Multivariate Time Series Anomaly Detection in Industry 5.0

TL;DR

The paper tackles anomaly detection in industrial multivariate time series under Industry 5.0, addressing data quality and unlabeled data challenges. It proposes a Time2Vec-inspired autoencoder to generate embeddings of MTS and applies multiple one-class classifiers to detect anomalies in the embedding space. In a Bonfiglioli gear-production case, the embedding-based approach demonstrates robustness to sporadic noise and often outperforms traditional reconstruction-based autoencoders, especially in noisy scenarios. The work holds practical value for real-time monitoring and predictive maintenance and points to future extensions, including broader datasets and remaining useful life estimation.

Abstract

Industry5.0 environments present a critical need for effective anomaly detection methods that can indicate equipment malfunctions, process inefficiencies, or potential safety hazards. The ever-increasing sensorization of manufacturing lines makes processes more observable, but also poses the challenge of continuously analyzing vast amounts of multivariate time series data. These challenges include data quality since data may contain noise, be unlabeled or even mislabeled. A promising approach consists of combining an embedding model with other Machine Learning algorithms to enhance the overall performance in detecting anomalies. Moreover, representing time series as vectors brings many advantages like higher flexibility and improved ability to capture complex temporal dependencies. We tested our solution in a real industrial use case, using data collected from a Bonfiglioli plant. The results demonstrate that, unlike traditional reconstruction-based autoencoders, which often struggle in the presence of sporadic noise, our embedding-based framework maintains high performance across various noise conditions.

Figures (1)

• Figure 1: Ad-hoc T2V-based AE Architecture.

Theorems & Definitions (1)

• definition \@thmcounterdefinition