Selected Journal of the Excellence Action Plan for China's Science and Technology Journals (Phase II),Selected Journal of Chinese Applied Core Journals (Extended Edition) / CACJ (Extended),Indexed in the Chinese Core Journals (Selection) Database
Article View

基于CAE-XGBoost模型的燃气调压器故障诊断研究

Issue Number Issue 06, 2025 [View other articles in this issue]
DOI
Receiving Time
Revision Time
Publish Date
Publication Date 2025.06.15
Manuscript ID
Author 王 强1,李泽明1,周梦琪2
Organization 1.无锡华润燃气有限公司
2.润智科技有限公司

Abstract

 燃气调压器是燃气输配系统中的关键设备,其安全稳定运行一直受到广泛关注。本文提出了一种基于CAE-XGBoost的中低压调压器故障诊断方法。首先,使用卷积自编码器(CAE)网络对调压器出口压力信号进行无监督学习,在重构误差允许范围内,通过编码器部分的输出,提取出口压力信号的深层次特征;再使用梯度提升集成学习模型(XGBoost)对提取的特征进行故障分类。使用121台在线调压器设备进行验证,针对正常、内漏、喘振和关闭压力过高4种故障类型进行识别。结果表明,基于CAE-XGBoost模型的故障识别准确率均达到95%以上,验证了模型的有效性。

Keywords

燃气调压器;卷积自编码器;XGBoost;故障诊断

References

Declaration:

The English version of this article is translated with the assistance of AI.

Related Articles

Contact Us
Urban Gas Magazine Co., Ltd.
Address:
Room 1207, Tower B, Financial Street Investment Plaza, Xicheng District, Beijing 100045, China
Tel:
+86-10-66020170
Advertising Department:
+86-18611817393
Circulation Department:
+86-18518361665

 

Gas800.com 版权所有 Copyright @ 2006-2026 京ICP备19057877号-1

京公网安备 11010802030632号