An online changepoint detection algorithm for highly correlated data

Maleki, Sepehr and Bingham, Chris and Zhang, Yu (2015) An online changepoint detection algorithm for highly correlated data. In: 3rd International Conference on Advances in Information Processing and Communication Technology - IPCT, 10 - 11 Dec 2015, Rome, Italy.

Documents
nm.php_id=7182
[img]
[Download]
[img] PDF
nm.php_id=7182 - Whole Document

1MB
Item Type:Conference or Workshop contribution (Paper)
Item Status:Live Archive

Abstract

An online 2-D changepoint detection algorithm for sensor-based fault detection, is proposed. The algorithm consists of a differential detector and a standard detector and can detect anomalies and meaningful changepoints while maintaining a low false-alarm rate. A new approach for determining a threshold is introduced and the efficiency of the algorithm is validated by an industrial example. It is thereby shown that the proposed algorithm can be used as an early warning indicator and prevent impending unit failures.

Keywords:Fault Detection, FDI, Changepoint Detection
Subjects:G Mathematical and Computer Sciences > G700 Artificial Intelligence
G Mathematical and Computer Sciences > G310 Applied Statistics
G Mathematical and Computer Sciences > G160 Engineering/Industrial Mathematics
Divisions:College of Science > School of Engineering
ID Code:22677
Deposited On:18 Mar 2016 09:40

Repository Staff Only: item control page