Use of Statistical Methods to Study Corrosion Aggressiveness at Nevada Mechanically Stabilized Earth Wall Sites
AuthorThornley, John David
StatisticsView Usage Statistics
Nevada has over 150 mechanically stabilized earth (MSE) retaining walls at 39 locations. Recently, high levels of corrosion were observed due to accidental discovery at two of these locations, specifically I-515/Flamingo Road and I-15/Cheyenne Avenue intersections. The resulting investigations of these walls produced direct measurements regarding the corrosion losses of the soil reinforcements, which included both bare steel and galvanized steel, and electrochemical properties of the MSE backfill in order to identify its aggressiveness. One of the three walls at the Flamingo intersection was replaced with a cast-in-place tie-back wall at great expense because of the significant metal loss due to corrosion. The initial Flamingo investigation focused on average uniform corrosion loss values from the direct reinforcement measurements and laboratory backfill test results based on a variety of test methods. The investigation results are reevaluated in this thesis, through the incorporation of statistical analysis in order to effectively undertake a prediction that includes the variability in electrochemical properties. In this thesis it was found that the original MSE backfill approval test results are significantly different from those measured in these investigations. A correlation has been developed between two distinctly different soil resistivity laboratory test methods, namely the Nevada T235B and AASHTO T-288 methods. The Nevada test method under predicts the corrosive nature of backfill soils when compared to the AASHTO test method. A Nevada test predicting mildly corrosive backfill would be evaluated as corrosive using the AASHTO procedure. As the Flamingo and Cheyenne investigations show, this has proved detrimental to the service lives of MSE structures.The internal stability analysis of the two remaining MSE walls at the Flamingo intersection were also analyzed using corrosion loss models developed from the statistical analysis of the direct measurements. The results of the analysis from these two intersections were subsequently extrapolated to other Nevada MSE walls. Through review of the backfill approval data, specific Nevada MSE walls have been ranked relative to estimated backfill aggressiveness and specific suggestions for future corrosion analysis are recommended. There are four groups of evaluation methods that will be identified in this research. Each of these methods has its own usefulness, but some will be more costly than others. The four groups of evaluation methods for existing walls include representative backfill soil testing, installation of non-stressed soil reinforcements, nondestructive monitoring methods, and destructive direct observational methods.