Multi-Level Decomposition Framework for Reliability Assessment of Assembled Stochastic Linear Structural Systems
T Chatterjee, S Adhikari & MI Friswell (Swansea University)
ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering, 7(1), March 2021
To reduce the computational cost of assembled stochastic linear structural dynamic systems, a three-staged reduced order model-based framework for forward uncertainty propagation is developed. Firstly, the physical domain is decomposed by constructing an equivalent reduced order numerical model which limits the cost of a single deterministic simulation. This is done in two phases: (i) reducing the system matrices of the sub-components using component mode synthesis, and (ii) solving the resulting reduced system with the help of domain decomposition in an efficient manner. Secondly, functional decomposition is carried out in the stochastic space by employing a multi-output machine learning model which reduces the number of eigenvalue analyses to be performed. Thus, a multi-level framework is developed which propagates the dynamic response from the sub-component level to the assembled global system level efficiently. Subsequently, reliability analysis has been performed to assess the safety level and failure probability of linear stochastic dynamic systems. The results achieved by solving a 2D building frame and a 3D transmission tower model illustrate good performance of the proposed methodology, highlighting its potential for complex problems.
This material has been published in ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering, Vol. 7, No. 1, March 2021, the only definitive repository of the content that has been certified and accepted after peer review. The copyright is held with ASCE.
Link to paper using doi: 10.1061/AJRUA6.0001119
ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering