摘要
针对钢管混凝土徐变的随机性,利用支持向量机回归拟合钢管混凝土徐变效应的显式函数计算随机变量的灵敏度系数,并结合蒙特卡洛法进行随机性分析;采用自适应混合粒子群法优化支持向量机相关参数的取值以提高计算效率;对2个钢管混凝土徐变模型试验构件进行徐变随机性分析,并将计算结果与蒙特卡洛法计算结果进行对比验证了该方法的可行性;同时分析了钢管混凝土徐变效应各影响因素的灵敏度。结果表明:基于支持向量机与蒙特卡洛法对钢管混凝土轴压构件徐变随机性的分析结果与蒙特卡洛法分析结果相比相对误差较小;钢管混凝土徐变效应呈现随机性,概率密度曲线近似于正态分布。
According to the randomness of creep for concrete-filled steel tube, the support vector machine (SVM) was used to regression fit the explicit function of creep effect for concrete-filled steel tube, and sensitivity coefficients of random variables were calculated. The randomness was analyzed combining Monte Carlo simulation (MCS). The adaptive hybrid particle swarm was used to optimize the parameters of SVM to improve the computational efficiency. Two creep model test specimens of concrete-filled steel tube were used for randomness analysis, and the calculated results were compared with the results of MCS to verify the feasibility of the method. Meanwhile, the sensitivity of various influencing factors of the creep effect for concrete-filled steel tube was analyzed. The results show that the relative error of the SVM-MCS results compared with the MCS results for creep random analysis o{ concrete-filled steel tube axial compression members is little. The creep effects for concrete-filled steel tube have the randomness characteristics, and the probability density curves approach to normal distribution.
出处
《建筑科学与工程学报》
CAS
北大核心
2016年第4期44-50,共7页
Journal of Architecture and Civil Engineering
基金
国家自然科学基金项目(51208431)
中央高校基本科研业务费专项资金项目(SWJTU12CX064)
贵州大学人才引进项目(201517)
关键词
支持向量机
蒙特卡洛法
自适应混合粒子群法优化
钢管混凝土
徐变效应
灵敏度系数
support vector machine
Monte Carlo simulation
adaptive hybrid particle swarm op-timization
concrete-filled steel tube
creep effect
sensitivity coefficient