Purdue University Graduate School
Thesis_Harsh.pdf (2.43 MB)


Download (2.43 MB)
posted on 2020-05-02, 11:47 authored by Harsh BohraHarsh Bohra
The author aims to investigate the current design provision for seismic and foundation settlement design of aboveground open-top storage tanks using finite element analysis. The thesis is divided into two independent but closely related studies: (1) seismic analysis of open-top storage tanks with flexible foundation and (2) fitness-for-service of open-top storage tanks subjected to differential settlement.

The present seismic design provisions in American Petroleum Institute’s storage tank standard API 650 (2013) assumes the tank foundation is rigid and therefore, ignores the effect of uplift during a seismic excitation. In the first study, the objective was to quantitatively critique rigid foundation assumption and conclude if the assumption is acceptable or not for a given tank geometry. Tanks with three different height to diameter ratio (H/D), i.e aspect ratios, of 0.67, 1.0 and 3.0 representing broad, nominal and slender geometry, respectively, were modelled having both rigid and flexible foundations. The flexible foundation was modelled with series of non-linear compression only springs. Additionally, for each tank model two different hydrodynamic pressure distribution suggested by (1) Housner and (2) Jacobsen-Veletsos were applied which are used by API 650 and Eurocode 8, respectively. Geometric non-linear analysis with non-linear material properties was conducted (GMNA) using Riks algorithm in Abaqus finite element analysis (FEA) program. The hoop stresses, longitudinal stresses, uplift and buckling capacity of each rigid foundation tank model were compared with its respective flexible foundation tank model and corresponding API 650 rule based provisions. It was observed that the assumption of rigid foundation from design point of view is acceptable for the broad tank, however, for the nominal and slender tanks this assumption is not acceptable. The buckling capacity of nominal and slender tanks having flexible foundation are significantly lower compared to rigid foundation. Therefore, the effect of uplift should not be neglected for design purposes for nominal and slender tank geometries.

In the second study, an alternative method for evaluating the structural integrity of storage tank subjected to differential settlement is proposed. The limitations of the existing method in API 653 (2014), currently used in the industry are highlighted. The tank settlement is measured underneath
the tank bottom along the tank circumference at discrete locations. The settlement can be transformed into a Fourier series by combining different harmonic components. In the existing API 653 method there is no distinction between the effects of different harmonic components whereas in the proposed method the effects of first five harmonic components are individually accounted and the cumulative damage is evaluated. The proposed method is formulated based on FEA conducted on twenty-one different tank models with each having different tank geometry. The limiting settlement value for each harmonic wave number is found for a given tank geometry by conducting GMNA using Riks algorithm, and a generalized trend is found for each harmonic wave number. The proposed method is further validated by performing numerous FEA simulations. The simulations were conducted for several tank models subjected to four representative actual measured settlement data. A set of tank models used in the validation was generated using random tank geometries and design parameters to have a blind test of the proposed method. Finally, a comparison is made between allowable settlement based on the API 653 method, the proposed method and the FEA. It was observed that the proposed method consistently results in conservative results compared to FEA. In contrast the API 653 method does not always result in conservative results. For some measured settlement data, the API 653 method gives overly conservative values and for others it gives non-conservative values. Moreover, the API 653 method is based on the beam theory which may not capture the true shell behavior. Therefore, the API 653 method requires modifications. The proposed method on the other hand is consistent and is based FEA which can capture the true shell behavior as it is formulated using shell theory. Therefore, it is recommended that the existing method in API 653 shall be replaced with the proposed method to determine the fitness of tank under differential settlement.


Goldberg fellowship

Baker consulting group, Inc

Lyles School of Civil Engineering


Degree Type

  • Master of Science in Civil Engineering


  • Civil Engineering

Campus location

  • West Lafayette

Advisor/Supervisor/Committee Chair

Sukru Guzey

Additional Committee Member 2

Mark D. Bowman

Additional Committee Member 3

Ayhan Irfanoglu

Usage metrics



    Ref. manager