Corrugated metal, reinforced concrete and high-density polyethylene jointed pipe culverts have been used as drainage structures for some time. Considerable research has been performed investigating the structural performance of the culvert pipes and their ability support loading applied at the ground surface by motor vehicles. This volume of research is reflected in the design codes. However, research into structural performance of culvert joints is far behind the research into structural performance of the culvert pipes, and design guidelines for culvert joints are lacking. Gaps in the joint can cause water to be released and to scour the soil beneath the pipe. This can cause structural failure of the culvert. This research is a step toward providing the ability to design culvert joints.
It is possible to model jointed pipe culverts using full 3-dimensional finite element analysis. However, this kind of modelling is very complicated, time consuming, and not suitable for design. Beam-on-springs modelling is much simpler than 3-dimensional finite element analysis and is much easier to use in design. The objective of this study is to provide a beam-on-springs model that can accurately represent buried pipe culverts and predict movement at the culvert joints. This objective is achieved by testing 5 in-service pipe culverts, 1 pipe culvert installation, and 3 pipe culverts in a laboratory under static and dynamic loading at the ground or roadway surface. Test culverts were selected for their relatively small diameter, shallow fill depth, and material. The test culverts are modeled using beam-on-springs modelling to predict the joint rotation and shear displacement across the joint.
The experimental program includes testing of two corrugated metal, two reinforced concrete, and two high density polyethylene pipe culverts in the field, and one culvert of each material in the laboratory. Experimental results showed no correlation between fill depth and spring stiffness despite the fact that maximum crown deflections were higher at shallower fill depths. Results also showed that the displacements under dynamic loading are smaller than displacements under static loading.
Several existing beam-on-springs models were evaluated to see if they were suitable for predicting the spring stiffness beneath buried pipe culverts. A modified version of the method originally proposed for flat beams by Terzaghi did the best job of predicting the spring stiffness. The beam-on-springs modelling did a good job of predicting the vertical displacements, joint rotations, and joint shear displacements measured in the three laboratory test culverts. The modelling had some difficulty in predicting the measured joint rotations and shear displacements in the field test pipes. This is likely due to years of repeated loadings, load spreading caused by the road surface, and years of deterioration. Using a conservative spring stiffness value in the beam-on-springs modelling should yield conservatively predicted joint movements, and conservative designs.