Failure prediction in defective pipelines using finite element simulation of fluid-structure interaction and neural network method
Author(s):
Article Type:
Research/Original Article (دارای رتبه معتبر)
Abstract:
The occurrence of pipeline failures can lead to significant damage to the environment and natural resources, as well as high repair costs. In this study, the finite element simulation is employed to model the fluid-structure interaction between the fluid flow and the damaged pipe wall to investigate stress distribution and failure in damaged pipes. Given the time-consuming nature of this simulation, an artificial neural network is also used to predict the behavior of the damaged pipe. This neural network is trained using a recurrent backpropagation algorithm. To this end, the maximum stress in the damaged pipe is considered as the objective function and is calculated by the finite element method for different values of the flow velocity, size, distance, and depth of the defects. The design parameters are selected by Taguchi method to optimize the neural network structure and increase its accuracy. The results have suggested that combining the finite element and artificial neural network methods is an effective approach for failure prediction in defective pipelines.
Keywords:
Language:
Persian
Published:
Journal of Solid and Fluid Mechanics, Volume:13 Issue: 5, 2023
Pages:
105 to 120
https://www.magiran.com/p2678519
سامانه نویسندگان