Diagnosis of Corrosion Process in Nuclear Power Plant Secondary Piping Structures [electronic resource]
Corrosion inside secondary pipe structures in Nuclear Power Plants is difficult to detect without intrusion. Statistical and machine learning techniques are used to predict corrosion using pipe vibration data. 3D accelerometers were used to record pipe vibration data in three directions using six se...
Saved in:
Online Access: |
Full Text (via OSTI) |
---|---|
Format: | Government Document Electronic eBook |
Language: | English |
Published: |
Oak Ridge, Tenn. :
Distributed by the Office of Scientific and Technical Information, U.S. Department of Energy,
2019.
|
Subjects: |
MARC
LEADER | 00000nam a22000003u 4500 | ||
---|---|---|---|
001 | b11323378 | ||
003 | CoU | ||
005 | 20200625000000.0 | ||
006 | m o d f | ||
007 | cr ||||||||||| | ||
008 | 200707e20190901||| o| f0|||||eng|d | ||
035 | |a (TOE)ost1616098 | ||
035 | |a (TOE)1616098 | ||
040 | |a TOE |c TOE | ||
049 | |a GDWR | ||
086 | 0 | |a E 1.99:inl/ext-19-55918-rev000 | |
086 | 0 | |a E 1.99:inl/ext-19-55918-rev000 | |
088 | |a inl/ext-19-55918-rev000 | ||
245 | 0 | 0 | |a Diagnosis of Corrosion Process in Nuclear Power Plant Secondary Piping Structures |h [electronic resource] |
260 | |a Oak Ridge, Tenn. : |b Distributed by the Office of Scientific and Technical Information, U.S. Department of Energy, |c 2019. | ||
300 | |a Size: 37 p. : |b digital, PDF file. | ||
336 | |a text |b txt |2 rdacontent. | ||
337 | |a computer |b c |2 rdamedia. | ||
338 | |a online resource |b cr |2 rdacarrier. | ||
500 | |a Published through Scitech Connect. | ||
500 | |a 09/01/2019. | ||
500 | |a "inl/ext-19-55918-rev000." | ||
500 | |a Araseethota Manjunatha, Koushik ; Mack, Andrea L. ; Agarwal, Vivek ; Koester, David ; Adams, Douglas | ||
500 | |a Idaho National Lab. (INL), Idaho Falls, ID (United States) | ||
500 | |a USDOE Office of Nuclear Energy (NE) | ||
520 | 3 | |a Corrosion inside secondary pipe structures in Nuclear Power Plants is difficult to detect without intrusion. Statistical and machine learning techniques are used to predict corrosion using pipe vibration data. 3D accelerometers were used to record pipe vibration data in three directions using six sensors. Vibration data were processed using the Hilbert-Huang Transformation, which is known for its adaptable basis function to non-linear and non-stationary signals. Hilbert-Huang Transformation using empirical mode functions generates multiple intrinsic mode functions of different amplitude and frequency levels from raw vibration data. From intrinsic mode functions features are extracted which include phase, energy, entropy, mean, standard deviation, skewness, and kurtosis. Logistic regression using only phase information correctly distinguished mass additions from baseline data in roughly 83% and 85% of the cases. Phase information varied substantially among different baseline runs and within trials of the same run. Random Forest resulted in >95% testing accuracy for both mass addition and mass removal data. Support Vector Machine had ̃98% testing accuracy for mass removal and ̃92% testing accuracy for mass additions when using Permutation Importance feature selection. When using Principle Component Analysis with Support Vector Machine, testing accuracy was ̃68% testing accuracy for mass additions and ̃85% testing accuracy for mass. | |
536 | |b AC07-05ID14517. | ||
650 | 7 | |a 46 instrumentation related to nuclear science and technology |2 local. | |
650 | 7 | |a Accelerometer, machine learning |2 local. | |
650 | 7 | |a Hilbert huang transform, support vector machine, principal component analysis, random forest, linear regression |2 local. | |
710 | 1 | |a United States. |b Department of Energy. |b Office of Scientific and Technical Information |4 dst. | |
856 | 4 | 0 | |u http://www.osti.gov/servlets/purl/1616098 |z Full Text (via OSTI) |
907 | |a .b113233784 |b 07-07-20 |c 07-07-20 | ||
998 | |a web |b 07-07-20 |c f |d m |e p |f eng |g |h 0 |i 1 | ||
956 | |a Information bridge | ||
999 | f | f | |i 35603507-fbd6-574e-96ce-dcb7d7ba253e |s 8c28041c-ccd8-5cce-929e-45b8eb83185c |
952 | f | f | |p Can circulate |a University of Colorado Boulder |b Online |c Online |d Online |e E 1.99:inl/ext-19-55918-rev000 |h Superintendent of Documents classification |i web |n 1 |