Lateral Load Testing Micropiles To Evaluate The Impact Of Threaded Joints And Casing Embedment On Short Micropiles In Shallow Rock

Jointed micropiles were investigated to support bridge foundations where shallow rock was present. In this application, micropiles would have some amount of unsupported length, and would be subject to nominal lateral loads. Five lateral load tests were conducted on micropiles to demonstrate the impact of casing joints and rock socket embedment on lateral behavior. Load tests were simulated with FB-Multipier, a hybrid-finite element program for analysis of deep foundations, for proper design of the loading apparatus and instrumentation. The micropile sections were 2 m (6.56 ft) long, 273 mm (10.75 in) in diameter, 13 mm (0.5 in) thick walled, with a 50-mm (2-in) long threaded joint. Micropiles were made of one to three casing sections and were embedded 0.3, 0.6, 1.5, or 3.0 m (1, 2, 5, or 10 ft) into rock. In each test, two identical micropiles were pulled together using a simple frame until the weaker pile failed. All micropiles were instrumented with inclinometer casings while two contained strain gages at 0.76 m (2.5 ft) intervals. Load and displacement were measured at the pile heads. Tests showed the embedment of 0.6 m, 1.5 m, and 3.0 m (2 ft, 5 ft and 10 ft) sufficiently carried lateral loads over 180 kN (20 tons). The piles embedded 0.3 m (1 ft) were not able to maintain lateral load. In general, all micropiles tested to failure exhibited a brittle failure mode where there is small lateral deflection followed by a sudden failure at the casing joint. This was demonstrated in three of the load tests. In most cases, the micropiles were stiffer and stronger in the load tests than originally predicted numerically. The final test, where piles were not loaded to failure, verified that embedding vibrating wire rebar strain meters in the grout captured the bending moment profile of the micropile. The results showed the maximum moment in this case being carried directly across the threaded joint as predicted by the preliminary models. The results of these tests were used to calibrate the FB-Multipier numerical model. Using a load test that included bending strain measurements for a match, the soil properties were adjusted to produce reasonable matches for all of the load tests.