Detailed analyses of dynamic and static errors in neutron radiography testing

摘要

Neutron radiography systems are being used for real-time visualization of the dynamic behavior as well as time-averaged measurements of spatial vapor fraction distributions for two phase fluids. The extraction of quantitative data on vapor-liquid flow fields is a significant advance in the methodology of fundamental two-phase flow experimentation. The data in the form of video images are typically recorded on videotape at 30 frames per second. Image analysis of the video pictures is used to extract time-dependent or time- averaged data. The determination of the average vapor fraction requires averaging of the logarithm of time-dependent intensity measurements of the neutron beam (gray scale distribution of the image) that passes through the fluid. This could be significantly different than averaging the intensity of the transmitted beam and then taking the logarithm of that term. This is termed the dynamic error (error in the time-averaged vapor fractions due t the inherent time-dependence of the measured data) and is separate from the static error (statistical sampling uncertainty). The results provide insight into the characteristics of these errors and help to quantify achievable bounds on the limits of these errors. The static error was determined by the uncertainties of measured beam intensities. It was found that the maximum static error increases as liquid thickness increases and can be reduced by increasing the neutron source strength. The dynamic error increased with large fluctuations in the local vapor fractions and with increasing liquid thickness. Detailed analyses of both sources of errors are discussed.