Fault Diagnosis of Shale Gas Compressor Based on MKMC-ResCNN

Wang Xin; Li Zhigang; Li Dingxia; Lü Zhuolun; Wang Hua; Chen Ming; Tang Wei

doi:10.12473/CPM.202412100

您当前的位置：

首页 >

文章列表页 >

Fault Diagnosis of Shale Gas Compressor Based on MKMC-ResCNN

THEMATIC PAPER | 更新时间：2026-01-19

- Fault Diagnosis of Shale Gas Compressor Based on MKMC-ResCNN
- China Petroleum Machinery Vol. 53, Issue 10, Pages: 64-71(2025)
- 作者机构：
  
  中国石油集团川庆钻探川庆页岩气项目经理部
- 作者简介：
- 基金信息：
- DOI：10.12473/CPM.202412100
  CLC： TE974
- Received：27 December 2024，
  
  Revised：2025-05-12，
  
  Published：10 October 2025
- 稿件说明：
移动端阅览
王欣，李治钢，李定夏，等.基于MKMC-ResCNN的页岩气压缩机故障诊断［J］.石油机械，2025，53（10）：64-71.

Wang Xin，Li Zhigang，Li Dingxia，et al.Fault diagnosis of shale gas compressor based on MKMC⁃ResCNN［J］.China Petroleum Machinery，2025，53（10）：64-71.
王欣，李治钢，李定夏，等.基于MKMC-ResCNN的页岩气压缩机故障诊断［J］.石油机械，2025，53（10）：64-71. DOI： 10.12473/CPM.202412100.

Wang Xin，Li Zhigang，Li Dingxia，et al.Fault diagnosis of shale gas compressor based on MKMC⁃ResCNN［J］.China Petroleum Machinery，2025，53（10）：64-71. DOI： 10.12473/CPM.202412100.

摘要

页岩气压缩机作为核心设备其智能故障诊断对保障能源安全生产具有重大意义。针对传统诊断方法在复杂工况下面临的特征提取能力不足、多源信号融合效率低等问题，提出基于多尺度核多通道残差卷积神经网络（MKMC⁃ResCNN）的页岩气压缩机智能诊断方法。通过构建融合

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=98441912&type=

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=98441897&type=

8.29733276

2.70933342

、

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=98441886&type=

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=98441927&type=

8.29733276

2.70933342

、

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=98441913&type=

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=98441898&type=

8.29733276

2.62466669

等3种卷积核的并行残差模块，实现多尺度振动特征的层次化提取；采用跨通道特征融合机制整合六维传感器数据，增强工况适应能力；引入余弦退火学习率调度以优化模型收敛过程。试验采集包含5种典型故障状态（活塞磨损、活塞环断裂、气阀弹簧片失效、气阀阀片故障及正常工况）的2 240组多通道振动数据，构建跨工况验证体系。试验结果表明：该方法在5种故障分类任务中达到98.76%的总体准确率，较传统SVM模型提升14.53个百分点；多通道融合策略使变工况下的诊断准确率提高23.85个百分点；与标准ResCNN的对比结果表明，所设计的多尺度模块将准确率提升了4.51个百分点。该成果可为复杂工业设备的智能运维提供具有强泛化能力的解决方案。

Abstract

With the expansion of shale gas exploitation scale， the intelligent fault diagnosis of compressor as core equipment is of great significance to ensure the safe production of energy. To address the challenges of insufficient feature extraction ability and low efficiency in multi-source signal fusion faced by traditional diagnostic methods under complex operating conditions， an intelligent diagnosis method for shale gas compressors based on a Multi-scale Kernel Multi-Channel Residual Convolutional Neural Network （MKMC-ResCNN） was proposed. By constructing a parallel residual module that integrates three convolutional kernels of ［3×3］，［5×5］ and ［7×7］， the hierarchical extraction of multi-scale vibration features was realized. The cross-channel feature fusion mechanism was used to integrate the six-dimensional sensor data to enhance operational adaptability. The cosine annealing learning rate scheduling was incorporated to optimize the model convergence process. Relying on the real industrial scene， 2 240 sets of multi-channel vibration data including 5 typical fault states （piston wear， piston ring breakage， valve spring failure， valve plate failure and normal operating conditions） were collected to construct a cross-operating condition verification system. The test results show that the proposed method achieves an overall accuracy of 98.76% in five fault classification tasks， which is 14.53 precentage points higher than that of the traditional SVM model. The multi-channel fusion strategy improves the diagnostic accuracy under variable conditions by 23.85 precentage points. The comparison with the standard ResCNN shows that the designed multi-scale module improves the diagnostic accuracy by 4.51 precentage points. The research results provide a solution with strong generalization ability for the intelligent operation and maintenance of complex industrial equipment.

