A Probabilistic Model to Simulate Slender Reinforced Concrete Columnsusing Quasi-Monte Carlo Methods within SSJ Technology

摘要

The design procedure for slender reinforced concrete columns under uni-axial bent has been wellrnestablished by Model Code 1990 [8]. The behaviour of bi-axially bent columns is a much morerncomplex problem, particularly if buckling or stability problems occur. There are more than fifteenrnparameters in the numerical simulation, which constitute the system failure.rnThe random based Monte Carlo method and derived sampling methods are useful techniques tornperform probabilistic analysis and to calculate probability of failure. If the observed problem isrnrather complex (slender columns) and the simulation of one single trail takes a lot of time, thesernmethods are not useful. This paper discusses other types of simulation, which are known as Quasi-rnMonte Carlo (QMC) or low-discrepancy methods used to simulate slender columns. QMC methodsrnhave the potential to increase convergence rate of the probabilistic model. The underlying conceptrnis to simulate the material and system parameters with quasi-randomized behaviour under givenrndistributions. Therefore highly uniform point sets in unit hypercube are used. Stochastic Simulationrnin Java technology (SSJ) provides the necessary methods for implementation.rnThe finite element model bases on 3D nonlinear finite beam elements including nonlinear stressstrain-rnrelationships, nonlinear geometry effects and any given boundary conditions.