Probabilistic Materials

Sample material-property distributions and explore the resulting factors of safety.

JSlope can evaluate uncertainty in supported material properties after the deterministic slip-surface search. A normal LEM search first establishes each enabled method's deterministic critical surface and factor of safety. If any material field has an enabled distribution, JSlope then samples those values with Latin hypercube sampling and re-evaluates those fixed critical surfaces on local workers.

Fixed-surface uncertainty check

The probabilistic stage does not repeat the global slip-surface search for every sample. It shows how the deterministic critical surfaces respond to sampled material properties; a different surface could govern under a sampled material realization.

This workflow is distinct from Parametric Analysis, which adds scenario groups with deterministic changes to the External Boundary.

Configure a distribution

Open Define Material Properties. Supported numeric fields have a distribution button beside the deterministic input:

Select a material and field, then open its distribution control.
Turn on Enable.
Choose Constant, Normal, or Gamma.
Enter the distribution values and, if needed, Absolute Min and Absolute Max bounds.
Apply the change. An enabled indicator remains beside that material field.

For a normal distribution, the ordinary material value is the mean and the popup sets the standard deviation. A gamma distribution uses shape and scale. Bounds are validated against one another and against physical limits such as a friction angle of 0–90 degrees.

Slope-analysis distributions are available for unit and saturated unit weight; Mohr-Coulomb and Generalized Hoek-Brown parameters; tensile and low-stress strength; elastic and hydraulic parameters; Hu; and the anisotropy add-on angles. Only fields showing the distribution control are sampled.

Sample count and random seed

Open the scenario's LEM Analysis settings. Probabilistic Sample Count must be at least 2. Probabilistic Random Seed makes repeated runs with the same model and settings reproducible.

Probabilistic evaluation activates automatically when at least one supported material field is enabled; there is no separate analysis-method checkbox.

Results and graphs

The result summary retains the deterministic FOS and reports valid and invalid sample counts plus P25, P50, and P75 FOS values. In the LEM Results panel, choose Graph... to open the Probabilistic Results Graph.

The graph reports, per series, the mean, standard deviation, range, probability of failure (the share of valid samples with FOS below 1) and reliability index. It can:

compare groups, scenarios, and computed LEM-method series;
plot sampled variables, sample index, or FOS on selectable axes;
show a Histogram of any sampled value as bars or lines, as sample counts or as a percent of each series;
show the Cumulative distribution of any sampled value, so the probability of a FOS at or below any level can be read off directly;
Scatter any two values against each other, with regressions;
show Convergence: the mean, standard deviation, probability of failure or reliability index as samples accumulate, to judge whether enough samples were run (probability of failure and reliability index apply to FOS values);
highlight or filter numeric ranges using Any/All rules;
add Point query callouts that read every curve at an entered x or y value, for example the probability of failure at FOS = 1 or the median FOS at 50% on the cumulative plot (scatter callouts read the trend lines); and
export the currently selected graph rows to CSV.

Where FOS is plotted, the part of each histogram or curve below FOS = 1 is shaded: bars are hatched and the area under each line is filled. A scatter plot shades the FOS < 1 band instead.

The graph button is disabled until at least one stored scenario result contains probabilistic graph data. Samples that cannot produce a valid FOS remain counted as invalid rather than being presented as valid results.