Evaluation of SRAT Sampling Data in Support of a Six Sigma Yellow Belt Process Improvement Project

Edwards, Thomas B.

Savannah River Site (S.C.)

关键词: Evaporators;    Sludges;    Chemical Composition;    Radioactive Waste Processing;    Sampling;   

DOI  :  10.2172/881441
RP-ID  :  WSRC-TR-2005-00292
RP-ID  :  DE-AC09-96SR18500
RP-ID  :  881441

美国|英语

来源: UNT Digital Library

【 摘 要 】

As part of the Six Sigma continuous improvement initiatives at the Defense Waste Processing Facility (DWPF), a Yellow Belt team was formed to evaluate the frequency and types of samples required for the Sludge Receipt and Adjustment Tank (SRAT) receipt in the DWPF. The team asked, via a technical task request, that the Statistical Consulting Section (SCS), in concert with the Immobilization Technology Section (ITS) (both groups within the Savannah River National Laboratory (SRNL)), conduct a statistical review of recent SRAT receipt results to determine if there is enough consistency in these measurements to allow for less frequent sampling. As part of this review process, key decisions made by DWPF Process Engineering that are based upon the SRAT sample measurements are outlined in this report. For a reduction in SRAT sampling to be viable, these decisions must not be overly sensitive to the additional variation that will be introduced as a result of such a reduction. Measurements from samples of SRAT receipt batches 314 through 323 were reviewed as part of this investigation into the frequency of SRAT sampling. The associated acid calculations for these batches were also studied as part of this effort. The results from this investigation showed no indication of a statistically significant relationship between the tank solids and the acid additions for these batches. One would expect that as the tank solids increase there would be a corresponding increase in acid requirements. There was, however, an indication that the predicted reduction/oxidation (REDOX) ratio (the ratio of Fe{sup 2+} to the total Fe in the glass product) that was targeted by the acid calculations based on the SRAT receipt samples for these batches was on average 0.0253 larger than the predicted REDOX based upon Slurry Mix Evaporator (SME) measurements. This is a statistically significant difference (at the 5% significance level), and the study also suggested that the difference was due to predictions of the formate and Mn concentrations in the SME product that were made at the time of the acid addition in the SRAT. For each of these analytes, the SRAT version was statistically different from the SME version (units are moles/kg SME product slurry): the SRAT values were, on average, 0.0914 larger than the SME values for formate and 0.0089 smaller than the SME values for Mn. A look at the signs of the terms corresponding to these two analytes in equation (2) indicates that both of these differences contribute to the calculated REDOX differences between the SRAT and SME product.