Autonomous mobile inspection system for detecting hidden voids in LNG carrier triplex bonding layers

Abstract

This paper describes the development of an autonomous mobile inspection system (AMIS) for detecting hidden voids in Liquefied Natural Gas (LNG) carrier triplex bonding layers. The triplex layer is the secondary barrier to prevent gas leaking from a Mark III type membrane LNG carrier tank. The triplex layer is composed of three composite layers: (1) a flexible secondary barrier (FSB), (2) an urethane-based bonding layer and (3) a resin-based rigid secondary barrier (RSB). During the installation of the triplex layers, air voids can be formed in the bonding layer, and gas leaking through these hidden voids can lead to failure of tightness of secondary barrier and reduction of ship's longevity. To prevent such accidents, the triplex layers are manually inspected 4 times during installation by triplex layer installer, manager of LNG tank, independent quality manager, and buyer. These manual inspections are very time consuming and labor intensive, because the voids invisible from the surface are only detected by rubbing a metal stick to the triplex layer surfaces. In this study, AMIS and void detection algorithms are developed to automatically detect and quantify the hidden voids in the triplex layers right after their installation. AMIS carries a thermography camera with multiple halogen lamps, and thermal images taken from the triplex layers are processed by a developed void detection algorithm. First, an amplitude image is obtained from raw thermal images after compensating uneven heating caused by halogen lamps. Second, binary imaging and noise canceling are performed to detect hidden voids. Third, only the voids, which are larger than 8 mm in diameter, are visualized according to the current void detection guideline. Finally, the effectiveness of the AMIS and void visualization algorithm are verified through multiple laboratory tests. The uniqueness of this study includes: (1) An autonomous inspection system is developed for real-time inspection of triplex during its installation, significantly saving the inspection cost and time; (2) A suite of new image processing techniques are developed, overcoming shortcomings of existing thermography NDT techniques; (3) The sizes as well as the locations of the hidden voids are quantified with high accuracy, reliability, and inspection speed.