中文摘要：

建立了随机变量参数区间变化的在役桥梁可靠性评估模型,从而验证参数的小变化对在役桥梁可靠性评估的影响.通过实践工程表明,随机变量参数的小偏差,对评估结果影响较大,若以工程界接受的5%的偏差为例,计算发现,最大失效概率是最小失效概率的6 248.8倍.因此,在役桥梁中,概率分布不明确时,不能按设计参数进行可靠性评估,而应根据实测资料进行概率分布统计.由于受经费及测量手段的限制,很难得到精确的概率分布函数,建议采用区间变量的非概率可靠性进行在役桥梁的评估.

英文摘要：

The paper establishes existing bridge reliability evaluation model of random variable parameters with interval changes to verify influences of small changes of parameters on reliability evaluation of existing bridge.Practical projects show that small deviation of random variable parameters is of great influence on evaluation results.To take the 5% deviation accepted by the engineering field as an example,calculation results show that the max failure probability is 6 248.8 times of the min failure probability.Therefore,for existing bridge,if the probability distribution is not clear,reliability evaluation shall not be made based on designed parameters but refer to measured data.Due to limitation of funds and measuring methods,it is difficult to obtain exact probability distribution functions.Therefore,we suggest using non-probability reliability of interval variables to evaluate existing bridge.