This paper presents a complete, reproducible MATLAB workflow for random-field (RF) reliability analysis of soil-nailed slopes. The implementation integrates four main software modules: a data-structure builder (inputData), a Karhunen–Loève random field generator (randomfield), a slice-based slope stability solver (soilCalc), and interchangeable reliability engines (FORM via slopeBeta; Monte Carlo Simulation via a custom MCS driver; Adaptive Radial Based Importance Sampling via ARBIS). The allData structure serves as the central data carrier, passing geometric, soil, nail, and probabilistic parameters between modules without global variables. Worked code excerpts from each module are presented alongside annotated pseudocode. Results demonstrate that spatially variable cohesion and friction angle, discretised over a 10-slice mesh, yield probability of failure estimates between 7% and 46%—values that conventional scalar-parameter FORM analysis cannot reproduce. The workflow is designed for transparency and replication; each function is self-contained with well-defined inputs and outputs.
Aduot Madit Anhiem (Sun,) studied this question.