Score: 0

LSTM VS. Feed-Forward Autoencoders for Unsupervised Fault Detection in Hydraulic Pumps

Published: January 16, 2026 | arXiv ID: 2601.11163v1

By: P. Sánchez , K. Reyes , B. Radu and more

Potential Business Impact:

Finds broken machines before they stop working.

Business Areas:
Machine Learning Artificial Intelligence, Data and Analytics, Software

Unplanned failures in industrial hydraulic pumps can halt production and incur substantial costs. We explore two unsupervised autoencoder (AE) schemes for early fault detection: a feed-forward model that analyses individual sensor snapshots and a Long Short-Term Memory (LSTM) model that captures short temporal windows. Both networks are trained only on healthy data drawn from a minute-level log of 52 sensor channels; evaluation uses a separate set that contains seven annotated fault intervals. Despite the absence of fault samples during training, the models achieve high reliability.

Page Count
10 pages

Category
Computer Science:
Machine Learning (CS)