The boundary layer flow on a continuous flat surface issuing from a slot in a wall into an infinite still fluid was analyzed first by Sakiadis (1961a). His solution consisted of solving the well-known Blasius equation for two-dimensional boundary layer flow (for example, White, 1974) subject to different boundary conditions. The resulting solution is truly a similarity solution for the two-dimensional problem and provides the necessary solution for the early development of the boundary layer on a continuous cylinder upstream of any location where the transverse curvature becomes significant. Sakiadis (1961 b) then solved the problem of boundary layer flow on a continuous cylinder issuing from a wall. This was a momentum integral solution incorporating a logarithmic velocity profile that was suggested first by Glauert and Lighthill (1954). Sakiadis modified the profile to account for different boundary conditions. Middleman and Vasudevan (1970) claimed to have obtained a similarity solution for this problem; however, as was pointed out by Fox and Hagen (1971), a similarity solution does not exist for this flow.A common method for transforming the boundary layer equations written in cylindrical coordinates was suggested first by Seban and Bond (1951). A drawback of this transformation is the large streamwise variations of the velocity profiles and boundary layer thickness that result. Wanous and Sparrow (1965) used this transformation with great success in their perturbation series solution for laminar flow longitudinal to a stationary circular cylinder with surface mass transfer. From their solutions, the dimensionless boundary layer thickness varied by a multiplicative factor of 15 over the entire solution domain. This was not a drawback for their method, but it is for other approximate methods that seek to neglect the prior development of the boundary layer flow. Sparrow et al. (1970) incorporated the transformation used both by Gortler (1957) and Meksyn (1 96 1 ) very successfully in the development of their multiple-equation boundary layer theory. Their choice of a suitable transformation was based on a desire to reduce streamwise variation in the solution, which accomplished the goal better than the previously mentioned transformation. Their three-equation local similarity model yielded solutions of high precision a t the expense of some calculational complexities. The results presented here represent the extension of the theory of Sayles (1984) for the viscous boundary layer flow over semiinfinite cylinders. The first term on the righthand side of Eq. 2 introduces the transverse curvature effect that precludes the possibility of obtaining a similarity solution for this boundary layer flow. The radial velocity component, u, can be expressed in terms of the axial velocity component, w, as
(3)A dimensionless axial velocity component can then be defined as:where wo is the velocity of the translating cylinder. Also, dimensionless independent variables q and s are defined as: