Underground power cables are set along the wall of a power tunnel. Therefore, a lot of time and labor are necessary for the structural check of the wall which is back of cables, due to the narrowness between cables and the wall. To improve the efficiency of structural checks, it is one of the best methods that automatic structural check based on images with artificial intelligence (AI) is used instead of visual check. In this paper, the crack detection of a power transmission tunnel by AI was tested. Additionally, the accuracy and efficiency of the result was discussed to use for the actual structural checks of a tunnel.
Generally, some new structural checks efficiency methods are developed in nowadays. According to Taniguchi and Nakamura et al, digitalization technique, such as laser measurement and photography from vehicles, are devised for the structural checks of road tunnels and subways whose inner diameter is larger i.e. over 6m. The inner diameter of power transmission tunnels, however, are much smaller than these tunnels i.e. 4m or less. Moreover, a lot of power cables are usually set along the wall. Therefore, the other methods should be needed for the structural check. According to this background, it was devised to utilize visual structural checks by photography, instead of the methods used on roads and subways. Also, we found that our developing method make the efficiency of crack structural checks extremely better.
The big problem of conventional structural checks work is to take much time. This method needs a lot of heavy duty for not only closing the cracks, but also marking with a chalk and sketching by hand. Moreover, it takes a lot of effort to make CAD data from hand drawn sketches, to make a structural check report.