Alternative Ultrafiltration Membrane Testing for the SRS Baseline Process

摘要

The ability to more rapidly process high-level waste sludge and supernate, without sacrificing cost savings, continues to be a crucial challenge facing the Savannah River Site (SRS). There has, to date, not been any extensive investigation of alternative filter technologies for the SRS baseline process. To address this problem, a focused investigation into alternative, state-of-the art filtration technologies to facilitate the strontium and actinide removal process, which can be cost effectively implemented in existing facilities and current equipment designs, was completed. Filter technologies manufactured by Mott (0.1 micron and 0.5 micron) Graver (0.07 micron), Pall (0.1 micron and 0.8 micron) and GKN (0.1 micron) were evaluated. Membranes had a nominal inside diameter of 3/8 inches and an active membrane length of 2 feet. The investigation was performed in two phases. The first phase of testing evaluated the consistency or variability in flux through the different membranes using water and a standard 5.0 wt percent strontium carbonate slurry. The second phase of testing evaluated the achievable permeate flux and clarity through the various membranes using the SRS average salt supernate simulant at solids loadings of 0.06, 0.29 and 4.5 wt percent. Membrane variation data indicate that membranes having an asymmetric ceramic coating (Pall 0.1 micron and Graver 0.07 micron), typically displayed the lowest variability with water. Membranes without a ceramic asymmetric coating (Mott 0.5 micron and GKN 0.1 micron) displayed the highest variability. This is most likely associated with the experimental uncertainties in measuring large volumes of permeate in a short amount of time and to the impact of impurities in the water. In general, variability ranging from 4-56 percent was observed when using water for all membranes. In the case of variation testing using strontium carbonate, variability decreased to 3-12 percent. In addition, membrane structure or composition had little effect on the variability. Data obtained from SRS simulant testing, indicate that membranes having a ceramic asymmetric coating (Graver 0.07 micron, Pall 0.1 micron), typically achieved the highest average steady state fluxes for all solution concentrations evaluated. In general, the Graver 0.07 micron and Pall 0.1 micron membranes achieved fluxes approximately 13 to 21 percent higher than those observed with the baseline Mott 0.1 micron membrane using the SRS simulant at solids loadings of 0.29 and 4.5 wt percent. Membranes without a ceramic asymmetric coating (GKN, Mott) achieved the lowest average steady state fluxes. It is postulated that small particles present in solution were unable to penetrate the ceramic layer, thus producing surface filtration where the filter cake acts as the filter medium. Conversely, membranes without the asymmetric ceramic coating were susceptible to the small particles present in solution penetrating into the internal pore structure of the membrane, thus producing depth filtration where the porosity is greatly reduced by particles trapped within the interstices of the internal structure. Turbidity data indicate that permeate from the alternative membranes provided reduced or equivalent turbidity measurements when compared to the baseline 0.1 micron Mott membrane.