Experimental Investigation to Local Settlement of Steel Cylindrical Tanks with Constant and Variable Thickness

2020 
Abstract Nowadays the use of thin-walled structures is widely accepted. The main reason for this is low weight and high resistance of these structures. In order to reduce the weight and cost of the structure, cylindrical steel tanks can be made in variable thickness. In general, foundation's settlement occur beneath the reservoir walls due to specific soil properties, which are divided into three general components: 1- Uniform Settlement 2- Tilt Settlement 3- Local Settlement. Among these components, Local settlement has the greatest impact on the reservoir shell. This component can cause large radial displacements, shell buckling and even tank failure. In this research, three steel cylindrical tanks, two specimens of constant thickness and one specimen of variable thickness, are first modeled in Finite Element and then made in the laboratory and subjected to local settlement at the edge of their floor. And the values of settlement, buckling load, final load, stress and radial deformation is compared. The data obtained from the eigenvalue analysis show that the effect of increasing the thickness on the critical buckling load is extremely high and the higher thickness at the lower part of the shell is highly influential. The results of Nonlinear analysis indicate that the behavior of tank with variable thickness is acceptable compared to other samples. Also in all samples, the radial deformation is initially negligible and then increases sharply due to the structural buckling, which is the same for all samples. According to experimental result the out of plane deformation recorded by the transducer mounted at a 45-degree angle indicates that in the samples of constant thickness the values are close together, whereas in the sample of variable thickness the amount of this deformation decreases, which is a positive result. Comparison of the results between numerical and experimental modeling also indicate good agreement between these two methods.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    12
    References
    1
    Citations
    NaN
    KQI
    []