2014 年 63 巻 1 号 p. 44-48
When the strength of earth is so low that the earth is not able to maintain the upper structure, soil improvement methods such as cement mixing are generally recommended. In an unconfined compression test for deciding cement volume and confirming the strength of soil improvement, the mean value of test results of three samples are generally used, although the dispersion of the three test results are left out of consideration. Recently, the estimation method of dispersion for test results is discussed internationally and the notation of dispersion is united under the term “uncertainty”, which is used in many fields of industry, medicine, agriculture and so on. As the Japanese Industrial Standard, the estimation methods of uncertainty are necessary to testing and calibration laboratories.
In this thesis, firstly the estimation method of the uncertainty of compression test results of cement mixing soil is presented by the analyses of verification experiments on causes of dispersion. Next, it makes clear that the uncertainty of test results is able to be estimated by the proficiency test that evaluates the test result of a laboratory in comparison with results of many laboratories. The uncertainties of the two methods above show that the measuring equipment and testing apparatus do not affect uncertainty, and the influence of the operator (laboratory) and the difference of samples are conversely large.