Load transfer in reticulated and non-reticulated micropiles from large-scale tests

Slope failures are significant hazards to public and private infrastructure, and their maintenance and repair is costly. Stabilization through the installation of micropiles is a relatively new technique that has been used successfully in several instances in the U.S. and abroad. Uncertainties involving the development of forces within the micropiles and a lack of fundamental understanding of the interaction between the soil and micropiles have prevented widespread use of the method. The objective of the research presented is to provide direly needed experimental data to improve prediction of limit loads for micropiles in slope stabilization applications.;The experimental data is obtained from tests of large-scale physical models of slopes stabilized with micropiles. The experimental apparatus includes a model container, a pore pressure control system, soil, model reinforcement, and an instrumentation system. The container accommodates model slopes that are up to 8 ft by 14 ft in plan view with slope heights of up to 5 ft. This scale is large enough to permit construction of model slopes and associated stabilization schemes using techniques that closely mimic common field procedures, thereby reducing issues with scale effects that are commonly encountered with smaller scale models. The reinforcement consisted of 1.5-in. diameter gravity-grouted micropiles reinforced with 0.75-in. diameter steel pipe with 0.1-in. wall thickness. The reinforcement was designed so that its flexural stiffness was appropriately scaled down from a common range of flexural stiffness values for micropiles. Instrumentation for the device include tensiometers for monitoring pore water pressures (or suctions), wire-line extensometers for monitoring deformation within the model slopes, as well as strain gages to monitor loads in the micropiles.;The testing program consisted of eight tests divided among three sets of models. For the first set, two tests were performed on models with micropiles installed in an A-frame arrangement through a capping beam at the slope surface. Successive members were installed 15 degrees upslope and 15 degrees downslope of perpendicular. The second testing set consisted of three models with micropiles installed in a reticulated fashion through a capping beam at the slope surface. Successive members were installed 30 degrees upslope and 30 degrees downslope of perpendicular. Successive members were also installed 22 degrees out of plane with the upslope members being battered left and downslope members being batter right (when facing model). For the final set, three tests were performed on models with micropiles installed in an A-frame arrangement through a capping beam at the slope surface. Within each set of tests, member spacing was varied to evaluate its effect in addition to that of member inclination and reticulation. Pore pressures, soil movement, and loads in the micropiles were monitored and recorded for each test.;Soil movement and micropile loading were analyzed using soil-structure interaction methods. Based on measured soil movement profiles, p-y analyses were performed to back-calculate parameters that would lead to the prediction of the measured bending moments in the micropiles. Similarly, t-z analyses were performed to back-calculate parameters that would lead to the prediction of measured axial loads. The back-calculated parameters were compared with values from literature and with one another to evaluate the effect of member spacing, inclination and end restraint on load transfer.;Results of analysis indicated a few observations. According to past research, when spacing decreased, loads in the micropiles also decreased. This was seen. However, when spacing became sufficiently small, the load in the micropile actually increased. There appeared to be no apparent connection between spacing of micropiles and tilt of the apparatus at failure. However, reticulated arrangements offered the greatest increase in stability when comparing tilts of the apparatus at failure. In general, soil-structure interaction methods worked well to predict the loads in micropiles, but further testing and analysis is necessary to improve the t-z method. Numerous t-z analyses performed, mostly for downslope members, yielded loads in the micropiles that are opposite to what is intuitively expected. The values of skin friction back-calculated are also much outside the range of commonly used values or values that are physically possible for soils. Half the back-calculated values of limit soil pressure were between estimates predicted by Ito and Matsui (1975) and Broms (1964). The other half were slightly above Broms (1964). Back-calculated values of limit side shear predicted using the Beta method varied between 7 and 575. As with loads, these back-calculated values generally decreased with spacing. However, at some point, a decrease in spacing begins to cause in increase in values.