Sediment erosion in a centrifugal pump

Abstract

AbstractWith a decrease in orders for their regular standard cargo-pump for chemical tankers, Frank Mohn Fusa started developing a new cargo-pump for the Offshore Supply Vessel market. This meant designing a pump which both should be able to pump high viscous oil-emulsion and a fluid called drilling-mud, which is a mixture of salt-water and small particles of barite stone. The first test performed when the thesis started showed that their initial design would not withstand the erosion from the particles for a long enough period of time; this because both the pump impeller and the pump volute casing suffered from heavy erosion.This master thesis has the objective of first quantifying the erosion in the original design and then comparing the CFD-simulations with real physical tests. From these results, together with analytical estimations of the erosion on the components and in compliance with erosion and fluid mechanical theory, a recommendation for a new design will be given. CFD-results for this new design will also be given. There is also a focus on finding the areas where erosion often occurs. Recommendations for material choice and coating are given for the original design.The results from the impeller simulations show that there are still some unknown variables in the prediction of erosion in CFD. While the steady state simulation differs quite much from the physical test, the transient simulation gives a better prediction for the pump impeller. It is found that the boundary layer in the grid has a large influence for the prediction of particle erosion. For the impeller it is found that the new design is a better solution than the original open impeller in regards to erosion rates, both at full-load and in the best efficiency point, while they are almost equal at part-load.The result from the pumps volute casing gives a good representation of the measured erosion from the physical test. The volute casings inlet tongue in the original design is a region where design changes are needed, since it has a high erosion rate caused by high velocity and high impinging angle. The new volute casing designed in this thesis, where the volume expansion is changed from radial to axial, gives better results, having approximately three times less maximum erosion rate at the inlet tongue. The analytical model used in this thesis gives a good estimation for some of the surfaces, while others differs to some degree from the CFD calculations. This is mainly caused by forces acting on the particles that are not accounted for in the analytical model.