摘要
提出了一种基于神经网络的燃气管道泄漏检测方法。首先将管道分成若干段,在考虑每段管道气流的输入/输出压力的前提下,利用基于物联网现象的无线传感器网络来收集泄漏检测点所需的信息。基于这些信息构建特征工程,并进一步将样本集划分为训练集和验证集,使用训练集对泄漏检测神经网络进行训练,来调整神经网络权值,使用验证集评价神经网络的故障检测系统的性能。利用实际管道数据对网络进行训练,以确保所提出的方法适用于实际工程。
A neural network-based method for leakage detection of gas pipeline by using gas flow pattern was proposed. Firstly, the pipeline was divided into several segments. On the premise of considering the input/output pressure of the air flow in each segment of the pipeline, the wireless sensor network based on the phenomenon of Internet of Things was used to collect the information required by the leak detection point. Based on these information,feature engineering was constructed, and the sample set was further divided into training set and verification set. The training set was used to train the leak detection neural network to adjust the weight of the neural network, and the verification set was used to evaluate the performance of the fault detection system of the neural network. Practical data gathered from a real life pipeline were used to train the network to make sure that the proposed method was applicable real projects.
作者
王晓敏
WANG Xiao-min(Xi’an International University,Xi'an Shaanxi 710077,China)
出处
《当代化工》
CAS
2022年第8期1932-1937,共6页
Contemporary Chemical Industry
基金
陕西省教育厅自然科学基金项目,秦岭北麓森林火灾无线传感器网络监测技术研究(项目编号:18JK1132)。
