Study on the application of robust statistical techniques in interlaboratory comparisons for oil emission spectrometers

In order to insure the reliability and the validity of the detection capability and detection data, 10 oil emission spectrometry laboratories which distribute in different areas tested 3 lubricating oil samples composed by different elements content, and interlaboratory comparisons for the detection capability of oil emission spectrometers were carried out. Based on robust statistical techniques, the detection data was evaluated and the evaluation results of robust Z-score were compared with the results of En. The results shows that the robust Z-score method had 86.67% hit rate, 7.33% failed rate, 6.00% omitted rate; while the En method had 31.33% hit rate, 66.67% failed rate, 2.00% omitted rate. the robust Z-score method is more suitable to be applied to interlaboratory comparisons for oil emission spectrometers than En method.