The K-stiffness Method is an empirically-developed working stress method used to compute reinforcement loads for the internal stability design of geosynthetic-reinforced soil walls under serviceability conditions. In this paper, additional data from Japanese case studies for five full-scale field and three full-scale laboratory geosynthetic-reinforced soil walls are added to the database that was used to calibrate the original K-stiffness Method. One more case study from an instrumented wall in the USA is also introduced. Measured loads are compared with predicted loads using the current AASHTO Simplified Method and a modified version of the K-stiffness Method that has been adjusted by back-fitting model parameters to the extended database. The AASHTO Simplified Method is shown to be excessively conservative (on average) with respect to accurate prediction of reinforcement loads and to correlate poorly with measured values. The modified K-stiffness Method is demonstrated by statistical analysis to give ratios (bias) of average measured to predicted reinforcement load values close to 1 and coefficient of variation (COV) values for the maximum reinforcement load in a wall that are less than 25%.

The paper describes measurements taken from a series of four full-scale modular block walls that were constructed with reinforcement layers having different stiffness. The walls were 3.6 m high and were reinforced with two different polypropylene geogrid reinforcement materials, a polyester geogrid and a welded wire mesh. Each wall was constructed with the same modular block facing and reinforcement spacing of 0.6 m. The influence of compaction effort on wall displacements and horizontal toe load measurements at the end of construction was detectable in this investigation. These values were adjusted to account for the influence of different compaction methods on end-of-construction wall response. However, during subsequent surcharging the effects of initial compaction effort were erased. Reinforcement loads are computed from strain readings and results of in-isolation constant-load (creep) tests. Computed maximum reinforcement loads are compared with values predicted using the current AASHTO Simplified Method and the K-stiffness Method. The predicted magnitude and distribution of reinforcement loads are shown to be more accurate using the K-stiffness Method for polymeric reinforcement materials. For the relatively stiff welded wire mesh product, the measured reinforcement loads fell between values predicted using both methods.

No abstract