With the rapid growth of demand for clean energy, the steel grade, diameter, wall thickness and transportation    pressure of steel pipes for pipeline construction have gradually increased in recent years. The technology of automatic    pipeline welding has gradually become the mainstream technology in pipeline construction due to its high matching strength    and toughness of girth welds, high welding efficiency and low labor intensity. This paper describes the development process    and engineering app   lications in China of welding robot technology for long-distance oil and gas pipelines. This is based on    all-position automatic welding of pipelines, as well as the engineering application characteristics and technical status of pipeline    welding robots. Several advanced pipeline welding robots are listed, including a multi-torch pipeline welding robot, laser-arc hybrid    pipeline welding robot, double-wire pipeline welding robot and friction stir welding robot. By combining and summarizing the    problems existing in the development of pipeline welding robots, it is pointed out that the welding trajectory planning, welding seam    tracking, online monitoring of welding quality, the adaptability of special working conditions and supporting welding power supply    are the technical bottlenecks in the applications in oil and gas engineering at present. The development of intelligent research and    development of pipeline welding robots in the future is examined, and it is believed that three directions will be the development    trend of intelligent pipeline welding robots in the future: (1) For the autonomous planning of welding path and adaptive seam track   ing technology, the groove and weld information of multi-layer and multi-pass welding of pipeline will be extracted, identified and    analyzed to obtain the ideal path point and welding torch posture, and the position of the welding torch can be controlled in real-time    according to the working conditions. It will realize flexible welding planning and adaptive seam tracking of pipeline girth welds    .(2) For remote fault diagnosis of welding, a complete welding expert system will be established relying on the data accumulation    of a large number of pipeline engineering applications, and remote monitoring and real-time fault diagnosis of welding robots will    be implemented. Networked real-time monitoring and fault processing systems are established for online monitoring of welding    quality; (3) For man-machine integration, the experience of skilled welders will be transformed into training samples for welding    robots through deep learning to guide welding training of non-skilled welders. Thus, the man-machine relationship will be changed    from master-slave to partnership that will complete pipeline welding work together.