Abstract Turbulent wall pressure statistics are computed from a direct numerical simulation (DNS) of channel flow for Rθ = 278 using methods common to physical experiments. The wall pressure database consists of 2,564 temporal data points at 128 streamwise and 128 spanwise locations. The time-averaged mean wall pressure field is shown to consist of streaklike regions of high and low pressure aligned in the streamwise direction and occurring at spanwise intervals of 1.3δ. The wall pressure field is also shown to consist of several spatial locations at which the rms is approximately 20 percent greater than the average rms value. This apparent spatial nonhomogeneity results from the limited temporal record length of 1730 viscous time units available for processing, which is on the order of the period of occurrence of high magnitude pressure events. The correlation functions are shown to agree well with experimental results and do not display unique characteristics directly related to the in-flow/out-flow boundary conditions used in the numerical solution. The streamwise spatial grid spacing determines the highest convective wavenumber resolved in the numerical simulation and thus limits the frequency bandwidth over which the autospectrum may be computed. The Corcos A and B functions are shown to have characteristics similar to recent experimental measurements. The methodology presented is appropriate for analysis of numerically generated databases of wall pressure or wall shear stress fluctuations.
Abraham et al. (Sun,) studied this question.