A fault detection algorithm for pipeline insulation layer based on immune neural network. (April 2022)

• Record Type: Journal Article
• Authors: Yu, Xiao; Tian, Xuesong
• Is Part Of: International journal of pressure vessels and piping. Volume 196(2022)
• Publication Date: 2022-04
• Subjects: Immune neural network -- Pipeline insulation layer -- Infrared image -- Fault detection; Pressure vessels -- Periodicals; Pipe -- Periodicals; Récipients sous pression -- Périodiques; Tuyaux -- Périodiques; Pipe; Pressure vessels; Periodicals; 681.76041
• DOI: 10.1016/j.ijpvp.2022.104611
• Languages: English
• ISSNs: 0308-0161

Abstract

Abstract: Images of the pipeline can be easily captured by the infrared thermal camera. However, factors such as uneven lighting or the complex and diverse outdoor environment can affect the pipeline infrared image. Under these circumstances, detection of the abnormally high-temperature areas of pipe insulation becomes very difficult because the contrast between the image target and background is reduced. This work aims to overcome these issues and presents an immune neural network algorithm. The existing artificial immune algorithm only reflects the function and characteristics of adaptive immunity, without considering the holistic nature of the biological immune system. This method draws on the synergy mechanism of the innate immune network and the adaptive immune network in the biological immune system and combines the holistic immune system with the neural network. The method can effectively detect abnormally high-temperature areas in the pipeline infrared images caused by insulation faults and has higher detection accuracy compared with other image detection algorithms. Highlights: An immune neural network algorithm is proposed to detect high temperature areas of pipeline. The mechanism relies on the cooperative mechanism of innate immune system and adaptive immune system. The algorithm combines the overall immune system with the neural network.