Sensitivity analysis and failure prediction of X80 pipeline under transverse landslide

Abstract

To investigate the mechanical response of X80 pipelines under lateral landslide conditions, finite element simulations of X80 pipeline under 375 lateral landslide conditions are conducted in this paper to examine mechanical behavior under varying pipe diameters and wall thicknesses, and landslide width and displacement on pipeline strain is assessed. The results indicate that under the influence of lateral landslides, The pipeline strain is predominantly induced by the bending moment, with axial strain being the most significant, constituting over 95 % of the total strain. The peak strain is primarily concentrated in the middle section of the pipeline's leading span. Based on extensive numerical simulation data, a grey relational analysis was conducted, revealing that the primary factors influencing the maximum axial strain in pipelines, in descending order of significance, are landslide displacement, landslide width, pipeline diameter, and pipeline wall thickness. Furthermore, to predict the safety of X80 pipelines under lateral landslides, a BP neural network prediction model and a fitting formula are developed based on the four influencing factors. Both the model and the formula were validated to accurately predict the maximum axial strain of X80 pipelines affected by lateral landslides. Moreover, a failure assessment method for X80 pipelines under lateral landslide conditions was established using the strain failure criterion. Results indicate that the prediction errors of the neural network model and the formula, compared to simulation outcomes, are within 10 %, the high accuracy of the failure prediction results is similarly demonstrated.