RESULTS AND DISCUSSIONLinearity and

RESULTS AND DISCUSSIONLinearity and RepeatabilityThe linear relation between measured color values and tracerconcentration has been tested. Blends with known amount oftracer ranging from 0.01% to 1% were prepared in a tumblingmixer (Turbula T2F, Willy A. Bachofen AG, Muttenz, Switzerland).Images of the blend were obtained with illumination at anangle of 458 to the camera. The obtained color values are shownin Fig. 6. The linear fit showed an R2 of 0.989, indicating a linearbehavior for this material. Repeatability of the RTD analysiswas assessed by measuring the RTD in triplicates during thesame extrusion run after 15, 30, and 45 min. Several experimentswere made, and Fig. 7 shows the triplicates with typicaland worst repeatability. The SD (standard deviation) and RSD(relative SD) within the triplicates are averaged over all experiments(Table 2) resulting in RSD values between 2.3% for T50and 19.7% for Pe. Thus, it can be concluded that the repeatabilityof the RTD measures (RTstart, RTmax, T50, RTrise, RTFWHM,and RTFW) is better with lower RSD compared to the fit parameters(h and Pe). Fluctuations were mainly due to changes in thepresentation of the strands to the camera and presumably due tosticking of material on surfaces or cavities during the extrusionprocess. This nonidealized material flow can be observed withseveral other measurement techniques as well (e.g., NIR [28]and Raman [16]).