Hydrological analysis and improved bridge scour prediction for selected streams in Hawaii

Abstract

Flood induced bridge scour is the erosion of soil sediments around a bridge structure caused by the increased velocity of a stream during a flood event and turbulent flows caused by any sub-structure in the stream bed. Bridge scour is the leading cause of bridge failure in the United States, responsible for 60% of all bridge failures. Since 1988 the Federal Highway Administration (FHWA) has required that all new bridges be designed to resist scour. An over prediction in scour depth can result in a significant, possibly prohibitive, increase in construction cost and difficulty. An under prediction in scour depth can result in a bridge that is unsafe over the life of the structure. The FHWA recommends using the Hydraulic Engineering Circular 18 (HEC-18) design manual to predict bridge scour. The HEC-18 manual includes a series of empirical equations to predict maximum scour depth based on laboratory experiments in non-cohesive soils. Researchers have shown that for certain situations, the HEC-18 method over predicts the depth of scour, sometimes by a large margin. Researchers from Texas A&M University have proposed the Scour Rate in COhesive Soils Erosion Function Apparatus (SRICOS-EFA) method that accounts the development of scour over time and a wider range of soil properties than the HEC-18 method. The SRICOS-EFA requires an erosion rate curve to define the soil properties, whereas the HEC-18 method only requires the median grain size. A hydrograph is also required for the SRICOS-EFA method whereas the HEC-18 method only requires a peak flow rate. The SRICOS-EFA method is still new and requires more data to evaluate and validate the accuracy of the scour predictions. The primary objectives of this study were to evaluate the HEC-18 design manual and SRICOS-EFA scour prediction methods and assess the hydrologic analysis required for each method. The pier scour predictions for both methods were evaluated for a single flood event, as only one scour event was available for use in this study. The hydrologic methods required to perform each scour calculation will be analyzed. This hydrologic analysis will include different methods to predict peak flow, determination of appropriate rainfall and stream flow duration for a design event and development of design hydrograph. Two case studies on the island of Oahu, Hawaii were conducted in this study. The first case study was conducted on the Kaelepulu Stream. A scour event was recorded on January 2nd, 2004. The HEC-18 method and SRICOS-EFA method were used to make predictions for the recorded event, and these predictions were compared with the recorded scour depth. The second case study was conducted on the Manoa-Palolo Stream. Flood frequency analysis was conducted on the Manoa-Palolo watershed to determine the 100 year peak flow rate using statistical analysis and rainfall-runoff models. The peak flow rates and design hydrographs generated by each of 6 different methods was compared and evaluated. The scour depth caused by the 100 year flood was predicted using both methods and these results were also compared. Both the HEC-18 method and the SRICOS-EFA method required a range of hydrologic and geotechnical parameters. Hydrologic analysis was conducted to determine how to reasonably represent local hydrologic conditions in a scour design calculation. The sensitivity of the final scour depth to variations in different parameters including the effect of hydrograph duration, hydrograph shape and soil type was analyzed with respect to local hydrologic conditions. The January 2nd, 2004 scour event in the Kaelepulu stream is utilized to validate the scour prediction methods. The SRICOS-EFA method provided a more accurate prediction of scour depth than the HEC-18 method for the January 2nd scour event in the Kaelepulu Stream. The SRICOS-EFA method and the HEC-18 method had similar maximum scour depths for soils that were highly erodible or for floods conditions that were sufficiently long. Soils with low erodibility were found to be more sensitive to variations in hydrograph duration and shape than soils with high erodibility.