A Novel Machine Vision Based Image Analysis Method for the Analysis of Mixing Elements in Rotary Drums

Abstract

Mixing performance in continuous rotary drums has been studied. The video analysis method was developed to evaluate different configurations of straight lifters in the rotary drum. The method converts a captured video into a single image, called stack image. The color marker tracking was estimated based on the color saturation of the stack image. Coefficients of variation and mixing indices were calculated from the color saturation profiles for different straight blade lifter configurations. The video analysis method was confronted to the impulse response of acid concentrations in water solutions. The developed analysis method has been superior with viscous fluids compared to traditional tracer impulse method in mixing evaluations. Water and 1% CMC-water solution were used in this mixing study for covering broadly different viscous materials. The drum lengthwise results for one lifter configuration were obtained from a single experiment due to the block representation of the image analysis method. It enables mixing analysis of axial segments and interaction analysis of mixer configurations. Thus, the axial mixing can be studied in more detail with rotary drums. The increase of lifters, residence time, and tip speed improved axial mixing in the studied experimental setup.

Challenge of Residence Time Distribution Curve in Tundish for Continuous Casting of Steel

“…The relationship between pH and hydrogen ion concentration is as shown in Equation (40). Koiranen [81] used HNO 3 solution as the tracer, and found that the acid tracer could lead to deviation between repetitive experiments. The experimental results based on acid tracer indicated no clear trends in the mixing at high viscous media when the mixing conditions were changed.…”

“…[17][18][19][20]43] In order to find a more accurate method, the digital image processing began to be applied in chemical engineering for RTD curves in recent years. [81][82][83] Nigam et al [84] first developed the measurement of RTD curve by digital image processing in the tundish water model. The experimental procedure is shown in Figure 18.…”

“…The experimental results based on acid tracer indicated no clear trends in the mixing at high viscous media when the mixing conditions were changed. [ 81 ] Moreover, the strong acid solution may lead to some issues about personal safety. …”

Novel Image Analysis-Based Method for Residence Time Distribution Analysis in Steelmaking Tundish

“…Accuracy of results obtained from proposed method has been validated against results of conventional RTD methods. For carrying out RTD analysis in tundish, the present work utilizes the wellestablished concepts of digital image processing having applications in different fields of research [18][19][20][21][22]. A novel velocity threshold-based Lucas-Kanade (VT-LK) method for measuring the residence time of incoming fluid in steelmaking tundish has been developed.…”