Flow evolution models were developed to evaluate the performance of the new advanced design mixer pump for sludge mixing and removal operations with high-velocity liquid jets in one of the large-scale Savannah River Site waste tanks, Tank 18. This paper describes the computational model, the flow measurements used to provide validation data in the region far from the jet nozzle, the extension of the computational results to real tank conditions through the use of existing sludge suspension data, and finally, the sludge removal results from actual Tank 18 operations.A computational fluid dynamics approach was used to simulate the sludge removal operations. The models employed a three-dimensional representation of the tank with a two-equation turbulence model. Both the computational approach and the models were validated with onsite test data reported here and literature data. The model was then extended to actual conditions in Tank 18 through a velocity criterion to predict the ability of the new pump design to suspend settled sludge. A qualitative comparison with sludge removal operations in Tank 18 showed a reasonably good comparison with final results subject to significant uncertainties in actual sludge properties.

Flow evolution models were developed to evaluate the performance of the new advanced design mixer pump (ADMP) for sludge mixing and removal operations in one of the large-scale Savannah River Site (SRS) waste tanks, Tank 18. This paper is the first in a series of four that describe the computational model and its validation, the experiment facility and the flow measurements used to provide the validation data, the extension of the computational results to real tank conditions through the use of existing sludge suspension data, and finally, the sludge removal results from actual Tank 18 operations using the new ADMP. A computational fluid dynamics (CFD) approach was used to simulate the sludge removal operations. The models employed a three-dimensional representation of the tank with a two-equation turbulence model, since this approach was verified by both test and literature data. The discharge of the ADMP was modeled as oppositely directed hydraulic jets submerged at the center of the 85-ft diameter tank, with pump suction taken from below. The calculations were based on prototypic tank geometry and nominal operating conditions. In the analysis, the magnitude of the local velocity was used as a measure of slurrying and suspension capability. The computational results showed that normal operations in Tank 18 with the ADMP mixer and a 70-in liquid level would provide adequate sludge removal in most regions of the tank. The exception was the region within about 1.2 ft of the tank wall, based on an historical minimum velocity required to suspend sludge. Sensitivity results showed that a higher tank liquid level and a lower elevation of pump nozzle would result in better performance in suspending and removing the sludge. These results were consistent with experimental observations.

URS Washington Group, Quality and Testing Division management and technical personnel provided use of the SRS Civil Engineering Laboratory and test equipment, and technical expertise and input that was essential for performing the tests and obtaining the results in this report. W. Pope, Jr., W. B. Mhyre, and J. T. Waymer, URS, are recognized as key contributors to this project.

No abstract

This third in a series of four papers (Parts I–IV) presents the equations used for the initial evaluation of a pump’s ability to suspend solids and extends those equations to establish the minimum local velocity required to suspend those solids. This minimum velocity was used in a finite difference model in Part I to predict the ability of a pump to suspend, or slurry, solids that had settled on the bottom of a nuclear waste tank. To slurry waste, the Advanced Design Mixer Pump (ADMP) discharges a fluid jet that impinges on, shears, and then suspends the waste. Prior to the pump’s installation in a waste tank, the local velocity at a point in the flow required to suspend solids was found from available equations, material properties, and empirical data for similar pumps. Also, the computational fluids dynamics (CFD) model was validated in Part II by comparing it to flow rates measured in a full scale test facility where the ADMP was operated. The CFD fluid model could then be used to predict flow rates throughout the actual waste tank where the pump was to be installed, and the ability of the pump to adequately slurry the waste could be shown. All that needed to be done was to compare the local velocity of the fluid required to shear the waste into suspension to the velocities modeled throughout the waste tank. In short, this paper validates the theoretical and experimental basis for the derivation of a minimum velocity required for the flow stream to shear the waste into suspension. The final installment to this series of papers (Part IV) validates the application of the CFD model, by concluding that a nuclear waste tank is effectively cleaned to the wall throughout most of tank, using the ADMP.