Experimental verification research of pipeline deflection deformation monitoring method based on distributed optical fiber measured strain

Distributed fiber optic sensing is suitable for long-distance water pipeline deformation monitoring. However, the distributed optical fiber measures the structural strain, which is difficult to reflect the deflection deformation of the pipeline directly. Hence, it is necessary to study how to transform the measured structural strain of distributed optical fiber into deflection deformation of pipeline. In this paper, the integral method and the Long Short-Term Memory neural network (LSTM) are used to realize the transformation from the measured strain of the optical fiber to the deflection of the pipeline. Based on the Euler-Bernoulli elastic foundation model, a set of test equipment was designed to simulate the deformation monitoring of underground water pipelines, and two optical fiber strain-deflection conversion methods were verified. The results show that the LSTM intelligent transformation model can effectively and accurately transform the strain-deflection of pipelines, and is suitable for practical engineering.