Subjective Odor Evaluation in Automotive Industry

This paper proposed a method to improve the consistency of the subjective odor testing for many automotive components. The variability in subjective odor test data has been very large due to lack of references, difficulty in maintaining reliable and consistent references, jury members’ bias, inconsistent jury members, misinterpretation of the odor evaluation scale, and lack of analytical methods for the data. To overcome these challenges, the proposed method used: 1) a selected and trained jury, 2) constant odor references or “anchors”, 3) a continuous data collection scale, and 4) statistical inference techniques. To reduce the variability from jury members, jury members were screened and selected based on a selection process and were trained using a consistent odor media. To reduce variability from bias originated by personal opinion or history, a consistent odor references or ‘anchors’ were introduced. Realizing the limitation of discrete scale, a continuous data collection scale and rating system was proposed. For the analysis of the data, using statistical inference techniques were suggested. When the proposed procedures were implemented, marked reduction in the variability in the subjective odor data was observed. Statistically significant conclusions could be drawn from the analysis of the data.