To support power generation infrastructure in northern Manitoba, 1,100 helical piles were installed. Fifty-seven axial compressive load tests from this work were evaluated for this study. Six of the tests were instrumented with strain gauges to evaluate the contribution from shaft adhesion. Several theoretical methods to predict the ultimate capacity of helical piles have been adopted from common shallow or deep foundations formulas and are investigated. Empirical methods to predict capacity from torque measurements obtained during install are also common and used to compare to other capacities, predicted or measured. Pile load tests are often completed to refine estimates of capacity and many interpretation methods are available to estimate a failure loads from the test results. Several methods are explored and the interpreted capacity is compared to theoretical and empirical methods.
Using the theoretical methods, the influence of bearing capacity factors and shear strength of the soil were found to have the largest influence on capacity. Based on the ultimate capacity obtained from load tests and failure criteria, the theoretical methods over predicted capacity in every case. This is attributed to the selection of unrepresentatively high shear strengths, high Nc factors or a combination thereof. The selection of shear strengths is further scrutinized based on variability in testing data and it was found that further conservatism in the selection of mean shear strength data is pertinent. Capacity to torque correlations predicted capacities from 400 kN to 7,000 kN, owing to the large range of KT factors available. From site specific data, an average KT value of 11.9 was back-calculated. Several failure criterion were reviewed and three were used to estimate capacity from load-deflection curves. From this exercise it was noticed that methods which were applicable in most cases underestimated the capacity whereas methods which predicted higher capacities were applicable is less cases. Based on the De Beer failure criteria, a modified method, denoted the Creep limit, was developed which utilizes graphical and mathematical approaches to interpret a failure load. The Creep limit was found to be applicable in more cases while interpreting relatively high capacities.