- Research Article
4
- 10.1080/14488388.2025.2517952
Novel multidimensional Gaidai hypersurface reliability design concept incorporating material manufacturing imperfections
- Jun 23, 2025
- Australian Journal of Multi-Disciplinary Engineering
- Oleg Gaidai + 8 more +8
ABSTRACT Imperfections within material manufacturing are defects or variations that arise throughout the production process and cause the intended functioning or design to change. Accompanied by geometrical manufacturing imperfections, these initial flaws may originate from various sources, e.g. poor craftsmanship, production procedures, and material quality issues. Existing structural reliability schemes, except CPU-expensive Monte Carlo (MC) based ones, are mostly limited to univariate (1D) or bivariate (2D) reliability models, e.g. FORM and SORM. This study advocates state-of-the-art multidimensional reliability approach that can be integrated early in the design phase, enabling structural and cost optimisation. The latter may alleviate overly conservative or non-conservative design values, currently based on existing 1D/2D design schemes, enhancing structural safety and cost-effectiveness within e.g. aerospace, deep ocean, material and offshore engineering. The primary goal of this investigation was to develop a generic state-of-the-art multi-variate reliability approach suitable for a high-dimensional dynamic system’s failure or damage risk assessment, allowing pertinent excessive dynamics information to be extracted from time-histories, that were recorded physically or Monte Carlo Simulated (MCS) numerically. Multi-modal Gaidai hypersurface reliability methodology provides for an accurate yet efficient prognostics of failure (damage or hazard) risks for a range of multi-modal, nonlinear dynamic systems, possessing multiple failure modes, accounting for time-variant memory effects. The presented generic multidimensional reliability methodology offers extended potential for big data processing, aerospace and other safety-related engineering applications, when the number of dynamic system’s cross-correlated dimensions is higher than two.
Read more