Nuclear fuel pellet inspection using machine vision and artificial neural networks

Author

Mark Le Roy Kelly

Abstract

"Nuclear fuel must be of high quality before being placed into service in a reactor. Nuclear fuel vendors currently use manual inspection to verify the quality of the nuclear fuel before the fuel pellets are inserted into the zirconium fuel rods and bundled into assemblies. This work examines the feasibility of using machine vision and artificial neural networks to perform the same task, with motivation being to improve accuracy, speed, costs, reduce employee radiation doses, and to provide defect statistics to the fuel manufacturer. Sample nuclear fuel pellets were photographed, scanned, and an appropriate feature extraction technique was developed and applied to the scanned images. The extracted features were then used as inputs to two neural network paradigms, ART2-A and backpropagation. The results of testing were then compared for the two networks. Both the ART2-A and backpropagation results are promising. Results indicate that a machine vision inspection system may be possible using a system of ART2-A networks or a single backpropagation neural network"--Abstract, page iii.

Advisor(s)

Keyvan, Shahla

Committee Member(s)

Moss, Randy Hays, 1953-
Kumar, A. S. (Arvind S.)

Department(s)

Nuclear Engineering and Radiation Science

Degree Name

M.S. in Nuclear Engineering

Sponsor(s)

National Academy for Nuclear Training (U.S.)
University of Missouri--Rolla. School of Mines and Metallurgy
University of Missouri--Rolla. Department of Nuclear Engineering

Publisher

University of Missouri--Rolla

Publication Date

1995

Pagination

x, 116 pages

Note about bibliography

Includes bibliographical references.

Rights

© 1995 Mark Le Roy Kelly, All rights reserved.

Document Type

Thesis - Citation

File Type

text

Language

English

Subject Headings

Nuclear fuels -- Quality controlNuclear fuels -- InspectionQuality control -- AutomationComputer vision -- Industrial applicationsNeural networks (Computer science) -- Industrial applicationsBack propagation (Artificial intelligence) -- Industrial applicationsNuclear fuels -- Defects

Thesis Number

T 7074

Print OCLC #

34220944

Recommended Citation

Kelly, Mark Le Roy, "Nuclear fuel pellet inspection using machine vision and artificial neural networks" (1995). Masters Theses. 5992.
https://scholarsmine.mst.edu/masters_theses/5992