An energy-based material model for the simulation of shape memory alloys under complex boundary value problems

C. Erdogan,T. Bode,P. Junker

2024-02-16

Abstract:Shape memory alloys are remarkable 'smart' materials used in a broad spectrum of applications, ranging from aerospace to robotics, thanks to their unique thermomechanical coupling capabilities. Given the complex properties of shape memory alloys, which are largely influenced by thermal and mechanical loads, as well as their loading history, predicting their behavior can be challenging. Consequently, there exists a pronounced demand for an efficient material model to simulate the behavior of these alloys. This paper introduces a material model rooted in Hamilton's principle. The key advantages of the presented material model encompass a more accurate depiction of the internal variable evolution and heightened robustness. As such, the proposed material model signifies an advancement in the realistic and efficient simulation of shape memory alloys.

Computational Engineering, Finance, and Science

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to develop an energy - based material model for simulating the behavior of shape memory alloys (SMA) under complex boundary value problems. Specifically, the paper aims to:

1. **Improve the accuracy of internal variable evolution**: By introducing Hamilton's principle, this model can more accurately describe the evolution of internal variables (such as volume fraction, rotation matrix, etc.) in shape memory alloys.

2. **Enhance the robustness of the model**: Especially when dealing with complex boundary value problems, ensure the convergence and stability of the model. To this end, the author proposes an improved algorithm to correct the Lagrange multiplier calculation error caused by time discretization.

3. **Achieve a balance between computational efficiency and accuracy**: Minimize the consumption of computational resources while ensuring the accuracy of simulation results, making it suitable for practical industrial applications.

4. **Avoid numerical instability in specific problems**: For example, use Euler - Rodrigues parameters instead of Euler angles to avoid the "Gimbal Lock" problem; use the penalty method instead of the active set method to improve the convergence speed.

5. **Accurately simulate the behavior of twinned martensite**: Ensure the consistency of material responses in different loading directions, thereby more realistically simulating the one - way effect of shape memory alloys.

### Key Formulas

- **Extended Hamilton's Functional**:
\[
H_{\text{ext}} := \int_{\tau} \left( -G - \int_{\Omega} \int_{\tau} \Delta_{\text{diss}} \, dt \, dV - \int_{\Omega} \int_{\tau} \mathbf{l} \cdot \mathbf{c} \, dt \, dV \right) dt
\]
where \( G \) is the total potential energy, \( \Delta_{\text{diss}} \) is the dissipation function, and \( \mathbf{l} \) and \( \mathbf{c} \) are Lagrange multiplier vectors.

- **Helmholtz Free Energy**:
\[
\Psi_i(\boldsymbol{\varepsilon}, \theta, \boldsymbol{\lambda}, \boldsymbol{\alpha}, \boldsymbol{\varepsilon}^{\text{pl}}) = \frac{1}{2} \left( \boldsymbol{\varepsilon} - \mathbf{Q}^T \cdot \boldsymbol{\eta}_i \cdot \mathbf{Q} - \boldsymbol{\varepsilon}^{\text{pl}} \right) : \mathbf{C}_i : \left( \boldsymbol{\varepsilon} - \mathbf{Q}^T \cdot \boldsymbol{\eta}_i \cdot \mathbf{Q} - \boldsymbol{\varepsilon}^{\text{pl}} \right) - c_i(\theta)
\]

- **Volume Fraction Constraints**:
\[
\sum_{i = 0}^{n} \lambda_i = 1 \quad \text{and} \quad \mathbf{\alpha} \cdot \mathbf{\alpha} = 1
\]

Through these improvements, this...