Random creep and shrinkage in structures: Sampling

This paper deals with uncertainty in the prediction of the effects of creep and shrinkage in structures such as deflections or stresses caused by uncertainties in the material parameters for creep and shrinkage including the effect of random environmental humidity This problem was previously analyzed using two-point estimates of probability moments Here the same problem is analyzed using latin hypercube sampling This has the merit that the number of required deterministic structural creep analyses is reduced from 2" to approximately 2n where n = number of random parameters A method of taking into account the uncertainty due to the error of the principle of superposition is also presented The mean and variance of creep effects in structures are calculated and scatter bands are plotted for seven typical practical examples and the results are found to be close to those obtained with two-point estimates Further it is demonstrated that the distribution of creep effects is approximately normal if the creep parameters are normally distributed Finally it is shown that good results may also be obtained with the age-adjusted effective modulus method which greatly simplifies structural creep analysis.