Experiments were performed to gather data on the effects of changing bulk temperature and power levels on the sound of boiling. This information was formatted and used to train a neural network to recognize boiling. The trained network was then used in an executable program to provide an indication of boiling. The trained network is an detector predictor of boiling under the experimental conditions. More work must be done before the detector can be reliably used in the reactor. With more development, the detector could possibly be used to provide important coolant information for Pressurized or Boiling Water Reactors. Thesis Supervisor: Dr. John E. Meyer Title: Professor of Nuclear Engineering Thesis Reader: Dr. David D. Lanning Title: Professor of Nuclear Engineering
[1]
J. Westwater,et al.
Sound of Boiling.
,
1955,
Science.
[2]
J. A. Koski,et al.
Experimental verification of subcooled flow boiling for tokamak pump limiter designs
,
1987
.
[3]
S. Usui.
Neural Computing
,
1989,
IFIP Congress.
[4]
Ye. I. Nesis.
Acoustic noise of a boiling liquid
,
1990
.
[5]
Om Pal Singh,et al.
New statistical features sensitive to sodium boiling noise
,
1990
.
[6]
D. Lübbesmeyer,et al.
Experimental reactor noise—A review on noise-analytic measurements of thermohydraulic parameters in operating BWRs and their interpretations
,
1984
.