Anomaly Detection for Industrial Big Data

TL;DR

This paper introduces a scalable, data-driven anomaly detection method tailored for industrial big data, leveraging independent ordinations and joint distribution analysis to identify anomalies in multivariate time-series data from IIoT sensors.

Contribution

The paper presents a novel anomaly detection technique that operates efficiently at industrial scale and generalizes across various multivariate datasets using bootstrapped partitions.

Findings

Effective in detecting anomalies in large-scale industrial data

Applicable to diverse multivariate datasets

Operates efficiently at industrial data scales

Abstract

As the Industrial Internet of Things (IIoT) grows, systems are increasingly being monitored by arrays of sensors returning time-series data at ever-increasing 'volume, velocity and variety' (i.e. Industrial Big Data). An obvious use for these data is real-time systems condition monitoring and prognostic time to failure analysis (remaining useful life, RUL). (e.g. See white papers by Senseye.io, and output of the NASA Prognostics Center of Excellence (PCoE).) However, as noted by Agrawal and Choudhary 'Our ability to collect "big data" has greatly surpassed our capability to analyze it, underscoring the emergence of the fourth paradigm of science, which is data-driven discovery.' In order to fully utilize the potential of Industrial Big Data we need data-driven techniques that operate at scales that process models cannot. Here we present a prototype technique for data-driven anomaly detection to operate at industrial scale. The method generalizes to application with almost any multivariate dataset based on independent ordinations of repeated (bootstrapped) partitions of the dataset and inspection of the joint distribution of ordinal distances.