Dynamic Analysis of Fluid Containing Cylindrical Tanks Using ANN

No Thumbnail Available

Date

2024-04

Journal Title

Journal ISSN

Volume Title

Publisher

Addis Ababa University

Abstract

Cylindrical containers that store liquid are among those delicate structures that are impacted through dynamic loads. These structures that contain fluid are affected due to impulsive and convective pressures created by the liquid inside. This study utilized an artificial neural networking (ANN) model to analyze the dynamic response of fluid-containing cylindrical tanks. Six input parameters were selected to characterize the geometric and mechanical attributes of the cylindrical tank, including the properties of the fluid it holds. A combined dataset of 6,912 samples from Housner's approximate method was utilized to train and test the ANN model. The training and testing sets yielded R2 values of 0.9997 and 0.9991, respectively. The ANN model obtained results that were comparable to the results of Housner's approximate method show that ANN simulations can accurately predict the dynamic response of fluid-containing cylindrical tanks. The model can be enhanced to investigate additional parameters that impact the dynamic response of fluid-containing cylindrical tanks, such as the second-order impact caused by axial load, the effect of baffles, etc. In addition, factors that affect the accuracy and precision of the prediction of the ANN model were investigated and directions were put forward to get a more accurate and precise prediction from available data.

Description

Keywords

Dynamic load, Training, Testing, Neural Network, Cylindrical Tank

Citation