Search
Filters

51315-5433-A Bayesian Network Based Approach to Corrosion Risk Assessment of Radioactive Waste Tanks

The Hanford site contains approximately 55 million gallons of radioactive and chemically hazardous wastes arising from weapons production. This paper will present a Bayesian framework to model the probability of tank failures at Hanford and illustrate its use for tanks that are already known to have failed and leaked waste.

Product Number: 51315-5433-SG
ISBN: 5433 2015 CP
Author: Narasi Sridhar
Publication Date: 2015
$0.00
$20.00
$20.00

The radioactive waste tanks arising from the U.S. nuclear weapons program have undergone a complex history of design construction and operations over the last seven decades. The Hanford site contains approximately 55 million gallons of radioactive and chemically hazardous wastes arising from weapons production beginning with World War II and continuing through the Cold War. The wastes are stored in 177 underground storage tanks of which 149 are single-shell tanks (SSTs) and the remaining tanks are double-shell tanks (DSTs). The U.S. Department of Energy Office of River Protection is responsible for retrieving the tank wastes treating them in order to encapsulate them in glass logs and then permanently close the tanks and associated facilities. Current plans call for transferring the wastes from the SSTs into the DSTs over the next 25 years or so retrieving wastes from the DSTs and vitrifying them and closing all tanks by approximately 2048. Such a time line places a great emphasis on maintaining the integrity of both types of tanks. The single shell tanks that were originally built have suffered from corrosion and stress corrosion cracking while the double shell tanks that were constructed later are showing signs of corrosion and possible SCC of their primary containment. Predicting the future behavior of these tanks requires a thorough understanding of the mechanisms of various failure modes and an accurate characterization of the internal environment and the metallurgical state of the tanks. However despite a continuous monitoring process there are many uncertainties about the internal environments of the tanks. There are also uncertainties in the metallurgical and stress state of the tanks due to the age of these tanks. Finally there are uncertainties in our understanding of the factors leading to corrosion and SCC failures. Therefore a probabilistic assessment of tank failure is appropriate and because of the complex interplay of many factors a Bayesian network approach will be valuable. This paper will present a Bayesian framework to model the probability of tank failures at Hanford and illustrate its use for tanks that are already known to have failed and leaked waste. The consequences of leaks (the other part of the risk calculus) are not addressed in this paper.

Key words: Radioactive waste, tank, probability, pitting, stress corrosion cracking,  conference papers, 2015 conference papers, downloadable

The radioactive waste tanks arising from the U.S. nuclear weapons program have undergone a complex history of design construction and operations over the last seven decades. The Hanford site contains approximately 55 million gallons of radioactive and chemically hazardous wastes arising from weapons production beginning with World War II and continuing through the Cold War. The wastes are stored in 177 underground storage tanks of which 149 are single-shell tanks (SSTs) and the remaining tanks are double-shell tanks (DSTs). The U.S. Department of Energy Office of River Protection is responsible for retrieving the tank wastes treating them in order to encapsulate them in glass logs and then permanently close the tanks and associated facilities. Current plans call for transferring the wastes from the SSTs into the DSTs over the next 25 years or so retrieving wastes from the DSTs and vitrifying them and closing all tanks by approximately 2048. Such a time line places a great emphasis on maintaining the integrity of both types of tanks. The single shell tanks that were originally built have suffered from corrosion and stress corrosion cracking while the double shell tanks that were constructed later are showing signs of corrosion and possible SCC of their primary containment. Predicting the future behavior of these tanks requires a thorough understanding of the mechanisms of various failure modes and an accurate characterization of the internal environment and the metallurgical state of the tanks. However despite a continuous monitoring process there are many uncertainties about the internal environments of the tanks. There are also uncertainties in the metallurgical and stress state of the tanks due to the age of these tanks. Finally there are uncertainties in our understanding of the factors leading to corrosion and SCC failures. Therefore a probabilistic assessment of tank failure is appropriate and because of the complex interplay of many factors a Bayesian network approach will be valuable. This paper will present a Bayesian framework to model the probability of tank failures at Hanford and illustrate its use for tanks that are already known to have failed and leaked waste. The consequences of leaks (the other part of the risk calculus) are not addressed in this paper.

Key words: Radioactive waste, tank, probability, pitting, stress corrosion cracking,  conference papers, 2015 conference papers, downloadable

Also Purchased