Estimating retained gas volumes in the Hanford tanks using waste level measurements

Abstract

The Hanford site is home to 177 large, underground nuclear waste storage tanks. Safety and environmental concerns surround these tanks and their contents. One such concern is the propensity for the waste in these tanks to generate and trap flammable gases. This report focuses on understanding and improving the quality of retained gas volume estimates derived fiom tank waste level measurements. While direct measurements of gas volume are available for a small number of the Hanford tanks, the increasingly wide availability of tank waste level measurements provides an opportunity for less expensive (than direct gas volume measurement) assessment of gas hazard for the Hanford tanks.Retained gas in the tank waste is inferred from level measurements --either long-term increase in the tank waste level, or fluctuations in tank waste level with atmospheric pressure changes. This report concentrates on the latter phenomena. As atmospheric pressure increases, the pressure on the gas in the tank waste increases, resulting in a level decrease (as long as the tank waste is "soft" enough). Tanks with waste levels exhibiting fluctuations inversely correlated with atmospheric pressure fluctuations were catalogued in an earlier study. Additionally, models incorporating ideal-gas law behavior and waste material properties have been proposed. These models explicitly relate the retained gas volume in the tank with the magnitude of the waste level fluctuations, &/a. This report describes how these models compare with the tank waste level measurements. Additionally, this report contains: New methodology for calculating dL/dP fiom tank waste levels -the methodology is applied to tanks with suitably available level measurements, and incorporates aspects of waste strength. The estimates are compared with previous estimates, which were based solely on the ideal gas law. Experimental work -Waste simulants containing retained gas were created, subjected to various pressure changes, and the waste levels recorded. For some experimental configurations, the behavior of the simulant tank waste was observed to be consistent with the simpler gas-law model. For waste simulants with sufficiently large shear strength, the level fluctuations were less than expected under the model. The 1eveUpressure hysteresis that is observed in some tanks' waste levels was not evident in the experiments. Work towards evaluating the uncertainty of tank gas volume estimates from direct measurements -For Tank A-101, the retained gas sampler measurements are compared with waste-level-based estimates related to retained gas volume. The two measurements compare well with each other for this tank.Comparison of dL/@ estimates fiom atmospheric pressure readings at weather stations around the Hanford site -Atmospheric pressure measurements fiom weather stations surrounding the tank farms are used to calculate dL/@ summaries like those used to evaluate tanks for flammable gas hazard. The choice of weather station has a very small effect onTank data summaries-Detailed summaries...

Baseline estimate of the retained gas volume in Tank 241-C-106

“…Two linear regression fitting methods are used to obtain the correlation of waste level on pressure, WdP: ordinary least squares fit and "robust" least squares fit. The robust linear fit is designed to accommodate possible deviation from linearity in the relation between level and pressure caused by the potential time lag effects between pressure and level changes and the potential for mis-specified 2~~ values (Whitney et al 1997).…”

“…A rigorous derivation of the BPE model is given in Meyer et al (1997). The resulting model for the gas volume fraction and in situ retained gas volume are described by the following equations:…”

“…The gas mixture is usually released continuously into the tank headspace at about the same rate as it is generated-where concentrations are kept low by passive or active ventilation (Wilkins et al 1997). However, if the waste accumulates gas and the tank has a deep layer of supernatant liquid, there is a potential for a buoyant displacement gas release that could create flammable concentrations in the headspace (Meyer et al 1997). After transfer, the waste in AY-102 will have a relatively high gas generation rate and a deep supernatant layer.…”

Retained gas inventory comparison

“…The minimum number of waste level measurements needed to qualitatively assess the presence of retained gas are approximately six to twelve values over a period of six to twelve months. However, it is subject to the uncertainties associated with water evaporation or condensation, waste transfers to or from the tank, and structural changes in the waste surface (Meyer, 1997).…”

“…In particular, comparisons of tank waste level data with atmospheric pressure did not show simple linear behavior as would be predicted by the gas law model. Whitney (1997) developed a refined model utilizing surface level data from tanks with ENRAF' level gauges connected to the Tank Monitoring and Control System (TMACS) data acquisition system. This refined model estimates dlidp based on a "linear parallelogram estimation" method (BPE 11) and is considered to be more accurate than the gas law model; the tank data used as inputs to the model are of higher quality because of more frequent and precise instrumentation measurements.…”