Issue link: https://www.e-digitaleditions.com/i/1197147
January 2020 / 43 force in a given data array, the data associated with a time period selected in reference to the process. FPM expresses the magnitude of the force variation and, unlike the raw data, is always positive, making it highly suitable for process monitoring. With a parame- ter that is always positive, monitoring simply involves comparing magnitude or difference relative to setpoint. When a parameter can be both positive and nega- tive, monitoring becomes more complex. For process monitoring and control, a signal that stays positive is therefore easier from the perspective of formulating control decisions, whether they be manual or auto- mated. Furthermore, as a differential measurement, FPM values are unaffected by drift in the raw signal. The FPM data shows high repeatability across the ve runs and indicates uniformity within the blended batches. All four grains are clearly differentiated on Figure 4, with the couscous, which has a spherical shape and the smallest particle size, generating the lowest FPM values. Particle size, shape, and density, along with the cohesivity of the grains, are all likely to contribute to the FPM rankings exhibited. In an extension of this trial, two new grains with small particles (SP) and large particles (LP) were blended alone and as binary mixtures, as shown in Table I. Blending conditions were the same as for the earlier trial with baseline data gathered for 20 seconds and samples then mixed for 120 seconds; ve repeat measurements were again made for each sample. Again, these trials produced highly repeatable data and evidence of blend uniformity, as shown in Figure 6, graph a. There is also a clear, positive, nonlinear trend between the percentage of large particles in the blend and the magnitude of the FPM signal, as summa- which was done with a simple impeller, was monitored with a 4-Newton (N) DFF sensor with ve repeat mea- surements made for each sample. Figure 4 shows measured data for each of the four samples expressed in the form of moving average force pulse magnitude (FPM) measurements. Figure 5 illustrates how FPM data is derived from raw DFF measurements, which are associated with the impact of individual particles on the sensor. The FPM is the difference between the minimum and maximum FIGURE 4 Real-time force pulse magnitude data exhibits high repeatability and clearly differentiates the four grains. 0.40 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 (moving average over ve runs) FPM (moving average, n = 50), N Time, seconds couscous green lentils pearl barley long-grain rice Test Number LP Content, %v/v SP Content, %v/v 1 0 100 2 25 75 3 50 50 4 75 25 5 100 0 TABLE I Blend compositions for a series of trials mixing grains with large and small particles in known volumetric ratios (%v/v). FIGURE 5 Raw ow force data measured by a drag force ow sensor (a.) is usefully converted into force pulse magnitude values for process monitoring (b.). Force, N Time MAX FPM MIN a. Time Force pulse magnitude, N b.