Ultimate pier and contraction scour prediction in cohesive soils at selected bridges in Illinois.

PDFTech ReportFHWA\uc2\ubfICT\uc2\ubf13\uc2\ubf025ICT\uc2\ubf13\uc2\ubf025UILU\uc2\ubfENG\uc2\ubf2013\uc2\ubf2026\uc2\ua0R27\uc2\ubf105Bridge piersContraction (Thermodynamics)Cohesive soilsScourField studiesMathematical predictionRegression analysisIllinoisIllinois Center for TransportationStraub, Timothy D.Over, Thomas M.Domanski, Marian M.University of Illinois at Urbana-Champaign. Dept. of Civil and Environmental EngineeringUnited States. Federal Highway AdministrationIllinois. Dept. of Transportation. Bureau of Materials and Physical ResearchUS Transportation CollectionThe Scour Rate In COhesive Soils-Erosion Function Apparatus (SRICOS-EFA) method includes an ultimate scour prediction that isthe equilibrium maximum pier and contraction scour of cohesive soils over time. The purpose of this report is to present the results of testingthe ultimate pier and contraction scour methods for cohesive soils on 30 bridge sites in Illinois. Comparison of the ultimate cohesive and non-cohesive methods, along with the Illinois Department of Transportation (IDOT) cohesive soil reduction-factor method and measured scour arepresented. Also, results of the comparison of historic IDOT laboratory and field values of unconfined compressive strength of soils (Qu) arepresented. The unconfined compressive strength is used in both ultimate cohesive and reduction-factor methods, and knowing how thevalues from field methods compare to the laboratory methods is critical to the informed application of the methods.On average, the non-cohesive method results predict the highest amount of scour, followed by the reduction-factor method results;and the ultimate cohesive method results predict the lowest amount of scour. The 100-year scour predicted for the ultimate cohesive, non-cohesive, and reduction-factor methods for each bridge site and soil are always larger than observed scour in this study, except 12% ofpredicted values that are all within 0.4 ft of the observed scour. The ultimate cohesive scour prediction is smaller than the non-cohesive scourprediction method for 78% of bridge sites and soils. Seventy-six percent of the ultimate cohesive predictions show a 45% or greater reductionfrom the non-cohesive predictions that are over 10 ft. Comparing the ultimate cohesive and reduction-factor 100-year scour predictionsmethods for each bridge site and soil, the scour predicted by the ultimate cohesive scour prediction method is less than the reduction-factor100-year scour prediction method for 51% of bridge sites and soils.Critical shear stress remains a needed parameter in the ultimate scour prediction for cohesive soils. The unconfined soilcompressive strength measured by IDOT in the laboratory was found to provide a good prediction of critical shear stress, as measured byusing the erosion function apparatus in a previous study. Because laboratory Qu analyses are time-consuming and expensive, the ability offield-measured Rimac data to estimate unconfined soil strength in the critical shear\ue2\u20ac\u201csoil strength relation was tested. A regression analysiswas completed using a historic IDOT dataset containing 366 data pairs of laboratory Qu and field Rimac measurements from common siteswith cohesive soils. The resulting equations provide a point prediction of Qu, given any Rimac value with the 90% confidence interval. Theprediction equations are not significantly different from the identity Qu = Rimac. The alternative predictions of ultimate cohesive scourpresented in this study assume Qu will be estimated using Rimac measurements that include computed uncertainty. In particular, the ultimatecohesive predicted scour is greater than observed scour for the entire 90% confidence interval range for predicting Qu at the bridges and soilsused in this study, with the exception of the six predicted values that are all within 0.6 ft of the observed scour