Bridge scour evaluation in cohesive soils using physical and geo-electrical soil properties

Abstract

Scour accounts for 60% of all bridge collapses nationwide. There are two recommended methods for evaluating scour in cohesive soils, and both are flawed. The most common method is to use the scour evaluation manual from the Federal Highway Administration, Hydraulic Engineering Circular-18. The empirical equations, however, are based on the results of flume tests using cohesionless soils and are typically over conservative. The second alternative, performing site specific erosion testing, can be cost prohibitive, as erosion testing using available apparatuses is highly specialized and time consuming. Therefore, this research seeks to provide a new methodology that gives more accurate information than the equations based on cohesionless soils and is more cost effective than erosion testing. Various soil characteristics that affect the erosion of soil also influence in situ bulk electrical resistivity measurements. The objectives of this research were to develop a rapid methodology to predict soil erodibility using electrical resistivity and build an empirical equation to predict critical shear stress for erosion in cohesive soils using various soil properties. A total of 26 sites were used for in situ testing and soil sampling. Soil samples were used for erosion testing with the Erosion Function Apparatus and measuring geotechnical properties. The results indicate that electrical resistivity works as an excellent binary classifier for identifying soil with high erodibility. An electrical resistivity over 50 Ωm has a 93% probability of classifying the soil as high erodibility. As such, electrical resistivity tomography can be utilized to rapidly prioritize existing bridges where soils near the surface are classified as highly erodible. Regarding the second objective, multiple variable screening criteria determined percent fines, liquid limit, and electrical resistivity as the statistically significant model variables for predicting critical shear stress. Electrical resistivity is an increasingly common measurement by transportation agencies and this is the first time it has been identified as a variable for predicting critical shear stress. The use of electrical resistivity reduces the need for uncommon geotechnical, geochemical, and biological soil tests to predict critical shear stress. Design factors for implementing the developed model for a practical design were also recommended based on probabilistic analysis. If adopted by transportation agencies, this research will reduce the need for cost prohibitive site specific testing and overconservative bridge scour designs.