基于UKF-WNN混合模型的海管段塞流预测与控制方法研究  

Research on the prediction and control method of offshore pipeline plug based on UKF-WNN hybrid model

作　　者：周波 廉晓龙 李斌[2] 孙楠 黄磊 王昆 ZHOU Bo;LIAN Xiaolong;LI Bin;SUN Nan;HUANG Lei;WANG Kun(Hefei General Machinery Research Institute Co.Ltd.,Hefei 230031,China;China Oilfield Services Limited,Tianjin 300456,China)

机构地区：[1]合肥通用机械研究院有限公司,合肥2300311 [2]中海油田服务股份有限公司,天津300456

出　　处：《流体机械》2024年第1期76-82,89,共8页Fluid Machinery

基　　金：国家重点基础研究发展计划(973计划)项目(2018YFC03102004-2)。

摘　　要：针对油气生产系统在深海情况下难以准确实时控制段塞流的问题,建立了基于UKF-WNN混合模型的海管段塞流预测与控制方法,并在不同阀门开度下,对不同的段塞流控制方案进行了比较研究。结果表明:UKF-WNN混合模型能够较好地拟合顶部压力、流量与阀门开度同底部压力的关系,与现场试验对比,其均方误差较UKF和WNN分别降低了44%、47%,在系统输入存在干扰的情况下能够为控制器提供较为精确的估计值,使得系统达到稳定。WNN的加入有效改善了控制系统的稳定性,体现了人工智能方法在段塞流控制领域获得应用的可能性。本文提出的方法对改善海管段塞流的预测和控制精度,保障海管安全具有一定的理论和实践意义。In order to solve the problem that it is difficult for oil and gas production system to accurately control slug flow in real time under deep sea conditions,a prediction and control method for slug flow in offshore pipelines based on UKF-WNN Mixture model was established.The results show that the UKF-WNN Mixture model can better fit the relationship between top pressure,flow,valve opening and bottom pressure.Compared with field tests,its Mean squared error is 44%and 47%lower than that of UKF and WNN,respectively.It can still provide more accurate estimates for the controller when there is interference in the system input,making the system stable.The addition of WNN effectively improves the stability of the control system,demonstrating the possibility of application of artificial intelligence methods in the field of slug flow control.The method proposed in this study has certain theoretical and practical significance in improving the prediction and control accuracy of slug flow in offshore pipelines,and ensuring the safety of offshore pipelines.

关 键 词：海管段塞流 状态估计 无迹卡尔曼滤波 小波神经网络 混合模型 

分 类 号：TH138[机械工程—机械制造及自动化] TE973[石油与天然气工程—石油机械设备]