当前位置：首页 > 过刊浏览->2025年第51卷第4期
基于BP神经网络的顺序输送混油界面跟踪
Mixed oil interface tracking of batch transportation based on BP neural network
浏览(746) 下载(13)
- DOI:
- 10.3969/j.issn.1001-2206.2025.04.003
- 作者:
- 梁博, 蔡亮, 罗衍涛
LIANG Bo, CAI Liang, LUO Yantao
- 作者单位:
- 1. 国家石油天然气管网集团有限公司甘肃分公司, 甘肃兰州 730070;2. 中航油西南储运有限公司, 重庆 401122;3. 国家石油天然气管网集团有限公司山东分公司, 山东菏泽 274000
- 关键词:
- 混油界面; BP神经网络; 混油长度; 到站时间; 界面跟踪
mixed oil interface;BP neural network;length of mixed oil;arrival time;interface tracking
- 摘要:
-
针对长距离输油管道混油界面跟踪难题，提出基于BP神经网络的混油界面跟踪方案，构建高效、准确的混油界面跟踪模型，通过历史数据对模型进行训练，优化模型的结构和参数，提高其对混油界面复杂非线性关系的学习和映射能力。以国内某输油管道为例进行了模型验证，结果表明所建模型对混油界面到站时间、混油界面位置和混油长度预测效果较好，且在跟踪精度和工况适应性方面具有显著优势。
To address the challenge of mixed oil interface tracking in long-distance oil pipelines, a mixed oil interface tracking scheme based on back propagation (BP) neural network was proposed. An efficient and accurate mixed oil interface tracking model was constructed and trained with historical data. Its structure and parameters were optimized. Its ability to learn and map the complex nonlinear relationships of the mixed oil interface was enhanced. The model was validated with a domestic oil pipeline as an example. The results show that the developed model effectively predicts the arrival time, position, and length of the mixed oil interface, significantly superior in tracking accuracy and adaptability to different operating conditions.