In the dredging industry, centrifugal pumps are most commonly used and selected based on their operational capacities and expected slurry characteristics. The handling capacity of a centrifugal pump is however supplied for handling water and cannot be applied to other slurry conditions without an applicable correction. For handling Newtonian fluids there exist a well established de-rating method by {HI1983} and further improved until most recently {ANSI2015} was obtained. For non-Newtonian fluids however, no such generally accepted method exist, especially not for large centrifugal pumps as encountered in the dredging industry. In order to obtain a de-rating prediction for non-Newtonian fluids, researchers have tried to predict pump de-rating by applying the {HI1983} method. This is only possible if all fluid characteristics of a non-Newtonian fluid are summarized into a single value as used to characterise a Newtonian fluid. Several approaches are suggested by \cite{Walker1984}, \cite{Graham2009} and others. Dependant on which non-Newtonian parameter was found to be governing, apparent viscosities, Bingham plastic viscosity, and limiting viscosity were used in time.

Because all methods developed until today are based on small centrifugal pumps handling a mixture with a constant rheological behaviour, validity on centrifugal pumps with larger impellers and additional features as cover flushing cannot be guaranteed. Therefore the aim of this thesis research is to obtain a valid input parameter for a Newtonian de-rating method to predict the pump performance when handling a non-Newtonian fluid and determine the validity for large centrifugal pumps. To achieve this research goal, a measurement campaign on board the Trailing Suction Hopper Dredger `Pallieter' of DEME - Dredging International NV, working in a viscous environment was carried out. Measurement data for pump head, power and efficiency were obtained by installing a torque measurement device. Additional a device was installed to measure the dissolved gas volume fraction in order to separate effects caused by increased viscosity from multi-phase flow problems. After measurement data is obtained, available de-rating methods are compared and conclusions are drawn with respect to applicability.

Dissolved air in the mixture is of significant influence on the complete pumping process. After filtering the data to remain with performance data only influenced by the higher viscosity, de-rating models are analysed. Head showed very little to none de-rating when a viscous mixture was handled. At lower viscosities the de-rating method showed little deviation from the theoretical pump head, only at higher viscosities a significant de-rating was predicted but measurements didn't confirmed this. The same effects as for head yields for power. Significant de-rating was expected at high viscosities but measurements contradict this phenomenon. Although the Bingham plastic viscosity is the most conservative characterisation of a non-Newtonian viscosity, over-prediction still exist at high viscosities. This is most probable caused by the absence of viscous material in the axial gap between impeller and casing which in turn result in much less disk friction losses as would be expected.