Proving strength of the fuel tanks which are used in busses by performing static and dynamic analysis

Mercedes Benz Türk A.Ş. Research and Development Center undertakes the development and testing responsibility of fuel tanks of all Mercedes and Setra Buses that are manufactured in Germany, Spain and Turkey. As a part of this responsibility, research and development process of transition from conventional steel fuel tank to plastic fuel tanks is also assigned to the R&D Center of the company. Within the scope of this thesis, it is aimed to develop a computer aided simulation method about the endurance of plastic fuel tanks that is one of the unknowns of fuel tank design process. Thus, design of product would be shaped during product development process. For that purpose, a finite element mesh is created by transferring the solid model of the fuel tank whose endurance tests have already accomplished from CATIA library to the finite element software, MEDINA. The finite element model used for static analyses has approximately 192500 nodes and 195000 elements. Shell (QUAD4 and TRIA3) and beam elements (BECOS) are used for modeling of the tubular structural elements, fuel tank and bolts, respectively. Steel material model is defined for bus structure and retaining straps. Polyethylene material model at room temperature is used for the fuel tank. Additionally, fuel is simulated by the structural solid elements (TET4) with low elasticity module for dynamic analysis. Initially, equivalent static loads are applied to the fuel tank. To do this, maximum and minimum gravitational (g) forces that are applied on the fuel tank during operation conditions are determined by using the acceleration-time data that Mercedes Benz Türk A.Ş. R&D Center has gathered from the torture track road in Germany. The deformations and stress distributions of fuel tanks are investigated under exposure of these loads. Afterwards, dynamic analyses are prepared to understand the contribution of dynamic effects that cannot be obtained by static analyses: • Natural frequencies and mod shapes are obtained via free vibration analyses. • Frequency-response analyses are completed by utilizing harmonic actuators that are applied in y and z directions. • Dynamic forced vibration analyses are accomplished by using the acceleration-time test signal that is utilized in the vibration flange of fuel tanks. All analysis are performed with the following assumptions and results are reviewed with the same assumptions, • A material property at room temperature is used for polyethylene material model. • Fuel tank has homogeneous wall thickness. • Sloshing effect of the fuel is ignored. • Fuel bonds with fuel tank walls and they are inseparable. Analyses that are progressed by using equivalent g-loads are completed for the fully filled fuel tanks. In this process, the effect of fuel on the walls is defined as a hydrostatic pressure distribution. Minimum and maximum acceleration values that occur in the x, y, z directions of the test signal are collected, and 3-axes loading packages are formed by using different fractions in each load condition. Finally pre-tension loads on the retaining straps are considered. So 24 different loading conditions are created. An exemplary loading package is formed as follows when the maximum acceleration in x direction is occurs: " %100 axMAX %40 ayMAX %40 azMAX %100 axMAX %40 ayMAX %40 azMIN %100 axMAX %40 ayMIN %40 azMAX %100 axMAX %40 ayMIN %40 azMIN " Contacts are defined between fuel tank – body structure and fuel tank – retaining straps. A visual basic macro is developed in Excel to automatically create input cards for performing analysis in PERMAS. Mentioned macro also calculates the parameters of the functions that are used to define hydrostatic pressure. Stresses and deformations on the fuel tank are investigated with analyses conducted using PERMAS software. For the locations on which the stresses and deformations are higher, the design team is informed about the requirement of design improvement. Initially, the inner part of the tank is meshed by structural solid elements to be able to get information about the dynamic behavior of the fuel tank. Contacts are removed and rigid coupling elements are used instead of contact definitions. With the help of PERMAS software, mode shapes are obtained for 5 different fullness fractions (empty, ¼, ½, ¾, full). Free vibration frequencies are found as 9.2 Hz and 25.1 Hz for full and empty fuel tanks respectively. If it is assumed that the natural frequencies of axles are around 10 Hz, it becomes obvious that the empty fuel tank is more affected by the frequencies coming from the axles with respect to the full one. On the other hand, during the free vibration tests those are done for all fullness fractions, obtained mode shapes on the front face of the fuel tank are similar to the ones obtained in the test bench. Frequency-response analyses are accomplished to investigate the variance of vibrations with high amplitudes with respect to the frequency. A sinusoidal wave in y, z directions with unit amplitude and between 0 Hz and 70 Hz bandwidth is applied, and acceleration response is read from the center of the front face. In the frequency-acceleration response graph for y-direction, two peak points around 25 Hz and 30 Hz are obtained, and for z-direction, two peak points around 20 Hz and 26 Hz are detected. The project proceeds to the last stage with the assumptions that there is a resonance effect at these peak points and, as a result of this resonance, there is an additional region along where a high stress value exists. The acceleration-time signal that has been applied to the vibration table during the tests is applied to the fuel tank. Large Mass Method that is often utilized in earthquake analyses is used to be able to apply the acceleration to a single node. Root Mean Squares (RMS) of all stresses, in each time step, are calculated for each element with PERMAS software. The results exhibit pretty similarity with the static analyses, and there is not any additional region with high stress attained. The analysis method that has been developed during this study based upon the equivalent static loading is then used in the development processes of other fuel tanks. Except for a few dynamic-related damages, the method produces pretty reliable outcomes. The ability to obtain results fast -thanks to the method's being based on static loadings- and the results' being able to be investigated at the end of the analyses make the designers prefer using it. In spite of these positive feedbacks, a small amount of damages that result from the vibration tests cannot be validated by analysis. It is supposed that one of the reasons of this situation is the modeling of the fluid inside the tank with structural solid elements. In the forthcoming studies, using alternative approaches (acoustic elements, etc.) in the modeling process of the fluid in the fuel tank and interdisciplinary methods (fluid-structure interactions, etc.) increases the accuracy of dynamic results. San-Tez, a university cooperation project with financial support from the Turkish Ministry, application was made for continuation of the project. Ministry of Labor and Industry has approved 0488.STZ.2013-2 numbered project conducted by Istanbul Technical University Faculty of Aeronautics and Astronautics and Mercedes Benz Türk A.Ş..