Graduate Thesis Or Dissertation
 

Classification Modeling of Nuclear Power Plant Outage Severities with Complement Naïve Bayes and Bidirectional Encoder Representations from Transformers

Public Deposited

Downloadable Content

Download PDF
https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/zg64tv347

Descriptions

Attribute NameValues
Creator
Abstract
  • Low-level nuclear power plant outages in the United States can lead to unanticipated costs, potentially compromising the expected operation lifetime of the plant. Nuclear power plants are complex systems of interfacing components and highly regulated processes. This inherent complexity makes predicting outages from system dependencies very challenging. When outages do occur, natural language is used to explain the cause, effect, and solution of the outage. The purpose of this study is to construct a classification model capable of accurately predicting the severity of an unseen nuclear power plant outage using historical natural language and machine learning algorithms. The construction of a classification model leveraging historical light water reactor text data can provide experts better insight into outages of varying severity. The performance of a Naïve Bayes algorithm is compared to that of the state-of-the-art Bidirectional Encoding Representations from Transformers (BERT) model. To improve the performance of the BERT model, external fast reactor text data is applied to the training methodology.
  • Keywords: Nuclear Energy, Transformers, Nuclear Engineering, BERT, Large Language Models, Nuclear Power Plants, Natural Language Processing
License
Resource Type
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Rights Statement
Publisher
Peer Reviewed
Language

Relationships

Parents:

This work has no parents.

In Collection:

Items

Thumbnail Title Date Uploaded Visibility Actions
file details CastilloMichael2024.pdf 2024-01-05 Public Download