Machine learning-based electrochemical monitoring for molten chloride salt reactors염화용융염원자로를 위한 기계 학습 기반 전기화학 모니터링

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 1
  • Download : 0
Radiation-resistant high-temperature molten salts are used in various nuclear industry fields, especially considered as liquid nuclear fuel and coolant for generation-IV reactors. Due to the corrosiveness of the molten salt, problems of structural material corrosion occur. Measurement of the concentration of corrosion products and corrosion potential are important technologies in various applications utilizing molten salts. In this paper, I developed a smart sensor that overcomes the electrode area problem and non-ideal electrochemical reaction, which are problems of existing electrochemical sensors in high-temperature molten salts, through electrochemical big data and multi-array electrodes. In this study, I designed an experimental apparatus by combining a multi-array electrode and an automatic vertical translator to build molten salt electrochemistry big data. The acquired voltammogram big data was extracted through an automatic feature extractor designed for machine learning. An automatic feature extractor extracted features that correlate with the surface area from the voltammogram big data. The extracted features were used as a train set, test set, and validation set of various machine learning models, and the optimized conditions were selected through minimum mean absolute percentage error (MAPE). Through the surface area prediction model, the MAPE of the predicted electrode area was 4.45% from single-component molten salt data and 5.45% from multi-component molten salt data that included non-ideal electrochemical reactions in the dataset. The electrochemical sensor using the predicted electrode area had a prediction accuracy of less than 19.9% for corrosion products for single-component and multi-component molten salts. The smart sensor combined with machine learning can measure the concentration of corrosion products and corrosion potential. The sensor can be used for operating molten salt reactors through structural material corrosion management and molten salt component measurement.
Advisors
성충기researcher
Description
한국과학기술원 :원자력및양자공학과,
Publisher
한국과학기술원
Issue Date
2024
Identifier
325007
Language
eng
Description

학위논문(박사) - 한국과학기술원 : 원자력및양자공학과, 2024.2,[v, 82 p. :]

Keywords

용융염▼a전기화학 센서▼a스마트 센서▼a부식 생성물 측정▼a기계학습; Molten salt▼aElectrochemical senso▼aSmart sensor▼aCorrosion product monitoring▼aMachine learning

URI
http://hdl.handle.net/10203/322105
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1099349&flag=dissertation
Appears in Collection
NE-Theses_Ph.D.(박사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0