Abstract:
Allowable Stress Design (ASD) and Load Resistance Factor Design (LRFD) are common frameworks for analyzing the loading of a structural element. Each method has advantages and drawbacks, and each is widely followed in the field of civil engineering. For geo-structural applications like support of excavation and foundation design, ASD has been the preferred approach for decades. ASD handles the inherent uncertainty in subsurface conditions with factors of safety that are dependent on the structure’ s function and quality of subsurface data. Conversely, LRFD approaches factors of safety differently: loads (demand) are scaled up and capacities (resistance) are scaled down, no traditional factor of safety is considered, and the method is assumed to be refined as more field data is gained.
In the latest AASHTO design specifications, LRFD method is prescribed for permanent retaining walls. However, inconsistencies arise when LRFD is used for retaining wall design versus ASD. Required toe embedment and wall loading demand vary greatly with varying subsurface conditions when the two design approaches are compared, all else being equal. This is due to application of linear load factors to properties that vary non-linearly in the LRFD approach, primarily active (Ka) and passive (Kp) earth pressure coefficients, resulting in uneven amplification of lateral earth pressures. These inherent inconsistencies are further intensified when introducing factors such as sloping grades, layered soil profiles, different wall types, and wall bracing.
When applied correctly, these two design approaches, in theory, should lead the designer to the same result with little discrepancy, as they do with steel structure design. However, use of the current LRFD method for various geo-structural retaining wall designs yields contradictory results, both conservative and unconservative, that suggest existing constructed retaining walls are unstable. Fundamentally, LRFD should be built on reliability-based calibrations; the successful past performance of existing retaining structures designed using ASD suggests that the LRFD approach requires considerable modification. The authors of this paper recommend that the current LRFD load factors be reviewed and revised based on more comprehensive wall performance data and wall type such that the difference between retaining structures designed using LRFD and those designed using ASD is eliminated.