Issue link: https://www.e-digitaleditions.com/i/1197147
44 / January 2020 powderbulk.com sensor, samples were also extracted at two locations, as shown in Figure 7, for of ine analysis. The liquid com- ponent contained appreciable levels of sodium chloride (NaCl), and therefore, its distribution through the powder could be ef ciently measured of ine by poten- tiometric titration. Potentiometric titration is a chemical method of analysis in which the material's homogene- ity is monitored with an indicator electrode; as testing methods go, it is generally not that expensive and is reliable and readily available. Moving average FPM data for the two runs shows highly repeatable trends, as indicated in Figure 8. Liquid addition causes an increase in FPM as particles coalesce and capillary bonds strengthen between particles, but FPM values then decrease rapidly as the liquid disperses through the powder. A clear plateau of FPM values establishes after approximately 25 seconds of blending, indicating that mixing for this length of time is suf cient to produce homogeneity in the powder. The potentiometric titration of ine analysis shows similar trends. Sampling location 2, which is closer to the liquid injection nozzle, initially has a higher sodium chlo- ride concentration than Sampling location 1, but from approximately 24 seconds and onward, the two sam- pling points generate closely similar values that change minimally with further mixing. The two monitoring techniques indicate almost identical times for even liq- uid dispersion to the point of blend homogeneity. Therefore, the DFF sensor closely replicates an estab- lished of ine technique, generating comparable data in real time with no requirement to disturb the process to determine progress toward the desired endpoint. rized by the integrated area under the FPM curve data in Figure 6, graph b. In summary, this study shows that the DFF sensor effectively differentiates a range of grains and robustly detects the nonlinear impact of varying the composi- tion of grain blends with large and small particles. Case Study 2: Monitoring liquid dispersion in a vertical twin-shaft mixer In the second study, a DFF sensor was used to monitor the dispersion of a liquid component in a powdered food product being mixed in a vertical twin-shaft mixer, as shown in Figure 7. Repeat measurements of the dispersion process were carried out with the liquid addition rate and the total liquid amount kept constant between the two runs. While physical changes in the process were monitored in real time using the DFF FIGURE 7 A vertical twin-shaft mixer is set up for dispersing liquid in a powdered food product. The drag force ow sensor is installed directly in the mixer. Liquid addition nozzle Sampling location 2 Sampling location 1 DFF sensor FIGURE 6 Force pulse magnitude data exhibits a positive, nonlinear trend with large particle content (%LP) and effectively differentiates the blends. Area under FPM curve, Ns Average area under FPM curve, n = 5 35 30 25 20 15 10 5 0 couscous green lentils pearl barley long-grain rice FPM (moving average, n = 50), N Time, seconds 0% LP 25% LP 50% LP 75% LP 100% LP 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 30 40 50 60 70 80 90 100 110 120 130 4 N DFF probe, FPM moving average data a. b.