the original process. The multi-objective optimization was carried out on the model by using the non-dominated sorting genetic algorithm with the concentration of H₂S and CO₂ approaching 2.5 mg/m3 and 2% respectively as the objective functions. In this way, the optimal process parameters were obtained. The real production data of a certain acid gas sweetening plant from January to December in 2014 were acquired for simulation experiments with the first 80% samples as the training set while the left as the testing set. It is shown that the dynamic model can better present the production rules of the sweetening process; that based on the optimization results, it is recommended to decrease the temperature of primary absorption column,increase the temperature of secondary absorption column, raise the pressure of flash drum and reduce the circulation rate of amine solution appropriately; that after the optimization, the desulfurization selectivity is improved significantly with H₂S concentration of the purified gas rising from 0.62 to 3.22 mg/m3, and the CO₂ concentration rising from 1.19% to 1.99%; and that the circulation rate of amine solution drops by 16.67%, the steam consumption decreases, and the production rate of purified gas increases by 0.8%. On the whole, the target of production increase and energy consumption decrease is reached.