Modeling Non-Newtonian Shear Thinning Fluid Dynamics

20 concepts16 papers1 proposal

When we think of a fluid, we typically imagine water. Water is a Newtonian fluid: its viscosity remains constant no matter how hard it is stirred, shaken, or squeezed. In the language of fluid mechanics, its shear stress scales linearly with its shear rate.

However, many of the fluids that drive our world—and flow through our bodies—defy this simple linear relationship. Blood, mucus, cell culture media, polymer solutions, and industrial slurries are non-Newtonian fluids. Specifically, most of these exhibit shear-thinning (or pseudoplastic) behavior: the faster the fluid layers slide past one another, the less viscous the fluid becomes. Imagine ketchup—stubbornly thick when stationary, but free-flowing once shaken.

For decades, computational models simplified these complex liquids by adopting the Newtonian fluid assumption to save on computational costs. But as simulations move from idealized geometries to patient-specific medicine and high-precision engineering, this simplification introduces significant errors. Modern research focuses on integrating non-Newtonian rheological models into computational frameworks to capture how variable viscosity fundamentally alters flow patterns, wall stresses, and transport dynamics.


The Failure of the Newtonian Assumption

Why does assuming a constant viscosity fail in complex environments? The error stems from the fact that complex geometries create highly non-uniform velocity gradients. When we assume viscosity is constant, we miscalculate the forces exerted by the fluid on its boundaries, known as Wall Shear Stress (WSS), as well as the fluid's internal resistance to flow.

Low-Velocity Stagnation

In regions where fluid slows down or stagnates, shear-thinning fluids become highly viscous. For example, in the left atrium of the heart, complex anatomical variations can create pockets of low-velocity flow. As Mill et al. demonstrate, a Newtonian model completely misrepresents these zones by underestimating blood's viscosity when it slows down. This underestimation obscures regions of true stagnation, which are critical to identify because stagnant, highly viscous blood is prone to forming dangerous thrombi (clots).

Micro-circulation and High-Stenosis Regimes

Conversely, when a fluid is forced through a tight restriction, the local shear rate spikes, causing the viscosity to plunge. In cardiovascular diseases where arteries become severely narrowed (stenosed), or within the microscopic vessels of myocardial tissue, the local flow resistance changes dynamically. Papamanolis et al. note that ignoring these shear-thinning effects in the micro-circulation leads to highly inaccurate perfusion simulations. Similarly, Wang et al. point out that constant-viscosity surrogates fail to capture the true physiological reality of wall shear stress in high-stenosis regions, which is vital for predicting arterial rupture risks.

Biological Barriers and Microfluidics

This issue extends beyond blood. In mechanical ventilation, the mucus lining the airways acts as a protective barrier. Treating mucus as a Newtonian fluid ignores its inherent shear-thinning and viscoelastic properties. Munir showed that this oversight leads to incorrect predictions of how hydrodynamic stresses amplify during microbubble propagation, potentially miscalculating cellular injury in the lungs.

Even in controlled laboratory environments like organ-on-a-chip devices, the assumption fails. Many experimental setups treat cell culture media as if it behaves like water. However, Chen et al. revealed that cell culture media actually exhibits non-Newtonian, shear-thinning characteristics. Failing to model this behavior introduces substantial errors when calculating the mechanical forces acting on tissue-engineered blood vessels, compromising the validity of vascular health models.


Mathematical Formulations of Variable Viscosity

To capture non-Newtonian behavior, researchers replace the constant viscosity term ($\mu$) in the Navier-Stokes equations with a dynamic function that responds to the local shear rate ($\dot{\gamma}$).

Newtonian:     Viscosity (μ) = Constant
Non-Newtonian: Viscosity (μ) = f(Shear Rate)

Several mathematical models have been developed to define this relationship, each balancing computational simplicity with physical accuracy.

The Power-Law Model

The simplest approach to non-Newtonian modeling is the Power-Law (or Ostwald-de Waele) model. It defines the relationship between shear stress ($\tau$) and shear rate ($\dot{\gamma}$) as:

$$\tau = K \dot{\gamma}^n$$

Which yields an effective viscosity of:

$$\mu = K \dot{\gamma}^{n-1}$$

Here, $K$ is the flow consistency index, and $n$ is the flow behavior index.

  • If $n = 1$, the fluid is Newtonian.
  • If $n > 1$, the fluid is shear-thickening (viscosity increases with shear).
  • If $n < 1$, the fluid is shear-thinning.

Foundational work by Kotake utilized capillary tube viscometry to establish these power-law constants, demonstrating that as flow velocity over a surface increases, shear-thinning fluids eventually converge toward standard Newtonian turbulent behavior once a specific boundary layer threshold is crossed.

Because of its mathematical simplicity, the power-law model is widely used across diverse engineering disciplines. The authors of Du et al. used it to simulate the mixing of carboxymethyl cellulose (CMC) solutions in wastewater treatment tanks. Sheng et al. implemented it to model how sand-laden fracturing fluids thin under high pressure, which directly dictates particle transport and tool erosion patterns downhole. It has also been successfully applied by Deng et al. to resolve complex electrokinetic slip flows in microfluidic channels.

