1938/03/18 by L. F. G. Simmons, C. Salter, Charles Salter · 1 citation
Engineering · Environmental Science · #Fluid Dynamics and Turbulent Flows #Fluid dynamics and aerodynamics studies #Wind and Air Flow Studies
paper · doi:10.1098/rspa.1938.0046
openalex publication_date 1938/03/18 · openalex created_date 2025/10/10 · openalex updated_date 2026/06/26
Abstract The experimental investigation of isotropic turbulent motion is most conveniently conducted in the air stream of a wind tunnel in which the turbulence is augmented by the addition of a grid of uniform mesh placed across it. At any point downstream, beyond the wind shadow of the grid, the velocity fluctuations are small compared with the mean speed of the stream, but vary irregularly with time. Records taken of u, the instantaneous value of the turbulent component in the direction of motion, show that the time variations follow the random law of errors (Simmons and Salter 1934; Townend 1934); but no successful attempts appear to have been made to analyse such a record into its harmonic components. On the other hand, by utilizing a hot-wire anemometer to produce a current proportional to u, and employing electrical filters to measure the contributions to u2‾ which arise from frequencies within the range 0 to n where n is varied, data may be obtained from which a spectrum curve of the variation in velocity u, at a fixed point can be plotted. The ordinate of this curve at frequency n represents F(n), the function denoting the probability of the existence of velocities between n and n + dn, whose significance is discussed by Professor G. I. Taylor in the paper which follows. In some experiments undertaken at Professor Taylor’s suggestion for the purpose of determining the values of F(n) for turbulence created by a grid of square mesh, the hot-wire technique was employed to measure u2‾ from the readings of a thermal milliammeter, placed in the output circuit of a valve amplifier. In order to effect the analysis different electrical filters were successively put in the measuring circuit. Two types of filter were used: one, a low-pass, allowed the passage only of currents below a certain frequency; the other, a high-pass, only those above a certain critical frequency, the limit in each case being settled by the electrical constants of the filter. The former type served to explore the range 0-325 c./sec.; the latter was used for frequencies above this figure. From the ratios of readings of the milliammeter taken, in each case with and without the filter, it was possible to obtain, at any given wind speed, a close approximation to the spectrum curve required.