中文摘要：

针对实际监测量的小子样分布类型可能不完全符合典型的分布函数，而最大熵密度函数可以直接根据样本的数字特征值进行计算，为此，选取经历过二期冷却的应变计组实测应变获得的最大主拉应力作为子样本，采用最大熵法计算获得子样本的最大熵概率密度函数，根据混凝土坝体已经抵御经历过拉应力的能力，来评估和预测抵御可能发生拉应力的能力，据此采用小概率事件法估计了混凝土实际抗拉强度，实例分析表明，最大熵密度函数包含的主观成分少，由此估计的混凝土抗拉强度和极限拉伸变形更可信。

英文摘要：

For small sample distribution type of the actual monitoring quantities may not fully comply with the typical distribution function ,while the maximum entropy density function can be calculated directly based on the numerical characteristic values of sam‐ples .Therefore ,the measured maximum principal tensile stress of strain gauge group under the second‐stage cooling was chosen as a sub‐sample ,the probability density function of the maximum entropy of the sub‐sample was obtained by maximum entropy method , the ability against the tensile strength that could occur which was evaluated and predicted based on the ability of the concrete dam to withstand and experience tensile stress ,so the small probability event method was applied to estimate the actual tensile strength of concrete .It is demonstrated by engineering practice that the less subjective composition w hich the maximum entropy density function contains ,thus the estimate of the tensile strength and ultimate tensile strain for concrete become more credible .