The Carreau and Carreau-Yasuda Models

While the power-law model is useful, it suffers from a severe physical limitation: as the shear rate approaches zero, the predicted viscosity approaches infinity. Conversely, as the shear rate approaches infinity, the viscosity drops toward zero. Real fluids do not behave this way; they possess distinct upper and lower viscosity limits.

The Carreau and Carreau-Yasuda models fix this by introducing two plateaus: $\mu_0$ (viscosity at zero shear rate) and $\mu_\infty$ (viscosity at infinite shear rate).

$$\mu(\dot{\gamma}) = \mu_\infty + (\mu_0 - \mu_\infty) \left[ 1 + (\lambda \dot{\gamma})^a \right]^{\frac{n-1}{a}}$$

This formulation is particularly vital for blood hemodynamics. At rest, red blood cells aggregate into structures called rouleaux, making blood thick ($\mu_0$). As blood accelerates through a vessel, these aggregates break apart and align with the flow, causing the fluid to thin until it reaches a baseline viscosity ($\mu_\infty$).

The authors of Fava-Sanches et al. leveraged the Carreau-Yasuda model to show how flow-diverter stents successfully alter local shear rates inside brain aneurysms, promoting thrombosis and reducing rupture risks. Similarly, Li et al. utilized the Carreau model within fluid-structure interaction (FSI) frameworks to evaluate bio-inspired vena cava filters, proving that inducing a swirling "tornado effect" alters local viscosity and wall shear stress to help dissolve trapped clots.

Multiphase Rheology: The Quemada Model

Blood can also be modeled as a multiphase mixture rather than a single continuous fluid. Red blood cells (RBCs) suspended in plasma make blood a concentrated suspension. To capture this, Si et al. employed an Euler-Euler multiphase approach using the Quemada model. This model defines viscosity as a function of both the local shear rate and the local volume fraction of red blood cells (the hematocrit level), providing a highly accurate method for predicting red blood cell damage (hemolysis) inside mechanical blood pumps.


Computational Frameworks and Acceleration

Integrating non-Newtonian rheology into 3D fluid solvers adds significant mathematical complexity. Because viscosity changes at every point in space and time based on the local velocity gradients, simulations require iterative, tightly coupled solvers.

Local vs. Global Transport Dynamics

Interestingly, the impact of non-Newtonian modeling depends heavily on what you are trying to measure. Vijayaratnam et al. conducted steady simulations comparing Newtonian and non-Newtonian blood models in stented arteries. They discovered a fascinating nuance: while non-Newtonian models drastically alter local flow patterns and dictate exactly where a drug deposits along the vessel wall, the total bulk amount of drug absorbed by the tissue remains surprisingly consistent across models.

This underscores that non-Newtonian modeling is non-negotiable when local spatial distribution matters—such as mapping drug absorption near complex geometric intersections like arterial bifurcations, a phenomenon explored in depth by Levine.

Unified Particle Solvers

Beyond traditional grid-based Computational Fluid Dynamics (CFD), computer graphics and physical animations require solvers that can handle transitions between different material states. Li et al. developed a unified particle-based solver that integrates multiple rheological models simultaneously. By defining viscosity as a direct function of the strain rate tensor within a smoothed-particle hydrodynamics (SPH) framework, their solver seamlessly transitions between shear-thinning behaviors (like pouring ketchup) and shear-thickening behaviors (like impacting cornstarch and water).

Bypassing the Computational Cost

Because full 3D non-Newtonian CFD and FSI simulations can take hours or days to compute, they are often too slow for real-time clinical decision-making or rapid engineering design loops. To solve this, researchers are turning to statistical and machine learning surrogates.

[High-Fidelity Non-Newtonian Data] ──> [Train Surrogate Model] ──> [Real-Time Prediction (ms)]

Bressloff et al. pioneered this approach by training a Bayesian Gaussian process emulator on a small set of high-fidelity hemodynamic simulations. Once trained, this statistical surrogate can predict non-Newtonian risk factors—like wall shear stress distributions—in milliseconds, enabling real-time sensitivity analysis. Following a similar paradigm, Wang et al. paired an immersed boundary-lattice Boltzmann method with a hybrid machine learning framework, achieving near-instantaneous predictions of complex, pulsatile, non-Newtonian flows through stenosed arteries.

By combining rigorous non-Newtonian physics with accelerated computational frameworks, modern fluid dynamics models are successfully moving past the oversimplifications of the Newtonian fluid assumption—yielding unprecedented accuracy where local fluid behavior matters most.

Key Foundational Work

Mining data from hemodynamic simulations via Bayesian emulation (2007) — contributed 1 concept elements to this direction.
Flow and Heat Transfer of Non-Newtonian Fluids on a Flat Plate (1970) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_newtonian_fluid_assumption_0e5735ef,
  title = {Newtonian fluid assumption},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/newtonian-fluid-assumption/rdr_0e5735ef8c7f491c8e4ecf61f67e7a32},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-22},
}

This exact version (version 1, frozen text):

@misc{lacuna_newtonian_fluid_assumption_0e5735ef_v1,
  title = {Newtonian fluid assumption},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/newtonian-fluid-assumption/rdr_0e5735ef8c7f491c8e4ecf61f67e7a32/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-22},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake