【摘要】 实验室间的能力比对及结果评价是判定实验室能力水平的重要技术手段，评价过程需要很好的结合比对项目的差异性、样品自身的固有特性以及参与实验室数量等多方面因素，选择适合的数据分析统计技术，才能精准的完成能力评价工作，获得较为公正的评价结论。本实验室组织《高频燃烧红外检测高温合金中微量碳硫的测定》比对项目，召集9家获取CNAS认可的实验室参与能力比对，获得检测数据，采用格拉布斯检验法先对数据中异常值（离群值）进行判别，如果出现异常值，则先行剔除异常值后再按照经典统计法进行统计，保留真值同时降低取伪的风险。经能力评价结果评估，除1家实验室的硫存在离群情况，其他实验室提供数据评价结果均为满意，证明参与的实验室在微痕量碳硫的检测能力方面具有较为稳定、优良的能力和水平。更多还原

【Abstract】 The ability comparison and evaluation results between laboratories are important technical means to determine the level of laboratory ability. In the process of capability evaluation, it is necessary to combine the differences of test items, the characteristics of samples, the number of laboratories and other factors, and select appropriate data analysis and statistical techniques to complete the capability evaluation and obtain a fair evaluation conclusion. Our laboratory organized a comparison project "Superalloy-Determination of carbon and sulfur content-Infrared absorption method after combustion in an induction furnace", gathered 9 qualified laboratories to participate in the comparison, obtain test data and Grubbs test was used to discriminate the data. In case of abnormal data, the abnormal value should be removed first and then counted according to the classical statistical method. In this way,the true value can be retained and the risk of falsification can be reduced. According to the capability evaluation results, all the data evaluation results provided by other laboratories meet the requirements except for the abnormal sulfur content in one laboratory,which proves that the laboratories involved in this capability evaluation have a high level of trace carbon and sulfur content detection.更多还原