The stability and reliability of the bolted state of the transmission pole tower is an important foundation to ensure the safety of transmission line operation, and the traditional bolt loosening detection method can no longer meet the needs of fast, efficient and intelligent operation inspection of transmission pole tower. Intelligent sensing, machine vision, laser measurement and other means combined with intelligent algorithm methods to achieve structural state detection, has been widely used in aerospace, railways, bridges and other industrial fields. This paper first briefly introduces the principle and shortcomings of conventional detection methods. Then, the working principle of the prior art is expounded from the two directions of contact and non-contact, and the characteristics of the two types of detection methods are compared and analyzed. In order to realize the intelligent diagnosis and accurate identification in field application, the existing identification algorithms are compared, and the above detection methods are synthesized to conclude that laser vibration measurement technology combined with convolutional neural network algorithm is the key direction of transmission pole tower bolting state detection and diagnosis and recognition research, and finally summarized and prospected.
|