Proceedings Article | 15 February 2022
KEYWORDS: Neural networks, Safety, Inspection, Feature extraction, Image processing, Environmental sensing, Artificial intelligence
In the long-term service process of underground power utility tunnel or underground power pipeline, due to the complex surrounding ground environment, it is affected by the comprehensive effects of different geological environment, ground buildings, rivers, expressways, railways, structural material properties of utility tunnel, as well as the long-term effect, fatigue effect and mutation effect of ground load. The accumulation of structural damage will inevitably occur in the pipe gallery or pipe wall, which would cause the settlement or cracks of the pipe gallery to a certain extent, and even lead to the structural failure and collapse of the pipe gallery in extreme cases, resulting in major safety accidents. Therefore, it is of great significance to monitor the settlement and wall cracks of underground power utility tunnel for the safe operation of underground power utility tunnel and to reduce the impact of settlement accidents. When the settlement of underground power utility tunnel occurs, cracks will appear in the wall structure of the tunnel. With the rapid development of industry 4.0 and intelligent manufacturing industry, artificial intelligence technology is more and more used in industrial safety production, intelligent inspection and other important fields. Among them, intelligent inspection robot for underground power utility tunnel safety monitoring is an important application of artificial intelligence technology in the power industry. Through the real-time video capture of pipe gallery wall cracks under complex and dangerous environment by intelligent inspection robot, the crack faults in the pipe gallery structure image can be identified in time by using relevant image processing and analysis algorithms, and the settlement warning of underground power utility tunnel can be realized, which is of great significance for maintaining the structural safety of the whole pipe gallery and the operation safety of the power system. In this paper, we employ the neural network technology, and proposed an improved back propagation algorithm, which is suitable for the detection of pipe gallery wall cracks in the complex environment. The experimental results in this paper have demonstrated that our proposed method has high recognition rate and detection efficiency for the settlement crack monitoring of underground power utility tunnel structure, which proves the effectiveness of the proposed method.
