Presentation Details
  1. Home
  2. Presentation List
  3. Presentation Details
Anomaly Warning Using LSTM with Canonical Correlation Analysis and Sequential Probability Ratio Test for Oil Pumps
ID：121 View Protection：ATTENDEE Updated Time：2025-11-10 15:42:50 Hits：126 Poster Presentation

Start Time：Pending（Asia/Shanghai）

Duration：Pending

Session：No Session »

Presentation File

Tips: Only the registered participant can access the file. Please sign in first.

Abstract
Oil pumps play a crucial role to provide the safe and reliable operation power in the oil transport system. They are usually monitored by various types of sensors with diverse parameters and variables to capture their working statues, so as to inhibit anomalies even faults by taking foresight actions, also as a prerequisite for lifetime and performance prediction, anomaly warning, as well as fault diagnosis and prediction. In this paper, we propose an anomaly warning method by using the canonical correlation analysis (CCA), sequential probability ratio test (SPRT), and long short-term memory network (LSTM) with fusing multi-type field data for oil pumps. First, the multi-type original data of the oil pump needs to be collected, cleaned and labeled as available and analyzable item in advance. Then, the key variable features are extracted by using the CCA with analyzing their correlations. Subsequently, the most relevant variables are input into the LSTM to predict the given variable signal trend. Additionally, the deviation errors between the predicted and real monitoring values are computed and input into the SPRT to garner the anomaly warning result. Finally, a case study of the oil pump with 13 on-site variables of four types is performed to demonstrate the effectiveness of the proposed method.

 
Keywords
Oil pumps,Anomaly warning,Canonical correlation analysis,long short-term memory network,Sequential probability ratio test
Speaker
Zhichao Wang
Assistant Researcher Tsinghua University

Submission Author
Zhichao Wang Tsinghua University
Ping Wang Tsinghua University
Tongyu Li Tsinghua University
Chao Liu Tsinghua University
Dongxiang Jiang Tsinghua University
Submit Comment
Verify Code Change Another
All Comments
Important Date

  • Conference Date

    Nov 21

    2025

    to

    Nov 23

    2025

  • Oct 20 2025

    Draft paper submission deadline

  • Dec 08 2025

    Registration deadline

Sponsored By
IEEE Instrumentation and Measurement Society
South China University of Technology
Organized By
South China University of Technology
Contact Information
Login to view full contact details
Previous Conferences