PSI - Issue 28

G.N. Gusev et al. / Procedia Structural Integrity 28 (2020) 2328–2334 Gusev G.N., Shardakov I.N/ Structural Integrity Procedia 00 (2019) 000–000

2329

2

differences. Nevertheless, the developed circuit of the strain gauge sensor allowed to implement all necessary experimental studies and compare the measurement results with the finite-element model. 2. Problem Formulation We consider a vertical steel oil tank of loading capacity of 5000 m3 that is a vertical cylinder, or shell, with an internal diameter of 20.92 m and a height of 14.9 m. The tank wall is made by welding together steel strip panels. The tank bottom is made of two steel sheets with ring edges. The roof of the tank is a framed roof of conical shape and is made of sixteen panels. The roof deck consists of four sheets. The foundation of the tank is a monolithic reinforced concrete ring of 2.08 m width and 0.3 m height (an inner diameter of 18.84 m), reinforced with space frames, the main elements of which are two-dimensional flat reinforcing meshes (Fig.1). The tank was installed with significant deviations from the design project so that the tank shell lost its original cylindrical shape. However, these repairing works did not yield the required results, and even produced new defects and deviations from the design documents. These deviations from the design parameters registered after the end of all repairing and assembling works did not allow us to put the tank into operation. Therefore, it was decided to investigate the causes of defect formation and their influence on the stress-strain state of the structure in order to find out whether the restored tank could withstand further operating loads (Tarasenko et al. , 2017). Analysis of the stress-strain state of the structure was performed in several stages. At the first stage, reliable information on the current state of the tanks was obtained. The second stage involved mathematical modeling of the tank-foundation-ground base system and analysis of the calculated results. The calculations were done by the finite element method using the software package ANSYS Structural 12.1. Numerical finite-element models were developed on the basis of the results of the above survey. For a solution of the problem, we defined three types of the models:  Elastic material models, which take into account the correct design installation of the tank and the absence of any deviations from its original form. This is a basic formulation of the problem under consideration and serves as a basis for subsequent verification of the analyzed models.  Elastic material models, which take into account the real installation of the tank assessed through visual observation and its deviations from the original shape registered at that time. No possible residual stresses that may occur during the repairing works were investigated; consideration was given to the geometry of defects only.  Elastic-plastic material models, which take into account the real conditions of the tank actual assessed using the results of up-to-the minute analysis and its deviations from the original shape registered at that time. In this formulation, the law of isotropic hardening for the steel of the reservoir wall is adopted and the prehistory of the formation of the buckling is ignored. No possible residual stresses that may occur during the repairing works were studied; consideration was given to the geometry of defects only. The material of the structure was considered in an elastic plastic formulation under the isotropic hardening law. The behavior of the soil was described in terms of the elastic-plastic model according to the law of associative flow with the Drucker-Prager strength criterion based on the results of engineering and geological surveys performed on the area where the tank was placed (Salnikov, 2016). The static and dynamic load combinations were analyzed: a) long-term oil storage; b) standard operating mode of this object. The tank under study is a commercial or transfer tank for petroleum products. During the operation, it experienced cyclic loads caused by periodic filling and emptying. We constructed the mathematical model of the tank with an emphasis on the defects detected at the first stage of our investigation. The areas with such defects were modeled using a fine grid and increasing the level of discretization of the examined part of the model.