Development of a novel approach for strain demand prediction of pipes at fault crossings on the basis of multi-layer neural network driven by strain data

摘要

The active fault is one of the commonest geological disasters in the pipeline construction. Excessive tensile or compressive strain may be induced in pipelines which may lead to the rupture or buckling when the two fault plates generate relative movement on the fault plane so that the pipeline cannot operate normally. In this study, a finite element model of pipelines subjected to fault displacements has been established for obtaining the database of strain demand which contains the prospective engineering conditions by considering the influencing factors, i.e. pipe geometrical parameters, internal pressure, fault parameters, and soil parameters. This process has been conducted by an effective hybrid methodology which can integrate the function of .inp documents generation, automatic calculation, and results extraction based on the technology of Python, Abaqus batch processing, and Matlab. The database can be utilized as training data in the multi-layer neural network for developing the strain demand prediction model which shows accurate results with high efficiency compared with general finite element methods.