M12 - A Computational and Experimental Investigation of Key Parameters Influencing Suction Device Performance

Introduction: : In emergencies, trauma, and battlefield medicine, rapid and reliable removal of fluids (blood, vomit, mucus) from patient wound sites and airway is often critical to save lives. Medical suction devices are universal, regulated tools for this task; they typically allow clinicians to set a target vacuum pressure. However, set-vacuum level alone does not fully determine suction performance. The actual fluid flow rate achieved depends on complex factors such as the suction tubing geometry, canister volume, fluid viscosity, and system compliance. In current practice, clinicians take vacuum pressure as the primary control parameter for specific clinical suction applications, yet this value does not fully account for how these other variables limit fluid flow in real scenarios. This gap underscores a need for a better understanding of suction dynamics beyond simple vacuum settings for effective clinical practice. Therefore, this study aims to experimentally characterize suction flow dynamics under realistic conditions to inform and validate a computational fluid dynamics (CFD) model which can predict fluid-specific suction device performance under the same vacuum settings.

Materials and

Methods: : An in vitro suction system was simulated to mimic routine clinical use, consisting of a 910 mm-long, 6.8 mm internal diameter medical-grade tube connected to a 1200 mL canister. A National Instruments NI-9215 DAQ system with a pressure transducer (OMEGA PX309-015V, 0–15 PSI) recorded vacuum pressure, while a calibrated load cell (OMEGA LC62SP, 6 kg, Class 3) measured water removal rates. Suction was applied using a commercial suction device (Zoll 330) set to 550 mmHg vacuum pressure. A controlled system geometry was used for measurements to isolate factors influencing fluid flow. Five distilled water trials (viscosity ≈ 1 cP) were conducted, and data processed in MATLAB (2024a) for analysis.

A computational fluid dynamics (CFD) model of the experimental setup was developed in COMSOL Multiphysics. The incompressible Navier–Stokes equations were solved using the k–ω SST (Shear Stress Transport) model to capture unsteady turbulent flow in the suction tube with canister pressure applied at the outlet. Boundary conditions were imposed to replicate the experimental test setup, to include: (1) a pressure outlet representing the vacuum (initially atmospheric, ramping toward 550 mmHg), and (2) an inlet = Atmospheric pressure + Pressure Head. No-slip conditions were applied at the walls. A mesh independence analysis using up to 1,043,230 elements achieved grid convergence ( < 1% flow rate deviation between normal and finer meshes). Simulation results for transient pressure and flow rate were validated against experimental data. The model was further used to simulate a higher-viscosity blood model (~5 cP) under identical vacuum conditions.

Results, Conclusions, and Discussions:: Under a 550-mmHg vacuum setting, canister pressure gradually rose to ~450 mmHg (±4.56, n = 5) in 7.2 minutes, resulting in a steady-state water flow rate of ~8.86 L/min (±0.5, n = 5). The CFD model closely replicated this behavior: A peak suction pressure of 439 mmHg (within 5% of the experiment) and a flow rate of 9.15 L/min (average velocity 4.2 m/s) were predicted; these simulated values fall within 3% of the measured value. Interestingly, these simulated flow rate and velocity values are roughly half of the idealized maximum (~17.5 L/min at 7.97 m/s) under instantaneous 550 mmHg vacuum. For blood, the model predicted an average flow rate of 6.60 L/min (velocity 3.03 m/s) after no settling time (zero minutes) and 6.55 L/min (velocity 3.01 m/s) after 60 minutes of fluid settling. These values differ sharply from idealized maximum flow rates of ~11.85 L/min and ~11.75, respectively, for a theoretical instantaneous 550 mmHg vacuum pressure.

The low error rate between experimental and simulated results—within 5% for peak pressure and 3% for flow rate— confirms the CFD model accurately represents suction dynamics. Such accuracy was achieved through fine mesh resolution and the use of the k–ω SST turbulence model, which successfully predicted flow behavior, including boundary layer separation at tubing curvature. Additionally, the CFD model captured the transient vacuum buildup and flow reduction observed experimentally— thereby highlighting the key roles of system dynamics and fluid inertia. Due to curved geometry and flow separation, the Hagen–Poiseuille equations—valid only for steady, laminar flow—overestimated flow, whereas the CFD while the CFD model closely matched experimental data. The validated CFD model was further applied to simulate higher-viscosity fluids. As fluid viscosity increased in the CFD model, it predicted reduced flow rates —due to increased flow resistance. These results indicate that increasing vacuum pressure alone is insufficient to compensate for thick fluids. Crucially, this study shows set-vacuum pressure is not a reliable indicator of suction performance. Accurate predictions require considering fluid viscosity, tubing geometry, and system compliance. It informs cost-effective next-generation suction device design, validation, and enhanced accuracy of CFD-based testing protocols.

Acknowledgements and/or References (Optional):: This work was supported by the AFWERX Direct to Phase II SBIR contract (FA864922P0928) and EmergenceMed, LLC, a San Antonio-based university startup led by Dr. Robert A. De Lorenzo and Dr. R. Lyle Hood. We extend our gratitude to the Medical Device Innovations (MDI) Laboratory and the Vascular Biomechanics and Biofluids Laboratory. The contributions of the National Science Foundation (NSF) I-Corps program are also gratefully acknowledged.