Particulate fouling of sensible heat exchangers

Fouling by a petroleum gas oil and a dilute suspension of sand in water was studied as a function of mass flow rate and wall temperature. The experiments were carried out by circulating the liquid through a single tube maintained at constant heat flux by electrical heating. The change in fouling resistance and pressure drop with time was measured. The fouling resistance of the water and of the oil at low heat fluxes grows to an asymptotic value. At higher heat fluxes the oil fouling resistance increased almost linearly with time after an induction period. The asymptotic fouling resistance of both the oil and the water decreased with increasing mass flow rate. At constant clean tube wall temperature the initial fouling rate of the oil decreased with increasing mass flow rate. The initial fouling rate of the water increased with increasing mass flow rate up to a critical mass flow rate, and then decreased with further increases in mass flow rate. At constant mass flow rate, the initial fouling rate of the oil depended exponentially on the clean tube wall temperature. An activation energy of 29 Kcal/mole was calculated for the oil fouling process by fitting the initial fouling rate data to an Arrhenius type of equation. The pressure drop increase showed the same general trends with mass flow rate and tube wall temperature as did the fouling resistance. Fouling resistances for heated Kraft cooking liquor, calculated from pulp mill operating data and from a single fouling experiment, appeared to follow similar trends to those, followed in common by the gas oil and the water. The experimental results of this study were compared to the mathematical model of Kern and Seaton. While the shape of most fouling curves was in agreement with that predicted generally by this model, dependence of the initial fouling rate and of the asymptotic fouling resistance of the gas oil on the mass flow rate were both in disagreement with the detailed predictions of the model. For low mass flow rates of the water, however, even the detailed predictions were borne out. It was, moreover, possible to remove part of the sand deposit by increasing the velocity of the water, in accord with the postulated removal mechanism of Kern and Seaton, but the coke-like deposit from the gas oil could not be similarly removed by increasing the oil velocity. Mathematical models are developed in which the deposition term is written as the product of a material flux to the wall region and a sticking probability, after Parkins, and the removal term depends on the shear stress, after Kern and Seaton. Specific cases are considered where deposition is controlled by transfer to the surface, adhesion at the surface, and a combination of both steps.. Where deposition is controlled partly by transfer and partly by adhesion, the model predicts mass flow rate and temperature dependence of the initial fouling rate in agreement with the experimental results found for the oil. The observed asymptotic fouling resistance of the oil, however, depended less strongly on the reciprocal of the mass flow rate than is predicted by the model. Where transfer alone controls the deposition process, the extended model reduces to a form similar to that of Kern and Seaton.Applied Science, Faculty ofChemical and Biological Engineering, Department ofGraduat