Dataset for Probabilistic Ultimate Strength Analysis of Cracked Marine-Grade Aluminum Stiffened Plates Using Non-Intrusive Chaotic Radial Basis Function
Description
This dataset supports the research article titled "Probabilistic Ultimate Strength Analysis of Cracked Marine-Grade Aluminum Stiffened Plates Using Non-Intrusive Chaotic Radial Basis Function (NICRBF)". The central hypothesis of the study is that the ultimate strength of marine aluminum stiffened plates with pre-existing cracks, subjected to uncertain operational and material parameters, can be effectively quantified through a probabilistic computational framework combining nonlinear finite element analysis (NLFEA) with NICRBF-based uncertainty modeling. The dataset includes input parameters, output results, finite element mesh files, and post-processed simulation data generated during the probabilistic analysis of aluminum stiffened plates with varying crack lengths, orientations, and locations. All simulations are based on a stiffened panel geometry representative of high-speed aluminum catamarans and consider realistic imperfections and mechanical properties of Aluminum 5083-H116. Key findings derived from this dataset include: Crack length significantly affects ultimate strength variability, with a coefficient of variation (COV) ranging from 0.27 to 0.35. Crack orientation exhibits a non-monotonic influence on mean strength and variability, revealing complex stress redistribution mechanisms. Crack location impacts failure probability, especially near mid-span positions. Under sagging and hollow landing conditions, the probability of structural failure escalates with vessel speed and wave height, reaching near-certainty at design conditions in severe seas. How the data was gathered: The ultimate strength simulations were performed using ABAQUS for NLFEA, with geometric imperfections introduced following established models (e.g., Paik’s approach). A non-intrusive uncertainty propagation framework using Chaotic Radial Basis Functions was developed in MATLAB to perform stochastic analysis on hundreds of scenarios. How to use and interpret the data: Files include input configurations (material properties, geometry, crack definitions), output stress-strain results, and probabilistic distribution files. Researchers can replicate or extend the analysis by varying key parameters or integrating alternative crack modeling techniques. Educators and students may use it to study the impact of uncertainty in marine structural design. Designers can reference the results to develop safety criteria or operational guidelines for high-speed vessels. This dataset provides a valuable resource for advancing reliability-based design and safety assessment of marine aluminum structures under real-world uncertainties.
Files
Steps to reproduce
This dataset was generated as part of a probabilistic ultimate strength analysis of marine-grade aluminum stiffened plates with pre-defined crack characteristics. The following steps outline the workflow used to produce the data: Define the Baseline Structural Model A stiffened plate representative of high-speed aluminum vessels was modeled in ANSYS Workbench (version 2022 R1). Geometric parameters such as plate dimensions, stiffener profiles, and material properties (Aluminum 5083-H116) were defined. Boundary conditions assumed simply supported edges under longitudinal compressive loads. Parametrize Critical Variables The ANSYS model was parametrized to accept variable inputs for crack characteristics: Crack length (as a fraction of plate width) Crack orientation (in degrees) Crack position (along plate length or stiffener location) Generate Random Input Samples A Latin Hypercube Sampling (or equivalent random generator) was used in MATLAB to create a sample space of random variables based on lognormal distributions defined in the study (e.g., for crack length, angle, yield stress, etc.). Each sample represents a unique damage scenario. Run Parametric Simulations in ANSYS The random input values were fed into the parametrized ANSYS model. A batch of simulations was executed using the built-in parametric design analysis tool. Each run calculated the ultimate strength of the structure under the defined crack configuration. Extract and Post-process Results For each scenario, ultimate stress or failure metrics were extracted and compiled into Excel (.xlsx) files. These outputs were statistically analyzed to generate probability density functions, coefficients of variation, and failure probabilities under different wave and speed conditions. Reproducibility Guidance Users can reproduce the dataset by: Opening the shared .wbpz project in ANSYS Workbench Adjusting or importing their own random variable sets Running the parametric batch analysis and exporting results Optionally re-integrating the NICRBF method (via MATLAB) for efficient uncertainty quantification