Peer-Reviewed PublicationProcesses 2026 · Open access · with HS Niederrhein and TU Graz

Vortex Formation in a Stirred Tank
Measured and Predicted

In the lab, a Rushton turbine tank (stirred by a six-blade radial disc impeller) ran with five different baffle setups while a camera filmed the vortex. Every setup was then simulated with the SimVantage free-surface model, which tracks the moving liquid surface.

Read the Paper Contact Our Team

Challenge

Vortex formation has to be designed for.

In most stirred tanks, a vortex that reaches the impeller means air in the impeller, lost power and a vibrating drive. Engineers can predict whether a vortex forms. Predicting its depth has been much harder, because the available correlations are fifty years old and ignore baffles entirely.

Some processes do want a surface vortex, for surface gassing or to draw down floating solids, and there its depth matters just as much.

Simulated surface vortex in a stirred tank, reaching down to the impeller and entraining gas bubbles into the liquid
Simulated vortex that has reached the impeller and is pulling air bubbles into the liquid. © SimVantage

Approach

Identical Tank and Evaluation in Lab and Simulation

The lab tank is small, 110 mm across. It ran without baffles and with four baffle designs, and at each stirrer speed the torque was measured and the vortex filmed. Our GPU-native free-surface model then simulated every run. Photos and simulated surfaces went through the same image evaluation, so the two data sets can be compared directly.

  • 01Baffle designs chosen in pairs with the same baffle index (a combined measure of baffle number and width), so any difference within a pair comes from baffle shape
  • 02Vortex depth, width and volume measured the same way on photos and on simulated surfaces
  • 03Nothing in the model was fitted or calibrated to the measurements, so every simulated value is a true prediction

Results

Vortex Geometry Predicted Without Fitting Factors

<5% Gap between simulated and measured slope of vortex depth, without baffles and with cylindrical DIN baffles
0.993 R² of the new swirl-number correlation for vortex depth, over all five configurations
0.3 Baffle index above which no vortex forms in turbulent flow
0 Fitting factors in the simulation

Main Findings

  • Baffle shape changes when a vortex starts. At the same baffle index, a tank with cylindrical baffles can run at much higher stirrer speeds than one with rectangular baffles before a vortex forms.
  • Vortex depth grows in a straight line with the Froude number (ratio of inertial to gravitational forces). Once you know the depth, you know the volume. Width stops growing at a limit set by the vessel.
  • The new depth correlation uses a swirl number, a measure of how strongly the liquid rotates, read from the simulated flow field. Because it comes from the actual flow, the swirl number already contains the effect of the baffles, and the correlation needs no constants fitted to one particular geometry.
  • If your process cannot tolerate a vortex, rectangular DIN baffles are the cheaper way to prevent one, because they use less material, are easier to build and draw less power.

Scope: single-stage six-blade Rushton turbine, flat-bottom vessel, D = H = 110 mm, water. The paper lists the deviation for each configuration and the limits of the study.

Plot of vortex depth over Froude number with experimental and simulated regressions, beside paired photographs and simulated vortex surfaces.
Measured and simulated vortex depth over Froude number, next to photos and simulated surfaces of the same vortices. Figure 6 in Lenters et al., Processes 2026, 14, 2942 (MDPI, CC BY 4.0).

Publication

Experimental and Computational Investigation of Vortex Formation in a Single-Stage Rushton Turbine Stirred Tank Reactor Under Standard and Non-Standard Baffle Configurations

Lenters et al. · Processes 2026, 14, 2942 · MDPI, open access (CC BY 4.0)
Joint study with the University of Applied Sciences Niederrhein and Graz University of Technology

Read the Paper All Publications

Facing a similar question? Let's talk.

If vortex formation, surface aeration or baffle design limits your process, we can simulate your own tank across your operating range before you change any hardware.

Book a Demo Contact Our Team All Case Studies