Applying open-path FTIR with computed tomography to evaluate personal exposures. Part 2: experimental studies

This paper presents the experimental evaluation results of using computed tomography coupled with OP-FTIR (CT-FTIR) measurement to estimate personal exposures. Experimental data were collected inside a ventilation chamber with a remote controlled robot as a surrogate for a real human. While the robot moved inside the chamber, a tracer gas (carbon monoxide) was released from a line source. A personal sampling device measured the true exposure on the robot. The estimated personal exposures were calculated from both the area sampling array data and the CT-FTIR measurements along with the information about the robot's locations in real time. The location information was obtained by applying image analysis on recorded digital videotapes. The average slopes of the regression lines between the true and estimated exposures was 0.76 with 1 included in the 95% confidence interval. The concordance correlation factor (CCF) between the true and the CT-FTIR estimated exposures was 0.52, which was similar to the findings from previous simulation studies. Kriging the area sampling array data with an exponential algorithm instead of a liner algorithm improved the CCF value from 0.60 to 0.75. This suggests that using a different basis function for the SBFM algorithm might improve the performance of our estimation approach. Based on the sensitivity and specificity analysis of the experimental data, we demonstrated that this approach is suitable as a warning system.