| ABSTRACT: |
QUANTITATIVE RUST-UNDER-PAINT DETECTION UTILIZING NEAR-FIELD MICROWAVE
NDE TECHNIQUES
M. AbouKhousa1 , N. Qaddoumi2, and T. Ibrahim2
1 Concordia University, Montreal, Quebec, Canada; 2 American University Of Sharjah, Sharjah, United Arab
Emirates
Near-field microwave NDE systems utilizing open-ended rectangular waveguides constitute a competent
candidate to detect and evaluate planner rust layers under paint coatings. Basically, the waveguide
illuminates the specimen with microwave signals and monitors the reflected waves. Minute variations in the
structure reflect in measurable variation in the reflection coefficient at the waveguide aperture. The
functional dependence of reflection coefficient on the rust layer physical propertiesÑi.e. thickness and
depthÑis exploited in the detection schema. Upon measuring the reflection coefficient, the inverse problem
of rust thickness and depth determination should be solved. This problem is ill-posed in nature and requires
sophisticated algorithm to be inverted quantitatively. In this paper, we will introduce a Maximum-
Likelihood algorithm to be applied in conjunction with multi-frequency measurements to solve the inverse
problem. As it will be shown, the multi-frequency measurements will provide diversity gain over the
uncertainties embedded in the system. The practical potential of the proposed algorithm will be
demonstrated in real life rust under-paint detection problem. Finally, the performance of the algorithm under
measurement noise will be simulated and analyzed. It will be shown that the proposed algorithm provides
significant accuracy with high sensitivity in determining the rust layer’s thickness and depth.
|