Pump Fault Classification based on Autoencoding Convolutional Neural Network Residuum

This paper deals with the fault classification of centrifugal pumps, based on the residuum between the output and the input of an Autoencoding Convolutional Neural Network previously trained for abnormal behaviour detection. The proposed classification method performs a dimensional reduction of the...

Full description

Saved in:
Bibliographic Details
Published in:2023 IEEE 28th International Conference on Emerging Technologies and Factory Automation (ETFA) pp. 1 - 7
Main Authors: Vasiliev, Iulian, Frangu, Laurentiu, Cristea, Mihai Lucian, Cristian Costea, Mihai
Format: Conference Proceeding
Language:English
Published: IEEE 12-09-2023
Subjects:
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper deals with the fault classification of centrifugal pumps, based on the residuum between the output and the input of an Autoencoding Convolutional Neural Network previously trained for abnormal behaviour detection. The proposed classification method performs a dimensional reduction of the residuum vector using Principal Component Analysis, and then, based on the first 3 principal components, classifies the data, using a simple rule-based algorithm, in one of the classes: normal, clogged filter, broken fan blade, detached rotor section and other fault source. The classification method proved to be reliable in an industrial application, providing a 90% correct identification of the machine condition.
ISSN:1946-0759
DOI:10.1109/ETFA54631.2023.10275399