均匀性试验抽样误差的蒙特卡罗模拟

Monte Carlo simulation of sampling error of homogeneous test

对现行均匀性试验抽样误差和扩大样本容量后的抽样误差,应用蒙特卡罗方法进行数值模拟。结果表明,当总体标准偏差为1.5MPa,样本容量为10,对应水泥28d抗压强度分别为40MPa、50MPa和60MPa时,以变异系数表示的抽样误差分别为1.7%、1.4%和1.1%。这个误差很大,可能导致对水泥均匀性的错误评价。根据扩大样本容量后的抽样误差模拟结果,提出了均匀性试验抽样方法的修改建议。

The current sampling error of homogeneous test and the sampling error with extended samples capacity were numerically simulated by Monte Carlo Method. The result showed when the population standard deviation was 1.5MPa, the sample capacity was 10, and the corresponding 28d compressive strength of cement was 40MPa, 50MPa, 60MPa, the sampling error denoted by coefficient of variation was 1.7%, 1.4%, 1.1%. The error was so big to lead a wrong evaluation on cement homogeneous index. According to the Monte Carlo simulation of sampling error after the sample was extended, the modification suggestion was presented on sampling method for homogeneous test.