Deep Anomaly Detection on Tennessee Eastman Process Data

This article evaluates the current state-of-the-art methods for anomaly detection on time series. The main focus is on chemical processes by running all experiments on the Tennessee Eastman process (TEP) data. After an introduction to anomaly detection and a short explanation of TEP, this research then provides a detailed evaluation of 27 methods.

Abstract

This paper provides the first comprehensive evaluation and analysis of modern (deep-learning-based) unsupervised anomaly detection methods for chemical process data. We focus on the Tennessee Eastman process dataset, a standard litmus test to benchmark anomaly detection methods for nearly three decades. Our extensive study will facilitate choosing appropriate anomaly detection methods in industrial applications. From the benchmark, we conclude that reconstruction-based methods are the methods of choice, followed by generative and forecasting-based methods.

Zum Volltext