本文研究了在相对误差和信度要求下试样的取样数目与测试结果离散程度的关系。根据长期的数据积累,可以引出临界离散系数的概念,从而决定了试验所必需的最少子样数,用以满足不同相对误差和信度的要求。 根据玻璃钢实质是一种结构的观点,按“失效”、“用坏”及“可以使用”三种受载情况,提出了假定的设计基准。由此,可按最少子样数的强度测试数据来决定设计值。本文提供的设计允许值的计算公式有明确的物理意义。文中列出了决定设计允许值的方法、步骤,供设计工作者参考。 This paper studies the relationship between the sampling number of the sample and the degree of dispersion of the test result under the relative error and reliability requirements. According to the long-term data accumulation, the concept of critical dispersion coefficient can be derived, thus determining the minimum number of subsamples necessary for the test to meet the requirements of different relative errors and reliability. According to the fact that FRP is a structural one, the hypothetical design basis was put forward according to the three loading conditions of “failure”, “worn out” and “can be used”. As a result, the design value can be determined based on the minimum number of subsample strength test data. The formula for the design allowable value provided in this paper has a clear physical meaning. The article lists the methods and steps for determining the design allowable values ​​for reference by the design worker.