关键词

Keywords

references

郑马嘉，欧志鹏，伍亚，等．深层页岩气井压窜特征与产能维护对策：以四川盆地泸203井区北部为例［J］．天然气工业， 2024 ， 44 （ 8 ）： 95 - 106 ．

ZHENG M J ， OU Z P ， WU Y ， et al . Frac⁃hit characteristics and productivity maintenance strategies of deep shale gas wells： a case study of northern Lu 203 well block in the Sichuan Basin ［J］. Natural Gas Industry ， 2024 ， 44 （ 8 ）： 95 - 106 .

杨柳，唐洋，杨鑫，等．页岩气开采往复式压缩机智能故障诊断算法优化［C］∥ 第33届全国天然气学术年会．南宁：中国石油学会天然气专业委员会， 2023 ： 669 - 679 ．

YANG L ， TANG Y ， YANG X ， et al . Intelligent fault diagnosis algorithm optimization for reciprocating compressors in shale gas extraction ［C］∥ The 33rd National Natural Gas Academic Annual Conference . Nanning ： Natural Gas Professional Committee of China Petroleum Society ， 2023 ： 669 - 679 .

唐洋，肖枭，关绵涛，等．多源信号融合往复式压缩机故障诊断方法［J］．山东大学学报（工学版）， 2023 ， 53 （ 5 ）： 10 - 19 ．

TANG Y ， XIAO X ， GUAN M T ， et al . The multi-source signal fusion reciprocating compressor fault diagnosis method ［J］. Journal of Shandong University （Engineering Science）， 2023 ， 53 （ 5 ）： 10 - 19 .

杨长华，张敦辉，滕阳．页岩气田往复式压缩机监测及故障诊断技术分析［J］．中国设备工程， 2021 （增刊1 ）： 50 - 53 ．

YANG C H ， ZHANG D H ， TENG Y . Monitoring and fault diagnosis technology analysis for reciprocating compressors in shale gas fields ［J］. China Plant Engineering ， 2021 （ Sup.1 ）： 50 - 53 .

巫庆辉，魏宇晨，侯元祥，等．基于MDF与改进SqueezeNet的泵机组故障诊断方法研究［J/OL］．控制工程， 2025 ： 1 - 7 （ 2024-09-09 ）［ 2025-04-29 ］． https：∥doi.org/10.14107/j.cnki.kzgc.20240197 https://doi.org/10.14107/j.cnki.kzgc.20240197 ．

WU Q H ， WEI Y C ， HOU Y X ， et al . Research on fault diagnosis method for pump units based on MDF and improved SqueezeNet ［J/OL］. Control Engineering of China ， 2025 ： 1 - 7 （ 2024-09-09 ）［ 2025-04-29 ］. https：∥doi.org/10.14107/j.cnki.kzgc.20240197 https://doi.org/10.14107/j.cnki.kzgc.20240197 .

LIU R ， CHEN G ， YANG B ， et al . Dislocated time series convolutional neural architecture： an intelligent fault diagnosis approach for electric machine ［J］. IEEE Transactions on Industrial Informatics ， 2017 ， 13 （ 3 ）： 1310 - 1320 .

GAN M . Construction of hierarchical diagnosis network based on deep learning and its application in the fault pattern recognition of rolling element bearings ［J］. Mechanical Systems and Signal Processing ， 2016 ， 72-73 ： 92 - 104 .

QIN Y ， WANG X ， ZOU J Q . The optimized deep belief networks with improved logistic sigmoid units and their application in fault diagnosis for planetary gearboxes of wind turbines ［J］. IEEE Transactions on Industrial Electronics ， 2019 ， 66 （ 5 ）： 3814 - 3824 .

