Document Type

Thesis

Degree

Master of Science (MS)

Major/Program

Electrical Engineering

First Advisor's Name

Ou Bai

First Advisor's Committee Title

Committee Chair

Second Advisor's Name

Shekhar Bhansali

Second Advisor's Committee Title

Committee Co-Chair

Third Advisor's Name

Aparna Aravellia

Third Advisor's Committee Title

Committee Member

Fourth Advisor's Name

Leonel Lagos

Fourth Advisor's Committee Title

Committee Member

Keywords

anomaly detection, piping systems, autoencoder, lstm, machine learning, fiber optics, acoustic sensors, structural health monitoring, nuclear

Date of Defense

7-8-2020

Abstract

Structural health monitoring (SHM) comprises multiple methodologies for the detection and characterization of stress, damage, and aberrations in engineering structures and equipment. Although, standard commercial engineering operations may freely adopt new technology into everyday operations, the nuclear industry is slowed down by tight governmental regulations and extremely harsh environments. This work aims to investigate and evaluate different sensor systems for real-time structural health monitoring of piping systems and develop a novel machine learning model to detect anomalies from the sensor data. The novelty of the current work lies in the development of an LSTM-autoencoder neural network to automate anomaly detection on pipelines based on a fiber optic acoustic transducer sensor system. Results show that pipeline events and faults can be detected by the MLM developed, with a high degree of accuracy and low rate of false positives even in a noisy environment near pumps and machinery.

Identifier

FIDC009175

Share

COinS
 

Rights Statement

Rights Statement

In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/
This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).