Electrochemical noise has been used at the Hanford Site for a number of years to monitor in real time in order to detect pitting corrosion and stress corrosion cracking (SCC) mechanisms in high level nuclear waste tanks. Currently the monitoring technique has only been implemented on three of the 177 underground storage tanks on the site. Widespread implementation of the technique has been held back for a number of reasons, including issues around managing the large volume of data associated with electrochemical noise and the complexity of data analysis. Expert review of raw current and potential measurements is the primary form of data analysis currently used at the Hanford site. This paper demonstrates the application of an on-line data filtering and analysis technique that could allow data from field applications of electrochemical noise to be managed by exception, transforming electrochemical noise data into a process parameter and focusing data analysis efforts on the important data. Results of the analysis demonstrate a data compression rate of 95%; that is, only 5% of the data would require expert analysis if such a technique were implemented. It is also demonstrated that this technique is capable of identifying key periods where localized corrosion activity is apparent.
The Hanford Site stores approximately 253 million litres of radioactive waste from 50 years of plutonium production in 177 underground tanks. One hundred and forty-nine of these tanks are constructed of a single shell of welded mild steel that was not stress relieved following fabrication. The remainder of the tanks are a double shell construction, the inner shells of which were stress relieved following construction. The waste contained in these tanks is in liquid, solid, and sludge forms and the tanks also have a vapour space. In general, the waste has a high pH (>12) and contains sodium nitrate, sodium hydroxide, sodium nitrite, and other minor radioactive constituents resulting from plutonium separation processes. The first leaks were detected in the single-shell tanks shortly after the introduction of nitrate-based wastes in the 1950s. Leaks are now confirmed or suspected to be present in a number of single-shell tanks.1 To date, there have been no leaks detected in the double shell tanks. There is currently a program underway to condense and transfer the waste from the single shell tanks to double shell tanks and no additional waste is being added to the single shell tanks still in service. Nitrate, nitrite and hydroxide levels are controlled in the double shell tanks so the dominant corrosion mechanism is general corrosion at a rate of less than 1 mpy. If off-normal waste chemistries occur in the double shell tanks or if areas were missed during the stress relief process, pitting, crevice corrosion and nitrate stress corrosion cracking (SCC) could become active in the double shell tanks.2
Corrosion monitoring and control in the double shell tanks are currently provided at the Hanford Site through a waste chemistry sampling and analysis program. In this process, tank waste is sampled, analyzed and compared to a selection of laboratory exposures of metal coupons in simulated waste. Tank wall corrosion is inferred by matching measured tank chemistries to the results of the laboratory simulant testing. This method is expensive, time consuming, and does not yield real-time data.
In 1995 a program was started to develop electrochemical noise (EN) based systems to improve Hanford?s corrosion monitoring strategy. Lab testing was conducted through 1995 and 1996 with the first prototype system installed in August of 1996. System design has progressed through a number of installations si
