A 2D vortex panel model with a viscous boundary layer formulation has been developed for the numerical simulation of a vertical axis wind turbine (VAWT), including the operation in dynamic stall. The model uses the ‘double wake’ concept to reproduce the main features of the unsteady separated flow, including the formation and shedding of strong vortical structures and the wake–blade interaction. The potential flow equations are solved together with the integral boundary layer equations by using a semi-inverse iterative algorithm. A new criterion for the reattachment of the boundary layer during the downstroke of a dynamically stalled aerofoil is implemented. The model has been validated against experimental data of steady aerofoils and pitching aerofoils in dynamic stall at high and low Reynolds numbers (Re = 1.5 × 10^6 and Re = 5 × 10^4). For the low Reynolds number case, time-resolved 2D particle image velocimetry (PIV) measurements have been performed on a pitching NACA 0012 aerofoil in dynamic stall. The PIV vorticity fields past the oscillating aerofoil are used to test the model capability of capturing the formation, growth and release of the strong leading edge vortex that characterizes the dynamic stall. Furthermore, the forces extracted from the PIV velocity fields are compared with the predicted ones for a quantitative validation of the model. Finally, the model is applied to the computation of the wake flow past a VAWT in dynamic stall; the predicted vorticity fields and forces are in good agreement with phase-locked PIV data and CFD-DES available in the literature

This paper reports an experimental investigation on centrifugal compressor surge. The compression system consists of a four-stage blower with vaned diffusers and a large plenum discharging into the atmosphere through a throttle valve. Measurements of unsteady pressure and flow rate in the plant, and of instantaneous velocity in the diffusers of the first and fourth compressor stage, are performed during deep surge, at several valve settings and three different rotation speeds. Additional tests have been carried out on a different system configuration, i.e., without plenum, in order to obtain the steady-state compressor characteristics and to collect reference data on stall in surge-free conditions. In this configuration, a fully developed rotating stall was detected in the compressor diffusers, while during surge it affects only a limited part of the surge cycle. The goal of the present experimental work was to get a deeper insight into unstable operating conditions of multistage centrifugal compressors and to validate a theoretical model of the system instability to be used for the design of dynamic control systems.

An experimental investigation of the turbulent flow downstream of a planar sudden expansion has been performed by means of a 2D particle image velocimetry ͑PIV͒ technique. Flow fields at the Reynolds number of 10 4 have been measured in several mutually perpendicular planes of a channel having an expansion ratio of 3 and an aspect ratio of 10. As usual for large expansion ratios, the separated flow exhibits a strong asymmetry about the expansion axis and, consequently, very different reattachment lengths on the two side walls of the channel. The mean flow turns out to be substantially symmetric about the midspan plane and strong three-dimensional effects are observed in wide portions of the separation bubbles adjacent to the upper and lower walls. The reattachment lengths exhibit significant spanwise variations that are particularly pronounced in the longer reattachment line. Measurements performed in a single flow plane at Re=4·10 4 show that the influence of the Reynolds number on the mean flow is not completely negligible in the considered variation range. Based on a careful analysis of the PIV data, a model of the three-dimensional mean flow structure in the separation bubbles has been conjectured and it is provided in the paper. The present investigation contributes to clarifying the controversial three-dimensional character of the turbulent flow in a planar sudden expansion and provides accurate and detailed reference data for numerical simulations.

This paper describes, from a theoretical point of view, the behavior of compression systems during surge and the effect of passive and active control devices on the instability limit of the system. A lumped parameter model is used to simulate the compression system described in Part I of this work (Arnulfi et al., 1999), based on an industrial multistage centrifugal compressor. A comparison with experimental results shows that the model is accurate enough to describe quantitatively all the features of the phenomenon. A movable wall control system is studied in order to suppress surge in the compressor. Passive and active control schemes are analyzed; they both address directly the dynamic behavior of the compression system to displace the surge line to lower flow rates. The influence of system, geometry and compressor speed is investigated: the optimum values of the control parameters and the corresponding increase in the extent of the stable operating range are presented in the paper.

No abstract

No abstract

The present paper reports a numerical–experimental study on the dynamic behavior of a compression system based on a multistage centrifugal blower and fitted with an innovative device for the dynamic suppression of surge instability. The control device is of passive type and is based on the aeroelastic coupling of the basic compression system with a hydraulic oscillator. The controlled system is modeled at first by using a nonlinear lumped parameter approach. The simulated system dynamics within a wide range of operating conditions allows a parametric analysis to be performed and the optimal values of the control parameters to be singled out. Such optimal values are then used to design the hydraulic oscillator, which results in a technically feasible and very simple configuration. Finally, experimental tests are carried out on the compression plant with and without the passive control device, which demonstrate the effectiveness of the proposed control system in suppressing surge instabilities, at least within the limits predicted by the numerical simulation.

This paper describes, from a theoretical point of view, the behaviour of compression systems during surge and the effect of passive and active control devices on the instability limit of the system. A lumped parameter model is used to simulate the compression system described in Part I of this work (Arnulfi et al., 1998), based on an industrial multistage centrifugal compressor. A comparison with experimental results shows that the model is accurate enough to describe quantitatively all the features of the phenomenon. A moveable wall control system is studied in order to suppress surge in the compressor. Passive and active control schemes are analysed; they both address directly the dynamic behaviour of the compression system to displace the surge line to lower flow rates. The influence of system geometry and compressor speed is investigated; the optimum values of the control parameters and the corresponding increase in the extent of the stable operating range are presented in the paper.