Analysis and design of incompressible rubber seals using experimental and finite element methods

Joint and crack seals are major problems in pavements. It is important to understand seal behavior and joint/crack geometry in order to design proper seals. A detailed numerical study was performed on incompressible rubber-like sealant material used in design of joints and cracks. This was done with the help of the finite element program 'ABAQUS' (FEM) using a Mooney-Rivlin formulation. The key issue is that FEM allows calculation of results for different seal configurations without destructive experimental testing which may be difficult and expensive to perform.; Laboratory experiments were performed on three different materials: Silicone Dow 888, Hi-Spec Rubber Asphalt, and Silicone Dow X3-5902. Rubber seal cross-sections were cast and tested in tension and compression to measure sag and bulge. The strain energy function was determined for Silicone Dow 888. Comparisons between FEM output and laboratory results were performed. Using FEM (ABAQUS), various seal cross-sections results were evaluated: namely rectangular, trapezoidal, trapezoidal-rectangular (Trap-Rec), and hollow sections. Analyses were also performed on rectangular sections under the combined effect of tension-shear and compression-shear. Design methods for rectangular and trapezoidal seals were developed and a method for determining a limiting design stress using Von Mises criteria is suggested.; Results indicate that, when sealing cracks, it is better to design for a trapezoidal and not for a rectangular cross-section. The use of a backup rod is recommended to eliminate the possibility of creating a Trap-Rec cross-section, which may be fatal to the seal structure. Hollow cross-section rubber seals should be considered for joints that tend to stay in tension throughout their life. Under static loads, shear displacement was found to have a minor effect when combined with tensile or compressive displacements. Deformations in incompressible seals were the same, for similar cross-sections, regardless of the rubber material used. All specimens tested failed in cohesive-adhesive mode. Silicone Dow 888, like any other material, was found to have a point where the material changes in performance; this point was named 'rubber monotonic yield point,' and was compared with equivalent Von Mises stress results obtained by FEM.