Experimental analysis of the influence of the number of rows of a tube bank on the FIV critical velocity

•This paper presents the influence of the number of rows on the FIV in tube banks.•Twenty-seven configurations tested: single cylinder, single row, two to five rows.•Results show that the critical velocity changes with the number of rows.•Fluid-elastic instability depends on number of rows and position of the cylinder. The present study investigated experimentally the influence of the number of rows and the position of a free-to-vibrate monitored cylinder in a tube bank on the critical reduced velocity and compared it with literature models. The experiments were conducted in an aerodynamic channel with a test section assembled with one cylinder free to vibrate in the transversal direction and additional fixed rigid cylinders. The tests were conducted for one, two, three, four, and five rows with the cylinder free to vibrate in different positions along the central line. Models presented in the literature for one row and five rows were used to compare with the instability response for the studied configurations. The experimental results showed that the position of the cylinders is relevant due to the number of rows that will be positioned downstream it and if it is in the first row. The models presented in the literature consider the damping and reduced velocity in the analysis without additional parameters, just the factor associated with the arrangement and space ratio. It was observed that the critical velocity changes with the number of rows: a reduction in the velocity of fluid-elastic instability results with the increase in the number of rows when the cylinder was free to vibrate in the second or third row. The positions in the second and third rows presented the lower critical reduced velocities. An increase in the rate of occurrence of fluid-elastic instability was observed with an increase in the number of rows for the same mass and damping parameters when the cylinder was free to vibrate in the first row. The comparison of the results with models to predict the critical reduced velocity showed that, in general, they did not well predict the instability occurrence due to the lack of parameters and the conservative approach adopted in the model development.