2016 conference paper

Modelling and experimental comparison of fluid structure coupling for thin sheet metal tanks using statistical energy analysis

Proceedings of the ASME International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, 2016, Vol 8.

By: K. Shende n & R. Keltie n

co-author countries: United States of America 🇺🇸
Source: NC State University Libraries
Added: August 6, 2018

Acoustic response of flat surfaces in contact with a fluid volume is of some interest for the design of automotive fuel tanks, fluid containers and underwater applications [1]. As this response can be related to the surface vibration response in the linear domain, the effect of fluid structure coupling on the vibration response of the structure is studied in this paper. Advances in the computational abilities have increased the focus of analysis-led approaches in the design of thin sheet metal tanks. Conventional finite element (FE) based approaches are useful at low frequencies but are highly sensitive to geometrical details and local effects at higher frequencies. With changing input parameters, finite element approaches could prove to be computationally expensive during the initial design phase of such structures. Statistical Energy Analysis (SEA) is an energy based approach and was used to study the fluid structure coupling effect on the vibration characteristics of a simple rectangular parallelepiped thin sheet metal tank. A thin steel tank (thickness/min. characteristic dimension <0.01) was excited by a broad band uniform power spectral density white noise signal and the spatial and frequency averaged acceleration responses were compared. Some parameters like the damping loss factor and the excitation force were calculated from the experimental measurements and used as input for SEA simulations. Coupling loss factors were calculated from tests and the trend lines were found to be in agreement with the theoretical calculations. The SEA simulation model results were compared with the conventional FE based approach for reference. Variance studies were used to compute the envelope for the SEA simulation response for a 90% confidence interval. The SEA and the test results comparison was quantified by a correlation coefficient which indicated a moderately strong correlation (>0.5) between the SEA and experimental results.