引用本文:韩万龙,范峥,薛岗,王文珍,刘子兵,葛涛. 利用BP人工神经网络预测天然气中重组分对净化装置的影响[J]. 石油与天然气化工, 2018, 47(6): 1-6.
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 487次   下载 289 本文二维码信息
码上扫一扫!
分享到: 微信 更多
利用BP人工神经网络预测天然气中重组分对净化装置的影响
韩万龙1,范峥2,薛岗1,王文珍2,刘子兵1,葛涛1
1.西安长庆科技工程有限责任公司;2.西安石油大学化学化工学院
摘要:
由于天然气中的重组分会对脱硫装置的运行效果及产品气气质造成影响,这一实际生产问题需要得到有效的解决。在MDEA溶液吸收性能评价装置上测定了不同条件下的MDEA溶液吸收性能,系统地研究了不同重组分对MDEA溶液吸收性能的作用规律,采用多因素方差分析筛选了关键因素,以判定其影响程度的大小,并采用人工神经网络建立了天然气中重组分不利影响的预测模型。结果表明:天然气中的重组分i-C5、C6、C7、C8和C10对MDEA溶液吸收能力具有十分显著的影响,它们均属于BP神经网络预测模型的有效输入信号,模型预测值与真实值较为近似,BP人工神经网络表现出良好的准确性和稳定性。因此,利用BP人工神经网络能够准确、可靠地预测天然气中重组分对MDEA溶液吸收性能的不利影响。
关键词:  天然气  脱硫  多因素方差分析  BP人工神经网络
DOI:10.3969/j.issn.1007-3426.2018.06.001
分类号:
基金项目:中国石油天然气股份有限公司重大科技专项“长庆油田5000万吨持续高效稳产关键技术研究与应用”(2016E-05);西安石油大学研究生创新与实践能力培养项目“基于大数据技术的长输管道分布式阴极保护系统研究”(YCS17212057)
Effects prediction of heavy components in natural gas on purification unit by BP artificial neural network
Han Wanlong1, Fan zheng2, Xue Gang1, Wang Wenzhen2, Liu Zibing1, Ge Tao1
1. Xi’an Changqing Scientific & Technological Engineering Co., Ltd., Xi’an, Shaanxi, China;2. College of Chemistry & Chemical Engineering, Xi’an Shiyou University, Xi’an, Shaanxi, China
Abstract:
As the heavy components in natural gas will affect both the operation effect of the desulfurization unit and the gas quality of the product gas, this actual production problem should be solved effectively. The absorption performance of MDEA solution under different conditions was determined with the MDEA solution absorption performance evaluation device. The effects of different heavy components on the absorption performance of MDEA solution was studied systematically. The key factors were selected by multiple factor analyses of variance to determine the influence degree, and an artificial neural network was used to establish a prediction model for the adverse effects of heavy components in natural gas. The results showed that heavy components i-C5, C6, C7, C8 and C10 in the natural gas had obvious influences on the absorptive capacity of MDEA solution. All of them were effective input signals of the prediction model of BP neural network. The predicted value approximated to the real value, and the BP artificial neural network showed good accuracy and stability. Therefore, the BP artificial neural network could predict the adverse effects of heavy components in natural gas on the absorption performance of MDEA solution accurately and reliably.
Key words:  natural gas  desulfurization  multifactor analysis of variance  BP artificial neural network