基于Bayesian-MCMC方法的深受弯构件受剪概率模型研究

考虑主观、客观不确定性因素的影响，以深受弯构件受剪分析模型为研究对象，基于引入马尔科夫链-蒙特卡洛（MCMC）高效采样方法，通过R语言对深受弯构件概率模型参数进行MCMC随机模拟，给出参数的最优估计值及其对应的可信度，在先验模型基础上建立钢筋混凝土深受弯构件受剪承载力概率模型，完成模型前后的对比分析，并根据不同置信水平确定了深受弯构件受剪承载力的特征值。结果表明:基于MCMC方法得到的受剪承载力概率模型是在50000次迭代分析后产生的结果，能合理地解释影响参数的不确定性，可信度较高；后验概率模型计算结果与试验结果吻合良好，较先验模型更接近试验值，且离散性小。

STUDY ON PROBABILISTIC SHEAR STRENGTH MODEL FOR DEEP FLEXURAL MEMBERS BASED ON BAYESIAN-MCMC

A shear analytical model for deep beams was investigated considering the influences of objective and subjective uncertainties, and parameters in the probabilistic model of deep beams were simulated based on the R programming language, the process of which introduced Bayesian posteriori parameter estimation theory and the Markov Chain Monte Carlo (MCMC) method. As a result, the most optimal values and the reliability of model parameters were presented, simultaneously a probabilistic shear model for reinforced concrete deep beams was established and a comparison of before and after the model was conducted. Finally, the characteristic shear strengths of deep beams were achieved on the basis of different confidence levels. The research results show that the MCMC method assumed a credible reliability owing to the 50000 times of iterations by which the calculation was obtained, and it was presented that the posteriori probability model possessed a better agreement with test results than that of a prior probability model, while the posteriori model shows less discreteness.