Abstract:Unsupervised anomaly detection (AD) methods, either reconstruction based or prediction based, determine anomalies based on residuals. Occasional mutations in a single variable can cause the residuals to exceed the limits. Indeed, such mutations are not variations in the operating mechanism of the system. Thus, system-level anomalies are challenging to characterize. Complex networks (or "graphs") are well adapted for modeling and characterizing the laws of evolution of complex systems. However, most industrial scenarios are without graphs. Hence, a self-supervised variational graph autoencoders (SS-VGAE) method is proposed. First, the multisource sensor dynamic graph is constructed through detrended cross-correlation analysis (DCCA). Second, target and self-supervised learning tasks are designed. The target task is to reconstruct the input graph structure to minimize the reconstruction loss. The self-supervised task is to learn the optimal center of the hypersphere in the latent space such that the mean features are gathered toward the center as much as possible. Multitask joint optimization allows high- and low-dimensional space features to be considered simultaneously, thereby improving the reliability of anomaly scores. Then, the distribution of anomaly scores is calculated and integrated into a system health indicator (HI). The system HI is more applicable to assist decision making. Finally, the superiority of the proposed method, namely, better detection accuracy and robustness, is demonstrated by nuclear personal computer transient analyzer (PCTRAN) simulation data and Skoltech anomaly benchmark (SKAB) data. Last but not least, systematic anomalies are found to make the correlation between the variables stronger.

Variational Autoencoder-Driven Adversarial Svdd for Power Battery Anomaly Detection on Real Industrial Data

Diagnosing Vehicles with Automotive Batteries

A Novel Battery Abnormality Detection Method Using Interpretable Autoencoder

MA-VAE: Multi-head Attention-based Variational Autoencoder Approach for Anomaly Detection in Multivariate Time-series Applied to Automotive Endurance Powertrain Testing

Detecting Electric Vehicle Battery Failure Via Dynamic-VAE.

Abnormal Data Detection for Industrial Processes Using Adversarial Autoencoders Support Vector Data Description

High-Precision Fault Detection for Electric Vehicle Battery System Based on Bayesian Optimization SVDD

Deep Convolutional Autoencoder for Assessment of Anomalies in Multi-stream Sensor Data

Self-Supervised Variational Graph Autoencoder for System-Level Anomaly Detection

Unsupervised Anomaly Detection using Batteries in eVTOL Vehicle Propulsion Test Bed

VAE-based Deep SVDD for anomaly detection

Data-Driven Fault Diagnosis and Early Safety Warning for Lithium-Ion Batteries

Multi-source self-supervised domain adaptation network for VRLA battery anomaly detection of data center under non-ideal conditions

Realistic fault detection of li-ion battery via dynamical deep learning

Enhanced Wavelet Transform Dynamic Attention Transformer Model for Recycled Lithium-Ion Battery Anomaly Detection

High-dimensional Data Abnormity Detection Based on Improved Variance-of-Angle (VOA) Algorithm for Electric Vehicles Battery

Sparse Representation GRU-AutoEncoder for Battery Fault Detection of Electric Vehicles

TeVAE: A Variational Autoencoder Approach for Discrete Online Anomaly Detection in Variable-state Multivariate Time-series Data

Self-Attention-Based Multivariate Anomaly Detection for CPS Time Series Data with Adversarial Autoencoders

DASVDD: Deep Autoencoding Support Vector Data Descriptor for Anomaly Detection

Sensing Anomaly of Photovoltaic Systems with Sequential Conditional Variational Autoencoder