ABSTRACT CTS-9500518 HANRATTY UNIVERSITY OF ILLINOIS This proposal deals with turbulence generated by flow over a solid boundary. The goals are the determination of how turbulence is produced, of how to control turbulence, of how to predict Reynolds stresses and the properties of small scale turbulence. An important feature of the research is that it uses direct numerical simulations (DNS), pulsed-image velocimetry (PIV), two-component laser-Doppler velocimetry (LDV), and a specially constructed facility for optical studies of turbulence. The chief thrust is to develop a conceptual framework for wall turbulence which can be used to analyze complex flows and to determine how turbulence can be modified. The approach taken is to consider the structure and dynamics of stress producing eddies. Studies will be made of (1) "simple" wall turbulence, such as fully developed flow in a channel or Couette flow, (2) flow over a wavy solid boundary, (3) flow of a solution of drag-reducing polymers. PIV and DNS will be used to study large scale eddies in the outer flow. Particular attention will be given to their origin and to their interaction with the flow oriented vortices close to the wall which are very large producers of turbulence. Proper orthogonal decomposition will separate large and small scale turbulence. Flow over a wavy boundary is fundamentally different from flow over a smooth boundary in that turbulence production close to the wall is associated with a separating shear layer rather than wall vortices. The addition of a long chain polymer, in amounts as small as 3 ppm, dramatically decreases drag by changing the structure in the viscous wall region, in ways which are poorly understood.