Drag Measurement on a Suspended Sphere and its Application to Corrosive Gas Viscosity Measurement.

The estimation of drag force exerted by fluid flow over a sphere is a fundamental problem and has received attention both for bounded and unbounded flow. An offshoot of this study on a sedimenting sphere in a cylindrical tube has provided an effective means to ascertain viscosity of fluids. However, low-drag measurement difficulties and other practical impediments had restricted the sedimentation study to the determination of liquid viscosity only. The current work incorporates an accurate microbalance to obtain the drag force exerted by gas flow over a suspended sphere. The drag measurements are carried out by an ultra precise thermal microbalance having a sensitivity of 0.1 microgram. The results obtained have been extended for viscosity measurements. The experimental findings provide enough impetus to propose a methodology that can assist in the development of a drag based viscometer with good potential to handle corrosive gases. In conjunction with the experimentation, a theoretical investigation afforded through dimensional analysis and numerical studies has been carried out. The examined phenomenon over a centrally suspended sphere. Previous studies in this area had come up with correlations that established association between the coefficient of drag on a sphere in unbounded flow and cylindrically confined flow. These relationships, however, do not corroborate the experimental data well. This prompts the development of a new correlation that can be explained physically. Finally, an attempt has been made to explain the observed experimental results by computational methods. Although the flow geometry is fundamental, it is difficult to numerically test it for drag measurements. Two different approaches have been taken in the computational fluid dynamics study. A body-fitted coordinate system using the streamfunction -vorticity method and a primitive variable technique in cylindrical coordinates were utilized. The results show that the streamfunction-vorticity method runs into some difficulty with the body-fitted coordinate system. The primitive variable approach appears more robust and can be recommended for future work with the body-fitted coordinate.