Above ground steel storage tanks are used worldwide for the storage of various liquids. EN 14015:2005, which has traditionally been used to design the tanks, does not necessarily fulfil the requirements of the Swedish Building Code. This has been underlined by hand calculation models in EN 1993-1-6:2007, EN 1993-4-2:2007 and numerical analysis using Finite Element Method (FEM). Therefore, this thesis investigates the differences between these design models and, preliminarily, the use of high-strength steel in tank shells.

A 10600 m3 cylindrical steel tank of diameter 26 m and height of 21 m located in Gothenburg, Sweden is studied. The study is limited to the assessment of the stress in the shell courses due to the hydrostatic pressure from the fluid action of a filled tank, and the buckling behaviour of the shell courses of an empty tank subjected to self-weight, snow and wind loads. Particularly, models of the tank shell with a yield strength of 355 MPa are investigated in detail, while the results of the 700 MPa model are considered as preliminary study, since the material is currently not used for tank shells.

An analysis of the fluid action on the tank shell courses in each of the three hand calculation models, showed that the EN 14015 model utilizes thicker courses than both Eurocodes. One benefit of the Eurocode models is that they do not limit the thickness of the shell courses, but it is still necessary to have thicker courses in the upper part of the tank in order to achieve sufficient resistance against buckling. EN 14015:2005, on the other hand, limits the minimum thickness to 6 mm for the investigated tank. Furthermore, only EN 1993-1-6 is applicable to the models with a yield strength of 700 MPa as per EN 1993-1-12 and this resulted in a uniform shell thickness of 6 mm. However, an increase in yield strength has no buckling benefits whatsoever. 

Buckling is the most critical aspect as observed in this study. EN 14015 has no specific buckling calculations but uses the approach of determining the number of stiffening rings which are deemed adequate against buckling. In this study, 3 secondary stiffening rings were found to be adequate. In comparison, the results of EN 1993-4-2 are very conservative and lead to a very high and uneconomical number of stiffening rings, ranging from 30 to 52 stiffening rings depending on the reliability class. EN 1993-1-6 resulted in 6-17 stiffening rings, for reliability classes 1-3 and fabrication classes A-C. Therefore, the so-called analytical models in the Eurocodes result in a much denser spacing of stiffening rings than 14015:2005. 

The buckling stresses due to the design loads were found to be lower than the yield strength of the tank shells for both hand calculation and FEM models. This means that the tank shells failed in buckling before the yield strength of the material was reached. Based on the parametric study of the EN 1993-1-6 (355 MPa) model regarding reliability class 1 and fabrication class A using FEM, the spacing of the stiffening rings can be increased up to 60 % (from 3825 mm to 6120 mm) with the variable loads also increased simultaneously up to 3.8 times before the shell buckles. Therefore, the design of future tanks using numerical analysis guarantee’s more reliability than all the aforementioned standards. 

The design for buckling according to EN 14015 is only valid for a design snow load and under-pressure ≤ 1.2 KN/m2. However, according to the standard itis possible to agree to use it for larger actions or use another design model for buckling.