A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering

Thesis (M.Eng.) - Asian Institute of Technology, 1996

A tube flow viscometer complete with data acquisition system was designed and developed for continuous measurement of pressure drop and flow velocity. Experiments were carried out with five fluids using different diameter stainless steel tubes to cover a wide range of test conditions. The viscometeric characterization of non-Newtonian fluid flow for all test samples followed a power-law relationship in the form of wall shear stress vs. shear rate. As the fluids were highly viscous, the flow remained in the laminar region. The fluid foods followed different power-law parameters when using the Brookfield and tube viscometers due to very low shear rates in the former case and wall-slip in the latter case. However, an empirical relationship was established between the slip corrected tube viscometer parameters and Brookfield parameters, and this formed the basis for the estimation of flow behavior in pipes. Finally, another approach based on the application of neural networks very accurately predicted the pressure loss in pipe flow from the information on fluid density, tube diameter, mass flow rate and the flow parameters (K and n) of a power-law relationship determined with a Brookfield viscometer. Three neural network architectures were used for the training. The generalized regression neural networks were most easy to train and predicted the pressure drop gradient in tube flow with average absolute error less than 5%.