surrounded by irrotational uid, the two circuits C 1 and C 2 drawn in Fig. 4 enclose the same vortex tube, such that the rst Helmholtz vortex law requires the circulation about these two circuits to be the same. De ning a variable » that measures distance along the vortex loop and has the value zero at the vortexloop tip, the strength0 S .»; t / of the vortex loop at a location » is related to the strength°.»; t / of the vortex sheet spanning the region in between the loop legs bywhere 0 0 is the vortex strength at the loop tip. To simplify the problem, we assume that the vortex sheet strength has a constant value°0 at the point where it is shed from the cylinder and that the vortex sheet is ejected outward from the cylinder face at a constant speed V E . This assumption is supported by the computational results of Ref. 9, which cover the time period during which the ejected vorticity wraps approximately one time around the primary vortex. The result [Eq.(2)] indicates that the vortex loop strength varies linearly both with time at a xed cross section and with distance along the loop at a xed time with slopes given by .where the vortex loop attaches to the cylinder.These slopes are compared to computational data of Ref. 9 for a case in which the vortex is inserted impulsively in a ow with no freestream velocity and S=D D 0:3. In the Ref. 9 computations the ejection velocity of the vortex sheet from the cylinder face is observed to maintain a nearly constant value V E » D 0:44 0=D, which is comparable to, but slightly larger than, the velocity 0:380=D induced by the columnar vortex (in isolation) evaluated at the ejection point. Estimating°0 using the correlation for the inviscid slip velocity W at the separation point gives°0 » D max.¡W / D 0:22 0=S. Lines having the theoretically predicted slopes for this case are shown in Fig. 5 to agree well with the computational data, which cover moderately short times after vorticity ejection from the cylinder. Over long times, it is likely that the increase in loop strength is limited by detachment and shedding of the loop from the cylinder surface.
V. ConclusionsAnalytical models are presented for aspects of the ow observed during normal interaction of a columnar vortex and a circular cylinder, including onset of outward ejection of secondary vorticity from the cylinder boundary layer and variation of the strength of the ensuing vortex loop as a function of time and position along the loop. It is found that vorticity is ejected from the boundary layer along the cylinder leading edge when the vortex-cylinder separation distance decreases below a critical value, which varies approximatelyin proportion to the ratio of vortex strength to normal freestream velocity. Once vorticity is ejected from the cylinder boundary layer, it forms a looplike shape that wraps around the columnar vortex. Another model is introduced for the spatial and temporal variation of the strength of the ejected vortex loop, which is controlled by roll-up of the vortex sheet that is continuallyejected fr...