| 曹峥,邓建强,王泽良,宣炳蔚,郭希健.基于Elman神经网络的流体管道泄漏点检测定位[J].装备环境工程,2020,17(4):8-13. CAO Zheng,DENG Jian-qiang,WANG Ze-liang,XUAN Bing-wei,GUO Xi-jian.Leakage Detection and Localization of Fluid Pipeline Based on Elman Neural Network[J].Equipment Environmental Engineering,2020,17(4):8-13. |
| 基于Elman神经网络的流体管道泄漏点检测定位 |
| Leakage Detection and Localization of Fluid Pipeline Based on Elman Neural Network |
| 投稿时间：2019-12-20 修订日期：2020-01-15 |
| DOI：10.7643/issn.1672-9242.2020.04.002 |
| 中文关键词: 水力输运 泄漏定位 负压波 神经网络 |
| 英文关键词:hydraulic transportation, leak localization, negative pressure wave, neural network |
| 基金项目:中央高校基本科研业务费专项(xjh012019022) |
|
|
| Author | Institution |
| CAO Zheng | Xi′an Jiaotong University, Xi′an 710049, China;Shaanxi Key Laboratory of Energy Chemical Process Intensification, Xi′an 710049, China |
| DENG Jian-qiang | Xi′an Jiaotong University, Xi′an 710049, China;Shaanxi Key Laboratory of Energy Chemical Process Intensification, Xi′an 710049, China |
| WANG Ze-liang | Xi′an Jiaotong University, Xi′an 710049, China;Shaanxi Key Laboratory of Energy Chemical Process Intensification, Xi′an 710049, China |
| XUAN Bing-wei | Xi′an Jiaotong University, Xi′an 710049, China;Shaanxi Key Laboratory of Energy Chemical Process Intensification, Xi′an 710049, China |
| GUO Xi-jian | Xi′an Jiaotong University, Xi′an 710049, China;Shaanxi Key Laboratory of Energy Chemical Process Intensification, Xi′an 710049, China |
|
| 摘要点击次数: |
| 全文下载次数: |
| 中文摘要: |
| 目的 通过Elman神经网络预测对泄漏点进行过检测定位。方法 基于流体压力波的负压波法及反馈型Elman神经网络方法，开展水力输运管道的泄漏定位研究。利用Flowmaster仿真软件中的水力输运模型建立长度为1100 m的一维管路系统，针对此系统开展不同管路状态参数下的数值仿真计算。结果 通过小波变换技术实现了数据降噪与奇异点捕捉，完成了泄漏点位置的估算。同时，借助反馈型Elman神经网络，开展了不同泄漏工况下的网络训练和预测，利用经过训练的神经网络对所选取的5组泄漏点完成了定位预测，最大测试误差为1.83%。结论 通过Elman神经网络预测得到的结果与实际泄漏位置进行对比，验证了反馈型神经网络方法在管路泄漏智能定位问题中的准确性与有效性。 |
| 英文摘要: |
| The paper aims to detect and locate the leakage points through the Elman neural network prediction test. Based on the negative pressure wave method and the feedback Elman neural network method of fluid pressure wave, the leakage localization of water pipelines were researched. By using one dimensional hydraulic model in Flowmaster, a pipeline system with a total length of 1100m was established, and numerical simulation for various leakage conditions was carried out. The leakage location was estimated after data noise reduction and singular point capture through wavelet transform. Meanwhile, with the help of the feedback Elman neural network, network training and prediction were carried out under different leakage conditions. Five groups of leakage locations were predicted by the trained neural network. The maximum error for the leak location prediction was 1.83%. The accuracy and effectiveness of the feedback neural network method for the pipeline leakage localization was verified through the comparison between the actual values and the results calculated based on Elman neural network. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
