Verification

This web page details the list of tests carried out using LimitState:GEO to verify its accuracy against known limit analysis solutions. These tests are run every time the software code is updated.

Except where indicated, all analyses are performed using the default coarse nodal density. Known exact benchmarks are used where available. Where unavailable, the mean of the best available upper and lower bounds are used. In these cases, the benchmark is not necessarily the exact solution and could in fact be higher than the LimitState:GEO result.

Results are presented in terms of deviation from benchmark collapse load and also in terms of change in input parameters required for the LimitState:GEO solution to match the benchmark solution. This latter approach matches a design approach that considers uncertainty in material parameters. It is particularly relevant where the collapse load is a highly non-linear function of the soil strength. For example a change of phi' by 5% can result in a change of footing bearing capacity by over 25% for certain simple footing problems on cohesionless soil

In general, even when using the coarse nodal density, accuracy in most of the below tests varies from exact to ~5% in terms of the input parameters. Increased accuracy is straightforwardly achieved by increasing the nodal density.

An animated image and further details of each test including the source of the benchmark can be obtained by clicking on the relevant hyperlink. The input file used to generate the test result may also be downloaded.