ADVANCED MATERIALS & PROCESSES | SEPTEMBER 2026 25 practical problems it should solve, have lagged behind. To address this concern, the authors have proposed several biofilm test methods. Among the variables these methods need to control, flow turned out to be an especially important one, and a laboratory- scale reactor that brings flow under control has been needed for some time. Two reactor designs are proposed based on this idea (Fig. 3). The first is a vertical column-type reactor, working as a plug-flow reactor, where a test coupon sits in a one-directional, upward flow. The authors have already built and used this design in the lab. But a column-type reactor can only hold a limited number of coupons at once, so running enough trials for solid conclusions takes a lot of time. The second design is meant to solve this problem: a horizontal, stirred-tank reactor, where a rotating impeller drives a circulating flow around coupons placed in the tank, instead of the one-directional flow of the column. The aim is for this stirred-tank design to give the same biofilm-forming ability as the column-type reactor, while allowing for many more coupons to be tested at the same time. The authors believe shear stress at the material surface is the variable that controls how easily a biofilm forms. So before building the physical design, they used particle-based simulation to calculate the wall shear stress for both reactor shapes (Fig. 4). In the column-type reactor, shear stress on the coupon surface increased toward the lower part of the flow path, because the flow becomes more turbulent there. In the stirred-tank reactor, shear stress was concentrated near the wall and increased with impeller speed— from about 5.1 Pa at 60 rpm to about 5.7 Pa at 120 rpm on the cylinder wall. From these results, an impeller speed and coupon placement can be chosen that give, in the stirred-tank reactor, a shear-stress environment close to the one already confirmed in the column- type reactor. The working idea is that matching the shear environment should also give comparable biofilm-forming behavior, and the next step is to check this directly against coupon data from both reactors. FOR THE FUTURE These two examples show how visualization on the digital side can be fed back into the real world. This is different from a simple simulation. The aim is a closed loop: feedback experiments in the real space, then feedback again to the virtual side, and a new round of study starting there. This closed loop is what a true digital twin means, and we are working on it now. This digital-twin approach makes it possible to repeat experiments in a way that was not possible before, and to reach the right answer in physical space through this new experimental style. As virtual space and artificial intelligence continue to develop, the authors have a strong feeling that this will become the standard way of doing things. ~AM&P Acknowledgments The authors thank Prometech Software Inc. for kindly providing the simulation images used in Figs. 1-4 of this article. They also are grateful to PacForce K.K. for arranging this collaboration. For more information: Hideyuki Kanematsu, Specially Appointed Professor of The University of Osaka, Japan, and President & CEO of BEL (Biofilm Engineering Laboratory) Inc., +81.90.8499.6124, h.kanematsu@mat. eng.osaka-u.ac.jp. Fig. 3 — Simulation model for the two candidate biofilm reactors: (left) a vertical column-type reactor with one-directional flow (inlet velocity 6.0 m/min) and (right) a horizontal stirred-tank reactor with rotational flow (60/120 rpm). Created with Prometech Software. Fig. 4 — Simulated wall shear stress: distribution on the jig in the column-type reactor (left) and on the cylinder wall in the stirred-tank reactor (right). Created with Prometech Software.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==