creator Gera, Mohit advisor Song, Yoonseok High-voltage leak detection (HVLD) is an emerging technology that has successfully been used to test the integrity of flexible food packages. It allows for nondestructive, fully automatic, and 100% on-line detection of pinholes as small as 10 μm in diameter. The applicability of this high-voltage technique to various semi-rigid packages, therefore, is of great interest to food industry. This research was conducted to evaluate and validate the effectiveness of the HVLD system and its processing variables on detection of pinhole leaks in semi-rigid packages through systematic factorial design and blind study experiments. The Student’s t-distribution and the corresponding P-value indicated that, among critical factors, package headspace had the greatest effect followed by dielectric constant, pinhole size, and food conductivity (p < 0.01). Strong interactions in between those variables were also observed (p < 0.05). In particular, two evident interactions between dielectric constant and food conductivity, and pinhole size and food conductivity contributed to the model at a significant level. Statistical Z-test analysis of results from blind studies (population size for each combination of package, food, and defect types = 90, defect rate = 0.05) showed that the HVLD technique is significantly effective in determining defective packages (p < 0.05). Results from worst case scenario study showed that there is no significant change in physical properties of plastic laminate lid films after inspection at highest voltage of 10kV. These results suggest that HVLD technique is a promising non-destructive method that can be safely applied to detect pinhole defects in a wide range of plastic based packages. Submitted by Dana Lamparello (dlampare@iit.edu) on 2012-08-23T16:30:48Z No. of bitstreams: 2 Final_thesis_gera.pdf: 1106287 bytes, checksum: 66a8a28d65b00e4ac119f01efadbb688 (MD5) Thesis Signed Title Page_gera.pdf: 286901 bytes, checksum: 4fbdf7ffe76815b7f8626393605c5e4c (MD5) Made available in DSpace on 2012-08-23T16:30:48Z (GMT). No. of bitstreams: 2 Final_thesis_gera.pdf: 1106287 bytes, checksum: 66a8a28d65b00e4ac119f01efadbb688 (MD5) Thesis Signed Title Page_gera.pdf: 286901 bytes, checksum: 4fbdf7ffe76815b7f8626393605c5e4c (MD5) Previous issue date: 2012-05 M.S. in Food Safety and Technology, May 2012 2012-04-23 2012-05 http://hdl.handle.net/10560/2840 en Thesis born digital application/pdf In Copyright http://rightsstatements.org/page/InC/1.0/ Restricted Access BIOL / Biology Illinois Institute of Technology Affiliated department