Modelling Material Injection Into Porous Structures Under Non-isothermal Conditions

This paper extends the Theory of Porous Media to model non-isothermal material injection into porous structures, specifically for percutaneous vertebroplasty, by incorporating local thermal non-equilibrium conditions and demonstrating thermodynamic consistency through numerical simulations.

The Gist

Here is an explanation of the paper, translated into everyday language with some creative analogies.

The Big Picture: Fixing a Broken Bone with "Hot" (or Cold) Cement

Imagine you have a cracked vertebra in your back. To fix it, doctors perform a procedure called vertebroplasty. They drill a tiny hole and inject a special liquid cement into the spongy bone to glue the cracks together. Once the cement hardens, your back is stable again.

However, there's a catch. The cement is injected at room temperature, which is much colder than your body (37°C). As it sits inside your warm body, it heats up. Also, the chemical process of the cement hardening actually creates its own heat (like how a hand warmer gets hot when you snap it).

The Problem: Previous computer models used to simulate this surgery assumed everything stayed at the same temperature (like a room with perfect air conditioning). But in reality, the cement is cold, the bone is warm, and the hardening process generates heat. This temperature difference matters because it changes how the cement flows and how the bone reacts.

The Solution: This paper introduces a new, more realistic computer model that accounts for these temperature changes. It treats the bone, the bone marrow (the fluid inside the bone), and the cement as three different "characters" that can have their own temperatures while they interact.

---

The Core Concept: The "Sponge" Analogy

To understand the math, imagine the vertebral bone not as a solid rock, but as a super-complex, microscopic sponge.

1. The Solid Skeleton: The actual bone structure (trabecular bone).
2. The Pores: The tiny holes inside the sponge.
3. The Fluids: Inside these pores, there are two liquids fighting for space:
   • Bone Marrow: The natural fluid already there (like water in a wet sponge).
   • Cement: The new liquid being injected (like pouring oil into that wet sponge).

The goal of the model is to predict: Where does the cement go? How fast does it move? And how does the temperature change as it moves?

The "Three-Body" Temperature Problem

In the old models, everyone agreed on the temperature instantly. If the cement was cold, the bone instantly became cold.

In this new model, the authors use a concept called Local Thermal Non-Equilibrium (LTNE). Think of it like this:

• The Cement is a cold swimmer jumping into a warm pool.
• The Bone is the warm water.
• The Marrow is the air bubbles in the water.

When the cold swimmer jumps in, the water right next to them cools down immediately, but the water a few inches away is still warm. The air bubbles might cool down at a different speed than the water.

The authors' model tracks the temperature of the Cement, the Bone, and the Marrow separately. It calculates how heat slowly travels from the warm bone to the cold cement, and how the cement's own hardening process adds extra heat.

The "Traffic Jam" of Fluids

The model also looks at how the fluids move. Imagine a crowded hallway:

• The Cement is a group of people trying to walk in.
• The Marrow is the people already standing there.

As the cement pushes in, the marrow has to squeeze out of the way. The model uses "permeability" (how easy it is for fluids to flow through the sponge) to figure out who gets pushed where.

They tested two different rules for this traffic:

1. The "Simple" Rule: The fluids move in direct proportion to how much space they have.
2. The "Complex" Rule (Brooks-Corey): The fluids interact in a more complicated way, creating a sharp "shockwave" front where the cement suddenly pushes the marrow aside.

What Did They Find?

The researchers ran computer simulations to see what happens in two scenarios:

Scenario A: The "Room Temp" Injection
They injected cement that was the same temperature as the body.

• Result: Not much happened. The temperatures stayed mostly the same. The model worked perfectly, proving it doesn't break when things are easy.

Scenario B: The "Cold" Injection
They injected cement that was significantly colder than the body.

• Result: A "cold wave" moved through the bone. The cement stayed cold for a while, the marrow cooled down slowly behind it, and the bone cooled down even slower.
• The Surprise: Even though the cement was cold, the difference in temperature between the three materials was actually quite small (less than 1 degree Celsius) in their specific test case. This suggests that for this specific type of slow injection, the old "simple" models might still be okay.

However, the authors warn that if the injection is faster (like in a real emergency surgery), the temperature differences could get much bigger, and this new model would be essential to get the physics right.

Why Does This Matter?

Think of this model as a flight simulator for surgeons.

• Old Simulator: Told the pilot (surgeon) "You are flying at 300mph." (It ignored the wind and temperature).
• New Simulator: Tells the pilot, "You are flying at 300mph, but there is a 20mph crosswind, the engine is overheating, and the fuel is freezing."

By understanding exactly how the cement flows and how the heat moves, doctors can:

1. Predict if the cement will leak out of the bone (which can be dangerous).
2. Ensure the bone doesn't get damaged by extreme heat or cold.
3. Plan the surgery better before the patient even gets on the table.

The Bottom Line

The authors built a sophisticated mathematical engine that treats the bone, marrow, and cement as three distinct characters with their own temperatures. While their initial tests showed the temperature differences were small, the model is now ready to handle more complex, real-world scenarios where heat and cold play a major role in saving lives.

Technical Summary

Here is a detailed technical summary of the paper "Modelling Material Injection Into Porous Structures Under Non-isothermal Conditions" by J.-S. L. Völter et al.

1. Problem Statement

The paper addresses the need for more accurate continuum-mechanical models for Percutaneous Vertebroplasty (PV), a medical procedure where acrylic cement is injected into damaged vertebral bone to stabilize it.

