Pipeline Leak Diagnosis Based on SDAE

To solve the problem that different oil and gas pipeline leakage signals are difficult to distinguish, a method of SDAE oil and gas pipeline leakage diagnosis was proposed. Firstly, the image preprocessing of one-dimensional signal is introduced to convert the original acoustic emission signal into...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2023 5th International Conference on Control and Robotics (ICCR) S. 221 - 224
Hauptverfasser: Lang, Xianming, Sui, Dongye, Lv, Yuanhao, Yuan, Kaixin, Zhang, He
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 23.11.2023
Schlagworte:
Acoustic emission
Autoencoder
Feature extraction
Leakage diagnosis
Noise reduction
Oil and gas pipeline
Oils
Pipelines
Robots
Robustness
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:To solve the problem that different oil and gas pipeline leakage signals are difficult to distinguish, a method of SDAE oil and gas pipeline leakage diagnosis was proposed. Firstly, the image preprocessing of one-dimensional signal is introduced to convert the original acoustic emission signal into two-dimensional feature gray map. Then, the feature extraction of two-dimensional feature gray image is carried out by the stack denoising autoencoder. Finally, the pipeline leakage diagnosis is carried out. The experimental results show that the method can accurately diagnose different types of pipeline leakage with an accuracy of 99.65%. Compared with one-dimensional original signal, the accuracy of two-dimensional gray image is improved by 27.72%. Compared with other methods, the accuracy of this method is improved by 3.08%. Compared with other methods, the superiority of this method in leakage diagnosis is further verified.
DOI:10.1109/ICCR60000.2023.10444746