ZHAO M H ， KANG M ， TANG B P ， et al . Multiple wavelet coefficients fusion in deep residual networks for fault diagnosis ［J］. IEEE Transactions on Industrial Electronics ， 2019 ， 66 （ 6 ）： 4696 - 4706 .

SHI C M ， PANOUTSOS G ， LUO B . Using multiple-feature⁃spaces⁃based deep learning for tool condition monitoring in ultraprecision manufacturing ［J］. EEE Transactions on Industrial Electronics ， 66 （ 5 ）： 3794 - 3803 .

HU H X ， TANG B ， GONG X J ， et al . Intelligent fault diagnosis of the high⁃speed train with big data based on deep neural networks ［J］. IEEE Transactions on Industrial Informatics ， 2017 ， 13 （ 4 ）： 2106 - 2116 .

SHAO H D ， JIANG H K ， ZHANG H Z ， et al . Electric locomotive bearing fault diagnosis using a novel convolutional deep belief network ［J］. IEEE Transactions on Industrial Electronics ， 2018 ， 65 （ 3 ）： 2727 - 2736 .

PAN J ， ZI Y A ， CHEN J L ， et al . LiftingNet： a novel deep learning network with layerwise feature learning from noisy mechanical data for fault classification ［J］. IEEE Transactions on Industrial Electronics ， 2018 ， 65 （ 6 ）： 4973 - 4982 .

YAO Y ， ZHANG S ， YANG S X ， et al . Learning attention representation with a multi⁃scale CNN for gear fault diagnosis under different working conditions ［J］. Sensors ， 2020 ， 20 （ 4 ）： 1233 .

XING S B ， LEI Y G ， WANG S H ， et al . Distribution⁃invariant deep belief network for intelligent fault diagnosis of machines under new working conditions ［J］. IEEE Transactions on Industrial Electronics ， 2021 ， 68 （ 3 ）： 2617 - 2625 .

YAMASHITA R ， NISHIO M ， DO R K G ， et al . Convolutional neural networks： an overview and application in radiology ［J］. Insights Into Imaging ， 2018 ， 9 （ 4 ）： 611 - 629 .

干逸飞，吕品，郑树泉．面向光伏功率预测的残差深度学习模型［J］．计算机应用与软件， 2024 ， 41 （ 11 ）： 101 - 107 ．

GAN Y F ， LÜ P ， ZHENG S Q . Residual deep learning model for photovoltaic power prediction ［J］. Computer Applications and Software ， 2024 ， 41 （ 11 ）： 101 - 107 .

IOFFE S ， SZEGEDY C . Batch normalization： accelerating deep network training by reducing internal covariate shift ［DB/OL］. arXiv ， 2015 ： 03167， 2015 （ 2015-02-11 ）［ 2025-05-10 ］. https：∥arxiv.org/abs/1502.03167 https://arxiv.org/abs/1502.03167 .

SUN W T ， SU Y P ， WU X ， et al . A novel end-to-end 1D-ResCNN model to remove artifact from EEG signals ［J］. Neurocomputing ， 2020 ， 404 ： 108 - 121 .

JAKKULA V . Tutorial on support vector machine（svm）［D］. Washington ： Washington State University ， 2006 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Study on Dual-tree Complex Wavelet Transform Used in Rotor Simulation Fault Diagnosis

Application of SAX-based Association Rule Mining on Fault Diagnosis

Artificial Intelligence in Production Performance Prediction for Rod Pumping Systems： Application and Prospects

A Sample Expansion Method for Beam Pumping Unit Dynamometer Cards Based on Convolutional Variational Autoencoder

Fusion of Frequency Domain Attention and Trend Decomposition Linear Network for Lost Circulation Prediction

Related Author

Duan Lixiang

Hu Zhi

Yang Dazhong

Li Richao

Ma Jun

Bai Tangbo

Zhang Laibin

Wang Xuduo

Related Institution

College of Mechanical and Transportation Engineering

Patent Examination Cooperation Center of SIPO

Tarim Oilfield Company

School of Computer Science and Technology， Shandong University

Heavy Oil Development Company， PetroChina Xinjiang Oilfield Company

AI问答

⁰