Investigation of Increased Wall Stiffness on Load Effect Equations for Soil Metal Structures

Abstract

The soil-steel bridge industry is expanding into new markets and more demanding applications. In the past, shallow corrugated plates (51mm by 152 mm) were replaced with deep corrugated plates (140 mm by 381 mm) to allow greater spans and covers to be achieved. In addition, stiffening rib products were also added to further improve moment capacity and structure stiffness. In 2010 the deeper corrugated plate (237 mm by 500 mm) was introduced into the market. The latest amendment to the Canadian Highway Bridge Design Code acknowledges this development. However, it further stipulates that the current simplified design equations may not be used as their validity for this application and product has not yet been verified In 2011, the first field structure was built using this new product to create a highway underpass on the Trans-Canada Highway in Newfoundland. The structure has a rise of 5.3_m, a span of 13.3 m and a height of cover of 2.7 m. The structure was effectively monitored from December 2011 until August 2013 using strain gauges and deflection prisms. Measured strain shows an increase of stresses during the backfilling process. At the end of the monitoring period a live load test was conducted using a loaded dump truck, structural responses varied slightly when applying the live load. A two dimensional non-linear finite element model was created in order to compare the monitoring results. The model was able to closely recreate bending moments; however, deflection and axial thrust results did not completely agree with the field results. The simplified method adopted by the Code is based on a parametric study using finite element models and flexibility number charts. In this research, a calibrated model allowed to conduct the same flexibility number analysis for various structures. The moment and axial loads derived from the monitoring program are compared to those predicted by the Code equations and recommendation are provided. In some instances it was shown that the finite element results were located above the Code limit. In other cases, particularly the results of large span structures, were shown to be very conservative.