Vibration of rotating shafts has been studied for different gap geometries, ranging from bearing configurations to pump systems. This paper deals with the rotor-flow dynamics of immersed shafts under moderate confinement-clearance gap about S=#=O.l( h w ere H is the average gap and R is the rotor radius). Following simplified assumptions, analytical models for the linearized forces, for both centered and eccentric immersed rotors have been developed as well as a theoretical nonlinear model which fully describes the nonlinear flow terms. These models were supported by encouraging results from preliminary experiments. In the present paper, we discuss some recent and representative results of an extensive series of tests performed on a small-scale model, in order to assert the validity of our theoretical models. From the overall experimental programma, the following conclusions emerged: (1) The linearized bulk-flow model is adequate, provided the dissipative effects are duly accounted for using an empirical friction coefficient to empirically model the turbulent stresses. Such predictions are quite accurate if the system is working at low rotor eccentricities and far from the instability boundaries. (2) However, for large rotor eccentricities and for dynamic regimes near the linear instability, the fully nonlinear model ‘Visiting researcher at Institute of Nuclear Technology. leads to better predictions. Obviously, these effects are instrumental to obtaining reasonable predictions for all post-stable motion regimes. (3) When discrepancies arise, the nonlinear model was usually found to be conservative. INTRODUCTION The effects of co-rotating annular flows on rotor vibrations have been studied extensively by many researchers, mostly in connection with bearings and seals [see for instance (Chills, 1993)]. However, flow structure interaction can also lead to significant effects in less common devices. In this paper we will address the dynamical behavior of immersed rotors, such as found in fast-breeder nuclear reactor pumps (for circulating the liquid sodium), and other ap plications. In such components, the clearance ratio 6 = H/R (where H is the average gap and R is the rotor radius) is typically about 0.1 -one or two orders of magnitude higher than the clearances found in bearings and seals. As a consequence, the flow is quite turbulent, inertial effects are then of prime importance and cannot be neglected as assumed in the basic Reynolds equation approach. Also, the shaft length subject to fluid forces is quite significant, these combined effects leading to specific rotordynamic properties. Using simplified assumptions for the flow, we have developed analytical models for the linearized forces, for both centered and eccentric rotors [ (Axisa and Antunes, 1992), (Antunes et al., 1996) and (Moreira et al., 2OOO)]. Because the flow was modeled as being two-dimensional, it was possible to extend our analytical solutions to fully account for the nonlin-