東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

FindBook

Google Book

Amazon

博客來

Auto Machine Learning Applications for Nuclear Reactors: Transient Identification, Model Redundancy and Security.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Auto Machine Learning Applications for Nuclear Reactors: Transient Identification, Model Redundancy and Security./
作者:	Mena, Pedro.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2022,
面頁冊數:	247 p.
附註:	Source: Dissertations Abstracts International, Volume: 83-12, Section: B.
Contained By:	Dissertations Abstracts International83-12B.
標題:	Computer science. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29162344
ISBN:	9798802715222

Auto Machine Learning Applications for Nuclear Reactors: Transient Identification, Model Redundancy and Security.
Mena, Pedro.

Auto Machine Learning Applications for Nuclear Reactors: Transient Identification, Model Redundancy and Security. - Ann Arbor : ProQuest Dissertations & Theses, 2022 - 247 p.

Source: Dissertations Abstracts International, Volume: 83-12, Section: B.

Thesis (Ph.D.)--Idaho State University, 2022.

This item must not be sold to any third party vendors.

Machine learning and AI are concepts that have had a large impact in daily life since 2000.It is unlikely that most people at this point in time do not have some sort of interactionwith an AI system on a daily basis. This research effort looked to contribute to the field ofnuclear safety and explore ways to expand the use of machine learning through the appli-cation of AutoML. This project consisted of four major phases. In the first phase, data wascollected from a GPWR simulator for five different reactor events, creating a dataset withover 30,000 points. Six different machine learning models were trained using the AutoMLpackage TPOT. The results from this test were positive with all models producing accu-racies in the high 90% range. The models were also able to perfectly distinguish a reactoroperating normally from one experiencing a transient. In the next phase, the datasetwas expanded using the GPWR, the number of classes was increased to 12 and the newdataset consisted of over 110,000 points. Models were retrained using TPOT and whilemany of models suffered in validation, three of the models were still able to score vali-dation results in the low 90% range. The models were then examined looking at modelredundancy by dropping key features, examine variation due to changes in random state,exploring ways to improve the model and identify the reasons behind misclassifications.The third phase of the project explored the use of autoencoders to identify GPWR datathat had been altered at random. The model was able to identify all points at high levelsof noise, but performance dropped off as the noise was decreased. Still, the techniquehas validity to help with security concerns and identify sensor malfunctions. The finalphase of the project was to explore different AutoML approaches and compare and con-trast their performance, ease of use and functionality. These were TPOT, H2O and GoogleCloud AutoML. Each of these approaches were found to have different advantages andissues, but all performed with models produced using GPWR data, with results in the midto high 90% range.

ISBN: 9798802715222Subjects--Topical Terms:

523869
Computer science.
Subjects--Index Terms:

Anomaly detection

Auto Machine Learning Applications for Nuclear Reactors: Transient Identification, Model Redundancy and Security.
LDR:03303nmm a2200373 4500 001 2352234
005 20221118093850.5
008 241004s2022 ||||||||||||||||| ||eng d
020 $a 9798802715222
035 $a (MiAaPQ)AAI29162344
035 $a AAI29162344
040 $a MiAaPQ $c MiAaPQ
100 1 $a Mena, Pedro. $3 3691855
245 1 0 $a Auto Machine Learning Applications for Nuclear Reactors: Transient Identification, Model Redundancy and Security.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2022
300 $a 247 p.
500 $a Source: Dissertations Abstracts International, Volume: 83-12, Section: B.
500 $a Advisor: Kerby, Leslie.
502 $a Thesis (Ph.D.)--Idaho State University, 2022.
506 $a This item must not be sold to any third party vendors.
520 $a Machine learning and AI are concepts that have had a large impact in daily life since 2000.It is unlikely that most people at this point in time do not have some sort of interactionwith an AI system on a daily basis. This research effort looked to contribute to the field ofnuclear safety and explore ways to expand the use of machine learning through the appli-cation of AutoML. This project consisted of four major phases. In the first phase, data wascollected from a GPWR simulator for five different reactor events, creating a dataset withover 30,000 points. Six different machine learning models were trained using the AutoMLpackage TPOT. The results from this test were positive with all models producing accu-racies in the high 90% range. The models were also able to perfectly distinguish a reactoroperating normally from one experiencing a transient. In the next phase, the datasetwas expanded using the GPWR, the number of classes was increased to 12 and the newdataset consisted of over 110,000 points. Models were retrained using TPOT and whilemany of models suffered in validation, three of the models were still able to score vali-dation results in the low 90% range. The models were then examined looking at modelredundancy by dropping key features, examine variation due to changes in random state,exploring ways to improve the model and identify the reasons behind misclassifications.The third phase of the project explored the use of autoencoders to identify GPWR datathat had been altered at random. The model was able to identify all points at high levelsof noise, but performance dropped off as the noise was decreased. Still, the techniquehas validity to help with security concerns and identify sensor malfunctions. The finalphase of the project was to explore different AutoML approaches and compare and con-trast their performance, ease of use and functionality. These were TPOT, H2O and GoogleCloud AutoML. Each of these approaches were found to have different advantages andissues, but all performed with models produced using GPWR data, with results in the midto high 90% range.
590 $a School code: 0320.
650 4 $a Computer science. $3 523869
650 4 $a Nuclear engineering. $3 595435
653 $a Anomaly detection
653 $a AutoML
653 $a Data Science
653 $a Machine learning
653 $a Nuclear safety
653 $a Nuclear simulation
690 $a 0984
690 $a 0552
710 2 $a Idaho State University. $b College of Science and Engineering - Computer Science. $3 3688460
773 0 $t Dissertations Abstracts International $g 83-12B.
790 $a 0320
791 $a Ph.D.
792 $a 2022
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29162344