• Limitations of Existing Models: Previous macroscopic models based on the Theory of Porous Media (TPM) assumed isothermal conditions (constant temperature). This contradicts clinical reality where:
  1. Injected cement is typically cooler than body temperature.
  2. The cement curing process is exothermic (generates heat).
• Gap: There was a lack of thermodynamically consistent models that account for Local Thermal Non-Equilibrium (LTNE), where the solid phase (bone), fluid phase 1 (marrow), and fluid phase 2 (cement) may possess different temperatures during the injection process.
• Objective: To extend existing TPM-based PV models to the non-isothermal case, incorporating energy balances and heat transfer mechanisms while maintaining thermodynamic consistency.

2. Methodology

The authors developed a multiphase continuum model using the Theory of Porous Media (TPM) framework.

A. Theoretical Framework

• Constituents: The model treats the system as three superimposed continua:
  1. Solid ($\phi^S$): Trabecular bone.
  2. Fluid 1 ($\phi^M$): Bone marrow (wetting fluid).
  3. Fluid 2 ($\phi^C$): Bone cement (non-wetting fluid).
• Governing Equations: The model derives from fundamental balance laws (mass, momentum, energy, entropy) for each constituent.
• Thermodynamic Derivation:
  • The authors employed the Coleman and Noll procedure to ensure thermodynamic consistency.
  • They utilized the Clausius-Duhem inequality to derive constitutive relations.
  • Key Innovation: Unlike standard approaches that assume Local Thermal Equilibrium (LTE), this model explicitly assumes Local Thermal Non-Equilibrium (LTNE). This requires three separate energy balance equations (one for each phase) and constitutive relations for heat exchange between phases.
• Constitutive Relations:
  • Mechanical: Hyperelastic Neo-Hookean model for the solid; Darcy's law for fluid flow (filter laws).
  • Thermal: Fourier's law for heat conduction within phases and a linear heat exchange term proportional to the temperature difference between phases ($\hat{\epsilon} \propto (\theta_S - \theta_\beta)$).
  • Capillarity: The Brooks-Corey model was considered but simplified to linear relative permeability factors because capillary forces were deemed negligible for this specific application.
  • Viscosity: Constant viscosities were assumed for simplicity, despite bone cement and marrow being non-Newtonian.

B. Numerical Implementation

• Discretization: The system was solved monolithically using the PANDAS finite element solver.
• Scheme: Petrov-Galerkin finite element method with Box discretization (full upwinding) for spatial stability and a Crank-Nicolson scheme for temporal integration.
• Geometry: A simplified tubular geometry representing a vertebral body was used for simulation.

3. Key Contributions

1. Non-Isothermal TPM Extension: The derivation of a fully coupled, thermodynamically consistent model for material injection in porous media under LTNE conditions.
2. Thermodynamic Consistency: The authors demonstrated that their model satisfies the second law of thermodynamics (entropy production $\ge 0$) even with simplifying assumptions like the Coleman and Noll procedure.
3. Constitutive Modeling: Explicit formulation of heat exchange coefficients and interface areas based on pore geometry, linking micro-structural parameters to macro-scale heat transfer.
4. Validation: The model was shown to reduce to known isothermal TPM models when LTNE conditions are relaxed to LTE, ensuring continuity with previous work.

4. Results

The authors simulated two scenarios to test the model:

• Scenario 1 (Isothermal-like): Cement injected at body temperature (308.15 K).
• Scenario 2 (Non-isothermal): Cold cement injected (298.15 K) into warmer bone.

Key Findings:

• Saturation and Pressure:
  • Using Brooks-Corey relative permeability factors resulted in a shock front in cement saturation and a non-linear pressure drop.
  • Using linear relative permeability factors resulted in a sigmoid-like saturation profile and a linear pressure drop.
  • The injection pressure remained constant over time with linear permeability but increased significantly with Brooks-Corey permeability (though the authors note this may be an artifact of neglecting non-Newtonian viscosity).
• Temperature Fields:
  • In Scenario 2, the injection of cold cement created a propagating temperature front.
  • LTNE Behavior: The cement temperature changed first, followed by the solid (bone), and finally the marrow. This "lag" confirms the model captures thermal non-equilibrium.
  • Magnitude: The temperature differences between phases were small (max deviation ~0.93 K in Scenario 2). This suggests that for the specific injection rates modeled, LTE assumptions might still be sufficient, but the model correctly captures the physics of the delay.
• Energy Dissipation: The temperature rise in Scenario 1 (due to viscous dissipation) was negligible (0.01–0.05 K), indicating that for slow injection rates, mechanical energy dissipation contributes little to heating.

5. Significance and Future Work

• Clinical Relevance: The model provides a tool to predict thermal effects during vertebroplasty. While the simulated temperature differences were small, the framework allows for investigating scenarios with faster injection rates or highly exothermic curing, where thermal damage to surrounding tissue could be a risk.
• Theoretical Rigor: The work bridges the gap between micro-scale pore physics and macro-scale clinical modeling, offering a rigorous thermodynamic basis for non-isothermal porous media flow.
• Future Directions:
  • Incorporating a curing model for the cement to account for the exothermic heat generation during polymerization.
  • Investigating the necessity of LTNE assumptions under high-flow-rate conditions.
  • Developing a two-scale approach to better capture micro-structural effects.

In conclusion, the paper successfully presents a thermodynamically consistent, non-isothermal extension of TPM for vertebroplasty. While the specific simulations showed minor thermal gradients, the framework is robust and ready for more complex simulations involving cement curing and higher injection rates.