OpenCossan 2.0: an efficient computational toolbox for risk, reliability and resilience analysis
Edoardo Patelli, Silvia Tolo, Hindolo George-Williams, Jonathan Sadeghi, Roberto Rocchetta, Marco de Angelis, Matteo Broggi · TU/e Research Portal · 2018
Many complex phenomena and the analysis of large and complex system and network can only be studied adopting advanced computational methods. In addition, in many engineering fields virtual prototypes are used to support and drive the design of new components, structures and systems. Uncertainty quantification is a key requirement and challenge for a realistic and reliable numerical modelling and prediction that spans across various disciplines and industry as well. The treatment of uncertainty required the availability of efficient algorithms and computational techniques able to reduce the computational cost required by the non-deterministic analysis and to interface with opensource and commercial model (e.g. FE/CFD) and libraries. In order to satisfy these requirements and allowing the inclusion of non-deterministic analyses as a practice standard routing in scientific computing, a general purpose software for uncertainty quantification and risk assessment, named COSSAN, is under continuous development. This paper presents an overview of the main capabilities of the recent release of the Matlab open source toolboxes OPENCOSSAN. The new release includes interfaces with 3rd party libraries allowing to couple OPENCOSSAN with the state-of-the-art tools in Machine Learning and Meta-modelling. In addition, new toolboxes for reliability and resilient analysis of system and network are also presented. OPENCOSSAN is released under the Lesser GNU licence. It is therefore freely available. It is also be package as a Python or Java library for distribution to end users who do not need MATLAB.